Contents

Title page
About the Author
Introduction

1) Centre Game (and other Second Moves)

2) Calabrese Counter-Gambit
3) Vienna Game
4) Bishop’s Gambit (and other King’s Gambits)
5) Wagenbach Defence
Part 1
Part 2
6) Three Knights Game (and other Third Moves)
7) Scotch Game
8) Two Knights Defence
Part 1
Part 2
9) Ulvestad Variation
10) Jaenisch Gambit
Part 1
Part 2
11) Jaenisch Gambit with 4.Nc3
Part 1
Part 2

Index of Variations

2
 A Disreputable Opening Repertoire

Jonathan Tait

www.everymanchess.com

First published in 2022 by Gloucester Publishers Limited, London.

Copyright © 2022 Jonathan Tait

The right of Jonathan Tait to be identified as the author of this work has been asserted in accordance
with the Copyrights, Designs and Patents Act 1988.

All rights reserved. No part of this publication may be reproduced, stored in a retrieval system or
transmitted in any form or by any means, electronic, electrostatic, magnetic tape, photocopying,
recording or otherwise, without prior permission of the publisher.

British Library Cataloguing-in-Publication Data

A catalogue record for this book is available from the British Library.

ISBN: 978 1 78194 606 0

Distributed in North America by National Book Network,

15200 NBN Way, Blue Ridge Summit, PA 17214. Ph: 717.794.3800.

Distributed in Europe by Central Books Ltd.,

Central Books Ltd, 50 Freshwater Road, Chadwell Heath, London, RM8 1RX.

3
All other sales enquiries should be directed to Everyman Chess.
email: info@everymanchess.com; website: www.everymanchess.com

Everyman is the registered trade mark of Random House Inc. and is used in this work under licence
from Random House Inc.

Everyman Chess Series

Commissioning editor and advisor: Byron Jacobs

Typeset and edited by First Rank Publishing, Brighton.

Cover design by Horatio Monteverde.
Printed by TJ International Limited, Padstow, Cornwall.

4
 About the Author

Jonathan Tait is a Senior International Correspondence Chess Master (2002) and editor for
Everyman Chess. They have been investigating and writing about opening theory for over 30 years
and blog at <200opengames.blogspot.com>.

Also by the Author

Nimzo-Larsen Attack (with Byron Jacobs)

5
 Introduction

WGM Jennifer Shahade (on Twitter): “Have you ever wasted time on an opening that wasn’t worth
your time?”
Me: “That’s virtually my whole chess life summed up in one question.”

I should therefore begin with a warning. If you want to play the very best openings, this book is not
for you because few of those herein are “best” for Black. What they represent is my own disreputable
repertoire in the Open Games after 1.e4 e5.
Hitherto many of these lines have been relegated to a single game, or a single paragraph, denoting
why they should be avoided, supposing they’re mentioned at all. The Calabrese Counter-Gambit
(2...f5; Chapter Two), for instance, was refuted by Carl Jaenisch in the 19th century; the Steinitz
Scotch (4...Qh4; Chapter Seven) is known to be extremely dodgy; the Wagenbach King’s Gambit
(3...h5; Chapter Five) is clearly ridiculous.
So why venture into this sort of territory? Just because! As IM Craig Hanley wrote in Chess
#9/2017 (when asked for a tip for the club player): “Play openings that you enjoy! A hobby is
supposed to be leisure time where you relax and have fun.” And what I most like in the opening is to
throw opponents off balance and onto their own resources. To force them to fight on unfamiliar turf –
preferably my turf. To dictate the terms of the contest – at once, and in a provocative, in-your-face,
suck-on-that kind of way.
Take the Bishop’s Opening (2.Bc4). This tends nowadays to be used as a route to strategic Spanish-
type positions, White setting up with d2-d3, Nf3, 0-0, c2-c3 and so forth. The insolent response
2...f5!? immediately disrupts that. Rather than being able to continue routinely, heading for a
systemic middlegame, White is faced with quite different problems, and at move three.
There are two pertinent points about playing such lines. Firstly, I think it’s important to want to
play them, to want defend their honour, to feel protective of them. This provides added motivation at
the board, leading to heightened concentration, which in itself should lead to better chess. Secondly,
the time limit and strength of opposition may be relevant. Disruptive chess is excellent for blitz and
rapid play, whereas you might have concerns about coming out worse against a strong opponent with
more time to think.
Nevertheless, my own experience – up to about 2350 Elo (equivalent to my highest OTB rating) –
is that you can pretty much play any old rubbish as long as you know what you’re doing. Stronger
players, too, can struggle against unexpected aggression and may choose to avoid critical paths,
hoping instead to outplay you later. Okay, they can try. And please note: this particular rubbish is not
complete rubbish. Some lines may teeter on the edge of soundness as the engines get ever stronger,
such that Stockfish 17 may refute the whole lot in five minutes, but they have stood up to computer-
assisted analysis pretty well thus far. The Wagenbach even featured in the 9th Computer
Championship Final (2019), scoring 1½/2 for Black.

6
 All this is based to a large extent on my own investigations and practice. A lot of the referenced
games are mine, in particular from online thematic tournaments at Tryfon Gavriel’s ChessWorld.net
website. Consequently, they will include players you’ve never heard of – players I’ve never heard of
either, knowing them only by their online sobriquets (or handles), apart from tsmenace, which is me.
(Other instances of “tsmenace” online are not me, though the source is likely to be the same, and I
still have the t-shirt.)
I have referred extensively to the literature too, so much so that I am not providing a bibliography.
A list of over 150 publications – twenty two books are piled up on my desk as I write this – doesn’t
seem like the best use of space. To compress things further I’ll be referring to some paired authors by
initials: B&B = Baker & Burgess; B&H = Botterill & Harding; I&K = Ivanov & Kulagin; K&S =
Khalifman & Soloviev; L&O = Lysyj & Ovetchkin; S&S = Shamkovich & Schiller; T&H = Taylor &
Hayward.
Please accept my apologies for the monstrous thickets of variations. It is the nature of such systems
that their efficacy (or otherwise) can only be demonstrated concretely. So while I accept Emperor
Joseph’s criticism: “There are simply too many moves, that’s all. Just cut a few and it’ll be perfect.” –
I’d also ask: “Which few did you have in mind, Majesty?”
In any case, I don’t really expect people to adopt this repertoire en masse. Perhaps they’ll be
interested in only two or three variations. Whatever and whichever, the very detailed information
required to play each one can be found in the appropriate chapter. If that detail seems overwhelming,
just playing through the main lines (in bold) may well prove enough to start off with. I shall probably
be doing that myself periodically, to try and remember what I’ve written.
Finally, I would be pleased to see anyone else’s games with, or analysis of, any of these openings. I
can be reached by email at <jon.statto@gmail.com> or via my blog <200opengames.blogspot.com>.

Jonathan Tait
Sherwood Forest, England
November 2021

Dedications
To my mother (1936-), sister (1961-2013), father (1924-2006), and gran (1907-2004).

Acknowledgements
• IM Byron Jacobs and Everyman Chess for allowing me to author another book after I failed to
deliver on the previous two; this time I wrote most of it before signing the contract.
• János Wagenbach, Kevin Simpson, Brian Oldham, Frank Noonan, Ken Morrison, David Holmes,
Chris Cantrill, Jim Burnett, David Amour, Peter Ackley, and other friends past and present from 36
years at Mansfield Chess Club.

7
• John Anderson, Donald Andrew, Michael Barnes, Stefan Bücker, Andrew Butterworth, Paul
Cumbers, John Elburg, John Emms, Bernard Hanison, Tim Harding, Otto Hardy, Volker Hergert,
Thomas Johansson, Richard Palliser, Gerard Welling – for various degrees of assistance over the
years, some so long ago they may have forgotten or even died. Apologies to whomever I’ve
neglected to mention.
• Everyone involved with the Stockfish open source chess engine. I’ve used other engines too (Fritz,
Houdini, and Rybka in various incarnations), but Stockfish has been my primary companion.
• My opponents (over-the-board, correspondence, and online) for their inadvertent, yet crucial
collaboration.

8
 Chapter One
Centre Game
(and Other Second Moves)

1.e4 e5 2.d4

The main subject of this chapter. Of the occasionals, White has played every legal move here bar
one (2.Ba6??). Most I’ve never faced and therefore have nothing to say about them, if indeed there’s
anything worth saying other than “Black has equalized”. Consequently, I’ll mention only a selection:
a) 2.Qh5!? Nc6 3.Bc4 g6 4.Qf3 is a joke opening but not a desperately bad one. I notice two games
in the databases: Ma.Carlsen-L.Dominguez Perez, St. Louis (blitz) 2019 and Stockfish-Houdini 6,
CCC 1 Rapid Rumble 2018, both drawn. In response and in the same happy spirit I like 4...f5!? and
then:

9
 a1) 5.exf5? runs into 5...Nd4 6.Qg3 (or 6.Qe4 d6) 6...Qf6! 7.Bd3 (or 7.Bb3 Ne7) 7...d5 8.Ne2 (not
8.fxg6? e4), N.Abarca Gonzalez-H.Lopez Silva, Chilean Ch., Iquique 2013, and now 8...Bd6!
(Lokander); e.g. 9.Nxd4 exd4 10.Qf3 Ne7 11.fxg6 Qxf3 12.gxf3 hxg6 with a big advantage.
a2) 5.Ne2 Nf6 6.Nbc3 (or 6.d3 f4 7.d4 Qe7 with ideas of ...g6-g5, ...Bg7, ...d7-d6, and ...g5-g4 –
Lokander) 6...d6 (or 6...Na5!? – Bologan) 7.d3 (here 7.exf5 Bxf5 8.d3 is fairly equal) 7...f4 8.h3?!
(8.Nd5 g5 9.d4 is a better solution; not 8.g3?? Bg4 9.Qg2 f3 and wins) 8...g5 (or 8...Na5 again –
Sasikiran) 9.g4 h5 10.gxh5 Rxh5 11.Nd5, E.Rios-A.Whatley, Dallas 2001, and now 11...Rh4!,
threatening ...Nxd5, ...g5-g4 or ...Bxh3, is good for Black.
b) 2.Qf3 might be met in similar fashion; i.e. 2...Nc6 3.Bc4 f5!?. For example: 4.Ne2 (either 4.exf5
Nf6, followed by ...d7-d5 or 5.Nc3?! Nd4 6.Qd1 d5; or 4.Bxg8 Nd4 5.Qd1 Qg5! 6.g4 Rxg8 7.c3,
P.Dimitrov-V.Belov, European Ch., Plovdiv 2008, and 7...Nc6 8.exf5 g6 9.d4 Qf6 gives Black
excellent play for the pawn) 4...Nf6 5.d3 f4 6.h3 (or 6.d4 d6 with ideas of ...g7-g5; not 6.g3?? d5
7.exd5 Bg4 and wins) 6...Na5 (or 6...g5 again) 7.b3 Nxc4 8.bxc4 Bb4+ 9.c3 Ba5 10.Nd2, P.Dimitrov-
M.Stoinev, Sunny Beach 2008, and now 10...0-0, intending 11.Nb3 (or 11.c5 d5) 11...c6, looks good
for Black too.
c) 2.d3 is non-theoretical, aiming perhaps for a reversed Philidor. I’ve had this a few times and have
usually gone for 2...Nc6. Then 3.Nc3 Bc5 4.g3 d6 5.Bg2 is a Vienna, specifically 5.d3 in line A of
Chapter Three. Instead, for 3.Nf3 see 3.d3 in Chapter Six; while 3.f4 Bc5 is a pleasant King’s Gambit
Declined, e.g. 4.Nf3 (or 4.fxe5 f6) 4...d6 5.Nc3 a6 6.Nd5 Nge7 7.c3 Ba7 8.Nxe7 Qxe7 9.Qe2 Bd7 and
Black is at least equal, D.M.Adams-J.Tait, Sheffield League 1999.
d) 2.Ne2 (Alapin’s Opening) I would probably answer with 2...Nc6 as well.

10
 Then 3.Nbc3 Bc5 is another sort of Vienna; e.g. 4.g3 d6 5.Bg2 is line A in Chapter Three again. As
there 4.Na4 can be ignored: 4...d6 (4...Be7 5.d4! was Alapin’s idea) 5.Nxc5 dxc5 6.d3 Be6 7.Be3
Qd6 8.Nc3 0-0-0 and Black has a more than acceptable position, K.Shanava-I.Beradze, Tbilisi 2016;
while 4.f4?! d6 5.Na4 Bg4! 6.Nxc5 dxc5 7.d3 Qh4+ (or 7...f5!?) 8.g3 Qh5 is already good for Black,
C.Schramm-A.Naiditsch, German Bundesliga 2016.
The immediate 3.f4?! is suspect on account of 3...Bc5! 4.fxe5 (or 4.c3 d6 5.d4 Bb6 6.Be3 f5!?)
4...d6! 5.exd6 (or 5.d4 Qh4+ 6.Ng3 Bxd4 7.exd6 Be5) 5...Qh4+ 6.Ng3 Bxd6 and White “is not very
far from disaster” (Bologan).
The only issue is 3.d4!, since 3...exd4 4.Nxd4 is then a Scotch. The issue is psychological rather
than theoretical: having started with 2.Ne2, does White really deserve to play a Scotch? You decide.
e) 2.c4 is a sort-of accelerated Botvinnik formation. The drawback in setting up so soon is that
Black has total freedom in reply. In my one game against this, I opted for 2...Nc6 3.Nc3 Bc5 4.Nf3 (if
4.g3 d6 5.Bg2 f5 6.d3 Nf6, “Black has achieved the optimal set-up: the ‘bad’ bishop is actively
placed on c5, and White’s centre is pressured with the lever on f5.” – Bologan) 4...d6 5.Be2 (or if
5.h3 f5 6.d3 f4!? 7.Be2, A.Van Osmael-V.Malinin, corr. 1998, then 7...Nge7 and ...Ng6 seems
appropriate) 5...f5 6.exf5 Bxf5 7.0-0 Nf6 8.d3 0-0 9.Be3 (or 9.a3 a5) 9...Nd4 10.Bxd4 Bxd4 11.Qd2
(after 11.Nxd4 exd4 12.Ne4 Bxe4 13.dxe4 c5 14.Bd3, M.A.Pintor-K.Foster, corr. 2008, Stockfish
likes 14...g5!? for Black) 11...c5 12.Nxd4 exd4 and had the edge when White defaulted, chess4fun87-
tsmenace, ChessWorld.net 2019.
f) 2.c3 (© Nicholas MacLeod, New York 1889) poses as an accelerated Ponziani, prompting a
similar response: 2...d5 3.Nf3!? (MacLeod’s original idea; otherwise, 3.exd5 Qxd5 4.d4 Nc6 5.Nf3
exd4 transposes to the Göring Gambit Declined in Chapter Six; while 3.d3 Nf6 4.Nf3 Nc6 5.Nbd2 a5
is another reversed Philidor where Black may follow with ...Bc5, as after 3.Be2 Nf6 4.d3 in Chapter
Six again) 3...dxe4 4.Nxe5 Qe7!? (in view of 4...Qd5 5.f4!) 5.Nc4 (not 5.Qa4+?! Nd7) 5...Nf6 6.Be2,
D.Van Elsen-G.Maxia, NATO Championship, Oslo 1990, and now 6...Nc6, intending ...Be6 and ...0-

11
0-0, or 7.d4 exd3 8.Qxd3 Be6 9.0-0 Nd5 and ...0-0-0, when White has the clumsier set-up.
g) 2.a3!? (Mengarini’s Opening) 2...Nf6 3.Nc3 is an interesting role reversal. I sometimes try and
surprise Alekhine’s Defence players with this (i.e. 1.e4 Nf6 2.Nc3 and if 2...e5 then 3.a3!?). White is
now playing the Black side of a standard Open Game, where intrusions at QN5 have been ruled out.
There’s no Ruy Lopez for starters, and the a-pawn move is quite useful in the Italian; while 3...d5 can
be met by 4.exd5 Nxd5 5.Qh5!?, playing a 4...Qh4 Scotch without having to worry about N-N5. Then
5...Nc6 6.Bb5 Qd6 7.Nf3 is equivalent to line A in Chapter Seven, which Black has likely never seen
before.

For instance, 7...Nxc3 8.dxc3 e4?! (here 8...Bd7 9.Bxc6 Bxc6 10.Nxe5 Qd5! is best) 9.Ng5 g6?
10.Qe2 f5 11.Qc4! Qf6 12.Be3 led to a quick win for White (with 24.b3 mate no less) in J.Tait-
A.N.Walker, Notts League 1997.
However, you can wipe White’s smug smile away with 7...Nf4! 8.Qxe5+ Qxe5+ 9.Nxe5 Nxg2+
10.Kf1 (or 10.Ke2 a6!) 10...Nh4 11.Nxc6 Bh3+ 12.Ke2 Bg2!, when there is nothing better than
13.Ne5+ c6 14.Bc4! Bxh1 15.Nxf7 Bf3+ 16.Kf1 Rg8 17.Ng5 with a draw, as in tsmenace-docjan,
ChessWorld.net 2019. Well played, my opponent there!
h) 2.Bb5 (Portuguese Opening) 2...c6 3.Ba4 d5!? fights for the initiative straight away; e.g. 4.exd5
(after 4.Nc3, I like Lokander’s 4...a5 5.a3 dxe4 6.Nxe4 Nf6 7.Nxf6+ Qxf6 with easy and active
development, while the white bishop looks silly on b5) 4...Qxd5 5.Qf3 (not 5.Nf3?? Qe4+ and
...Qxa4) 5...Qa5 6.Nc3 Nf6 7.Bb3 Na6 8.Qg3!? (an attempted improvement on 8.a3 Bd6 9.d3 Nc5
10.Ba2 Ne6 11.b4 Qc7, K.Poulheim-W.Brodda, corr. 2002, which looks like an excellent Centre-
Counter for Black) 8...e4! (partly to cramp, partly to enable ...Nc5 again) 9.f3 Bf5 10.fxe4 Nxe4
11.Nxe4 Bxe4 12.Nf3 Nc5 13.Qe5+!? (otherwise White is just worse) 13...Be7 14.Qxg7 0-0-0
15.Bxf7 Bxf3 16.gxf3 Rhf8 with superb play for the pawn, K.Poulheim-W.Pawlowski, corr. 2016.

2...exd4

12
 Not completely forced: my game against Pawnhunter1 in Chapter Seven actually started 1.d4
Nc6!? 2.e4 e5 3.Nf3, which would be 2...Nc6!? etc here. Nonetheless, taking the d-pawn is certainly
best.

A: 3.c3
B: 3.Qxd4
Other moves:
a) 3.Nf3 is a popular way to avoid the Petroff, offering instead a Scotch Game or Gambit after
3...Nc6. If Black doesn’t feel like either, there is 3...Bc5 4.Nxd4 (or 4.Bc4 d6 5.Nxd4; while after
4.c3 dxc3 5.Nxc3 d6 6.Bc4 Nf6 7.0-0 0-0 8.Bg5 Nbd7 or 7.e5 dxe5 8.Qxd8+ Kxd8 9.Nxe5 Re8
10.Bf4 Be6 11.0-0-0+ Kc8 12.Bxe6+ fxe6, it’s up to White to try and justify the pawn sacrifice –
which the engines have mostly managed in their own contests, only losing once) 4...d6 5.Bc4 Nf6
6.Nc3 0-0 7.0-0 Re8 8.Bg5 (Bologan mentions the nice trap 8.Re1? d5! 9.Nxd5 Nxd5 10.Bxd5 Bxd4
11.Qxd4 c6, winning a piece) 8...Nbd7, which was never anything other than equal in Gull 3-
Komodo, TCEC 9 Rapid 2016.
b) 3.Bc4 can be met by 3...Nf6, when 4.e5?! d5 is good for Black, 4.Qxd4 Nc6 5.Qe3 transposes to
5.Bc4 in line B, and 4.Nf3 is the Urusov Gambit, which it’s acceptable to accept; e.g. 4...Nxe4
5.Qxd4 Nf6 6.Nc3 c6 7.Bg5 d5 8.0-0-0 Be7 9.Rhe1 Be6 is roughly equal. Alternatively, 4...Nc6 is the
Scotch Gambit again; while Bologan likes 4...Bb4+!? 5.c3 (or 5.Bd2 Bxd2+ 6.Nbxd2 Nc6 7.0-0 0-0)
5...dxc3 6.bxc3 (or 6.0-0 0-0 7.bxc3 Be7) 6...d5! 7.exd5, though here I’d prefer 7...Be7 (just to avoid
7...Bd6 8.Qe2+ Qe7, even if Black is fine) 8.0-0 0-0 9.Ba3 Bxa3 10.Nxa3 Bg4 with at least equality,
T.Sarkisian-V.Antoshin, Yerevan 1981.
c) 3.f4?! only makes sense – as a reversed Elephant Gambit – if Black plays ...Nf6. So Black should
refrain from that: 3...Nc6 4.Nf3 d5! 5.e5 (or 5.Nxd4 dxe4 6.Bb5 Bg4) 5...Bc5 6.Bd3 (or 6.Nbd2 Nh6
7.Nb3 Bb6 8.Nbxd4 Nxd4 9.Nxd4 0-0, intending ...f7-f6 and/or ...c7-c5) 6...Nge7 and White is just a

13
pawn down; e.g. 7.a3 (or 7.0-0 Bf5 8.a3 a5 9.Qe2 Qd7 10.Ne1 0-0 11.Nd2 Rae8 12.Kh1 f6,
A.Lantos-T.Banusz, Hungarian League 2014) 7...Bf5 (or 7...Nf5) 8.b4 Bb6 9.Bb2 a6 10.0-0 Qd7
11.Nbd2 Bxd3 12.cxd3 Nf5 13.Rf2 0-0 14.Nb3 f6 and Black is clearly better, S.Soto Perez-M.Marin,
Seville 1992.

A: 3.c3

The so-called Danish Gambit. Personally I’d give GM Jacques Mieses the credit, since it was his
some three dozen games with 3.c3 in the nineteen-noughties that established the theory – but there
you go.

3...Qe7!?

I like this move with its shades of 4...Qh4!? in the Scotch (see Chapter Seven). Black sends the
queen out early to nab the white e-pawn. Alternatively, 3...d5 looks for a Göring Gambit Declined
after 4.exd5 Qxd5 5.Nf3 Nc6 (see Chapter Six).

4.cxd4

White’s best response is to sacrifice the e-pawn thus.

Other ways of doing so are no good: 4.Nf3? Qxe4+ 5.Be2 d3!; or if 4.Bc4? Qxe4+ 5.Ne2,
K.Glacner-M.Kominek, Valasska Bystrice 2017, then 5...d5! and Black is pretty much winning.
Nor is trying to defend it: 4.f3? d5; or 4.Nd2? dxc3; or 4.Bd3?! d5 5.e5 Qxe5+ 6.Ne2, N.Cook-
B.Winkel, corr. 2004, and 6...Bd6 7.cxd4 Qf6; or 4.Qe2?! Nf6 5.e5 (or 5.Nd2 d5 6.e5 d3! 7.Qxd3
Qxe5+ 8.Be2 Bd6 9.Ngf3 Qe7, R.Réti-R.Spielmann, Baden 1914) 5...d3! 6.Qxd3 Qxe5+ 7.Be2 d5
8.Nf3 Qf5 9.Qd1 Bd6, Fr.Arnold-L.Van Foreest, Paleochora 2020. In all cases White has nothing for
the pawn.

14
 Finally, 4.Qxd4 can be met by 4...f5!? (Bologan’s suggestion; the standard 4...Nc6 5.Qe3 Nf6
6.Nd2 d5 looks good too) 5.Bf4 (or 5.Bd3 Nf6 6.Bg5 Nc6 7.Qe3 Ne5 8.Be2 fxe4 9.Bxf6 gxf6
10.Qxe4 c6 11.Nd2 d5, D.Fricke-D.Kornblum, corr. 2008) 5...Nc6 6.Qa4 Qxe4+ 7.Qxe4+ fxe4 8.Nd2
(or 8.Bxc7 d6 9.Bb5 Bd7 10.Bxc6 Bxc6 11.Ba5 Nf6) 8...Nf6 9.0-0-0 Bc5 10.Bxc7 d6 11.b4 Bxf2
12.Bxd6 Be3 and Black is better, even without the extra pawn.

4...Qxe4+

5.Be3

The best way to block the check.

a) 5.Ne2?! is clumsy, so that something like 5...d5 6.Nbc3 Qe6 7.Be3 Nf6 8.Nf4 Qd7 9.Bd3 c6
10.Qd2 Bd6 11.0-0-0 Qc7 reorganizes effectively.
b) 5.Be2?! is too speculative, since Black can safely take the second pawn: 5...Qxg2 6.Bf3 Qg6
7.Ne2 (or 7.Nc3 Bb4 8.Ne2) 7...Bb4+ (not 7...Ne7 8.Nbc3 c6? 9.Nf4 Qd6? 10.0-0 f5? 11.Re1 and
White soon won, A.Alekhine-A.Cheron, Vevey 1925 – not M. Cheron’s finest hour) 8.Nbc3 Ne7
9.Rg1 (or 9.Nf4 Qa6 10.Qb3 Nbc6 11.Be3 d6 12.Rg1 0-0 13.0-0-0 Bxc3 14.bxc3 Na5 15.Qa3 Qb5
16.Nh5 Ng6 17.Rg5 Nc4 18.Rxb5 Nxa3 19.Rb3 Nc4 20.Bxb7 Rb8 21.Bxc8 Rfxc8 and Black is just a
pawn up, A.Vlasov-L.Klykow, corr. 2009) 9...Qf6 10.Qd3 (if 10.Nf4 Nbc6 11.Be3 g6 12.Qd3,
I.Firnhaber-V.Belov, corr. 1990, Black might even go 12...g5!? 13.Nfd5 Qxf3 14.Nxc7+ Kd8
15.Nxa8 h6 and will be winning when the a8-knight drops) 10...h6 11.Bd2 Nbc6 12.Qe3 Kd8!?
(Black really is playing a Scotch) 13.Ne4, J.Barrios Troncoso-F.Smalcl, corr. 2002, and now 13...Qe6
keeps a big advantage; e.g. 14.Bxb4 (or 14.Rxg7 Nf5) 14...Nxb4 15.Qd2 a5 16.Nc5 (if 16.a3 then
16...Qb3!, or 16.Rxg7 d5) 16...Qf6 17.Bxb7 Rb8 18.Bh1 g5 and White’s initiative has run out still a
pawn down.

15
5...Bb4+

Here Bologan opts for 5...d5!? 6.Nc3 Qe6 7.Nf3 c6 8.Bd3 Be7 9.0-0 Nf6 (De Firmian), and if
10.Qc2, as in P.Gara-Is.Horvath, Hungarian League 2003, then 10...Qd6. Stockfish thinks White is
“0.00” after 11.Rae1 0-0 12.Ne5, but it would be up to a human player to prove it.

6.Nc3 Nf6 7.Nf3 Nd5

Taking the time to remove the enemy bishop. After 7...0-0 8.Bd3 Qe7 9.0-0 d5 10.Bg5, as in
J.Mieses-He.Wolf, Coburg 1904, the pin is a bit annoying.

8.Qd2 Qe7 9.Bd3

Not 9.Bc4?! Nxe3 10.fxe3 c6 11.0-0 (or 11.d5 d6) 11...d5 12.Bd3 Nd7, as in S.Tabak-G.Schnider,
Prague 2020, as Black’s extra ...c7-c6 prevents the Nxd5 trick in the next note.

9...Nxe3 10.fxe3 d6

If 10...d5 11.0-0 Nd7, then 12.Nxd5! Bxd2 13.Nxe7 Bxe3+ 14.Kh1 Kxe7 15.Rae1 Kd6 16.Rxe3
Nf6 17.Bc4, Y.Cordel-Y.Simonet, corr. 2015, probably gives White enough play for the pawn.

11.0-0 Nd7!?

Old theory goes 11...Bxc3 12.bxc3 Nc6 13.e4 0-0 14.Rae1, A.Romashkevich-P.Saburov, Russia
1889, when 14...f6 is equal according to GM De Firmian.

12.e4 Nf6 13.Rae1

After 13.Rac1 c6 14.a3 Bxc3 15.bxc3 0-0 16.Rce1 Be6, White has merely wasted a move with the

16
rook and stands worse, Je.Andersen-Y.Anokhin, corr. 2011.

13...0-0

Neither do men put new wine into old bottles: else the bottles break, and the wine runneth out, and
the bottles perish. Or not, as the case may be.
For example: 14.a3 Bxc3 15.bxc3 Be6 16.e5 dxe5 17.Rxe5 (if 17.Nxe5 c5 18.Qf2 h6 19.Qh4 Qd8
20.Rf3, I.Nowak-V.Liamtsau, corr. 2012, then 20...Nd5; or if 17.dxe5 Nd5 18.c4, D.Andrae-R.Boles,
corr. 2012, then 18...Nb6 19.Qf4 Bxc4 20.Bxc4 Qc5+, or even 17...Nd7, heading for f8) 17...Rae8
18.Ng5 h6 19.Nxe6 fxe6 20.Bc4 Qd6 21.Qe2 Qxa3 22.Bxe6+ Kh8 with an edge, H.Fabig-K.Gräfen,
corr. 2012.
Although all those games were drawn, the onus was on White to demonstrate how.

B: 3.Qxd4

“The Centre Game” has a nice archaic ring to it; indeed it dates back to the 16th century according
to Hooper & Whyld. Well, the queen is there to be harried, so:

3...Nc6

17
4.Qe3

Her Majesty has gone elsewhere too, including:

a) 4.Qc3?? Bb4 and White is 1/47 in the databases, with the “1” looking like a default, perhaps
from annoyance that White hadn’t resigned.
b) 4.Qd3 gives Black extra options, not least ...d7-d5 and ...Nb4, which can even be played at once:
4...d5!? 5.Qxd5 (similarly 5.exd5 Nb4 6.Qe2+ Qe7 7.Na3 Nxd5; or 5.Nc3 d4 6.Nd5 Nf6 and if
7.Bg5?! then 7...Nb4! 8.Nxb4 Bxb4+ 9.c3 dxc3 10.Qxd8+ Kxd8 11.0-0-0+ Ke8 12.bxc3 Bxc3 with
advantage – L&O) 5...Qxd5 6.exd5 Nb4 7.Bb5+ (or 7.Na3 Nxd5) 7...Bd7 8.Bxd7+ Kxd7 9.Kd1 Nxd5
and Black is certainly not worse.
c) 4.Qd2 is credited to GM David Bronstein, with the idea of Bd3, f2-f4, Nf3, 0-0, b2-b3 and Bb2.

18
 As GM John Emms says: “This works very well if Black has no moves, but otherwise 4...Nf6
5.Bd3 d5 is a reasonable antidote”; e.g. 6.exd5 Qxd5 7.Nf3 Bg4 8.Be2 0-0-0 9.Qxd5 Nxd5 with a
useful lead in development, J.Ruano Marco-F.De la Fuente Gonzalez, Galapagar 2004.
d) 4.Qd1 wastes time, such that 4...Nf6 5.Bd3 (or 5.Nc3 Bb4) 5...d5 6.exd5 Qxd5 7.Nf3 Bg4 is
even better for Black; e.g. 8.Be2 Qh5 9.Bf4 Nd5 10.Bg3 0-0-0 11.Nbd2 Bb4 12.0-0, Ma.Lange-
L.Paulsen, 2nd matchgame, Leipzig 1864, where 12...Ne3! 13.fxe3 Bxd2 wins.
e) 4.Qa4 is a reversed Centre-Counter, but White’s extra e2-e4 is not optimal because the logical
adjunct Nc3 then rules out c2-c3, leaving White’s queen stuck on the side and vulnerable to ideas of
...b7-b5 or ...Bd7. There are also issues with the f2-square since White doesn’t have e2-e3 either. For
example, following 4...Bc5 and then:

19
 e1) 5.c3?! Nf6 is already quite bad for White, since 6.Bg5? runs into 6...Bxf2+.
e2) 5.Nc3 d6 6.Nd5 (or 6.Nf3 Bd7 as below) 6...Bd7 (or 6...Nf6, intending 7.Bg5?! Bxf2+! 8.Kxf2
Ng4+ 9.Ke1 Qxg5 10.Nxc7+ Kd8 11.Nxa8 Re8 with strong play for the rook; e.g. 12.Bd3 Ne3 13.g3
Qc5 14.Qa3 Nb4 15.Qc3 Nexc2+ and so on) 7.c3? (here 7.Be3 may equalize) 7...Nf6 8.Qc2 Nxd5
9.exd5 Ne5 10.Bf4 Qf6 11.Qd2 0-0 12.0-0-0 a5 and Black stands very well, J.Mieses-P.Leonhardt,
Vienna 1908.
e3) 5.Nf3 d6 6.Bg5 (lines like 6.c3 Nf6 7.Be2 0-0 8.0-0 Re8 9.Bg5 h6 10.Bxf6 Qxf6, M.Steadman-
P.Briggs, British Ch., Scarborough 2001; or 6.Nc3 Bd7 7.Qc4 Nf6! – Sv.Johnsen; or 6.Bb5 Bd7 7.0-0
Nf6 8.Nc3 0-0 9.Bg5 a6 10.Bxc6 Bxc6 11.Qc4, D.Lardot-K.Lehtosaari, Oulu 2002, and 11...h6 are at
least a little better for Black) 6...Nf6 7.Nc3 (not 7.c3?! h6 8.Bh4 g5 9.Bg3 Qe7 10.Nbd2 Bd7 11.Qc2
d5! with a clear advantage – Lokander) 7...h6 8.Bh4 Bd7 9.Bb5 (if 9.0-0-0 0-0 10.Nd5?! g5 11.Bg3,
G.Van Perlo-M.Schroeder, corr. 2000, then 11...Nxd5 12.exd5 Nb4 13.Qb3 a5 is good for Black; e.g.
14.a3 a4 15.Qc4 b5 16.Qe2 Qf6 17.axb4 a3 18.bxc5 Qxb2+ 19.Kd2 a2 and so forth) 9...a6 10.Bxc6
Bxc6 11.Qc4? (White had to play 11.Qb3, intending 11...g5 12.Bg3 Nxe4 13.Nxe4 Bxe4 14.0-0-0 0-0
15.h4 with compensation) 11...g5 12.Bg3 b5 13.Qd3 b4 14.Nd5 Nxd5 15.exd5 Bb5 16.Qe4+ Qe7
17.Qxe7+ Kxe7 18.h4 f5! 19.0-0-0 f4 20.Bh2 Bxf2 and Black won, Ed.Levi-Gu.West, Melbourne
2002.
f) 4.Qc4!? has been seen in numerous high-level blitz games recently, having received a fillip when
the Chinese GM Xu Xiangyu scored 6/6 with it in 2019.

The idea seems to be that, after 4...Nf6 5.Nc3 (not 5.Bd3?! d5 6.exd5 Qxd5 – L&O; or 5.Bg5?! h6,
intending 6.Bh4? g5! 7.Bg3 d5 8.exd5 Nxd5 – Dhopade) 5...Bb4 6.Bd2 (here 6.Nge2 d5! 7.exd5
Nxd5 8.Bd2 Nxc3 9.Nxc3 Qe7+ 10.Qe4 Qxe4+ 11.Nxe4 Bf5 was dead level in 32Stockfish-
42Stockfish, TCEC 20 Swiss Test 2021) 6...0-0 7.0-0-0 Re8, White can safely play 8.f3 since Black
doesn’t have ...d7-d5 (compare 8.f3 in the main line). That apart, after 8...d6 the queen will soon have
to retreat from c4 and Black can use the time to generate counterplay:

20
 f1) 9.Qe2, as in Wei Yi-S.Karjakin, FIDE Grand Prix, Jerusalem 2019, can be answered by 9...Ne5,
with ideas of ...c7-c6 and ...d7-d5, or even 9...d5!? at once; e.g. 10.a3 (or 10.Qe1 d4 11.Nce2 a5,
while 10.Qf2 transposes to 8.f3?! d5 9.Qf2 in the main line) 10...Bf8!? 11.Bg5 d4 12.Nd5 Be6
13.Nxf6+ gxf6 14.Bh4 Ne5 15.f4 Ng6 16.Qf2 Bh6 17.Bg3 c5 18.Kb1 f5 and Black stands well.
f2) 9.a3 a5 10.Qe2 (or if 10.Qa2?! Be6 11.Bc4, J.Ehlvest-An.Tang, Titled Tuesday blitz 2020, then
11...Bxc3 12.Bxc3 Nd7, intending ...b7-b5) 10...Ne5 11.g4 (not 11.axb4? axb4 12.Nb1 c5, “followed
by ...Ra1, with a nasty attack” – Dhopade) 11...Be6 (the immediate 11...b5!? looks tempting too)
12.Qg2 Nc4 13.Bxc4 Bxc4 14.g5 Nd7 15.h4 b5 16.axb4?! axb4 17.b3 bxc3 18.Bxc3 b4 19.Bb2,
P.Ponkratov-M.Matlakov, World Blitz Ch., Moscow 2019, and now 19...Ra2! wins: Black threatens
...Qa8, ...Rxb2 and so on, or if 20.Qd2 then 20...c5! (preventing Qd4) 21.Qxd6 Be6 22.Qe5!? Nxe5
23.Rxd8 Rxd8 24.Bxe5 Rda8 with a decisive advantage.
f3) 9.g4 Be6 10.Qe2 (not 10.Nd5? Bxd2+ 11.Rxd2 Bxd5 12.exd5 Re1+ 13.Rd1, K.Rathnakaran-
N.Batsiashvili, Titled Tuesday blitz 2020, when 13...Qe7 threatening ...Qe3+ is close to winning)
10...Nd4 (again 10...d5!? looks logical) 11.Qf2 c5 12.Nge2 Qb6? 13.a3 Bxa3 14.bxa3 Nxc2 15.Kxc2?
(15.Nf4 is good for White) 15...Qb3+ 16.Kd3?? (16.Kc1 Qxa3+ is a draw) 16...Bc4+ 17.Ke3 Nd5
mate, Ma.Carlsen-An.Tang, 9th bullet game, lichess.org 2021.
Using Stockfish to critique blitz games, even those by the sixteenth world champion, is the
proverbial barrel shoot – but theory has to be based on something.

4...Nf6

21
5.Nc3

A few other moves must be mentioned:

a) 5.Bd2 avoids the pin in the main line, though Black can play 5...Bb4 anyway; e.g. 6.a3 (more
usually White accepts the transposition after 6.Nc3, whereas adding in 6.c3 Be7 7.c4 0-0 8.Nc3 Bb4
doesn’t fit with White’s intended long castling) 6...Bxd2+ 7.Nxd2 0-0 8.0-0-0 d5 9.exd5 Qxd5
10.Nc4 (10.Ne2 Qe5 11.Qxe5 Nxe5 was dead level in D.Anton Guijarro-W.So, Carlsen Invitational
rapid 2021) 10...Qh5 11.Ne2 (or 11.Be2 Bg4) 11...Ng4 12.Qg3 Qc5 13.Ne3 Nxe3 14.Qxe3 Qxe3+
15.fxe3 Bg4 16.h3 Bxe2 17.Bxe2 Ne5 and Black is not worse, R.Bancod-O.Korneev, Johor 2019.
b) 5.Bc4 also prompts 5...Bb4+!

22
 6.c3 (now 6.Nc3 0-0 7.Bd2 runs into 7...Bxc3! 8.Bxc3 Nxe4, followed by ...Nxc3, and White is just
a pawn down; while 6.Bd2 0-0 7.a3 allows 7...Re8 8.f3 Bxd2+ 9.Nxd2 d5) 6...Ba5 7.Nf3 (or 7.Qg3 0-
0, since 8.Bh6 Nh5 is fine for Black) 7...0-0 8.0-0 Re8 9.Qd3 Nxe4 10.Qd5 Re7 11.Ng5 Qe8 and
Black has safely won a pawn, D.Durner-F.Schmid, corr. 2012.
c) 5.Nf3 Bb4+! (once again) 6.c3 (everything else is worse here) 6...Ba5 7.Bd3 (relatively best;
7.Bc4 is note ‘b’ above) 7...0-0 8.0-0 Bb6 9.Qf4 d5! (simpler than 9...d6 10.Na3 Nh5 11.Qg5 Qxg5
12.Bxg5 h6 13.Bd2 Bg4 14.Nc4 Bxf3 15.gxf3, when White had two bishops vs. two knights and
eventually won in Stoofvlees II-Rodent III, TCEC 16 Qualification 2019, despite the “0.00”
assessment at this point) 10.exd5 (or 10.e5 Nh5 11.Qa4 Bd7) 10...Nxd5 11.Qg3 (or 11.Qa4 h6)
11...Nf6 12.Bc2 Nh5 13.Qg5 Qxg5 14.Bxg5 h6 15.Bc1 Re8 and Black is at least equal.
d) 5.e5?! is risky because of 5...Ng4!

23
 6.Qe2 (after 6.Qe4 Ngxe5! 7.f4 d5 8.Qe3 d4 9.Qe2 Bd6 10.fxe5 Nxe5, followed by ...0-0 –
Dhopade, or 8.Qe2 Bg4 9.Nf3 Bc5 10.Kd1 0-0 11.fxe5 Re8 12.Qb5 Nxe5 13.Be2 Bxf3 14.Bxf3 Bb6,
Black has two pawns and an ongoing initiative for the piece, M.Soszynski-S.Nichols, corr. 2010)
6...d6! 7.exd6+ (inserting 7.f3 Nh6 isn’t a huge improvement, e.g. 8.exd6+ Be6 9.Bxh6 Qh4+ 10.g3
Qxh6 11.d7+ Kxd7 12.Qd2+ Qxd2+ 13.Nxd2 Re8 14.0-0-0 Bxa2 safely wins a pawn; while if 7.h3
Ngxe5! 8.f4 Nd4 9.Qe4, J.Taubenhaus-I.Gunsberg, Hamburg 1885, then 9...c5! 10.fxe5 d5 or 10.Na3
Qh4+ 11.Kd1 Be7 12.g3 Qf6 more or less wins for Black) 7...Be6 8.h3 Nge5 9.f4 Nc4 10.dxc7 Qxc7
11.Nf3 0-0-0 12.Nbd2 Bb4 13.c3 and the database says “½-½” M.Gagunashvili-T.Sanikidze,
Georgian Ch., Tbilisi 2016, whereas in view of 13...Rhe8 White might well have resigned.
e) 5.Be2 (preparing e4-e5) can be met by 5...Qe7 (Stockfish goes for 5...Nb4!? 6.Na3 d5 7.exd5+
Be7 8.c4 0-0 with compensation) 6.Nc3 d5 (Schlechter); e.g. 7.Nxd5 (7.exd5 Nb4 8.Bd1 should
transpose) 7...Nxd5 8.exd5 Nb4 9.Bd1 Nxd5 10.Qxe7+ Bxe7 11.Nf3 Bf6 with a pull for Black whose
bishops will be more active, M.Aymard-M.Boulanger, corr. 2019.

5...Bb4

The most natural and logical move. Black eyes three targets on the e-file and plans to follow with
...0-0, ...Re8, and potentially ...d7-d5. In light of that, White’s next two moves are almost forced.

6.Bd2 0-0 7.0-0-0

Nothing else is worth considering: 7.f4, 7.Bd3, and 7.Qg3 all run into 7...d5!, while 7.Bc4 Bxc3!
returns to 5.Bc4 above.

7...Re8

24
 Black has played as planned and is now ready to take on c3 and e4. What can White do about that?
Nothing, since trying to defend the pawn allows Black to break powerfully with ...d7-d5. So the only
option is to offer it as a sacrifice. To that end there are two main continuations, the second of which
should deter Black from material gain.
B1: 8.Qg3
B2: 8.Bc4
Instead:
a) 8.Bd3?! is duly answered by 8...d5! 9.Qf4 Bg4 10.f3 Bd6 11.Qg5 h6 12.Qh4 dxe4 13.Bxe4
Nxe4, unklesubz-jatait47, Chess.com 2021, and even after 14.Qxd8 Raxd8 15.Nxe4 (not 15.fxg4?
Nxc3 16.bxc3 Ba3+ 17.Kb1 Rxd2) 15...Rxe4 16.fxe4 Bxd1 17.Kxd1 Re8, White will be a pawn
down for less than nothing.
b) 8.Nge2?! also fails to 8...d5! 9.Nxd5 Nxd5 10.Qf3 Bxd2+ 11.Rxd2 Ncb4 12.exd5 Nxa2+ 13.Kb1
Nb4 and Black is clearly better, ETC_Chess-Jpv, playchess.de 2003.
c) 8.f3?! meets yet again with 8...d5! 9.Qf2 (neither 9.Be1 Qe7! 10.Nxd5 Nxd5 11.Rxd5 Be6 –
L&O; nor 9.Qg5 h6 10.Qh4 d4 11.Nd5 Nxd5 12.Qxd8 Rxd8 13.exd5 Rxd5, S.Bunkov-A.Karayev,
Ekaterinburg 2008, improves for White) 9...dxe4 10.Bg5!? (or 10.Nxe4 Nxe4 11.fxe4 Qe7 – Keres;
e.g. 12.Nf3 Bxd2+ 13.Qxd2 Rd8 14.Bd3 Nb4 with advantage, C.Alonso Gonzalez-E.Gonzalez Alcon,
corr. 2008) 10...Bd7 11.Nd5? (11.Nge2 is necessary) 11...e3! 12.Qh4 (taking the pawn is no better:
12.Bxe3 Nxd5 13.Rxd5 Qe7 14.Ne2 Be6, or 12.Nxe3 Bc5 13.Qd2 Ne4!? 14.fxe4 Qxg5 – Bologan)
12...Bd2+ 13.Kb1, P.Grott-U.Vetter, corr. 1999, and now 13...Bf5! 14.Nxf6+ gxf6 15.Bxf6 Qd6
(L&O) is very good for Black; e.g. 16.Qh6 Qf8 17.Qg5+ Bg6 18.f4 Re6 19.Nf3 (not 19.f5? e2)
19...h6 20.Qh4 Bxc2+ 21.Kxc2 Qc5+ 22.Kb1 Qf5+ 23.Ka1 Qxf6.
d) 8.Qf4 offers the pawn with tempo to tempt Black to take with the rook. There’s no reason to
decline: 8...Bxc3! 9.Bxc3 Rxe4 10.Qg3 (not 10.Bxf6? Rxf4 11.Bxd8 Nxd8 and White is just a pawn

25
down; or if 10.Qd2 then 10...Re8 11.f3 d5 12.Ne2 Ne5 and ...c7-c6, or even 12...a5!? 13.Nf4 Nb4,
while after 11.Qf4 d6 12.Bxf6 Qxf6 13.Qxf6 gxf6 Black’s damaged kingside doesn’t negate the
majority on the other flank) 10...d5 11.f3 Re6! 12.Ne2 (or if 12.a3 Rd6 13.Bb5 d4 14.Bd2 as in
I.Nepomniachtchi-L.Aronian, Chess.com blitz 2020, then 14...Bf5 with a big advantage) 12...Rd6
(Bologan’s idea, threatening ...d5-d4) 13.Qf4 (after 13.Nd4 Nxd4 14.Bxd4 Bf5, White has no
particular compensation) 13...d4 14.Nxd4 Nd5 15.Nxc6 (or 15.Qe4 Nxc3 16.bxc3 Be6 and Black is
clearly better) 15...bxc6 16.Rxd5 cxd5 17.Be5 Re6 and Black won, I.Nepomniachtchi-V.Anand, St.
Louis (blitz) 2017.

B1: 8.Qg3

Frank Marshall’s idea, offering the e-pawn for... well, it’s not clear exactly what now. For the lols
perhaps.

8...Nxe4!

Simplest and strongest. 8...Rxe4!? is flashy, based on 9.Nxe4? Nxe4 and ...Bxd2+, but is far less
straightforward and may not offer as much anyway.

9.Nxe4 Rxe4

Okay, that’s the pawn. How is White going to justify the sacrifice? Clearly 10.Bh6 Qf6 and 10.Bd3
Bxd2+ are no good. Nor are 10.Bxb4 Nxb4 11.a3 Nc6 and 10.Bg5 Be7 a whole lot better, as White is
basically just a pawn down. That leaves:

10.Bf4

The only significant alternative is 10.c3, but this is well met by 10...Be7! (improving on 10...Bf8

26
from F.Marshall-H.Pillsbury, Buffalo 1901). Although the rook on e4 looks a little loose, it turns out
there’s no real cause for concern:

a) 11.Bd3 is negated by 11...Bh4! 12.Qf3 Re8, securing the rook and now threatening ...Ne5. White
has nothing for the pawn here and has lost virtually every game; e.g. 13.Qf4 (or 13.Qh5 g6 14.Qh6
Ne5 15.Be2 d6 16.h3 Bf6 17.Nf3 Bg7 18.Qf4 Be6 19.Kb1 Qf6, B.Nadj Hedjesi-Mis.Pap, Serbian
League 2011) 13...Ne5 14.Bc2 d5 15.Nf3 Nxf3 16.Qxf3 Be6 17.Be3 Qf6 with a clear advantage,
W.Belka-H.G.Koch, corr. 2011.
b) 11.f4 d5 12.Bd3 (or 12.Nf3 Bf5, when 13.Bd3 transposes) 12...Bf5! (offering up the exchange
for the light squares; after 12...Bh4 13.Qf3 Re8 14.g4 or 12...Ra4 13.Kb1 White has a bit of play for
the pawn) 13.Nf3 (or if 13.Bxe4 Bxe4 14.Ne2 Qd7 15.c4 Bg6 16.Qh3, A.Karagollu-C.Ali Marandi,
Turkish League 2016, then just 16...Qxh3 17.gxh3 dxc4 with a big advantage) 13...Qd7 14.Bxe4
Bxe4 15.Rhe1 (hoping for 15...Na5 16.c4! Qa4 17.Rxe4! dxe4 18.Bc3 with a defence; the seminal
game saw 15.Be3 Na5! 16.Bd4 Bf8 17.b3 Qb5 18.Kb2 c5 19.Be5 Nc4+ 20.Ka1 Ne3 21.Ng5 Qe2 and
Black soon won, J.Rudd-B.Avrukh, London Classic 2010; other moves are no better, e.g. 15.Ng5
Bxg5 16.fxg5 Ne5! 17.Qxe5 Qa4 and wins – Lokander; or 15.c4 Bf6) 15...Bg6! (keeping the bishop)
16.Qh3 Bf5 (and the queen) 17.Qg3 Bf6 18.Be3 Na5 (finally this move) 19.Bd4 Qa4 20.Re2 Qxa2
21.Kd2 Nb3+ 22.Ke1 Nxd4 23.Nxd4 Be4 and Black was back, now with the two bishops and two
pawns for the exchange, and went on to win in M.Michalek-S.Renard, corr. 2018.
c) 11.Nf3 (IM Andrew Greet’s hope in 2008 to revive the variation) 11...d6 12.h3 (neither 12.Bd3
Rg4 – Greet, nor 12.Bf4 Ra4 13.a3 Bf5 14.Bb5 Ra5, S.Desquiens-L.Nouveau, corr. 2012, is any
better for White) 12...Ne5! (the improvement) 13.Nxe5 (if 13.Ng5 Ra4 14.a3 Bf5 15.f4,
D.Langschmidt-A.Hahner, corr. 2013, then 15...Qd7! 16.fxe5 Rxa3! 17.bxa3 dxe5 18.Kb2 Qa4 wins
– Dhopade) 13...Rxe5 14.f4 Ra5! and Black is just a pawn up; e.g. 15.a3 Bd7 16.Bd3 Ba4 17.Bc2
Bxc2 18.Kxc2 d5 19.Qd3 Qd7, A.De Blasio-S.Diaz, corr. 2009, with ideas of ...Rd8, ...Ra6, ...c7-c5
and suchlike.

27
10...Qf6! 11.Nh3

Not 11.Bxc7? d6 12.Bxd6 Qh6+! and wins (Keres); i.e. 13.Kb1 Bxd6 and White can’t take due to
the weak back rank.

11...d6 12.Bd3

Black was threatening to win a piece with ...Bxh3; and 12.f3? Rd4 is hopeless for White.

12...Nd4!

Stronger than 12...Re8 13.Bg5!, to which Stockfish gives its “0.00”; e.g. 13...Qe6 (or 13...Qe5
14.Nf4) 14.a3 Ba5 15.b4! Bb6 (15...Ne5 16.bxa5 Qa2 is only a draw) 16.Rde1 Ne5 17.Kb1! Qd7
18.f4 Nxd3 19.Bf6! g6 20.Qxd3 and White has sufficient play for the pawn.

13.Be3

Everything else is very much worse:

a) 13.c3? Ne2+ 14.Bxe2 Rxe2 15.Qf3 Bxh3 wins (Pickett).
b) 13.Bg5? Qe6 14.c3? (or 14.Kb1 Rg4) 14...Qxa2! 15.cxd4 Rxd4! 16.Bxh7+ (or 16.Nf4 Bf5
17.Bxf5 Rc4+ and mates) 16...Kxh7 17.Rxd4 Qa1+ 18.Kc2 Bf5+ 19.Kb3 Be6+! (Vallejo Pons)
20.Kc2 Qa4+ 21.Kd3, Caissa-Goofy, playchess.de 2002, and now 21...Re8! leaves White helpless;
e.g. 22.Qh4+ Kg8 23.Be3 Qb5+ 24.Kc2 c5 and wins (L&O).
c) 13.Kb1? Ne2! 14.Qf3 (or 14.Bg5 Nxg3 15.Bxf6 Nxh1 16.Bxe4 Bxh3 17.Rxh1 Be6 – Dhopade)
14...Ba3! 15.bxa3 (not 15.c3? Nxc3+ – Korchnoi; and 15.Bc1 Qxf3 16.gxf3 Re7! is terrible too,
L.Freund-Y.Dothan, corr. 2014) 15...Nc3+ 16.Kc1 Nxa2+ 17.Kb1 Nc3+ 18.Kc1 Bxh3 19.Bxe4
Bxg2! (Black had to see this far) 20.Qxg2 Qxf4+ 21.Kb2 Nxe4 22.Rd4 Qf6 23.Qxe4 c5 24.Re1 h5

28
25.Qf4 Qxd4+ 26.Qxd4 cxd4 with an easily won rook endgame, I.Nepomniachtchi-F.Vallejo Pons,
Moscow 2007.

13...Nf5!

Not 13...Bg4? 14.Ng5 Rxe3 15.Qxg4 Ne2+ 16.Bxe2 Rxe2 17.Ne4 Rxe4 18.Qxe4 and White was
winning, J.Mieses-J.R.Capablanca, 2nd matchgame, Berlin 1913. Capablanca’s own improvement,
13...Rg4!? 14.Bxd4 Rxd4 15.c3 Bxc3 16.bxc3 Rg4 17.Qe3 Qxc3+ 18.Bc2 Qxe3+ 19.fxe3 Rxg2, is
not so clear after 20.Nf4 Rf2 21.Rd2 Rxd2 22.Kxd2, as it’s not that easy to consolidate and activate
the pawns. All the same, only Black can win.

14.Qf3 Re8 15.Bg5 Qe5

In this possibly critical position, White has two possibilities:

a) 16.c3 Bxc3 (forced, since Rhe1 was threatened) 17.bxc3 Qxc3+ 18.Bc2 (not 18.Kb1? Qb4+
19.Ka1 Nd4) 18...Qa1+ 19.Kd2 (and not 19.Bb1? Re5) 19...Qxa2, when “there is no doubt that in the
near future [White] will have to trade the queens and defend a rather unpleasant endgame.” (L&O);
e.g. 20.Rhe1 Bd7 21.Qd3 h6 22.Ra1 Qb2 23.Rab1 Qd4 24.Be3 Qxd3+ 25.Bxd3 b6 and Black is
clearly better, since the four pawns – as an unopposed queenside mass – are more dangerous than in
Capablanca’s line above.
b) 16.Bc4 is the only move to have been tried, after which 16...Be6 17.Bxe6 Qxe6 18.Qxb7
regained the pawn at the cost of conceding the initiative completely: 18...Rab8 19.Qd5 Qg6 20.g4
Ne7 21.Qc4 d5 22.Qd3 Qb6 23.b3 (or 23.Be3 c5 24.Ng5 g6 25.a3 Ba5 26.b4 Bxb4 27.axb4 Qxb4,
threatening ...c5-c4 and/or ...d5-d4; e.g. 28.Rhe1 c4 29.Qd4 Nc6 30.Qc5 Qb2+ 31.Kd2 d4 32.Bxd4
Nxd4 33.Rxe8+ Rxe8 34.Qxc4 Re7 35.Qc8+ Kg7 36.Qc3 Qxc3+ 37.Kxc3 Rc7+ and Black gets the
pawn back again) 23...c5 24.Bxe7 Rxe7 25.Qxd5 Qa6 26.Kb1 c4! 27.Nf4 (not 27.Qxc4?? Qxc4
28.bxc4 Bd2+ 29.Ka1 Bc3 mate) 27...Bc3 28.Qc5 Ree8 29.Nd5 cxb3 30.Nxc3 bxc2+ 31.Kxc2 Rec8

29
32.Qd4 h6! (taking time out since White can do nothing) 33.a4 Qe2+ 34.Rd2 Rxc3+! 35.Qxc3 Qe4+
36.Rd3 Qxh1 and Black won, Han.Wagner-B.Näter, corr. 2018.

B2: 8.Bc4

Probably best, in that it neutralizes the attack on e4. On the other hand, the bishop is exposed on c4
and Black may use it to generate a queenside initiative with moves like ...Ne5, ...a7-a5 and ...b7-b5.

8...d6

There is no advantage to be gained from taking the pawn here, except in one instance: 8...Bxc3
9.Bxc3 Nxe4 10.Qf4 Nf6 11.Nf3 d6 12.Ng5 Be6 13.Bd3 h6 14.h4?! (instead, 14.Nxe6 is better and
equal, either on this or the previous move) 14...Ne5! (improving on 14...Nd5, as in S.Winawer-
W.Steinitz, Nuremberg 1896, and others) 15.Bxe5 dxe5 16.Qxe5 Bd5 17.Qg3 Qe7 18.Kb1 Qe5
19.Qxe5 Rxe5 20.Nf3 Bxf3 21.gxf3 Rae8 does give Black an edge, W.Raasch-H.Buchmann, corr.
2005.

9.Nf3

Deterring ...Ne5 looks like the prudent course. Other options are worse:
a) 9.h3? makes the e-pawn a lot more appetizing: 9...Bxc3 10.Bxc3 Nxe4 11.Qf4 Qd7!; e.g. 12.Nf3
Nxc3 13.Rhe1 Rxe1 14.Rxe1 d5 and Black won, I.Elishev-J.Jirka, Bol 2010.
b) 9.Qg3 Ne5 10.Bd3 (forced, since 10.Bb3? Bxc3 and ...Nxe4 wins the pawn and removes the
dark-squared bishop as well) 10...Nxd3+ 11.cxd3 d5 is also good for Black.
c) 9.Nge2 obstructs the queen, which Black can exploit by 9...Ne5 10.Bb3 Bc5! 11.Qg3 (or 11.Qg5
Bxf2 12.Rhf1 h6 13.Qf4 Bc5) 11...Nh5! 12.Qg5 Qxg5 13.Bxg5 Bxf2 14.Rdf1 Bb6 with a safe extra

30
pawn, M.Biava-J.Morgado, corr. 1994.
d) 9.Nh3 Ne5 (or 9...Bxc3 10.Bxc3 Nxe4 11.Qf3 Re7 12.Bxf7+ Rxf7 13.Qxe4 Bxh3 14.gxh3 Qg5+
15.Kb1 Qf5 with a slight advantage based on structure, R.Hase-R.Servat, Santafesina 2004; not
11.Qf4? Nxc3! 12.Bxf7+ Kh8 13.bxc3 Rf8 or 12.Qxf7+ Kh8 13.bxc3 Ne5) 10.Bb3 a5 11.a3 (if 11.a4
then 11...Be6; or 11.Ng5 Bxc3 12.Bxc3 a4 13.Bd5, D.Itkin-I.Mahoney, Toronto 2010, and now
13...c6! 14.f4 cxd5 15.fxe5 dxe5 16.exd5 a3 17.b3 Bg4 – Bologan) 11...h6!? 12.Nd5 (after 12.axb4?!
axb4 13.Nb1 c5 Black has a strong attack, with ideas of ...Ra1 and ...Qa5, similar to the earlier 4.Qc4
line) 12...Bxd2+ 13.Rxd2 a4 14.Ba2 Nxd5 15.Bxd5 c6 16.Ba2 Be6 17.Bxe6 fxe6 with an edge for
Black in view of the strong centre.
e) 9.f3 Ne5! (preferable to 9...Na5, when the bishop can retreat down its own diagonal, leaving the
black knight misplaced and a hindrance to the queenside march)

10.Bb3 (now if 10.Be2 then 10...Bc5 11.Qg5 c6 prepares the charge with ...b7-b5; e.g. 12.h4 h6
13.Qg3 b5 14.Kb1, P.Groot-L.Hlavacek, corr. 2017, and perhaps 14...Qb6 15.Nh3 a5 16.h5 Kh8
17.f4 Nc4 18.Bxc4 bxc4 19.Ng5 Qc7 with ideas of ...Rb8 and ...a5-a4, as well as ...d6-d5; note that
15.Bxh6? fails to 15...Nh5 16.Qg5 f6 17.Qxh5 g6 and the queen is unexpectedly trapped) 10...a5!
(more accurate than 10...c6 11.g4 a5 12.g5 Nfd7 13.f4 Bxc3 14.Bxc3 a4, G.Cardelli-T.Koskela, corr.
2011, where Stockfish throws out 15.Be6! as a spoiler and claims equality)

31
 I think White is already in trouble here as Black’s attack comes much more swiftly, In particular,
...b7-b5 can often be played without any preparation. For instance:
e1) 11.g4 meets with 11...b5! at once, since 12.Nxb5 loses to 12...Bc5 13.Qc3 a4 14.Bc4 Nxc4
15.Qxc4 Ba6 etc, and 12.a4 bxa4 13.Bxa4 Nc4 or 13.Nxa4 c5 is also strong.
e2) 11.a4 c6 12.g4 b5 is very promising too; e.g. 13.h4 Be6 14.Bxe6 Rxe6 15.Nge2 Bc5 16.Qf4 b4
17.Nb1 b3 and Black was much faster, D.Padurean-M.Dumitru, Calimanesti 2018.
e3) 11.a3 Bc5 12.Qe1 b5! (without ...c7-c6) 13.Bg5 (if 13.Nxb5 a4 14.Ba2 c6 15.Nc3 Qb6 16.Bf4
then 16...d5! – Bologan, or 16.Nge2? Be6 17.Bxe6 Reb8! and wins, T.Puleston-V.Perevozchikov,
corr. 2008; while after 13.a4 bxa4 14.Nxa4 Bb4 15.Qg3 c5 16.Bg5 c4 17.Ba2 Bd7 18.Bxf6 Qxf6
19.Nb6 Bc5 White gave the game up as hopeless in U.Jansen-P.Larwinski, corr. 2017) 13...h6
14.Bxf6 (or 14.Bh4 Ng6 15.Bg3 a4 16.Ba2 b4 17.axb4 Bxb4 18.Rd4 c5 19.Rxd6 Qa5, threatening
20...a3 and wins, J.Rudd-An.Bird, Crawley 2013) 14...Qxf6 15.Nxb5 a4! 16.Nxc7 (or 16.Ba2 Rb8
17.Nxc7 Nxf3 etc) 16...axb3 17.Nxe8 Qg5+ 18.Kb1 (if 18.Rd2, as in J.Rudd-A.Balaji, 4NCL Online
2020, then 18...Bxa3! 19.bxa3 b2+ 20.Kb1 Rxa3 or 19.Nc7 Nc4 20.Nxa8 Bxb2+ is one way to finish)
18...bxc2+ 19.Kxc2 Qxg2+ 20.Ne2 Ba6 21.Nf6+!? Kh8 22.Rd2 Be3 and Black wins.

9...Be6

32
 Opposing the opposing bishop. Not now 9...Ne5?! 10.Nxe5 dxe5 (or 10...Rxe5 11.Rhe1) 11.Nb5
Bxd2+ 12.Rxd2 Qe7 13.f3 with an edge for White. Nor 9...Bxc3?! 10.Bxc3 Nxe4 11.Rhe1! Bf5
12.Qf4 Bg6 13.Bxf7+! Bxf7 14.Rxe4 Rxe4 15.Qxe4 and White has whatever chances are going.

10.Bxe6

If White wishes to minimize risk, then 10.Bd3!? may suffice; e.g. 10...Bc5 11.Qe2 Bg4 (or 11...Nd4
12.Nxd4 Bxd4 13.f3) 12.Be3 Bb4 13.Nd5, S.Porkolab-M.Murittu, corr. 2008, when 13...Nxd5
14.exd5 Ne5 15.h3 Bxf3 (or 15...Bh5 16.Be4) 16.gxf3 Bc5 17.h4 looks just about okay.

10...Rxe6 11.Ng5 Re8

Inserting 11...Re5?! 12.f4 Re8 is inferior because f2-f4 is a useful free move, preventing ...Ne5
again; e.g. 13.h4 Qd7 14.a3 a5 15.Qd3! eyes h7, so that 15...h6 16.Nd5 Bxd2+ 17.Rxd2 Nxd5?
18.exd5 wins for White, while 17...Kf8 18.Nxf6 gxf6 19.Nh3 is roughly equal.

33
 In this position Black looks to have the better chances without necessarily being better, if that
makes sense. The primary continuations are:
a) 12.f4?! is no good without the extra tempo; e.g. 12...h6 13.h4 Qc8! 14.Qf3 (or 14.Qd3 Qg4)
14...Nd4! (improving on 14...Kf8?!, S.Tartakower-S.Reshevsky, Stockholm Olympiad 1937) 15.Qd3
c5 16.Nf3 Bxc3 17.Bxc3 Nxe4 18.Bxd4 cxd4 19.Rhf1 Ng3 20.Rfe1 Qg4 21.Kb1 Qxf4 with a clear
advantage, H.Warzecha-N.Eremin, corr. 2013.
b) 12.f3 allows 12...Ne5 13.Qe2 (not 13.Nh3?! Nc4 14.Qd4 b5! 15.Bg5? c5! 16.Qd3 Qa5 17.Bxf6
gxf6 18.Nb1 Qxa2 and Black won, A.P.Smith-M.J.Turner, London Classic 2018) 13...h6 14.Nh3 c6
and White’s attacking dreams are over, so there is only defence; e.g. 15.Nf2 (after 15.Nf4 a5 16.h4
d5! 17.exd5 cxd5 18.Qb5 Rc8! 19.Ncxd5 Nxd5 20.Bxb4 Qc7! 21.Qxe8+ Rxe8 22.Nxd5 Qb8 23.Bxa5
Nc6 Black was clearly better with ideas of ...Re2 and ...Qg3, S.Renard-M.Boulanger, corr. 2018)
15...Bc5 16.Nd3 Bd4 17.Nxe5 Bxe5 18.Qd3 Qe7 19.Kb1 Rad8 20.Bc1 b5 21.g4 (looking for
counterplay) 21...b4 22.Ne2 d5! 23.f4 dxe4 24.Qc4 Bc7 25.g5 hxg5 26.fxg5 Nd5 27.Nd4 Nb6 28.Qb3
Rd5 29.Nxc6 Qd7 30.Rxd5 Nxd5 31.Nxb4 Nxb4 32.Qxb4 e3 33.Re1 Bb6 and I had to work very
hard to draw, tsmenace-lysol, ChessWorld.net 2019.
c) 12.a3!, asking the question, appears to be the best option and may be equal if White can defend
like an engine; e.g. 12...Ba5 (or 12...a5 13.axb4! axb4 14.Nb1 h6 15.Nf3 Nxe4 16.Rhe1 d5 17.Qb3!
Qd6 18.Be3 Red8 19.Nfd2 Nf6 20.g3! Na5 21.Qd3 c5 22.Qb5! d4 23.Bf4 Qc6 24.Qxc6 Nxc6, and
now Black had to show its capabilities in rofChade 2-Stoofvlees II, TCEC 18 Premier 2020) 13.Qg3
h6 14.h4 Rb8 15.Kb1 b5 16.b4! Bb6 17.Nd5 Nxd5 18.exd5 Ne5 19.Bc3 Qd7 20.Rde1! Qf5 21.Ne4
and White equalized, Stoofvlees II-rofChade 2, TCEC 18 Premier 2020. Both those games were
drawn.
Sorry, that was a lot of proper theory, wasn’t it – and none of it at all disreputable from Black’s
point of view. But from that point of view, if White really wants to play the Centre Game, I’m more
than happy to let them.

34
 Chapter Two
Calabrese Counter-Gambit

1.e4 e5 2.Bc4 f5!?

So Black is effectively playing a King’s Gambit a move behind. Can this be even slightly sound?
We’ll see. On the plus side, there’s no need to worry about the gambit being Accepted (if 3.exf5 Nf6,
then ...d7-d5 will hit the bishop), while the Declined (2.f4 Bc5) is not the most testing response to the
KG, so White’s extra tempo may not matter all that much. Perhaps.
Incidentally, the opening is named after the 17th century master, Gioacchino Greco (“Il
Calabrese”), who included a game in his manuscripts (see the notes to line B).
A: 3.exf5
B: 3.Bxg8
C: 3.Nc3
D: 3.Nf3
E: 3.d4
F: 3.d3
A few rarer options:
a) 3.f4!? exf4 is a bizarre route to the Nordic Counter-Gambit (see Chapter Four). Alternatively,
3...Nc6 was considered by GM Miles (via 2.f4 Nc6 3.Bc4 f5!?), who noted briefly 4.fxe5 (4.exf5? d5;
4.Bxg8 Rxg8) 4...Nxe5, all of which are good for Black. Hence White should play 4.d3, when 4...Nf6
is line F1 below.
b) 3.Qh5+?! g6 4.Qe2 is a reversed KG Keene (2.f4 Qh4+), which the extra bishop move does not

35
improve: 4...fxe4 5.Qxe4 (the standard 5.d3 idea is no good because of 5...d5) 5...Nc6 6.Bxg8 Rxg8
7.Qd5 Rg7 and Black is already better, D.Bennion-T.Kett, Welsh Ch., Swansea 2003.
c) 3.Qf3!? is a reversed KG Nordwalde (2.f4 Qf6): then the thematic 3...Nf6 (for 3...Nc6!? see
2.Qf3 in Chapter One) 4.Qxf5!? (if 4.d3 Nc6 5.Ne2, as in Panza-tsmenace, ChessWorld.net 2014, I
might play 5...f4!?, given a second chance) 4...Nc6 5.Qg5 (not 5.Bb5? Bc5! and White is playing the
Nordwalde main line which is unsound) 5...Nxe4 6.Qxd8+ Kxd8 7.Nc3 is about equal.
d) 3.Nh3!? Nf6 (not 3...fxe4?? 4.Qh5+) 4.Ng5 (for 4.d4 exd4 see line E2) 4...Nxe4 5.Bf7+ (or
5.Nf7 Qh4 6.0-0 Nd6! 7.Nxh8 Nxc4) 5...Ke7 6.d3 Nf6 7.Bb3 d5 8.c4 h6 9.Nf3 e4 10.Ne5 Qe8
11.cxd5 Kd8 12.d4 Nbd7 was quite unclear in Langbein-Honeybunch, ChessWorld.net 2016.

A: 3.exf5

The Gambit Accepted! Since Morphy once played this way (albeit only against his dad), we
shouldn’t dismiss it too casually.

3...Nf6

4.g4

Aiming to keep the pawn in proper KGA fashion. Others:

a) 4.d4 exd4 5.Qxd4 (or 5.Nf3 d5 6.Bd3 c5) 5...d5 6.Bd3 c5 offers perfectly good compensation for
the pawn; e.g. 7.Qe5+? (or 7.Qe3+ Be7 8.c3 0-0) 7...Be7 8.c4 Nc6 9.Qf4 Bd6 10.Qg5? Qe7+ 11.Qe3
Ne5 12.Be2 d4 13.Qb3 Bxf5 14.Bg5 0-0 and Black soon won, C.Mayet-W.Hanstein, Berlin 1837.
b) 4.Nf3 d5 (or 4...e4 first) 5.Bb3 e4 6.Nh4?, as in B.Ferrandi-M.Lyell, Calvi 2005, runs into 6...g5!
7.fxg6 Ng4 8.g3 hxg6 and White is in huge trouble; e.g. 9.h3 Rxh4! 10.gxh4 Ne5 11.d4 Nf3+ 12.Ke2

36
Nc6 13.c3 b6 and so on.
c) 4.Be2!? is intriguing, making it a true reversed KG Cunningham. Still, Black can hardly be
worse; rather, after 4...Bc5, it’s White who has to prove equality.
d) 4.Nc3! is probably critical and is best answered by 4...Nc6, transposing to the Deferred version
(see Chapter Three).

Not 4...d5? 5.Nxd5!, P.Morphy-A.Morphy, New Orleans 1848, and White is already winning in
view of 5...Nxd5 6.Qh5+ (which even a young Paul Keres fell for); 4...c6?! 5.d4 exd4 6.Qxd4 d5
7.Bd3 also offers White an edge.

4...d5 5.Bb3 h5

Alternatively, 5...Bc5!? 6.Nc3 (not 6.d3? h5 7.Be3 Bxe3 8.fxe3 Nxg4 9.Qf3 Qh4+ 10.Ke2 Rf8 and
Black was better, De Rives-T.Von der Lasa, Brussels 1860; while 6.g5 Bxf5 7.gxf6 Qxf6 looks like a
very promising reversed Muzio, e.g. 8.Bxd5?? Be4 and wins) 6...h5 (not 6...c6?! 7.h3 h5 8.Qf3) 7.g5
Ng4 8.d4 Bxd4 9.f6 gxf6 10.Nxd5 is unclear, except to Stockfish, which says “0.00”.

6.g5 Ne4

6...Ng8!? is possible too.

37
 We’ve now reached a strange sort of reversed Kieseritzky. White should continue 7.Nc3 (both
7.d3?! Nc5! and 7.Nf3 Bxf5 8.Nxe5?! Nd7 are good for Black) 7...Bc5 (or just 7...Bxf5 8.Bxd5 Nxc3
9.dxc3 Nc6) 8.d4 exd4 (not 8...Bxd4? 9.Nxe4 dxe4 10.c3) 9.Nxd5 Bxf5 and Stockfish says “0.00”
again; e.g. 10.Qe2 Nc6 11.Bf4 d3 12.cxd3 Bxf2+ 13.Kf1 Nd4 14.Nxc7+ Ke7 15.Nd5+ Ke8 with a
draw.

B: 3.Bxg8 Rxg8

Given the issues with 3.exf5 Nf6 and ...d7-d5, swapping the bishop off first is tempting, especially
as a further Qh5+ will then win a pawn. Some of the earliest games went this way, including Greco’s
own. The cost is the strong light-squared bishop and – if White does go for the pawn – the time spent,
which allows Black to develop a powerful initiative.

38
4.exf5?!

It turns out that the exchange on g8 doesn’t strengthen this move at all. Instead:
a) 4.Nc3! (best) is again best met by 4...Nc6! with a Calabrese Deferred.
b) 4.d4!? exd4 is covered in the notes to line E. Black might consider 4...d5!? as well; whereas
4...Qh4?! 5.Nc3 Bb4, F.Accardo-M.Bianca, Catania 1991, and 6.Ne2! is good for White.
c) 4.Qh5+ g6 5.Qxh7 Rg7 6.Qh8 (hoping for something with Bc1-h6; anything else and White is
just worse) 6...Qg5 (or 6...Nc6!, N.Ralphs-P.Kendall, Welsh League 2007, when White has to try
7.d4!?; e.g. 7...fxe4 8.d5 Nb4 9.Bh6 Rf7 10.Bxf8 Rxf8 11.Qxe5+ Qe7 12.Qxe7+ Kxe7 13.Na3 Nxd5
14.0-0-0 c6 15.f3 with rough equality) 7.Qh3?! (7.Ne2 Qxg2 8.Rg1 Qf3 9.d4 keeps the game level)
7...fxe4 8.Nc3 Qf5 9.Qe3?! Rf7 (9...Nc6! is stronger, intending 10.Nxe4 Nd4 or 10.Qxe4 d5!)
10.Nh3? (White’s last chance was 10.Nxe4 d5 11.Ng5 – Larsen) 10...d5 11.Nxd5 Nc6 12.c3? Be6
13.c4 Nd4 14.Qc3 Qg4 15.0-0 Ne2+ and 0-1 in Greco’s original game.

4...d5!

There’s no need to fear the check.

5.Qh5+?

Having said “A”. Well, no. White should prefer 5.d4 and try to draw. Note that 5.g4?, as in T.Von
der Lasa-P.Von Bilguer, Berlin 1839, gets smashed up straight away by 5...g6!.

5...g6 6.fxg6 Rxg6

39
7.Ne2

Grabbing another pawn is no better: 7.Qxh7 (not 7.Qxe5+?? Re6) 7...Qf6 8.Qxc7?! (but 8.g3 Nc6
and 8.Nc3 Be6 are terrible for White anyway) 8...Nc6 9.Qh7 Nd4 10.Kd1 Bg4+ 11.f3 Nxf3 and Black
soon won, Pentz-G.Maróczy, Nagyteteny 1901.

7...Nc6 8.0-0?! Bg4 9.Qxh7 Rh6 10.Qd3 e4 11.Qe3 Qh4 12.h3 Bxe2 13.Qxe2 Nd4 14.Qd1 Nf3+!
15.Kh1 Qxh3+!?

Here 15...Qg4! and ...Rxh3+ (Lane) wins more quickly.

16.gxh3 Rxh3+ 17.Kg2 Rh2+ 18.Kg3 Bd6+ 19.Kg4 Rh4+

My main reason for showing this game in full is because of a Stockfish intervention: 19...Rg2+!
20.Kh3 Kf7! (just giving the rook away for a faster win) 21.Kxg2 Rg8+ 22.Kh3 Rh8+ 23.Kg4 Rh4+
24.Kf5 Rf4 mate. A splendid variation.

20.Kf5 Rh5+ 21.Kg6 Rg5+ 22.Kh6 Bf8+ 23.Kh7 Kf7 24.Rh1 Bg7

And ...Rh8 mate follows, hence 0-1 B.Fisher-W.Steinitz, London (simul) 1872.

C: 3.Nc3

A useful developing and semi-waiting move. If Black responds in the most obvious way, we can
end up in a tricky Latvian sideline.

3...Nf6

40
 I prefer 3...Nc6, going into the Deferred Calabrese again.

4.Nf3!

And this is it. We’re suddenly defending Mlotkowski’s variation (2.Nf3 f5 3.Nc3 with 3...Nf6
4.Bc4). Instead, 4.exf5 was mentioned in line A, while 4.d3 transposes to 4.Nc3 in the notes to line F.

4...fxe4

Black has nothing better:

a) 4...Nc6 5.d4! exd4 (or 5...Bb4 6.dxe5! Nxe4 7.0-0 – Kosten) 6.Nxd4 fxe4 (or 6...Nxd4?! 7.Qxd4)
7.Bg5 Bb4, H.Szudra-J.Olschok, Hamburg 2005, and now 8.0-0 Bxc3 9.bxc3 Nxd4 10.Qxd4 c6 11.f3
d5 12.Bb3 is good for White.
b) 4...Bb4 5.d4! (analysed by CraigEvans on the ChessPublishing forum) 5...fxe4 (or 5...exd4
6.Nxd4 fxe4 7.Bg5) 6.dxe5 exf3 7.exf6 Qxf6 8.0-0 Kd8?? (8...Bxc3 9.bxc3 was necessary) 9.Nd5 1-0
tsmenace-Raffzahn, ChessWorld.net 2015.
c) 4...Nxe4 5.Nxe5 Qe7 6.Bf7+ Kd8 7.Nd5! Qxe5 8.d4 Qd6 9.Bf4 Qc6 10.Bxc7+ Qxc7 11.Nxc7
Kxc7, G.Niemand-F.Destrebecq, corr. 1999, is “very promising for White” (Kosten), who might play
12.0-0 again.

5.Nxe5 d5 6.Nxd5!

The same trick as after 4.Nc3 in line A, although it’s not decisive this time.

6...Nxd5 7.Qh5+ g6 8.Nxg6 hxg6

8...Nf6? 9.Qe5+ Be7 10.Nxh8 is no good.

41
9.Qxg6+

If 9.Qxh8?!, Black is quite okay. For instance, Stockfish analyses 9...Qg5!? 10.0-0 (or 10.g3 Bg4)
10...Nf4 11.g3 Qf5 12.f3 (not 12.gxf4? Qg4+ 13.Kh1 Qf3+ 14.Kg1 Bh3) 12...b5! 13.fxe4 Nh3+
14.Kg2 Qxe4+ 15.Rf3 Nf4+! 16.Kf2 (and not 16.gxf4? bxc4) 16...Nh3+ 17.Kg2 with a draw.

9...Kd7 10.Bxd5

Not 10.Qf5+? Kc6 11.Bxd5+ Qxd5 12.Qf6+ Be6 13.Qxh8 Na6 and Black is much better,
G.Niemand-J.Nilsson, corr. 1991.

Here, with soon-to-be four pawns (including three connected passed pawns) for the piece, White
can play for a win at no risk whatsoever, and Stockfish duly proclaims a clear advantage. On the other
hand, making real progress is by no means straightforward. I’ve tried and failed more than once.
For example: 10...Qe8 (instead, 10...c6 and 10...Qe7 have also been played) 11.Bxe4 Qxg6
12.Bxg6 Nc6 13.c3 Ke6 14.d4 Kf6 15.Bd3 Bf5 16.Bxf5 Kxf5 17.h3 Re8+ 18.Kd2 Bd6 19.Kd3 Reg8
20.g4+ Kf6 21.Be3 Ne7 22.c4 Ng6 23.Raf1 Nf4+ 24.Bxf4 Bxf4 25.Ke4 Bd6 26.f4 Re8+ 27.Kf3 c6
and Stockfish says the position is more or less equal now, tsmenace-MarcShaw, ChessWorld.net
2012.

D: 3.Nf3

42
 Transposing into a Latvian is a critical continuation since 2.Nf3 f5 is too disreputable even for me.
Fortunately, 3.Bc4 is one variation where Black seems to be okay. But this can only be demonstrated
via tactics, so there’ll be a lot of moves below.

3...fxe4!

Nothing else is worth considering. In particular, 3...d6?! 4.d4 exd4 (4...fxe4 5.Nxe5 d5 6.Qh5+ just
gives White an extra d2-d4 on the main line) 5.Ng5 Nh6 6.0-0 is virtually a forced win for White; e.g.
6...Nc6 7.exf5 Bxf5 8.Re1+ Kd7 9.Be6+ Bxe6 10.Nxe6 Qh4 (Steinitz) 11.Bxh6! gxh6 12.Qf3 Be7
13.Qf5 Ne5 14.Nxd4+ Kd8 15.Ne6+ Kc8 16.Rxe5 dxe5 17.Nc5+ with an unusual epaulette
smothered mate to follow, V.Zemitis-T.Svendsen, corr. 1998.

4.Nxe5

43
4...d5!

Svedenborg’s variation. It is unlikely that the bonkers 4...Qg5?! 5.d4 Qxg2 6.Qh5+ g6 7.Bf7+ Kd8
8.Bxg6! Qxh1+ 9.Ke2 will ever be repaired for Black, given the huge amount of work that has gone
into it already.

5.Qh5+

Inserting 5.Bb5+ c6 doesn’t improve White’s chances: 6.Qh5+ g6 7.Nxg6 hxg6 8.Qxh8 cxb5!
(better than 8...Kf7 9.Be2 Bg7 10.Qh7 Qg5 11.g3 Nf6 12.Qh4, as in H.Tompson-J.Morgado, corr.
1974) 9.Qxg8 Qg5 10.Rg1 (or 10.g3 Nc6) 10...b4 and Black has quite enough for the exchange.

5...g6 6.Nxg6 hxg6

Black really has to sacrifice here. After 6...Nf6?! 7.Qe5+ Be7 8.Bb5+! c6 (or 8...Bd7!? 9.Nxe7
Qxe7 10.Bxd7+ Kxd7 11.Qxe7+ Kxe7 12.b3, MarcShaw-tsmenace, ChessWorld.net 2013) 9.Nxe7
Qxe7 10.Qxe7+ Kxe7 11.Bf1! c5 (or 11...Rg8 12.d4 exd3 13.cxd3, G.Gunderam-J.Hempel, corr.
1970) 12.d3 exd3 13.Bxd3 Rg8 14.0-0 Bh3 15.Re1+ Kf7, G.Niemand-I.Budovskis, corr. 1998, and
now 16.Bf1 again, Black doesn’t have anything much for the pawn.

44
7.Qxh8

White cannot hope for any advantage by refusing the rook with 7.Qxg6+.
a) 7...Kd7 is the usual move, but there’s a whole load of theory on it. One critical line runs 8.Bxd5
Nf6 (8...Qf6!? 9.Bxe4 Qxg6 10.Bxg6, E.T’Jong-I.Cirulis, corr. 2007, is not unthinkable as White’s
four pawns are not easy to advance; c.f. the notes to line C) 9.Nc3 Qe7 10.d3! exd3+ 11.Be3 c6
12.Bb3 Bh6 13.0-0! Bxe3 14.fxe3 Rf8 (14...Qxe3+ 15.Kh1 Nh5 16.Ne4 looks more dangerous)
15.cxd3 Kc7 16.Qg3+ Qd6 17.Qxd6+ Kxd6 18.e4 and White had the somewhat better chances,
though Black held on in Jü.Mack-F.Borrmann, corr. 2011.
b) 7...Ke7!? may be simpler, if Black is content with a draw:
b1) 8.d4 sees Stockfish throw out 8...e3! (not yet 8...Qd6?, since 9.Bg5+ Kd7 10.Qf5+ Kc6 11.Qxc8
dxc4, R.Schwibbe-E.Szchorn, corr. 1970, and 12.Qe8+! Kb6 13.Nd2 wins for White – Kosten)
9.Bxe3 Qd6, and then 10.Bg5+ (10.Qg5+ Nf6 is good for Black) 10...Kd7 11.Qf7+ (the point of
jettisoning the e-pawn earlier is that 11.Qf5+ Qe6+! now forces the queens off) 11...Ne7 12.Be2 Qe6
(12...Nbc6 13.Nc3 Nd8 is another option) 13.Qxe6+ (or 13.Qf4 Bh6) 13...Kxe6, assessing the
position as “0.00”.
b2) 8.d3 Bh6! (here 8...e3? 9.Bxe3 Qd6 is ineffective due to 10.Qg5+ Nf6 11.Nc3! and Black can’t
take the bishop because of 11...dxc4?? 12.Bc5, winning the queen) 9.Bg5+ (9.Bxh6! Nxh6 10.Qg5+
Ke6 11.Qg6+ Ke7 is a draw; clearly 11...Qf6?! 12.Bxd5+ Ke7 13.Qxf6+ Kxf6 14.dxe4 doesn’t give
Black any winning chances) 9...Bxg5 10.Qxg5+ Ke8 11.Qxd5 (not 11.Qe5+? Qe7 12.Qxh8? exd3+
13.Kd2 Qe2+ and Black wins – Kosten) 11...Qxd5 12.Bxd5 exd3 13.cxd3 Nf6 and Black certainly
shouldn’t be worse with a knight for three pawns, A.Radoszta-R.Mayo, Buenos Aires 1992.

7...Kf7

45
8.Qd4

Retreating before ...Bg7 confines her to the kingside. Either way, Black will gain time against the
queen in order to develop sufficient counterplay.
a) 8.Be2 Bg7 9.Qh7 Nc6 (9...Qg5!? is also good and led to a draw in LcZero-Stockfish, TCEC 20
Superfinal 2021)

a1) 10.d3? Nd4 11.Bd1 exd3 12.cxd3 Qe7+ 13.Be3 Nf6 14.Qh4 Nf5 15.Qf4 Bh6 16.Qf3 d4 and
Black won, F.Borrmann-V.Kozlov, corr. 1979.
a2) 10.f3 Nd4 11.fxe4, S.Szilagyi-F.Pescod, corr. 1976; and now Kosten gives 11...Nf6 12.Rf1 Bf5,

46
when 13.Qh4 Qe7 14.Nc3 Rh8 offers Black good play.
a3) 10.c3 prevents ...Nd4 at the cost of leaving White weak on the light squares; after 10...Qg5
11.g3 Ne5 12.0-0 Nf6 13.Qh4 Qxh4 14.gxh4 Bh3 15.Re1 Re8 (preventing d2-d4), Black has
excellent play for the exchange.
a4) 10.0-0 Nd4 11.Nc3 (or 11.Bd1 Nf3+! 12.Bxf3 exf3 13.d4 Bg4 14.Nd2 fxg2 15.Kxg2,
D.Kraszewski-P.Mulleady, corr. 1995, when 15...Qf6! is very good for Black; e.g. 16.f3 Bh5 17.Ne4
dxe4 18.fxe4 Qxf1+ 19.Kxf1 Nf6 etc) 11...Nf6 12.Qh4 Nf5 13.Qf4 Bh6 14.Qe5 c6 (threatening
...Nd7) 15.h4 (or similarly 15.h3 Bf8) 15...Bf8 and here, for no obvious reason, White resigned in
Krambambuli-tsmenace, ChessWorld.net 2016. Otherwise the game might have continued 16.Qf4 (or
16.Qh2 Bd6 17.Qh1!? Qb6) 16...Bd6 17.Qg5 Qh8 18.d3 Nh7 19.Qd2 Qe5 20.g3 e3 21.Qe1 exf2+
22.Qxf2 Bc5! 23.Qxc5 Qxg3+ and draws.
b) 8.Bb3 Bg7 9.Qh7 Be6 (threatening ...Nd7-f8)

b1) 10.0-0?! Nd7 11.f3? (11.Bxd5 Bxd5 12.Qh3 was necessary) 11...e3! 12.Bxd5 Bxd5 13.Qh3 e2!
14.Re1, D.Rosner-J.Malmström, corr. 2001, and now 14...Bd4+! 15.Kh1 Be6 16.Qg3 Ne7 17.Rxe2
Nf5 wins by force.
b2) 10.Nc3 Nd7 11.Nxd5 (not 11.Bxd5?! Bxd5 12.Nxd5? Nh6 and Black wins) 11...Nf8 12.Nxc7
Nxh7 13.Nxe6 Qb8 and White’s assortment of material doesn’t quite add up to the queen,
M.Michalek-M.Heilmann, corr. 2013.
b3) 10.f4 Nh6! (not yet 10...Nd7? due to 11.f5! gxf5 12.Qh5+) 11.Nc3? (11.f5 is still correct, even
if 11...Nxf5 now gives Black full compensation) 11...Nd7 12.Nxe4 Qe7 13.Ng5+ Kf6 14.0-0 Bg8 and
Black won, M.Sandford-P.Corbat, corr. 2001.
b4) 10.f3 e3! 11.dxe3 (if 11.g4? Nh6 12.h4 Qd6, J.Gierke-P.Leisebein, corr. 1984, White is again in
serious trouble, even after 13.f4 Nd7 14.h5 gxh5) 11...Nd7 12.Bxd5! (the only way to extricate the
queen) 12...Bxd5 13.Qh3 Bc4 (or 13...Be6 14.Qg3 Be5) 14.Nd2 Ba6 15.Ne4 Qe7 gives Black

47
promising play for the slight material deficit, J.Velasco Blasco-R.Guerrero Sanmarti, Catalunya 1997.
b5) 10.d3 Nd7 11.g3?! (instead, 11.g4 exd3 12.0-0 Ngf6 13.Qh3 Nxg4 is unclear) 11...Ne5 (or
11...c5!?) 12.Nd2 was MarcShaw-tsmenace, ChessWorld.net 2014, and here Stockfish gives 12...Nf6
13.Qh4 Qd6! 14.Qf4 c5 as good for Black.

8...Nf6!?

An interesting and little-seen replacement for 8...Be6, which is nearly always played. One line then
runs 9.Bb3 (the primary alternative is 9.Be2, when 9...Nc6 10.Qe3 Bh6 11.f4 d4! or 11.Qg3 Nd4
gives Black good play; and 9.Bb5?! a6 10.Qa4 Qg5 11.Bf1 Nc6 12.h3?! b5 13.Qb3 Nd4 14.Qc3 c5
was already difficult for White in juliangon-tsmenace, ChessWorld.net 2013) 9...c5! (better than
9...Nc6 10.Qe3 Bh6 11.f4 Nge7 12.Qf2! and White slowly consolidated in juliangon-Reprimand,
ChessWorld.net 2015) 10.Qe3 c4 11.Ba4 (or 11.Bxc4 dxc4 12.Nc3 Nf6 13.Nxe4 Nxe4 14.Qxe4 Qd5)
11...Na6!? 12.b3, A.Holzhauser-J.Malmström, corr. 2001, and now 12...Nc5! 13.Ba3 (or 13.Bb2 Nh6
and ...Nf5) 13...d4 14.Qf4+ Nf6 15.Bb5 Bd6 looks fine for Black.

48
9.Be2

As with 8...Be6, White has a choice about where to put the bishop:
a) 9.Bb5 a6 10.Qa4? (10.Be2 is correct, meaning ...a7-a6 was for free) 10...axb5! (sacrificing
another exchange) 11.Qxa8 Nc6 and White’s extra material is meaningless with everything out of
play; e.g. 12.a4 Nd4 13.Na3 b4 14.c3 Nh5! 15.cxd4 Nf4 wins for Black.
b) 9.Bb3?! c5 10.Qe3 (or 10.Qe5? Nc6 11.Qg5 c4 12.Ba4 b5! 13.Bxb5 Nd4 14.Ba4 Nf5 and ...Bh6
traps the queen) 10...c4 11.Ba4 a6 12.c3 Ng4 sees Black attacking straight away; e.g. 13.Qe2 Bc5
14.h3 Qh4 15.Rf1 b5 16.Bc2 Nc6, followed by ...Nce5 and so on.
c) 9.Nc3 Be6 10.Bb3 (or 10.Bb5?! a6 11.Qa4 c6!) 10...c5! 11.Qe5 (or 11.Qe3 c4) 11...c4 12.Ba4 a6
13.d3 b5 14.dxc4 Nbd7 15.Qf4 bxa4 16.cxd5 Bf5 and Black is at least equal.

9...Nc6 10.Qe3

Or 10.Qc3 d4 11.Qb3+ (11.Qg3 transposes to the next note) 11...Be6 12.Bc4 Nd5, intending ...Na5.

10...d4

49
11.Qg5

After 11.Qg3 Bf5 12.0-0 (or 12.d3 Bd6 13.Bf4 exd3 14.cxd3 Qe7) 12...Bd6 13.Qh4, Stockfish
suggests 13...g5 14.Qh6 Qh8! 15.Qxh8 Rxh8 and Black’s powerful initiative continues without the
queens.
Instead, one of the three games in the databases to reach this position saw 11.Bc4+ Kg7 12.Qg5
Qe8! 13.0-0, H.Begjanian-Al.Gusev, Bashanta 2007, and now Black might play 13...Ng4!, followed
by ...Be7; e.g. 14.d3 (similarly 14.h3 Be7) 14...Be7 15.Qd2 (here 15.Qd5 Nf6 16.Qg5 Ng4 17.Qd5
Nf6 is a draw by repetition, and 15.Qb5 a6 doesn’t change anything) 15...Bd6 16.h3 (not 16.f4? Qh8
17.g3 Ne3) 16...Qe5 (or 16...e3!?) 17.Qf4 Qxf4 18.Bxf4 Bxf4 19.hxg4 Bxg4 seems fine for Black
with ...Rh8 coming next.

11...d3 12.cxd3

12.Bd1 Nd4 13.cxd3 exd3 transposes. Black also the option of 12...Nb4!?, as in Al.Holzhauser-
He.Beutel, corr. 2001, which continued 13.c3?! (White should prefer 13.Na3 Be7 14.c3 Kg7,
intending 15.cxb4 Nd5 16.Qe5+ Bf6 17.Qg3 Bh4 with perpetual on the queen) 13...Nc2+ 14.Bxc2
dxc2 15.Na3 Qd3 16.Qe3 Nd5 17.Qxd3 exd3 and the intrusive clamp on the light squares provides
more than enough compensation for the material.

12...exd3 13.Bd1 Nd4

50
 White’s negative development and exposed queen gives Black excellent compensation for the
exchange. Note that the d4-knight can go both ways. For example:
a) 14.Na3? Nf5! 15.Qf4 (or 15.Bb3+ Kg7 16.0-0 Qd4 17.Nb5 Qe5) 15...Qe7+ 16.Kf1 Be6 and
White is unlikely to survive.
b) 14.0-0?! is again met by 14...Nf5! (threatening ...Bh6) 15.Bb3+ (or 15.Qf4 Bd6 16.Qf3 Nd4
17.Qxd3 Bf5 18.Qc4+ Be6 19.Qd3 Nb3!, intending 20.Bxb3 Bxh2+ or 20.Bc2 Qh8!) 15...Kg7
16.Qf4? (but 16.Nc3 Qd4 17.g3 Bd7 is good for Black anyway) 16...Bd6 17.Qc4 b5! 18.Qf7+ (the
classic sacrifice 18.Qxb5 Bxh2+! 19.Kxh2 Ng4+ 20.Kg1 Qh4 is the point) 18...Kh6 is pretty much
decisive; e.g. 19.Re1 (or 19.Bd1 Bb7) 19...Ng4 (19...Bxh2+ works too) 20.g3 Bb7 21.Be6 Qe8! and
so forth.
c) 14.Nc3 Qe8+ (not now 14...Nf5? 15.Ne4) 15.Kf1 (not 15.Qe3 Bf5 16.Qxe8+? Rxe8+ 17.Kf1
Nc2, and neither 16.Ba4? b5 nor 16.Bb3+? Kg7 helps White at all; therefore 16.0-0 Qd7 and
Stockfish says “0.00”, though I doubt whether a human player would succeed in defending this)
15...Nc2 16.Bxc2 dxc2, followed by ...Bf5 with good play on the light squares; e.g. 17.f3 Bf5 18.d4
Qc6 19.Kf2 Rd8 20.Ne2 Re8 21.Re1 Qc4 22.g4 Be7 23.Qd2 Bxg4 24.Qc3 Be6 25.Bd2 Rh8 26.h4
Qxc3 27.bxc3 Rxh4 led to a draw in Stockfish-Komodo, TCEC 13 Superfinal 2018, where Black was
never in any difficulty.

E: 3.d4

51
 The most aggressive response – regarding ...f7-f5 as merely weakening, White rushes to open
things up. Although 3.d4 is only the fifth most popular move statistically, three of my stronger
opponents reacted this way over the board (all following up with 4.e5), so it’s an important line to
examine.

3...exd4

Pretty much forced. Obviously 3...fxe4?? loses at once to 4.Qh5+.

E1: 4.e5
E2: 4.Nh3
Others:
a) 4.Nf3 is duly met by 4...fxe4, when 5.Ne5? (5.Nxd4 Nf6 6.Nc3 c6 7.Nxe4 Qe7 leads to equality;
e.g. 8.0-0 Nxe4 9.Re1 d5 10.f3 Qe5 11.Bb3 Bg4!?) 5...d5 (this is essentially line D with White having
given the d-pawn away) 6.Qh5+ g6 7.Nxg6 Nf6! (stronger than 7...hxg6 here) 8.Qe5+ Be7 9.Bb5+ c6
10.Nxh8 (after 10.Nxe7 Qxe7 11.Qxe7+ Kxe7, White is the one down a pawn) 10...cxb5 11.Qg3 Be6
gave Black a superb position, Lei.Williams-T.Kett, 4NCL 2003.
b) 4.exf5 Nf6 is a reversed KGA Modern with an extra Bc4 for White, which is less than useful
because Black’s ...d7-d5 now comes with tempo: 5.Nf3 d5 6.Bd3 c5 and Black is better, while 5.c3
d5 6.Bd3 Nc6 7.Ne2 dxc3 8.Nbxc3 is actually a true KG Nimzowitsch (3...c6) with reversed colours.
c) 4.Bxg8 Rxg8 doesn’t really improve this for White: 5.exf5 (or 5.Qxd4 Qf6! 6.Qd5 Qe6 with
equality, and not 6.e5? Nc6!) 5...d5 (or 5...Nc6) 6.Ne2 Nc6 7.Nxd4 Qf6 8.Nxc6 bxc6 9.0-0 Bxf5 and
Black is fine, J.Sande Edreira-I.Villalta Bustillo, Ourense 2006.
d) 4.Qxd4 Nc6 5.Qe3 (or 5.Qd5 Qe7) 5...fxe4 6.Nc3 (similarly 6.Qxe4+ Qe7) 6...Nf6 7.Nxe4 Nxe4
8.Qxe4+ Qe7 9.Qxe7+ Bxe7 10.Nf3 d6 is quite equal, hunterthepunter-tsmenace, Yahoo.com blitz
2010.

52
 e) 4.c3!? is another reversed KG Nimzowitsch. Black should reply thematically 4...Nc6, since
5.Qb3 (or 5.cxd4 fxe4 6.Nc3 Nf6 7.Bg5 Bb4, while 5.exf5 Nf6 6.Ne2 d5 7.Bd3 is the same as note
‘b’) 5...Nf6 (or 5...Na5) 6.Bf7+ Ke7 7.Bd5 fxe4 is nothing to worry about.

E1: 4.e5

Pushing on produces a reversed Falkbeer, where the extra Bc4 deters the standard plan of
eliminating the cramping e-pawn; i.e. 4...d6?! 5.Nf3 dxe5?? 6.Nxe5 wins for White. On the other
hand, ...d7-d5 now becomes more attractive:

4...d5 5.Bb3

Instead:
a) 5.Bd3?! c5 6.Nf3 Nc6 7.0-0 Be6 8.c3?! c4 9.Bc2 d3 10.Ba4 Bc5 11.Nd4? Bxd4 12.cxd4 Nge7
13.Qh5+ Bf7 14.Qg5 0-0, D.M.Adams-J.Tait, Sheffield League 2001, is a game Dave won’t want to
see again.
b) 5.exd6 Bxd6 6.Ne2 Nc6 7.0-0 and now 7...Na5? 8.Bxg8 Rxg8 9.Nxd4 is good for White,
J.Pietrasanta-K.Shirazi, Pau 2008 (as given by Emms in Beating 1.e4 e5). Black can improve with
7...Qh4!? (or 7...Qf6 at once) 8.g3 (or 8.Nf4 Nge7) 8...Qf6 9.Nd2 (or 9.Re1 Nge7) 9...Bd7 (or
9...Ne5!? 10.Nxd4 Bd7) 10.Nf3 h6, followed by ...0-0-0, and the best White can do is regain the
pawn.

5...c5

Defending the d4-pawn and threatening to attack the bishop again.

6.c3

53
 Or 6.Nf3 Nc6 (6...a6!? may be better) 7.c4!? dxc4 (and here 7...Nge7) 8.Bxc4 b5 9.Bxb5 Qa5+
10.Nc3! (having missed this resource I failed to put up any further resistance) 10...Bb7? (after
10...Nge7 11.Qa4 Qxa4 12.Nxa4 Rb8 Black is still okay) 11.Qa4 Qxa4 12.Nxa4 0-0-0? 13.0-0 h6
14.Bd2 g5 15.Rac1 and White won easily, T.Hinks Edwards-J.Tait, Notts League 1999.

6...Nc6 7.Nf3

Instead, 7.Ba4!? dxc3 8.bxc3 Be6 9.Nf3 h6 10.Bxc6+ bxc6 11.Qa4 Qb6 12.Nbd2 Qb5 13.Qh4 Qd3
14.Rb1 Ne7 15.Qa4 was wonderfully messy in R.J.Phillips-J.Tait, Blackpool 1998. Stockfish favours
Black, but anything might happen over the board and in fact I later lost this game as well.

7...d3

I would go for 7...c4!? 8.Ba4 d3. After 9.b3 (or 9.0-0 Nge7) 9...Qa5 (or 9...Be6) 10.Nd4 Nge7 11.0-
0 a6 12.bxc4 dxc4 13.Nd2 Qd5 the onus is on White to prove compensation.

8.c4 Be6?

Here 8...d4 was essential, to keep the position closed, so that Black has time to develop with
...Nge7, ...h7-h6, and ...Ng6, ...Be7, or even ...g7-g5 and ...Bg7.

9.cxd5 Bxd5 10.Qxd3

This was V.Bologan-S.Bednarek, European Rapid Ch., Warsaw 2011. White has a definite
advantage, having regained the pawn and with easy development to follow, whereas Black has to
struggle with both an airy position and a GM opponent.
The game concluded 10...Bxb3 11.Qxb3 Qb6 12.Qe6+ Nge7 13.0-0 Nd8 14.Qxb6?! (keeping the

54
queens on with 14.Qc4 looks stronger) 14...axb6 15.Na3 Rxa3?? (presumably time pressure;
15...Nec6 16.Nb5 Kf7 or 16.Nc4 b5 is not so terrible) 16.bxa3 Ne6 17.Rd1 Nc8 18.Rb1 Be7 19.a4 g5
20.a5 g4 and either Black’s flag fell or he just threw in the towel.

E2: 4.Nh3!

A very dangerous continuation, now recommended by GM Martyn Kravtsiv in The Italian

Renaissance. The knight heads for g5 to probe the weak f7-square; it goes to the edge (rather than f3
– see note ‘a’ at the top) to leave the d1-h5 diagonal open for the queen, such that 4...fxe4?? 5.Qh5+
again wins for White.

4...Nf6

The natural-looking 4...Nc6 5.Ng5 Ne5 runs into trouble after 6.Qxd4 (or 6.Bxg8! Rxg8 7.Qxd4
Qf6 transposing, since 7...d6 8.Nc3 c6 9.Bf4 is equally good for White) 6...Qf6 (or 6...Nxc4 7.Qxc4
Qe7 8.0-0 h6 9.Nh3 c6 10.Nc3, ETC_Chess-Reddog, playchess.de 2005; while Kravtsiv goes for
8.Nc3 c6 9.Qe2) 7.Bxg8! (not 7.Nc3?! c5!) 7...Rxg8 (or 7...f4!? 8.Bf7+! Nxf7 9.Qxf6 gxf6 10.Nxf7
Kxf7 11.Bxf4, Lc0-Stockfish, CCC 13 Final 2020) 8.Nc3 (not 8.Nxh7?? Nf3+) 8...c6 9.exf5! Be7 (if
9...h6 10.Nce4 Qxf5 11.0-0 c5!? 12.Qd1 hxg5 13.Re1 Kd8 14.Nxg5 Kc7 15.Qd5 Nf3+ 16.Qxf3 Qxf3
17.Nxf3, White is a clear pawn up) 10.Nce4 Qxf5 11.0-0 Nf7 12.Re1 d5 13.Ng3 Qf6 14.Qh4 h6
15.Nh5 Qf5 16.Qb4 Qd7 17.Nf3 Kd8 18.Qb3 Bd6 19.Bf4 with a dominant position for White,
Stockfish-Lc0, CCC 13 Final 2020.

55
5.Ng5

As intended, at least to my thinking then. White has another possibility in 5.e5!?, effectively
offering a second pawn for an ongoing initiative:
a) 5...Ne4? 6.0-0 Nc6 7.Re1 (or just 7.Nd2, when Kravtsiv says that “White already has a crushing
advantage”; and indeed tsmenace-bobby fissure, ChessWorld.net 2017, ended very quickly: 7...Nxe5?
8.Nxe4 fxe4 9.Qh5+ Ng6 10.Bg5 Be7 11.Nf4 1-0) 7...Nxe5 8.Bd5? (White can improve with the
unlikely sequence 8.Qh5+ g6 9.Rxe4! d6 10.Rxe5+! dxe5 11.Qe2 or, more simply, 8.Rxe4! fxe4
9.Qh5+ Ng6 10.Bg5 Be7 11.Nf4 and Black is still in trouble) 8...d6 9.f4 Ng4 10.Ng5 c6 11.Bf7+ Ke7
12.Bh5 Ngf6 13.Nf7 Qe8 worked out well for Black in S.Paridar-Y.Gozzoli, Abu Dhabi 2004.
b) 5...Ng4 (going after the e5-pawn; Kravtsiv doesn’t mention this, perhaps because he doesn’t
consider it to be any good either) 6.0-0 Nc6 7.Re1 Be7 and now my games have seen:

56
 b1) 8.Bd5 Ncxe5! (the right knight; 8...Ngxe5? 9.c3 d3 10.Nd2 d6 11.f4 Ng4 12.Ng5 Nb4!?
13.cxb4 c6 14.Bf7+ Kf8 15.Nc4! b5 16.Ne3 Bxg5 17.Nxg4 fxg4 18.fxg5 Kxf7 19.Qb3+ d5 20.Rf1+
Kg8 21.Be3 Be6 was close to winning for White in draco69-tsmenace, ChessWorld.net 2016, even if
I did later hold an opposite-coloured bishop ending) 9.Qxd4 (or 9.f3 Ne3 10.Qxd4 Nxd5 11.Qxe5
Nf6) 9...d6 10.f4 (or 10.Nc3 c6) 10...c6 11.Bb3 Ng6 12.Qxg7 Kd7 is quite unclear.
b2) 8.Nf4 Ngxe5! (this time 8...Ncxe5? 9.Qxd4 Nxc4 10.Nd5! is good for White) 9.Bb3 d6 (not yet
9...Nf7?! 10.Nd5 0-0 11.Nxe7+ Nxe7 12.Bg5 Re8 13.Bxf7+! Kxf7 14.Qh5+ Kg8 15.Nd2 and White
has a very strong attack; e.g. 15...h6 16.Nf3 d6 17.Bxh6!? gxh6 18.Qxh6 Rf8 19.Ng5 Rf7 20.Qh8+
Kxh8 21.Nxf7+ Kg7 22.Nxd8 Kf6 23.Re2, followed by Rae1, which extricates the knight and wins,
tsmenace-Pawnhunter1, ChessWorld.net 2017) 10.Nd5 (or 10.c3 d3 11.Nd2 Nf7 etc) 10...Be6!
11.Nxc7+ Qxc7 12.Bxe6 Rf8 is unclear again; e.g. 13.Qh5+ Kd8 14.Qxh7 Bf6 15.Bxf5 Qf7 with
strong counterplay after ...Kc7 etc.

5...Qe7

The typical Two Knights’ move 5...d5 doesn’t work so well here: 6.exd5 h6 (or 6...Bb4+ 7.c3 Qe7+
8.Kd2!, tsmenace-Langbein, ChessWorld.net 2016) 7.Nf3 Bb4+ 8.Bd2 Qe7+ 9.Qe2 Bc5 10.b4 Qxe2+
11.Kxe2 Bd6 12.Nxd4 and White is just a pawn up, RubiChess-SlowChessBlitz, TCEC 20 League 1
2020.

57
6.Nc3!?

Although this leads to very interesting play, the engines are probably right to prefer the alternatives:
a) 6.Bf7+ Kd8 7.Qxd4 (not 7.Bb3?! d5) 7...Nc6 (7...Qb4+ 8.Qxb4 Bxb4+ 9.c3 Bc5 10.Bg6! wins a
pawn) 8.Qd1 d6 9.Bb3 Ne5 10.f4 Nxe4! 11.fxe5 Nxg5 12.0-0 Ne4 and White has decent
compensation but not a definite advantage; e.g. 13.Bf4 g5! 14.exd6 cxd6 15.Be3 Bg7 16.Nd2 Nc3!
17.Qf3 g4 18.Qf2 f4! 19.Bxa7 (not 19.Qxf4? Ne2+ or 19.Bxf4? Bd4) 19...Ne2+ 20.Kh1 Rxa7!
21.Qb6+ (or 21.Qxa7 Bd4 22.Qa5+ b6 23.Qh5 Rf8) 21...Qc7 22.Qxa7 Bd4 23.Qa4 Qg7 now gives
Black good play for the exchange.
b) 6.0-0 fxe4 7.Re1! (the most direct move; 7.Nd2?! d5 8.Bxd5 is well met by 8...Bg4! 9.f3 Nxd5
10.Ndxe4 Nc6 11.fxg4 0-0-0 and Black is at least equal; if 7.Bf7+ Kd8 8.Qxd4 Nc6 9.Qd2 then
9...e3! 10.fxe3 d6 11.Bb3 Ne5 12.Nc3 h6 13.Nf3 c6 and ...Kc7 looks okay; while 7.Nxe4 Nxe4 8.Re1
c6 9.f3 d5 10.Nd2 Nxd2!? 11.Rxe7+ Bxe7 12.Bxd2 dxc4 gives Black three good pieces for the queen,
even if they aren’t out yet) 7...c6 8.Nd2 d5 9.Ndxe4 dxe4 10.Qxd4 (or 10.Nxe4 Nxe4 11.Qf3 Qf6
12.Rxe4+ Kd8 13.Rf4 Qd6, followed by ...Bd7) 10...Bf5! 11.Bf7+ Qxf7 12.Nxf7 Kxf7 13.Qd8 Rg8
14.Qc7+ Nbd7 15.Qxb7 Bd6 16.Qxc6 Be5 reaches another unclear position, where Black has three
active minor pieces for queen and two pawns.

6...dxc3!

The precautionary 6...c6?! wastes time Black can hardly afford: 7.Bf7+ (stronger than either 7.0-0
dxc3! 8.Bf7+ Qxf7 9.Nxf7 Kxf7, or 7.Qxd4 d5 8.Bf4 h6 9.0-0-0 Qc5 10.Qe5+ Be7 11.Ne6 Qxc4
12.Rd4 Nbd7 13.Qxf5 Nf8 14.Rxc4 Bxe6 15.Qe5 dxc4, when Black again has a trio of pieces for the
queen) 7...Kd8 8.Qxd4 d5 9.Be3! (instead, 9.0-0 fxe4 is okay for Black, but not 9...h6? because of the
following remarkable Stockfish variation: 10.exd5 hxg5 11.dxc6+ Kc7 12.Nb5+ Kxc6 13.Rd1 Qxf7
14.Qc3+ Kxb5 and now 15.Rd6!! Bxd6 16.a4+ Ka6 17.Qd3+ Kb6 18.Qxd6+ Nc6 19.Be3+ Ka6

58
20.Qd3+ leads to mate) 9...h6 (or 9...f4 10.exd5 fxe3 11.0-0-0 Nxd5 12.Nxd5 cxd5 13.Qf4) 10.exd5
hxg5 11.dxc6+ Nbd7 12.0-0-0 Qxf7 13.cxd7 Bxd7 14.Bxg5 Be7 15.Rhe1 and White has a probably
decisive attack for the piece.

7.Bf7+ Qxf7

Not 7...Kd8?? 8.Ne6+ and wins.

8.Nxf7 Kxf7 9.e5

After 9.exf5?! d5 10.g4 h6 Black is at least equal.

9...Ng8

The knight has to retreat, since 9...Ng4?! (or 9...Ne4? 10.g4!) 10.h3 Nxe5 11.Qd5+ Kf6 12.h4 Nbc6
13.Bg5+ Kg6 14.f4 is just asking for trouble.

10.Qd5+ Ke8

I played for this position as White thinking it should be favourable, given that the entire black army
is now back on the back rank. However, apart from the rather meaningless c7-pawn, there are no
weaknesses for the queen to attack, and there is no obvious way to create any. Meanwhile Black can
slowly and carefully mobilize.
The game tsmenace-draco69, ChessWorld.net 2016, continued 11.b3 Ne7 12.Qc4 b6 13.Qxc3
(13.Qxc7 Nbc6 14.Qd6 Nd8 is no better for White) 13...Nbc6 14.0-0 (14.Bb2 Bb7 15.0-0-0 Nd8 is
much the same) 14...Bb7 15.Rd1 Nd8! (heading for the fine e6-square) 16.Qd3 (or 16.Qxc7 Bc6)
16...Bc6 17.b4 a6! (making it harder for White to dislodge the c6-bishop) 18.a4 Ne6 19.c4 g5

59
 20.Bb2? (mistakenly trying to improve on 20.b5 axb5 21.cxb5 Be4 22.Qxd7+ Kf7 23.Ba3 Rd8
24.Qxd8 Nxd8 25.Rxd8 Ng6 26.Bxf8 Rxf8 27.Rd7+ Ke6 28.Rxc7 with a level endgame; by inserting
Bb2 White hopes to recapture on b5 more strongly with the a4-pawn, but Black’s next move defuses
this whole plan and the knights have free access to the f4-outpost) 20...Rd8! (now 21.b5? axb5
22.cxb5 runs into 22...Bxg2!) 21.f3? (temporarily controlling the c6-bishop, at the high cost of
providing a hook on the kingside which Black immediately exploits) 21...g4! 22.b5 Ba8 23.a5 (still
hoping for a queenside breakthrough) 23...gxf3 24.axb6 Rg8 25.g3 cxb6 26.bxa6 Rg6 (preparing both
...f5-f4 and ...Nf4, with further options of ...Rh6) 27.Kf1 (ruling out ...Nf4, but not...) 27...f4! 28.Ba3
(28.gxf4 Nc5 is quickly decisive) 28...fxg3 29.hxg3 Ng5 (heading for e4 or h3) 30.c5!? (as good as
anything) 30...Be4 31.Qd2 Bc6 32.cxb6 (online games can be prolonged without cost, even of time if
moves are made quickly; hence I sometimes play on when completely lost if there’s a variation or
position I want to see appear on the board rather than “in the notes” – here I wanted my queenside
pawns to arrive on the seventh rank; and I had that little victory, before being mated by a pawn)
32...Ne4 33.Qh2 Rh6 34.Qxh6 Bxh6 35.Bc1 Bxc1 36.Rdxc1 Nd5 37.b7 Ne3+ 38.Ke1 Bb5 39.a7 f2
mate.

F: 3.d3

60
 By far the most popular move and the one usually recommended in books on the Bishop’s Opening.
White proceeds with the intended d2-d3 as if nothing has really happened. It’s also the natural follow-
up in the King’s Gambit Declined, of which this is a reversed version. Can White make the extra
tempo count for something?

3...Nf6

The natural and best reply. 3...d6?! 4.Nf3! is an awkward Philidor; while if 3...Nc6?! 4.Nf3!, Black
has committed the b8-knight too soon (c.f. 4...Nc6 in line F2).
Now White has two main tries:
F1: 4.f4
F2: 4.Nf3
Instead:
a) 4.Bg5?! is premature due to 4...h6! 5.Bxf6 Qxf6 and Black is already fine; e.g. 6.Nc3 c6 7.Bb3
Na6 8.Nf3 d6 9.0-0?! f4!, followed by ...g7-g5 with a typical kingside attack, boa-tsmenace,
ChessWorld.net 2017.
b) 4.Nh3 (threatening Ng5 or f2-f4) is defused by 4...fxe4 (4...c6 5.f4!? is more tricky) 5.dxe4 c6,
when 6.Qe2 (here 6.0-0 d5 7.exd5 cxd5 8.Bb3 Bxh3 9.gxh3 Qd6 is equal) 6...Be7 7.Nc3 d5 8.exd5
cxd5 9.Nxd5 Nxd5 10.0-0 Bxh3 11.Qh5+ g6 12.Qxh3 Nb6 13.Bb3 Qd7 14.Be6 Qc6 worked out well
for Black, JubaJuba-tsmenace, ChessWorld.net 2016.
c) 4.a3!? is a useful extra move on a KGD. There are no practical examples, but 4...c6 is logical.
After 5.Nc3 (5.f4!? and 5.Ba2!? are options) 5...d5 6.exd5 cxd5 7.Ba2 Nc6 8.Nf3 e4 9.dxe4 fxe4
Stockfish suggests a piece sac with 10.0-0!? though claims no particular edge at the end of it:
10...exf3 11.Re1+ Be7 12.Nxd5 Bg4 13.Bg5 Kf8 14.Nxe7 Qxd1 15.Raxd1 Nxe7 16.Bd2 a5 17.Bf4
Re8 18.Bd6 h5 19.h3 Bf5 20.Rxe7 Rxe7 21.Re1 Ng8 22.Bxg8 Kxg8 23.Rxe7 fxg2 24.Rxb7 Bxc2 and

61
the result should be a draw.
d) 4.Nc3 is a major option too. I think the best reply is 4...Nc6, transposing to the Calabrese
Deferred in Chapter Three.
Instead, 4...Bb4 (Larsen) is very natural and I’ve played that without thinking more than once.
Those games have generally gone something like 5.Nf3 fxe4 6.dxe4 (if 6.Nxe5 then 6...d5 7.Bb3 Qe7
8.d4 Nbd7) 6...Qe7 7.0-0 Bxc3 8.bxc3 d6 with a decent enough position for Black.
The problem is 5.Ne2!, and if 5...fxe4 (or 5...c6 6.f4!) 6.dxe4 d6, as in L.Donovan-R.Schuermans,
Belgian League 2014, then simply 7.0-0 and a2-a3 gives White an easy game. From the other side
Black is behind in development, several moves from castling, and has no structural compensation
with the c3-knight defended by its colleague.

F1: 4.f4

This challenging continuation was Carl Jaenisch’s proposed refutation of 2...f5 in the 19th century;
it is still given in ECO Vol.C, at least up to the 4th edition (2000).

4...Nc6

Other moves were part of the old analysis:

a) 4...d6 5.Nf3 fxe4 (or 5...exf4 6.0-0 fxe4 7.dxe4; while for 5...Nc6 see 5...d6 in the main line)
6.dxe4 Bg4 7.fxe5 Bxf3 8.Qxf3 dxe5 9.Qb3 Qc8 10.Bg5 Bc5 11.Bf7+ (Jaenisch).
b) 4...d5 5.exd5 exf4 (or if 5...e4, from Bilguer, then simply 6.dxe4 fxe4 7.Be3) 6.Qe2+ Be7 7.Nc3
Nbd7 8.Bxf4 Nb6 9.0-0-0 (Staunton).

62
5.Nf3

Taking either pawn is bad: 5.fxe5?! Nxe5 or 5.exf5?! d5 6.Bb5 exf4, as is 5.Qe2? Nd4; while 5.Nc3
can be met by 5...Bb4 (or 5...fxe4 6.dxe4 Bb4) 6.Nf3 (or 6.Nge2 fxe4 7.dxe4 exf4 8.Bxf4 Qe7)
6...fxe4 7.dxe4 Qe7 8.Qe2 d6 9.0-0, WaterDragon-jana84, ChessWorld.net 2011, and now 9...Bxc3
10.bxc3 exf4 11.Bxf4 Bg4 12.Rab1 0-0-0 with a sound position even if it looks a little risky. Stockfish
shows that 13.Ba6!? bxa6 14.Qxa6+ Kd7 15.Bxd6! Qxd6 16.Rbd1 Nxe4 17.Rxd6+ Nxd6 is fine for
Black.

5...fxe4

In Chess Mail I concentrated on 5...exf4, having noticed a transposition to analysis by GM Matthias

Wahls in New in Chess YB38; specifically 2.f4 Nc6 3.Nf3 f5!? with 4.Bc4 exf4 (Wahls rightly prefers
4...fxe4!) 5.d3 Nf6, which comes to the same thing. The critical line continues 6.Bxf4 fxe4 7.dxe4,
when Wahls writes: “Future analysis will have to assess the value of the positions after 7...Qe7 or
7...Nxe4.” Unfortunately, it’s assessed that 7...Qe7 (or 7...Nxe4?! 8.0-0 Bc5+ 9.Kh1 Rf8 10.Qd5!)
8.Nc3! (Bücker) 8...d6 (or 8...Qb4 9.Qd3) 9.0-0 Bg4 10.Bb5 is just very good for White.
The cautious 5...d6 also works out in White’s favour after 6.fxe5 dxe5 7.Ng5 Bc5 8.Nf7 Qd4 9.Rf1
Rf8 10.c3 Qd7 11.Ng5 f4 12.Ne6.

6.dxe4

If 6.Nxe5?! Bc5! (or just 6...d5 7.Bb5 Bd7) 7.Nc3 Qe7 8.dxe4 d6 9.Nxc6 bxc6 10.e5 Bg4 11.Be2
dxe5 12.Bxg4 exf4+ 13.Qe2 (or 13.Kf1?! Rd8 14.Qe2 Nxg4 15.Qxg4 0-0) 13...Nxg4 14.Qxe7+ Bxe7,
Black is at least equal.

63
6...Nxe4

The critical position, identified by FM Mark Lyell 25 years ago. White has tried:
a) 7.Bd5 Nf6 8.Bxc6 (or 8.fxe5 Nxd5 9.Qxd5 d6) 8...e4 9.Bxe4 Nxe4 10.Be3 Bc5 11.Qe2 0-0 and
Black already stood well, K.Bowden-M.Lyell, British Ch., Southampton 1986.
b) 7.fxe5 Nxe5 (not 7...Qe7?, D.M.Adams-J.Tait, Sheffield League 2008, because 8.Qe2! Nxe5
9.Nxe5 Qxe5 10.Nc3 Bb4 11.0-0! pretty much wins by force) 8.Bd5 (or 8.Nxe5 Qh4+) 8...Nxf3+
(and not 8...Nf6? 9.Nxe5 Qe7 10.Qe2 Nxd5, since 11.0-0 Nf6 12.Bg5 Qe6 13.Rxf6! again gives
White a decisive attack according to Stockfish) 9.Qxf3 Nf6 10.Bg5 (if instead 10.Nc3 then 10...Qe7+
11.Kf2 Kd8 or 11.Be3 Ng4 equalizes; and the piece sacrifice 10.0-0 c6 11.Nc3!? cxd5 12.Nxd5 Qa5
is inconclusive, e.g. 13.Nxf6+ gxf6 14.Be3 Be7 15.Rae1 Kd8 16.Bd4 Re8 17.Kh1 f5) 10...c6?! (here
10...h6! is correct; e.g. 11.Bh4 Qe7+ 12.Kd1 Qd6 13.Re1+ Be7 14.Nc3 Kd8 15.Bg3 Qb6 with a very
unclear position) 11.Bb3? (now 11.Nc3! cxd5 12.0-0-0 was much stronger) 11...Qa5+ 12.Bd2 Qe5+
13.Kd1 Qh5 (or 13...d5 14.Re1 Ne4) 14.Re1+ Kd8 15.Qxh5 Nxh5 and Black emerged with a clear
advantage, J.Emms-M.Lyell, British Ch., Southampton 1986.
c) 7.Qd5 Nd6 8.Bb3

64
 8...e4! (stronger than 8...exf4 9.0-0 Qf6 10.Nc3 Kd8, as in alapin-tsmenace, ChessWorld.net 2014,
where Stockfish suggests 11.g3!? Ne7 12.Qd3 as promising for White) 9.Ng5 (or 9.Ne5 Nxe5,
intending 10.fxe5 c6! 11.Qd4 Qh4+ 12.g3 Nf5 or 10.Qxe5+ Qe7 11.Nc3 c6 12.Be3 Nf7) 9...Qe7
10.Nc3 h6 11.Ngxe4 Nxe4 12.Nxe4 Nb4 13.Qh5+ Kd8 14.Qe5 b6 15.0-0 Bb7 16.Re1 Qxe5 17.fxe5
a5 and Black had no problems, tsmenace-Panza, ChessWorld.net 2014.

F2: 4.Nf3

White continues in KGD fashion, with 0-0, Nc3 and Bg5 likely to follow. Just going along with this
– i.e. 4...Nc6 or 4...c6 – can lead to trouble because, a tempo behind, Black lacks the resources to
make such set-ups work (see the next note). So what else?

65
4...fxe4!?

Exchanging on e5 early is a weak idea in the KGD and cedes any chance of an advantage. In this
reversed version equality will do just fine and resolving the central tension reduces the tactical
element, giving Black more time to manoeuvre and organize.
Whereas:
a) 4...Nc6?! 5.0-0 Bc5 6.Nc3 d6 7.Bg5 is good for White as neither of the standard moves work out
well for Black: 7...Na5 (or 7...h6 8.Bxf6 Qxf6 9.Nd5 Qd8 10.d4!, blasting open the centre with the
black king still in the middle) 8.Bxf6 Qxf6 9.Nd5 Qd8 10.b4 Nxc4 11.bxc5 fxe4 12.dxc4 exf3
13.Qxf3 c6 (or 13...dxc5? 14.Rae1 Qd6 15.Qh5+ Kf8 16.Rxe5 and wins – Lane) 14.Nc3 Qf6 15.Ne4
Qxf3 16.Nxd6+ Ke7 17.gxf3 and White is clearly better, with an extra pawn, a big knight, and the
initiative, D.Fryer-M.Lyell, Hastings 2003/04.
b) 4...c6?! 5.0-0 d6 (if 5...d5 6.exd5 cxd5 7.Bb3 e4 8.dxe4 fxe4, then 9.Nc3! is very strong: 9...exf3
10.Re1+ Be7 11.Nxd5 Nxd5 12.Bxd5 Nd7 13.Bg5 Nf6 14.Bxf6 gxf6 15.Qxf3 and White soon won,
conde7-bulent_2010, ChessWorld.net 2012) 6.Re1 fxe4 7.dxe4 Qe7 (or if 7...Be7 8.Ng5 d5 9.exd5
cxd5, S.Milliet-K.Shirazi, Cap d’Agde 2002, then 10.Bb3 h6 11.Nf3 e4 12.Nh4 and Black’s problems
are nowhere near solved) 8.a4 Be6 9.Nbd2 Nbd7 10.b3 Nc5 11.Ba3 Nfd7 12.a5 with a big advantage
for White, J.Magem Badals-I.Argandona Riveiro, Donostia 2009.
c) 4...d5!? 5.exd5 Bd6 is an interesting option, a sort-of Elephant Gambit. Steinitz once played this
way. Stockfish claims a clear advantage for White but was unable to beat itself in four games.

One example: 6.0-0 0-0 7.Nc3 h6 (7...a6 8.a4 h6 was Lenhof-W.Steinitz, Vienna 1859) 8.Re1 Nbd7
9.Nb5 a6 10.Nxd6 cxd6 11.a3 f4 12.Ba2 Re8 13.c4 Nf8 14.h3 g5 15.c5 Ng6 16.cxd6 g4 17.d4 gxf3
18.dxe5 Rxe5 19.Rxe5 Nxe5 20.Bxf4 Qe8 21.g4 Bd7 was eventually drawn, StockfishClassical-
StockfishClassical, TCEC 20 Test Game 2021.

66
5.dxe4 Bb4+!

Patzer sees a check. The point is to deny White the possibility of Bg5 and Nc3-d5 as in note ‘a’
above. 5...Nxe4?, as in L.Bledow-P.Von Bilguer, Berlin 1839, is just asking for trouble; e.g. 6.0-0 c6
7.Re1 d5 8.Rxe4! dxe4 9.Bf7+ and wins.

6.c3

Played in every game bar one. After anything else Black’s plan is accomplished with little
consequence: 6.Nc3 Bxc3+ 7.bxc3 d6; or 6.Nbd2 Bxd2+ 7.Bxd2 d6; or 6.Bd2 Qe7 (or 6...Bxd2+)
7.0-0 Bxd2 8.Nbxd2 d6 and Black has equalized already, Reprimand-tsmenace, ChessWorld.net
2015.

6...Bc5

I’ve defended this set-up eighteen times since 2011 (W5, D12, L1). Black intends ...d7-d6 and
...Qe7, aiming to castle short after ...Be6 or first (and more usually) ...Nc6-d8, reinforcing key squares
on the a2-g8 diagonal as well as enabling ...Nf7. It is very important to avoid an untimely Ng5 or Nh4
and f2-f4. Black’s slow plan only works because of White’s lack of an effective pawn break.

7.0-0

Direct tactical approaches don’t give White much:

a) 7.Nxe5 Qe7 8.Nf3 Qxe4+ regains the pawn at once.
b) 7.b4 Bb6 8.Nxe5 (8.0-0 d6 transposes to the main line, and 8.a4 a6 is likely to) 8...Qe7 9.Nf3
Qxe4+ 10.Be2 0-0 is much the same.
c) 7.Ng5 Qe7 8.Bf7+ Kf8 is no inconvenience for Black, who can follow up with ...h7-h6, ...g7-g5

67
and ...Kg7.

7...d6

8.b4

Gaining space with tempo is the most natural choice. Others:

a) 8.Nbd2 Qe7 9.Nb3 Bb6 10.Qe2 Nc6 11.a4 a6 12.a5 Ba7 13.h3 Nd8 14.Rd1 Be6 (at last) 15.Ng5
Bxc4 16.Qxc4 h6 17.Nf3 Ne6 was good for Black, Reprimand-tsmenace, ChessWorld.net 2013, and
17...g5 looks even better.
b) 8.Qb3 Qe7 9.Bg5 Nbd7!? (or 9...Nc6) 10.Nbd2 Nb6 11.Be2 Be6 12.Qb5+ c6 13.Qd3 Nbd7
14.b4 Bb6 15.a4 0-0 16.a5 Bc7 17.h3 h6 18.Be3 d5 saw Black take the initiative, AndyAndyO-
tsmenace, ChessWorld.net 2012.
c) 8.Qd3 Qe7 (or 8...a5!?, Panza-conde7, ChessWorld.net 2012) 9.Bg5 a6!? (or 9...Nc6 10.b4 Bb6
as in the next note) 10.b4 Ba7 11.Nbd2 Be6 12.Bxf6 Qxf6 13.Bxe6 Qxe6 14.Ng5 Qe7 15.Qd5 c6
16.Qe6 Nd7 was equal, Panza-tsmenace, ChessWorld.net 2012; e.g. 17.Nc4 Qxe6 18.Nxe6 Kf7
19.Ng5+ Ke7 20.Rad1 Bb8.

8...Bb6

68
9.a4

And once more. Instead, 9.Nbd2 Qe7 10.a4 a6 or 10.h3 Nc6 11.a4 a6 transposes below; while
9.Qd3 Qe7 10.Bg5 Nc6 11.Nbd2 Nd8 12.a4 a6 13.Nh4 Nf7 14.Bxf6 (or 14.Nf5 Bxf5 15.Bxf6 gxf6
16.exf5 c6) 14...Qxf6 15.Nf5 0-0 16.Ne3 c6 was again roughly equal, nii_sama-tsmenace,
ChessWorld.net 2014.

9...a6

I prefer this to 9...a5, which creates a potential weakness on the queenside: after 10.Nbd2, White
can continue with Bb3 and Nc4, or else b4xa5, Nb3 and a4-a5. Even if Black’s position is playable, it
seems sensible to keep things as solid as possible while trying to sort the pieces out.

69
10.Nbd2

Defending the e-pawn for the moment. Later the knight may go on to c4 and possibly e3.
Other moves:
a) 10.a5 Ba7 11.h3 Qe7 12.Bg5 Be6 (an early game; I’d probably play 12...Nc6 now) 13.Bxe6
Qxe6 14.Bxf6 Qxf6 15.Qd5 c6 16.Qd3 Nd7 17.Rd1 0-0-0! 18.Nbd2 (not 18.Qxd6?? Nb8; while if
18.Na3 Rhe8 19.Qxd6 then 19...Qf4 20.Qd2 Nf6 21.Qxf4 exf4) 18...Nf8 19.c4 Ne6 20.b5 g5 with
mutual chances, stormytlc-tsmenace, ChessWorld.net 2011.
b) 10.Qd3 Qe7 11.Na3 (or 11.Nbd2 Nc6 12.Bb3 Nd8 13.Nc4 Ba7 14.Ne3 Be6 15.Bc2 0-0 16.Ng5
Kh8, MarcShaw-tsmenace, ChessWorld.net 2012) 11...Nc6 12.Nc2 Nd8 13.a5 Ba7 14.Be3 Be6
15.Bxa7 Rxa7 16.Ng5 Bxc4 17.Qxc4 Nh5 18.Nh3 Qf7 19.Qxf7+ Nxf7 and Black’s problems were
over, Pawnhunter1-tsmenace, ChessWorld.net 2017. Both games were soon drawn.
c) 10.Na3 Qe7 11.Bb3 Be6 12.Nc4 Ba7 (12...Nxe4!? may be possible; e.g. 13.Nxb6 cxb6 14.c4 0-0
or 13.Qd3 Nxf2 14.Qe2 Ba7 15.Rxf2 Bxf2+ 16.Qxf2 0-0) 13.Be3 Bxc4 14.Bxc4 Bxe3 15.fxe3 Nbd7
16.a5 0-0-0 17.Ng5 (or 17.Qd3 Kb8 18.Bxa6!? bxa6 19.Qxa6 d5) 17...Rde8 18.Qd3 h6 19.Nf3 d5!
20.Bxd5 Nxd5 21.Qxd5 (not 21.exd5?? e4) 21...Nf6 22.Qa2 Nxe4 23.Qc2 Nd6 24.c4 Qe6 25.Rac1
Rhf8 26.c5 Nf5 saw Black equalize in SlowChess Blitz-RubiChess, TCEC 20 League 1 2020.

10...Qe7

70
11.h3

Not yet 11.Ba3 Nc6 12.b5 axb5 in view of 13.Bxb5 0-0, or else 13.axb5 Nd8; while 11.Bb3 Be6
(rather than 11...Nc6?! 12.Nc4 Ba7 13.Be3 Nd8 14.Bxa7 Rxa7, bulent_2010-tsmenace,
ChessWorld.net 2012, when 15.Nh4! and f2-f4 is strong) 12.Nc4 transposes to 10.Na3 etc.

11...Nc6

Better here than 11...Be6?! 12.Qb3 Bxc4 13.Nxc4 Ba7 14.Be3, when Black still has issues with
Ng5 or Nh4 and f2-f4.

12.Ba3

Blocking the a-file before pushing the b-pawn, which will also uncover the bishop towards the
black queen.

12...Nd8

Still not 12...Be6?! due to 13.Ng5 Bxc4 14.Nxc4, when any of Nxb6, b4-b5, f2-f4 and Qh5+ are
potential threats.

13.b5

My one loss in this line came about because, after 13.Qb1 Be6 14.a5 Ba7 15.b5 axb5 16.Bxb5+
Bd7 17.Qb3 Bxb5 18.Qxb5+ Qd7 19.Qb3, I didn’t think it mattered where the queen went: 19...Qf7??
(19...Qe6 is fine) 20.Nc4! Qe6 (or 20...0-0 21.Ng5) 21.Ng5 Qe7 22.Rad1 (threatening to sac on d6
with anything) 22...Bb8 23.f4 and White soon won, Raffzahn-tsmenace, ChessWorld.net 2013.

71
13...a5

Now 13...axb5?! 14.Bxb5+ Bd7 15.Nc4 is good for White.

I’ve reached this position three times (via different routes). Black needs to be a little careful since
putting a piece on e6 might allow Nxe5.
Here 14.Qc2 Nh5 15.Bb3 Be6 16.Nc4 (not 16.Nxe5?? Bxb3) 16...Bxc4 17.Bxc4 Rf8 18.Qd1 Nf4
19.Bc1 Nde6 20.Kh2 Nc5 21.g3 Nfe6 22.Nh4 0-0-0, alexkhesin-tsmenace, ChessWorld.net 2019; and
14.Re1 Nd7!? (or 14...Nf7) 15.Bb3 Bc5 16.Nc4 Bxa3 17.Rxa3 Ne6 18.Ne3 Nb6 19.Nd2 0-0, Quinoa-
tsmenace, ChessWorld.net 2015, led to two more draws.
However, 14.Rc1!? is more cunning, intending 14...Nf7 (or 14...Be6? 15.Nxe5) 15.Bxf7+ Qxf7
16.c4 and c4-c5. So I tried 14...Ba7!? 15.Re1 Nd7 (Stockfish prefers 15...Nf7 16.Bxf7+ Qxf7 17.c4 b6
18.Nf1 Bb7 19.Ne3 0-0-0 20.Nd5 Kb8 and says “0.00” despite the entombed bishop) 16.Nf1 Ne6?!
(16...Nf7 17.Ne3 0-0 18.Nd5 Qd8 was safer) 17.Ne3 Nf4 18.Nd5 Nxd5 19.Bxd5 Nc5?! (and here
19...Rf8) 20.c4 g5 21.Nd2 g4, when 22.hxg4? h5! 23.g5 Bg4 24.Nf3 0-0-0 led to a win for Black in
Tihomir Glowatzky-tsmenace, ChessWorld.net 2015. Stockfish is not impressed, throwing out
22.Nb3! h5 (not 22...gxh3? 23.Qh5+ or 22...Nxb3? 23.Qxb3 Bd4 24.Rcd1 gxh3 25.Rxd4! exd4 26.e5
and wins) 23.Nxc5 Bxc5 24.Rc3, assessing it as good for White. So Black must go back to one of the
earlier bracketed suggestions.
specifically 2.f4 Nc6 3.Nf3 f5!? with 4.Bc4 exf4 (Wahls rightly prefers 4...fxe4!) 5.d3 Nf6, which
comes to the same thing.

72
 Chapter Three
Vienna Game

1.e4 e5 2.Nc3 Nc6

Rather than being disreputable, 2...Nc6 is actually a very good move. I prefer it to 2...Nf6, which is
seen between twice and thrice as often and is usually recommended for Black; so much so that its
supposed superiority has become a factoid.
To my mind, there are only two reasons for developing the king’s knight first. One, to meet 3.Nf3
with 3...Bb4, avoiding the Four Knights. Two, to go for a Frankenstein-Dracula after 3.Bc4 Nxe4!
4.Qh5 Nd6 5.Bb3 Nc6! 6.Nb5 g6 7.Qf3 f5 8.Qd5 Qe7 9.Nxc7+ Kd8 10.Nxa8 b6. In those cases,
2...Nf6 makes sense. Otherwise, start with the queen’s knight.
The main justification arises from 3.f4. With 2...Nf6, sure, 3.f4 d5! is a routine equalizer; but after
2...Nc6 3.f4 exf4! it’s White who must think about equalizing.
Against 3.g3 too – unless you want to push 3...d5 – there are advantages to leaving the g8-knight at
home: the f-pawn is still free to advance and/or Black may opt for deployment with ...Nge7 or ...Nh6,
depending on circumstances. As for 2...Nc6 3.Bc4, well, 3...Nf6 is just as good as it was the other
way round (and 4.d3 Na5! equalizes there), though I’m offering something else in line B.
Finally, checking results in the TWIC database: super-GM (Elo 2700+) games over the past 15
years (mostly rapid or blitz) show 2...Nf6 scoring +28 =19 -12, which is 63.6% for White. Whereas
2...Nc6 scores +6 =7 -14, which is 64.8% for Black. Yes, that surprised me too. So 2...Nc6! it is.
The main continuations then are:
A: 3.g3

73
 B: 3.Bc4
C: 3.f4
Or 3.Nf3, transposing to Chapter Six. Other moves can be regarded as “non-theoretical”. A few
random examples:
a) 3.a3 Bc5 4.d3 Nge7 5.Na4 Bb6 6.Nxb6 axb6 7.Nf3 0-0 8.g3 d6 9.Bg2 f5 10.exf5 Bxf5 11.Nh4
Be6 12.0-0 Qd7 with a pleasant position for Black, V.Moiseenko-Yu Yangyi, World Rapid Ch.,
Moscow 2019.
b) 3.Bb5 Nd4 4.Ba4 c6 5.d3 Nf6 6.Bb3 d6 7.Nge2 a5 8.Bg5 b5 is even more pleasant, B.Certic-
N.Karaklajic, Belgrade 1991.
c) 3.d3 Bc5 4.g3 d6 5.Na4 (for 5.Bg2 f5 see line A) 5...Nf6 (or 5...h5!? as in line A) 6.Nxc5 dxc5
7.h3 (if 7.Bg2 Qd6 8.f4 Bg4 9.Qd2, K.Badev-M.Amerkeshev, Sofia 2007, then 9...0-0-0!, intending
10.f5 c4 11.h3 cxd3 12.hxg4 Nb4 13.cxd3 Qc5 with a strong attack for the piece) 7...Qe7 8.Nf3 h6
9.Bg2 Be6 10.Qe2 0-0-0, when Black stands well and can take the initiative on either side of the
board. A.Dohr-M.Franic, Frohnleiten 1999.

A: 3.g3

It seems White wants to avoid confrontation. What can we do about that? The half-dozen times I’ve
ever faced this line, I’ve generally opted for ...Bc5 and ...h7-h5, usually with ...a7-a6 thrown in to
preserve the dark-squared bishop. Its bark is worse than its bite.
I’m now more inclined to the ideas outlined in Bologan’s Black Weapons, where Black shows
complete disregard for the bishop’s safety. If White wants to spend time with the knight exchanging it
off, well, let them. GM Bologan’s suggestions, involving ...f7-f5 or ...h7-h5, have naturally had some
testing since his book appeared in 2014.

74
3...Bc5

If you’re happy to play the White side of line C2, then 3...f5!? 4.exf5! (otherwise Black captures on
e4) 4...Nf6 is a definite option. After 5.g4 g6!, I think Black is just about okay (as with 5.g3! for
White in C23), but it isn’t worth the space to try and prove that when the whole thing is extremely
rare: just four games with 5.g4 in all my databases.
Also, there’s a potential spoiler: 5.d3 d5 6.Bh3! g6 (not 6...Bc5?! 7.Nf3 0-0 8.0-0 Qe8 9.Bd2 Bd7
10.Re1 Qh5 11.Nh4 and White is clearly better, F.Apaydin-A.Karagollu, Turkish League 2010)
7.Bg5 (7.fxg6?! Bxh3 8.Nxh3 hxg6 9.Ng5 Qd7 gives Black good play) 7...Bxf5 8.Bxf5 gxf5 9.f4
(after 9.Bxf6?! Qxf6 10.Nxd5 Qf7, White has to find 11.Qh5! Qxh5 12.Nf6+ Kf7 13.Nxh5 Kg6
14.g4! fxg4 15.Ng3 just to be equal) 9...exf4 10.Qe2+ Kf7 11.Bxf6 Qxf6 12.0-0-0, when White has a
useful initiative for the pawn. Then again, it’s not too frightening.

4.Bg2 d6

So what? So what if the bishop dies...? So what about the...? I don’t give a...!

5.Nge2

Other variations:
a) 5.Nf3 f5!? 6.exf5 (or 6.d3 Nf6 7.0-0 0-0 8.exf5 Bxf5 9.Be3 Bb6, A.Tsirulnik-V.Narizskaya,
Ukraine 2000, intending ...Qd7 and ...Bh3 – Bologan; if instead 7.Bg5 h6, G.Lagvilava-T.Martynuk,
Minsk 2003, then 8.Bxf6 Qxf6 9.Nh4 0-0 10.Nd5 Qf7 11.Nxf5 Nd4! is equal) 6...Bxf5 7.Nh4 (or if
7.0-0 Qd7 8.d3 0-0-0 9.Be3 Bxe3 10.fxe3, M.Bitalzadeh-A.Slotboom, Rijswijk 2006, then 10...Nf6 –
Bologan) 7...Be6 8.Ne4 Nf6! (characteristically not caring about the c5-bishop) 9.Nxc5 dxc5 10.0-0
Qd7 11.a3?! Bd5 12.Re1 Bxg2 13.Kxg2 0-0-0 14.d3 c4 and Black is clearly better, D.Kopylov-
Rob.Stein, Magdeburg (blitz) 2019.

75
 b) 5.d3 f5!? 6.Na4 (if 6.Nd5, J.Mieses-F.Dus Chotimirsky, Karlsbad 1907, then 6...fxe4 7.dxe4 a5
negates White’s queenside plans) 6...Nf6 7.Nxc5 dxc5 8.exf5 Bxf5 9.Be3 (not 9.Bxc6+? bxc6 10.f3
c4 with a clear advantage, M.Matlakov-A.Chandra, PNWCC blitz 2020) 9...Qd7!? 10.Bxc5 0-0-0
11.Ne2 Bh3 (or 11...e4!?) 12.Bxh3 Qxh3 13.Qd2 Qg2 14.0-0-0 Qd5 15.Qc3 Qxa2 and Black is no
worse, R.Mamedov-R.Hovhannisyan, World Blitz Ch., St. Petersburg 2018.
c) 5.Na4 h5!? (you play on your side...) 6.Nxc5 (if 6.Nf3, M.Turov-P.Acs, Budapest 1997, then
6...a5!? provokes the exchange on c5 anyway – Bologan) 6...dxc5 7.d3 (or if 7.h4, L.Holzinger-L.Van
Wely, Kuppenheim blitz 2005, then 7...Be6 8.d3 Qd6 and ...0-0-0, with ideas of ...c5-c4, or ...Nd4, or
...f7-f6 and ...Nh6) 7...h4 8.Be3, G.Markotic-D.Kuljasevic, Sibenik 2010, and now 8...Qd6 9.Nf3 (or
9.h3 hxg3 10.fxg3 Be6) 9...hxg3 10.fxg3 Be6, followed possibly by ...Nh6, or 11.Ng5 Bd7 and ...f7-
f6.

5...f5!?

6.Na4

Here 6.exf5 Bxf5 7.Ne4 Nf6 8.Nxc5 was ½-½ in M.Kuijf-N.Mitkov, Sitges 1999; play might
continue 8...dxc5 9.d3 Qd7 10.Be3 0-0-0 11.Bxc5, as in note ‘b’ above.
If instead 6.d3 Nf6 7.0-0 (for 7.Na4 see the next note) 7...fxe4 8.dxe4, Stockfish suggests 8...0-0
9.Na4 b6!? 10.Nxc5 (or 10.a3 a5) 10...dxc5 with ideas of ...Ba6.
My clubmate Ken Morrison has caused me problems in blitz games with lines like 6.0-0 Nf6 7.d4!?
Nxd4 (or 7...exd4 8.Nd5 Nxe4 9.b4!) 8.Be3 Nxe2+ 9.Qxe2 Bxe3 10.fxe3 fxe4 11.Nxe4, when
White’s initiative is enough for the pawn and counts for more at blitz. At home 11...0-0 12.Nxf6+
gxf6 13.Rf2 Kg7 14.Raf1 c6 15.Qh5 d5 16.e4 d4 17.Bh3 Bxh3 18.Qxh3 Qe7 19.Qg4+ Kh8 20.Qh4
Kg7 21.Qg4+ is one route to a draw.

76
6...Nf6 7.Nxc5

On 7.d3 fxe4 8.dxe4, Black might go for 8...Bb4+!? 9.Bd2 (or 9.c3 Ba5 10.b4 Bb6 11.0-0 Be6
12.a3 Qd7 13.Qd3 Qf7, probing the weakened light squares, S.Ravichandran-N.Mitkov, Las Vegas
2010) 9...Be6 10.Bxb4 Nxb4 11.Qd2 a5 12.Nac3 Na6 13.0-0 Nc5 14.b3 0-0, followed by ...Qd7,
W.Nyberg-H.Gulbis, corr. 2019.

7...dxc5 8.Nc3 0-0

This position has been played out a few times post Bologan:
a) 9.0-0 f4! (Bologan) is good for Black: if 10.d3, as in Dar.Moran-M.Magnusson, Reykjavik 2018,
then 10...g5.
b) 9.d3 Nd4 (or 9...fxe4, intending 10.Nxe4 Bg4! 11.f3 Nxe4) 10.0-0 f4 is still good enough for
equality after 11.gxf4 exf4 12.Bxf4 (not 12.f3?, D.Popovic-B.Murtazin, Palic 2016, because of
12...Nh5!) 12...Nxe4.
c) 9.exf5 Bxf5 10.d3 Nd4 11.0-0 (not 11.Bxb7? Bh3, when the white king is trapped in the centre)
11...Qd7 12.Be3 Rad8 13.Qd2 Bh3 14.f3 Bxg2 15.Qxg2 b6 16.Ne4 h6 17.Kh1 Nd5 18.Bg1 a5 and
Black stands quite well, T.Rosenhöfer-T.Schwetlick, corr. 2020.

B: 3.Bc4 f5!?

77
 I call this the Calabrese Counter-Gambit Deferred. Surprisingly (to me), hardly any of my books on
the Vienna even mention this variation. MegaBase 2020 lists a mere five games (which includes one
via 2.Bc4 f5), yet Black scored 4/5 in those (and the loss was an under-10s game). Even after 4.d3
Nf6 (which can arise via multiple routes) Black’s results are positive: P15, W9, D2, L4. So I think
3...f5!? certainly deserves more attention.

4.d3

The most natural response. White has also tried:

a) 4.Nf3?? (this position has been seen 53 times, albeit only once by this move order) 4...fxe4
5.Bxg8 (if 5.Nxe4 d5 6.Nxe5 Nxe5 7.Qh5+ Ng6 8.Bxd5 Nf6 9.Nxf6+ Qxf6, White’s two pawns are
worth nowhere near the piece, while 6.Bb5 is the main line from Chapter Eleven with Black to
move!) 5...exf3 6.Bd5 fxg2 7.Bxg2 Qh4 and Black is just a pawn up.
b) 4.Bxg8?! Rxg8 5.Qh5+?! (5.exf5?! d5 6.Qh5+ g6 7.fxg6 Rxg6, blackcrowman-tsmenace,
ChessWorld.net 2008 is even worse; White should play 5.d3 d6, though I still prefer Black) 5...g6
6.Qxh7 Rg7 7.Qh8 Nd4 (I’d given this far in Chess Mail and had actually analysed further) 8.Kd1 (or
8.d3 f4) 8...Qg5 9.Qh3 (or 9.g3 fxe4 10.Nxe4 Qf5) 9...d5! 10.Qg3 (or 10.Nxd5 fxe4) 10...Qh5+ 11.f3
fxe4 12.Nxd5 Rf7 and wins, all according to my 1997 notes. This was backed up some 20 years later
by 13.c3 exf3 14.Nxf3 Nxf3 15.gxf3 Rxf3 16.Ke1 Be7 0-1 J.Shepley-J.Tait, corr. 2019.
c) 4.exf5!? Nf6 is more interesting and can arise via 3.exf5 in Chapter Two.

78
 White should probably play 5.Nf3 (the immediate 5.d4!? Nxd4 6.Nf3 Nxf3+ 7.Qxf3 c6 8.Bg5 d5
9.0-0-0 Be7 10.Kb1 Qc7 11.Bxd5!? cxd5 12.Nxd5 Nxd5 13.Qh5+ Kf8 14.Rxd5 Bd7 15.Bxe7+ Kxe7
16.Rhd1 Bc6 didn’t get me anywhere in tsmenace-juliangon, ChessWorld.net 2015; and 5.Nd5 d6
6.Nxf6+ Qxf6 7.Qh5+ Kd8 is nothing to worry about; for 5.d3 see note ‘b’ below) 5...d5 6.Bb5 Bd6
7.d4 exd4 8.Nxd4 0-0 9.Bxc6 (not 9.Nxc6?! Qe8+ and Black is better, jana84-tsmenace,
ChessWorld.net 2011) 9...bxc6 10.0-0 Re8 (Stockfish prefers 10...Bd7! 11.Bg5 Qe8 12.Re1 Qf7 and
says “0.00”; there are ideas of ...Rab8, or ...Rae8 and ...Be5, or ...c6-c5) 11.Bg5 c5!? 12.Ne6 Bxe6
13.fxe6 c6 14.Re1 and White held an edge in Clay56-tsmenace, ChessWorld.net 2014, though I
managed to draw.

4...Nf6

79
 Now White has two main options:
B1: 5.Nf3
B2: 5.a3
Other moves:
a) 5.f4!? transposes back to Chapter Two (see 5.Nc3 in the notes to line F1).
b) 5.exf5 d5!? (or 5...Bb4 6.Bg5, when 6...d5!? 7.Bxf6 gxf6 8.Qh5+ Ke7 9.Bb3 Nd4 10.0-0-0 c6
11.f4 Qe8! 12.Qh4 Nxf5 was quite unclear, M.Kivisto-A.Pyhala, Finnish Ch., Pori 1986) 6.Bb5 (if
6.Nxd5!? Nxd5 7.Qh5+ Ke7, Stockfish gives White no more than a forced draw; i.e. 8.Bg5+ Nf6
9.Bxf6+ gxf6 10.Qf7+ Kd6 11.Qd5+ Ke7 etc) 6...Bxf5 7.Bg5?! (7.Nf3 Bg4 8.h3 Bxf3 9.Qxf3 is
equal) 7...Bb4 8.Bd2?! 0-0 and Black was clearly better, P.Mercs-J.Tait, Notts Handicap rapid 2006.
c) 5.Nge2 is answered by 5...Na5! 6.Bb3 (Stockfish indicates 6.Bg5 c6 7.f4 Qb6 from its silicon
fantasy world) 6...Nxb3 7.axb3 Bc5!? (7...fxe4 8.dxe4 d6 is safe and equal) 8.0-0 (White should
accept the pawn: 8.exf5) 8...0-0 9.h3?! (here too) 9...d6 10.Kh1?! f4 and Black was again clearly
better, S.Brethauer-P.Lasinskas, Griesheim 2002.
d) 5.Bg5 requires more caution:

80
 5...Na5? 6.Nd5 Nxc4? 7.Bxf6 gxf6? 8.Qh5 mate would be embarrassing.
The routine 5...Bb4?! can also land Black in trouble: 6.Ne2! (intending Nd5 as soon as possible)
6...h6 (or 6...Na5 7.0-0 Nxc4 8.dxc4) 7.Bxf6 Qxf6 8.0-0 Ne7 and now 9.exf5! (rather than 9.Nb5?
Ba5 10.b4 Bb6 11.a4 a6, as in A.George-J.Tait, Sheffield League 2010) 9...Qxf5 (not 9...Nxf5?
10.Nd5) 10.d4, opening up with the black king stuck in the middle.
Probably 5...d6 is necessary, covering the f5-pawn and planning ...Be7 to break the pin; e.g. 6.Nf3
(here 6.Bxf6 Qxf6 7.Nf3 Na5 8.Bd5 c6 9.b4 Qd8! 10.bxa5 Qxa5 11.Qd2 cxd5 12.Nxd5 Qxd2+
13.Kxd2 Rb8 worked out okay for Black, WaterDragon-tsmenace, ChessWorld.net 2013; while 6.a3
Be7 7.Nf3?! allows 7...Nxe4! 8.Bxe7 Nxc3 9.Bxd8 Nxd1 10.Bxc7 Nxb2 11.Bxd6 Nxc4 12.dxc4 e4
with equality) 6...Be7 (or just 6...Na5 7.Bb3 Nxb3 8.axb3 Be7, Sha.Smith-L.Zilbermintz, Parsippany
2010) 7.Nh4 fxe4 8.Bxf6 Bxf6 9.Qh5+ Kd7 10.Nxe4 Nd4 11.Qf7+ Qe7 12.Nxf6+ gxf6 13.0-0-0 b5
14.Qxe7+ Kxe7 15.Bd5 c6 16.Bf3 Rg8 and Black came out well, R.Boudesseul-M.Pellen, corr. 2008.

B1: 5.Nf3

This has been White’s most popular reply – overwhelmingly so if transpositions to this point are
included – and is the one on which GM Ovetchkin and IM Soloviov (O&S) concentrate in their The
Modern Vienna Game – the only book hitherto to give proper space to 3...f5!?.

5...Bb4

81
 Now that Nge2 is no longer an option for White, there is no reason to refrain from pinning the c3-
knight. Inserting 5...fxe4 6.dxe4 is also acceptable. The plan in any case is to exchange on e4 and c3
(doubling White’s pawns) then play ...d7-d6, ...Qe7, ...Be6, and castle. Once achieved, Black should
obtain a decent position, so White should try and interfere with this in some way.

6.0-0

Other moves:
a) 6.exf5!? d5 7.Bb5 0-0 8.0-0 (not 8.Bxc6 bxc6 9.Nxe5?? d4) 8...Bd6 9.Bxc6 (not 9.Nh4?! Nd4 or
9.Bg5?! Ne7) 9...bxc6 10.Re1 (not 10.Nh4?! Ne8) 10...Re8 11.Nh4 Rb8, when the two bishops and
central majority give Black good compensation for the pawn, which White may struggle to keep
anyway.
b) 6.Bg5 d6 (or 6...Bxc3+ etc) 7.a3 Bxc3+ 8.bxc3 fxe4 9.dxe4 Qe7 (or 9...h6) 10.Qe2 Nd8 11.Nh4
Nf7 12.Bd2 0-0 and Black was fine, F.Roeberg-F.Grafl, Frankfurt 2000.
c) 6.Bd2 d6 7.Qe2 (or 7.Ng5 Qe7) 7...Qe7 (or 7...Na5) 8.Nd5 Bxd2+ 9.Qxd2 Nxd5 10.Bxd5 Be6
11.Bxe6 Qxe6 and Black can castle either way, L.Bourcier-K.Shirazi, Issy les Moulineaux 2007.
d) 6.Ng5 Qe7

82
 7.0-0 (if 7.Bf7+ Kf8 8.Bb3 Nd4 9.exf5!? d5 10.0-0, samurai-tsmenace, ChessWorld.net 2011, just
10...Nxb3 11.axb3 Bxf5 should be fine; even 8...f4!? is possible) 7...Bxc3 8.bxc3 h6 9.Nf3 fxe4
10.Nh4 d5 11.dxe4? (instead, 11.Ng6 Qd6 12.Nxh8 Bg4 13.f3 exf3 14.gxf3 Bh5 15.Bb3 0-0-0 is
unclear; Stockfish says “0.00”) 11...dxc4 12.Ng6, F.Roeder-L.Henris, Ostende 1992, and now
12...Bg4! 13.f3 (or 13.Nxe7 Bxd1 14.Nxc6 Bxc2 15.Nxe5 Nxe4) 13...Qc5+ 14.Kh1 Bd7 15.Nxh8 0-
0-0 16.Nf7 Rf8 is good for Black, who picks up the second piece.

6...Bxc3

Not 6...d6? 7.Nd5! Bc5? 8.Bg5 and Black is in huge trouble, playing a KGD two tempi down,
I.Rabinovich-A.Flamberg, Triberg 1914.

7.bxc3 fxe4

Still not 7...d6?! 8.Ng5 Qe7, R.Liyanage-T.Beerdsen, Titled Arena (blitz) 2021, because of 9.exf5!
d5 (or 9...Bxf5 10.f4) 10.Bb5 (Stockfish likes 10.Bxd5 Nxd5 11.Qh5+ Kd7 12.Re1 even more) 10...0-
0 11.a4 (threatening Ba3) 11...Re8 12.g4 with an extra pawn.

8.dxe4

83
8...d6

Taking the e4-pawn would be suicidal: 8...Nxe4?? 9.Qd5 Nd6 10.Bg5 Ne7 11.Qxe5 Nxc4 12.Qxg7
Rg8 13.Qxh7 and wins, as shown by O&S. So the choice is between two planned moves: 8...d6 and
8...Qe7. They can transpose if Black plays both. The question is whether 9.Nh4 or 9.Ng5 in response
is more troublesome.
I have generally gone for 8...Qe7, directed against 9.Ng5, which can now be met 9...h6 10.Nf3 g5.
Instead, 9.Bg5 h6 10.Bxf6 Qxf6 costs a tempo, but Black is fine anyway; e.g. 11.Qd5 d6 12.a4 Nd8
13.Qb5+ Bd7 14.Qb3 a5 15.Rad1 b6, when Black has consolidated the queenside and is ready to put
something on e6, Hullplayer50-tsmenace, ChessWorld.net 2019.
Therefore 9.Nh4 is critical: 9...d6 (not 9...Na5? 10.Nf5) 10.Nf5 (if 10.f4 then 10...Bg4 and ...0-0-0;
while 10.Bg5 Na5 11.Bb5+ Nc6! 12.Bc4 Na5 led to a draw by repetition in Hullplayer50-tsmenace,
ChessWorld.net 2020) 10...Bxf5 (10...Qf8!? is not unthinkable) 11.exf5 0-0-0 12.Be6+ (not 12.f4?
d5, followed by ...Qc5+) 12...Kb8, with ideas of ...Rdf8 and ...Nd8. I think this should be okay for
Black, who of course has the two knights.

9.Ng5

Probing f7 and clearing the way for f2-f4. Instead:

a) 9.Nh4 can now be met by 9...Na5, since the f5-square is covered.
b) 9.Bg5 makes little sense in view of 9...h6.
c) 9.h3 Qe7 10.Bg5 Be6 11.Bxf6 Qxf6 12.Bd5 Nd8 13.Nh2 0-0 completes the plan, and after
14.Ng4 Qf7 15.Ne3 Bxd5 16.Qxd5 Qxd5 17.Nxd5 Ne6 Black is at least equal, R.Van Kemenade-
J.D.Turner, Welsh Ch., Cardiff 2019.

84
 d) 9.a4 Qe7 10.Qd3 (or 10.Nh4 Bg4 and if 11.Qd3 then 11...Be6!, now that 12.Nf5 can be answered
by 12...Bxc4) 10...Be6 11.Rb1 Bxc4 (or 11...Rb8 12.Ng5 Bxc4 13.Qxc4 Kd7 14.Qd3 Kc8 15.Nf3
Qe6 16.Re1 h6, which was level in R.M.Allen-J.Tait, Derbys League 1999) 12.Qxc4 Nd8 13.Re1 Qf7
14.Qb4 0-0 and, job done, Black is ready to fight for the initiative; e.g. 15.Ng5 Qd7 16.Qc4+ Kh8
17.f3 h6 18.Nh3 b6 19.Nf2 a5 20.Be3 Ne6 21.Rbd1 Rae8 22.Nd3 Qf7 23.Qc6 g5 24.Nf2 h5 25.Rf1
Qg6 26.Kh1 g4 with a strong attack, WaterDragon-tsmenace, ChessWorld.net 2011.

9...Qe7

Not 9...Na5? 10.Bf7+ Kf8 11.f4 (O&S) or 9...Rf8? 10.Nxh7! Rh8 (or 10...Nxh7 11.Qh5+) 11.Ng5
and White has won a pawn, S.Kosmo-E.Auvinen, Vantaa 2008.

10.Bf7+ Kd8

Not 10...Kf8? 11.Bd5, followed by f2-f4 (O&S).

11.Bd5

Ovetchkin & Soloviov stop here, giving White a slight edge. Stockfish plays:

11...Rf8

And says “0.00” again. I think I’d be happy to defend this as Black; e.g. 12.f4 Bg4 13.Qd3 h6
14.Bxc6 bxc6 15.Nf3 (or 15.fxe5 Qxe5 16.Bf4 Qc5+ 17.Be3 Qe5) 15...exf4 16.e5 Nd7 17.exd6 Qxd6
18.Nd4 g5 and White has sufficient compensation, but no more than that.

B2: 5.a3

85
 This useful little move takes the game into a reversed King’s Gambit Declined (KGD), where the
corresponding main idea of ...Na5 has been neutralized.

5...fxe4

Black trades pawns before Nxe4 becomes a good reply. It also avoids the possibility of a later e4xf5
or Nh4xf5. For example: 5...Bc5 6.Nf3 d6 (or if 6...h6, B.Sikora Gizynska-Z.Wiliczkiewicz, Polish
League 1993, then 7.Na4 Be7 8.Nh4 is good for White, even with two knights on the edge) 7.Bg5
(not 7.Ng5?! Ng4!) 7...h6 (or 7...f4 8.Nd5 Bg4 9.c3 a6 10.b4 Ba7 11.Ba2 h6 12.Bxf6 gxf6 13.Qb3
and White is clearly better, especially after 13...Bxf3? 14.gxf3 Rf8 15.Rg1, D.M.Adams-M.Lyell,
British Ch., Scarborough 2001) 8.Bxf6 Qxf6 9.Nd5 Qg6 is well met by 10.0-0! (whereas 10.Nxc7+
Kd8 11.Nxa8 Qxg2, followed by ...f5xe4, transposes to the main line; and 11.Nh4 Qg5 12.Nxa8
Qxh4 is also unclear) 10...Kd8 (or 10...Bb6 11.Nh4 Qg5 12.Nxf5 Bxf5 13.exf5 0-0-0 14.a4!) 11.Nh4
(or 11.exf5) 11...Qg5 12.Nxf5 Bxf5 13.exf5 Qxf5 14.b4 Bb6 15.c3 Ne7 16.a4 c6 17.Nxe7 Kxe7 18.a5
Bc7 19.b5 with a strong initiative for White.

6.dxe4 Bc5 7.Nf3 d6

86
 I’ve reached this position on half a dozen occasions and, as yet, none of my opponents has taken
one of the critical paths.

8.Bg5

The KGD standard move is probably best. Instead:

a) 8.h3 Nd4! 9.Nxd4 Bxd4 10.Qd3 Qe7 11.Nb5 Bb6 12.Bg5 Be6 13.0-0-0 Bxc4 14.Qxc4 Qf7
15.Qxf7+ Kxf7 16.f3 h6 17.Bd2 Ke6 was equal in bulent_2010-tsmenace, ChessWorld.net 2012.
b) 8.Na4 Bg4 (or 8...Bb6) 9.c3 (or 9.Nxc5 dxc5 10.Qxd8+ Rxd8 11.Bb5 Nxe4) 9...Bb6 10.0-0 Qd7
11.Nxb6 axb6 12.Re1 Be6 13.Qd3 Bxc4 14.Qxc4 Na5 15.Qa2 Qf7 16.Qxf7+ Kxf7 was at least equal
in view of White’s weak light squares, Reprimand-tsmenace, ChessWorld.net 2014.
c) 8.0-0 Bg4 9.Qd3 Qd7 10.b4 Bb6 11.Bg5 Rf8 12.Nd2 Nd4 13.h3 Be6 14.Rab1 Nh5 15.Nd5 c6
16.Nxb6 axb6 17.f4 b5 18.fxe5 dxe5 19.Rxf8+ Kxf8 20.Bxe6 Qxe6 21.c3 Qg6 22.Rf1+ Kg8 23.cxd4
Qxg5, when Black finally consolidated and the game ended with a repetition: 24.Rf5 Qg3 25.Rf3 Qg5
26.Rf5 Qg3 27.Rf3 ½-½ Hullplayer50-tsmenace, ChessWorld.net 2018.
d) 8.Ng5!? is another critical option:

87
 8...Qe7 (not 8...Rf8?! 9.Nxh7! Bxf2+ 10.Kf1 Nxh7 11.Qh5+ Ke7 12.Qxh7 Bh4+ 13.Ke2 Nd4+
14.Kd3 and White is much better, tsmenace-Tihomir Glowatzky, ChessWorld.net 2015) 9.Nd5! Nxd5
10.exd5, when Black should try 10...Nd4 (if 10...Nd8?! 11.Bd3 Nf7 then 12.Bb5+ is annoying, since
12...c6? 13.dxc6 0-0 loses to 14.Bc4) 11.c3 Nf5 12.Bb5+ c6!? (12...Bd7 allows the knight into e6,
and 12...Kf8 13.Ne4 Bb6 14.g4 is good for White) 13.dxc6 0-0 14.0-0 Kh8 with some, if perhaps not
really enough, compensation for the pawn; e.g. 15.Qd3 a6 16.cxb7 (or 16.Ba4 Ba7) 16...Bxb7 17.Bc4
h6 18.Ne4 Ba7 19.Bd5 Bxd5 20.Qxd5 Qh4.

8...h6

Obviously 8...Na5 (the thematic KGD reaction) would be pointless here; but either 8...Bg4 or
8...Rf8 might be worth investigating, as in note ‘c’ above.

9.Bxf6 Qxf6 10.Nd5 Qg6

88
 So we have an old KGD 6...Bg4 7.h3 main line with the extra move ...a7-a6. Or rather a2-a3. How
much influence does that have? We’ll see. At the very least the positions are quite fascinating to
study, especially if White grabs the rook in the corner.

11.Nxc7+!?

Not now 11.Nh4?? Qxe4+, while 11.Qe2 Kd8 (or 11...Bb6) 12.Nh4 (after 12.0-0-0 Rf8 or 12.c3
Rf8 Black has good play) 12...Qg5 13.Nf3 Qg6 is just a draw.
But as in the 5...Bc5 line above, the strongest defence is 11.0-0!. Now continuing in KGD fashion
with 11...Kd8 12.b4 Bh3 13.Nh4 Qxe4 leads to equality, while 13.Ne1 fails to 13...Nxb4 – except of
course it doesn’t. The earlier extra a2-a3 means Black can’t take the b4-pawn, while after 13...Bb6
14.Qd3 Be6 15.Nf3 White is clearly better, intending c2-c3 and a3-a4-a5, or 15...Ne7? 16.Nxe5 and
wins.
Instead, 11...Bg4 12.Nxc7+ Kd8 (suggested by Tartakower in the KGD) fails – as it does there – to
13.Nxa8 Nd4 14.Be2 Nxe2+ 15.Qxe2 Rf8 16.Kh1 Qh5 17.b4! Rxf3 18.Qc4.
It seems Black is obliged to go into 11...Bb6 12.a4! (surrendering the extra tempo in order to target
the b6-bishop – this would be good in the KGD too; only the equivalent of 12.Qd3 has been seen in
practice, which allows an escape with 12...Ne7!) 12...Bh3 13.Ne1 (after 13.Nh4 Qxe4 14.a5 Qxc4
15.axb6 0-0 16.Nxc7 Qxh4 17.Nxa8 Nd4, things can get a bit messy; e.g. 18.Ra3 Bf5 19.c3 Bg4 20.f3
Nxf3+! 21.gxf3 Bh3 22.bxa7 Qg5+ 23.Kf2 e4) 13...Be6 14.c3 (intending b2-b4 and a4-a5 again)
14...Qf7 15.b4 a5 16.Nxb6 cxb6 17.Bxe6 Qxe6 18.Nc2 0-0 and, despite having managed to castle,
Black is worse in view of the wrecked pawn structure.

11...Kd8 12.Nxa8 Qxg2

89
13.Rf1

13.Kd2 can be met by 13...Rf8 (or 13...Bg4 14.Rg1 Qxf2+ 15.Qe2 Be3+! 16.Kd3 Qxe2+ 17.Kxe2
Bxg1 18.Rxg1 Nd4+ and equalizes) 14.Rg1 Qxf2+ 15.Kc1 Rxf3 (not 15...Qe3+?? 16.Kb1 Qxf3
17.Qxf3 Rxf3 18.Rxg7 as there are no back rank mating threats) 16.Rxg7 Re3 17.Rg8+ Ke7 18.Rg7+
Kd8 with a draw.

13...Bg4 14.Be2 Rf8

In the KGD, theory assesses this as good for the attacker (there White), with the best option being
14...Qd4. Following GM Nigel Davies’ analysis (in reverse) gives us 15.Qd5 Bxf3 16.0-0-0 Qg5+
17.Kb1 Bxe2 18.Qxc5 Kc8 19.Rxd6 Bxf1 20.Rxc6+ bxc6, when 21.Qxc6+ Kb8 22.Qd6+ Kxa8
23.Qxf8+ Kb7 24.Qb4+ Kc6 25.Qa4+ Bb5 26.Qxa7 may be okay for White, since 26...Qg1+ is not
mate.
However, the earlier a2-a3 allows a far more testing continuation, where Black has to be extremely
precise:

15.b4! Bxf3 16.Bxf3

16.bxc5 Nd4 17.Bxf3 Rxf3 transposes; while 16.c3 has to be answered by 16...Bd4! 17.Rc1
(17.cxd4?? Nxd4 wins) 17...Rf4 18.Bxf3 Rxf3 19.cxd4 Rxa3 20.dxe5 (not 20.f3?? Re3+) 20...Qxe4+
21.Qe2 Qxb4+ 22.Qd2 Qe4+ with another draw.

16...Rxf3 17.bxc5 Nd4

90
 Despite the missing rook (or exchange, looking at the stranded a8-knight), it seems that Black has
sufficient play to hold.
For example: 18.Qd2 (not 18.cxd6? Qg4!) 18...Qh3! (enabling ...Rc3) 19.0-0-0 (or 19.c6 bxc6
20.Qa5+ Ke8 21.Nc7+ Kf7 22.0-0-0 Rc3 23.Kb1 Rb3+! and the king has to go back) 19...Rc3 20.Kb1
(or 20.Kb2 Rxc2+ 21.Qxc2 Nxc2 22.Kxc2 Qxa3) 20...Rxc5 21.f3 (similarly 21.Qe3 Qd7! or 21.Qd3
Qc8!) 21...Qd7! (coming round another way) 22.Rc1 Qa4 23.c4 (alternatively, 23.Qd3 Nb5 24.Qb3
Nxa3+ 25.Ka2 Qa6, or 23.Kb2 Qb5+ 24.Qb4 Qe2 25.Rfe1 Rxc2+ 26.Ka1 Qd3) 23...Qxa3 24.Rc3
Qa4 25.Qb2 (or 25.Qa2 Rb5+) 25...Rxc4 26.Ra3 (or 26.Rxc4 Qxc4 27.Rd1 Nb5 28.Rc1 Qa4)
26...Qc6 27.Rc1 Rxc1+ 28.Qxc1 Qb5+ and so forth.
In all these variations Black creates enough annoyance for a draw, according to Stockfish anyway.
The chances of a human finding the right path over the board are, I’d guess, close to zero. Then again,
no one has ever gotten anywhere near this as White in any game ever. So you pays your money...

C: 3.f4

The Vienna Gambit. Just take the pawn!

3...exf4

91
 White now has four main systems: three ancient and one modern (line C23).
C1: 4.d4
C2: 4.Nf3
Other moves are worse:
a) 4.h4? was, according to Tim Harding, “described in Chess (1938) as ‘a move approved by
Emanuel Lasker though scorned by others!’” The others were right: 4...Nf6! 5.Bc4 (if either 5.d4 or
5.Nf3 then 5...Bb4) 5...d5! 6.Bxd5 (or 6.exd5 Nd4) 6...Nd4 7.Nce2 Bg4 is terrible for White.
b) 4.Bc4 is a bad Bishop’s Gambit:

4...Qh4+ 5.Kf1 Bc5! 6.Qe1 (not 6.g3? fxg3 7.Kg2 gxh2 8.Rxh2 Qf2+ 9.Kh1 Qxg1+ 10.Qxg1 Bxg1

92
11.Kxg1 Nf6 12.Nb5 Kd8 and Black won, H.Asauskas-M.Olszewski, Krakow 2005) 6...Qxe1+
7.Kxe1 Nd4 8.Bb3 Ne7 and even if White gets the pawn back, Black stands better, aiming to break
with ...d7-d5 or ...f7-f5 soon.

C1: 4.d4

The Steinitz Gambit might be regarded as a hyperclassical opening, where White prioritizes a big
centre ahead of any lesser considerations such as a wandering king or missing pawn. Let’s see how it
works out.

4...Qh4+ 5.Ke2 b6!

Black threatens a decisive check on a6. I’d always thought this was good for White – and winning
an easy game confirmed me in that view, until I looked at it with the engines. They’re monsters.

6.Qd2

Making room for the king to drop back.

The alternative is 6.Nb5 (blocking the diagonal) 6...Ba6 7.a4 0-0-0 8.Nf3, when Black has two
main options:

a) 8...Qe7 9.Kf2 (threatening Nxa7+) 9...Bb7! (not 9...g5? 10.Nxa7+ Nxa7 11.Bxa6+ Kb8 12.Re1,
W.Steinitz-Falk, Moscow 1896; nor 9...Kb8? 10.Bxf4 d6 11.Bd3 Nb4 12.Re1 and White is already
winning, J.Tait-D.M.Adams, Sheffield League 2011) 10.a5! (and not 10.Bxf4?, An.Martin-
Mi.Adams, London 1992, because of 10...Nf6! 11.Bd3 Nxe4+ 12.Bxe4 Qxe4 13.Bxc7 Nxd4!
14.Qxd4 Bc5 15.Nd6+ Kxc7 16.Nxe4 Bxe4 and Black has regained the extra pawn) 10...g5 11.axb6
axb6 12.Bd3 g4 13.Bxf4! (a belated Rosentreter type sacrifice) 13...gxf3 14.Qxf3 d6 15.d5! Ne5
16.Na7+ Kb8 17.Bb5! c6 (the point was 17...Nxf3 18.Nc6+ Kc8 19.Na7+ and draws; not 18...Bxc6??

93
19.Bxc6 with Ra8 mate to follow) 18.Nxc6+ Bxc6 19.Qa3 Bb7 20.Qa7+ Kc7 21.Ra6 Rb8 22.Qxb6+
Kc8 23.Ra3 Qc7! 24.Rc3 Bc6 25.Qxc7+ Kxc7 26.Bxe5 dxe5 27.Rxc6+ Kd8 28.Ra1 eventually led to
a draw in Redfish-LcZero, TCEC 15 Bonus 2019 (½-½, 78).
b) 8...Qg4! is stronger:

b1) 9.Kf2 (threatening Nxa7+) 9...Bb7! 10.Bd3 (if 10.a5 then 10...Nf6 11.Bd3 Nxa5 12.Bd2 d5
13.e5 Ne4+ 14.Kg1 a6; or 10.Be2 Nf6 11.Rf1 Nxe4+ 12.Kg1 Re8 13.Ne5 Qe6) 10...f5! (IM
Bangiev’s one move suggestion, holding ...a7-a6 in reserve; Bologan’s 10...Nf6 looks good too; but
not 10...a6 11.Nc3 Nf6?, A.Gavrilov-A.Potapov, Perm 1997, because of 12.h3! Qg3+ 13.Kg1,
threatening Ne2) 11.a5 (if 11.Re1, F.Wagner-P.Ackermann, corr. 2004, then 11...a6, or 11.e5 a6)
11...fxe4! (stronger than 11...Nf6, T.Tocklin-B.Ducoulombier, corr. 2014, or even 11...a6 here)
12.Bxe4 Nf6 13.Bxc6 dxc6! 14.Nxa7+ Kb8 15.axb6 Ne4+ 16.Kf1 Bd6 is winning for Black in view
of the stranded white knight; e.g. 17.Qe2 (or 17.c4 cxb6 18.Qa4 Qd7, or 17.bxc7+ Bxc7 18.b4 Bb6)
17...Rhe8 18.Ne5 (or 18.Qc4 Qd7) 18...Qxe2+ 19.Kxe2 Bxe5 20.dxe5 g5.
b2) 9.h3! Qg3 10.Qd2! is the engines’ improvement (on 10.Qe1, B.Lonyuk-J.Morau, corr. 2012,
where 10...Nf6! 11.Qxg3 fxg3 12.e5 Re8 13.Kd1 Ne4 14.Be3 Bb7 or 14.Nd6+ cxd6 15.Bxa6+ Kc7
16.Re1 Nf2+ 17.Kd2 dxe5 18.dxe5 d6 is good for Black), intending either Qxf4 or Kd1 (as after
6.Qd2 below).

94
 For example: 10...Nf6 (after 10...g5 11.Kd1 Nf6 12.Nxa7+! Kb8 13.Nxc6+ dxc6 14.Bd3! or
10...Re8 11.Kd1 Nf6 12.Nxa7+! Kb7 13.Nxc6 Bxf1 14.Nd8+ Rxd8 15.Rxf1 Nxe4 16.Qe2, White is
okay) 11.Qxf4 Nh5 12.Qe3 (not 12.Qg5?! Qxg5 13.Bxg5 Ng3+ 14.Ke3 Re8 15.Nxa7+ Kb7
16.Bxa6+ Kxa7 17.Bd3 Nxh1 18.Rxh1 f6 and White doesn’t have quite enough for the exchange;
while if 12.Qxf7 Nf6 13.e5 then 13...Bxb5+ 14.axb5 Nxd4+ 15.Nxd4 Qxe5+, or 13.Kd1 Bb7 14.Bg5
a6) 12...Re8 13.Kd1 Bb7 14.Bd2 Bb4!? (investing two tempi to lure the c-pawn forward; 14...f5 15.e5
a6 would be more human) 15.c3 Bf8 16.Rg1 a6 17.Na3 Nf6 18.Bd3 Na5 19.b4 Nxe4 20.Kc2 Nc6
21.d5 Nd8 22.Bxe4 f5 23.Be1 Qg6 24.Nh4 Qf6 25.Nxf5 Bxd5 26.Bxd5 Rxe3 27.Nxe3 Bxb4! 28.Rf1
Qg6+ 29.Kb2 Bxa3+ 30.Rxa3 Re8 31.c4 c6 32.Bf3 Nb7 was StockfishClassical-Komodo 14, TCEC
20 Bonus 2021. I’d rather play White here with rook and two bishops for queen and two pawns.
Stockfish itself says Black is clearly better, but it still made the draw (½-½, 92).

6...Ba6+ 7.Kd1 Bxf1 8.Nf3 Qh5 9.Rxf1 g5

95
 After a few fairly forced moves Black has managed to prop up the f4-pawn, whereas the big centre
doesn’t seem like enough for White.

10.Qd3!

A recent try, discouraging Black from castling long. White may follow with Ne2, breaking the pin
on the f3-knight (thus threatening d4-d5) and preparing g2-g3 and/or h2-h4 to challenge Black’s
kingside wedge.
Previous moves are clearly worse:
a) 10.d5 Ne5 11.Nb5 Bd6! 12.Nxd6+ (or 12.Qc3 Nf6 13.Nxc7+ Bxc7 14.Qxc7 Ng6 15.Qd6 Nxe4
16.Re1 f5 17.Qf6 g4 18.Qg5 Qxg5 19.Nxg5 0-0 20.Nxe4 fxe4 21.Rxe4 Rae8 and, despite having just
regained the pawn, White resigned in S.Korzun-M.Pech, corr. 2011) 12...cxd6 13.b3 (if 13.Qb4 then
13...g4, or 13.h4 h6 14.Qb4 Nf6! 15.Qxd6 Neg4 16.Qb4 Qg6 17.Qd4 0-0 18.e5 Ne4 – Bologan)
13...Nf6 14.Qe2 Qg6 15.Nxe5 dxe5 16.Re1 0-0 and Black went on to win, A.Hutton-M.Lewis, corr.
2017.
b) 10.Nb5 0-0-0 11.d5 can be ignored: 11...Nf6! (even better than 11...a6, as in Y.Meister-
A.Aleksandrov, USSR Ch., Moscow 1991) 12.dxc6 (or 12.Qe2 Bc5 13.Nc3 Rhe8) 12...Bc5! 13.c3
dxc6 14.Nbd4 Rhe8 with a crushing attack; e.g. 15.Kc2 Bxd4 16.cxd4 Rxe4 (threatening ...g5-g4)
17.Re1 (or 17.Qd3 g4 18.Re1 Rdxd4 19.Nxd4 Rxe1) 17...Rxe1 18.Nxe1 Re8 (threatening ...Re2)
19.Nf3 Nd5 (threatening ...Qg6+) 20.a3 g4 21.Ne5 f6 22.Nd3 g3 (threatening ...Re2 again) 23.Nxf4
Nxf4 24.Qxf4 gxh2 wins.
c) 10.Nd5 0-0-0 is good for Black. In chronological order: 11.c3 (W.Steinitz-W.Shipley, corr.
1899) 11...f5! 12.e5 (or 12.exf5 Bd6) 12...Qf7 13.c4 h6 consolidates. Or 11.Qe2 (M.Chigorin-I.Zybin,
corr. 1900) 11...f5! (Bologan) 12.e5 (or 12.exf5 Nge7) 12...Nge7 consolidates. Or 11.a4!? (J.Polasek-
M.Jurcik, Czech League 2018) 11...f5! 12.e5 (or 12.exf5 Qf7) 12...Nge7 13.Nxe7+ Bxe7 14.Qd3 a5

96
and Black is better.

10...Nge7

After 10...Bg7 11.Ne2 h6 12.g3 fxg3 13.Nxg3 Qg6, T.De Vassal-A.Sumets, corr. 2007, Stockfish
sees only a slight edge for Black.

11.Ne2

If 11.Nb5 Rc8 12.d5, then 12...Ne5 13.Qe2 a6 14.Nxe5 Qxe2+ 15.Kxe2 axb5 16.g3 d6 17.Nc6
Nxc6 18.dxc6 fxg3 and Black keeps the pawn.

11...f6

97
 Supporting the pawn chain in advance and ensuring control over e5 in case d4-d5 should follow.
The only practical examples are three high-level clashes at TCEC:
a) 12.g3 0-0-0 (Black considers it safe to castle now that the c3-knight has retreated; if 12...fxg3
13.Nxg3 Qg4 14.Bd2, White seems to have reasonable compensation) 13.gxf4 g4 14.Ng3 (or 14.Ne1
f5) 14...Qh3 15.Ne1 h5 16.Rh1 h4 17.Ne2 Qxd3+ 18.Nxd3 d5 19.e5 Nf5 20.c3 Be7 21.Bd2 Rdg8
22.Kc2 Nd8 23.Rag1 and White managed to hold, Houdini 6-Komodo 1970. TCEC 10 Superfinal
2017 (½-½, 135).
b) 12.h4 h6 13.Bd2 0-0-0!? (far more risky here; 13...Bg7 looks preferable, leaving a decision with
the king until later; e.g. 14.g3 fxg3 15.Nxg3 Qg4 16.Ne2 Qh3 17.Neg1 Qg2 18.c3 and then perhaps
18...0-0-0) 14.c4 Kb8 15.a4 f5 16.e5 d6 17.e6 d5 18.c5 bxc5 19.Qb5+ Ka8 20.Qxc5 Rc8 21.Nc3 Bg7
22.Nb5 Qg6 23.Ra3 Rhe8 24.Rf2 Bf6 25.Rb3 with strong play for the pawn, requiring Black to be
accurate to hold. Komodo 14-StockfishClassical, TCEC 20 Bonus 2021 (½-½, 61).
c) 12.Bd2 Bh6 (instead, 12...Qf7 is an option, in order to answer 13.g3 with 13...g4, or 13.h4 h6
14.c4 Bg7) 13.h4 g4 14.Ne1 d5 15.Nxf4 Bxf4 16.Rxf4 0-0-0 17.c3 Qxh4 18.exd5 Rxd5 19.Nc2 Qg5
20.Rf2 Qg7 21.Ne3 Rdd8 22.Kc2 Kb8 23.Raf1 Rhf8 24.Qe4 h5 25.b3 and White’s control of the f-
file and the backward pawn provided sufficient compensation, ClassicAra-Seer, TCEC 21 League 2
2021 (½-½, 83).
Five engine games, five draws, so maybe the Steinitz Gambit is playable after all. All the same, the
onus is on White to prove it after 5...b6!.

C2: 4.Nf3

By developing the knight White prevents the queen check of line C1 but allows Black an equally
strong option in ...g7-g5, supporting the pawn on f4.

4...g5

98
 There are now three main continuations, of which the comparatively rare 5.g3!? is certainly the best
try – in that White can draw with that one.
C21: 5.d4
C22: 5.h4
C23: 5.g3
Others:
a) 5.Bc4 allows Black a favourable Hanstein Gambit, which was favourable anyway; e.g. 5...Bg7
6.d4 d6 7.0-0 h6 8.Nd5 Nce7 and White had nothing, P.Faulkner-J.Tait, Notts Championship 1996.
However, Black can also go for more: 5...g4! 6.0-0 (6.d4 transposes to line C21) 6...gxf3 7.Qxf3
(instead, 7.d4 is line C21 again; while 7.Bxf7+ Kxf7 8.Qxf3 Qh4 9.d4 Nxd4 10.Qd3 Ne6 11.Bxf4
Nxf4, M.Vachier Lagrave-H.Nakamura, Chess.com blitz 2020, leaves Black with far too much after
12.Rxf4+ Qxf4 13.Rf1 Qxf1+ 14.Qxf1+ Ke8, even if it’s all sitting at home) 7...Qh4 8.d4 (or if
8.Nd5, J.Nebel-S.Ter Sahakyan, Titled Arena blitz 2021, then 8...Bc5+ 9.Kh1 Ne5 10.Qxf4 Qxf4
11.Rxf4 Kd8 or 10.Qc3 Ng4! wins) 8...Nxd4 9.Qd3 Ne6 10.Bxe6 (or 10.Nb5 Qe7 11.Bxf4 Nxf4
12.Rxf4 Kd8 13.Rxf7 Qc5+ 14.Rf2 Ne7 and so on) 10...dxe6 11.Bxf4 Qe7 12.Nb5 e5 13.Be3 a6
14.Nc3 Be6 and Black won, O.Westermann-D.Van Donk, corr. 2017.
b) 5.h3 takes the sting out of ...g5-g4 but is hardly useful otherwise: 5...Bg7 (5...d6 6.d4 h6 may be
more accurate; e.g. 7.d5 Ne5 8.g3 Qf6, as in Houdini 6-Stockfish, CCC 1 Rapid Rumble 2018) 6.d4
d6 7.g3!? (rather more promising than 7.Bc4?! h6 8.a3 Nf6 9.0-0 0-0 10.Re1 Nh5, when White is
virtually lost already, J.Murey-Y.Kosashvili, Israeli League 1999) 7...g4 and White defaulted in
nii_sama-tsmenace, ChessWorld.net 2017. I thought Black had the edge in any case; e.g. 8.hxg4
Bxg4 9.Bxf4 Nxd4 10.Bg2 Ne6 or 9.Bb5 fxg3 10.Rg1 h5.

C21: 5.d4

99
 The Pierce Gambit. It’s unsound, and all Black needs to know is a few key moves.

5...g4! 6.Bc4

If instead 6.Bxf4 gxf3 7.Qxf3 Nxd4 8.Qf2 Ne6 9.Be5, P.Zelbel-M.Geenen, Belgian League 2010,
then 9...Bc5! 10.Qf3 (if 10.Qf5 then 10...Ne7, or 10.Qd2 f6) 10...Bd4 and White can take on h8 if
they like.

6...gxf3

7.0-0

There is nothing better:

a) 7.Qxf3 Qh4+! (not yet 7...Nxd4? due to 8.Bxf7+ Kxf7 9.Qh5+ Kg7 10.Qg4+ Kf7 11.Qh5+ with
a draw) 8.g3 Nxd4 9.Qf2 Qf6 10.Bxf4 (or 10.Nd5 Qe5 11.0-0 Bc5 12.Bxf4 Nf3+ 13.Kg2 Bxf2
14.Bxe5 Nxe5 15.Nxc7+ Kd8 16.Nxa8 Bc5 – Sakaev) 10...Bb4 11.0-0 Bxc3 12.bxc3 Ne6 and Black
won, F.Kreideweiss-M.Kahl, corr. 1974.
b) 7.Bxf4!? is a blitz player’s trap: 7...fxg2?? (happy face) 8.Bxf7+! Kxf7 9.Qh5+ Kg7 10.Rg1 and
wins. But: 7...f2+! (sad face) 8.Kxf2 Bg7 (McDonald) 9.Rf1 (or 9.e5 d5! 10.Bxd5 Nge7 – Bologan)
9...Bxd4+ 10.Ke1 Qh4+ 11.g3 Bxc3+ 12.bxc3 Qe7 13.Qf3 d6 14.Bg5 Ne5 and White soon resigned,
J.Polasek-L.Cernousek, Prague 2015.

7...Nxd4!

This is Black’s key trick.

8.Bxf4

100
 White can’t take the knight because 8.Qxd4 Qg5 threatens both ...Bc5 and ...Qxg2 mate; 8.Bxf7+
Kxf7 9.Qxd4 Qg5 doesn’t help; and 8.Rxf3 Nxf3+ 9.Qxf3 Bh6 10.e5 Qe7 is hopeless, P.Sorenfors-
Th.Johansson, Lund 1998.

8...Bg7!

Simpler than 8...Bc5, when 9.Bxf7+! Kxf7 10.Be3 Ke8! 11.Bxd4 Bxd4+ 12.Qxd4 Qf6 13.Qd3 isn’t
straightforward to defend, even if Black managed to do so in J.Polasek-T.Karolyi, Prague 1988.

9.Be3

Again, White has nothing better:

a) 9.Bxf7+ Kxf7 10.e5 (10.Be3 c5 is clearly no good) 10...Nf5! (stronger than 10...d5, as given by
Konstantinopolsky & Lepeshkin; the knight provides a sturdy barricade on the f-file) 11.Qd5+ (or
11.Qxf3 d6 12.Rad1 h5 13.Ne4 Qe8) 11...Kf8 12.Qxf3 d6 13.Rad1 (or 13.g4 dxe5) 13...h5 14.exd6
Nf6 15.dxc7 Qe7 and Black is winning; e.g. 16.Nd5 Qe4, or 16.Rd8+ Kf7, or 16.Rfe1 Qc5+ 17.Kh1
Kf7, or 16.Bd6 Nxd6 17.Nd5 Bg4! 18.Nxe7 Bxf3 19.Ng6+ Kf7 20.Nxh8 Ke6 21.Rxf3 Bxh8.
b) 9.e5 d5! 10.exd6, as in Mi.Tseitlin-G.Ciolac, Wattens 1992, can be met by 10...cxd6 11.Be3 Qh4
12.g3 Qg4 (Bologan); e.g. 13.Nb5 Nxb5 14.Bxb5+ Bd7 15.Bxd7+ Qxd7 16.Qxf3 Be5 with a solid
extra piece. Taking on d5 is no better: 10.Nxd5 Ne2+ 11.Kh1 fxg2+ 12.Kxg2 Nxf4+ 13.Rxf4,
M.Bilalic-M.Zerdo, Bosnian League 2003, and now 13...Be6; or if 10.Bxd5 c5 11.Nb5, S.Kosmo-
J.Kivijarvi, Finnish League 2005, then 11...Ne2+ 12.Kh1 fxg2+ 13.Bxg2 Qxd1 14.Raxd1 Nxf4
15.Rxf4 Ke7.

9...Qf6!

Not yet 9...Ne6? 10.Qxf3 Qf6 11.Qe2 Qg6, as in Gy.Horvath-Pa.Petran, Hungarian Ch., Budapest

101
1995, because 12.Nb5! then gives White strong counterplay; e.g. 12...Be5 (or 12...Kd8 13.Bxe6 fxe6
14.Rf3) 13.Rf5! a6 14.Rxe5 axb5 15.Bxe6 dxe6 (or 15...fxe6? 16.Rg5) 16.Qxb5+ c6 17.Qb6 Nf6 (or
17...f6 18.Rd1 Kf7 19.Bg5!) 18.Rd1 Bd7 (or 18...Nd7 19.Qd4) 19.Qxb7 Rd8 20.Bb6. If Stockfish had
a concept of fun it would realize it was enjoying itself very much.

10.Nd5 Qg6

This is from Bangiev & Hergert’s 1993 book on the 3.Nc3 King’s Gambit. Note that neither Nxc7+
nor Bxd4 is possible because of the mate threat at g2, and White’s pieces are somewhat obstructing
each other, which means Black is able to consolidate.
For example, 11.Rf2 (not 11.Nf4? Qxe4) 11...Ne6! now defends everything: 12.Rxf3 (or 12.Qxf3
Be5 and ...c7-c6, or else 13.Nf4 Ng5 14.Qg3 Qh6) 12...Be5 13.Rf5 f6 14.Be2 h5 15.Bxh5 Rxh5
16.Qxh5 Qxh5 17.Rxh5 Bxb2 with two pieces for the rook.
Even giving an exchange to remove the key knight fails to trouble Black: 11.Rxf3!? Nxf3+ 12.Qxf3
Kd8 13.Rf1 (or if 13.Bf4 d6 14.e5 then 14...c6 15.Nc3 dxe5 16.Rd1+ Bd7) 13...d6! (simpler than
13...c6 14.Nf4 Qd6, as in Y.Komissarov-Pa.Jones, corr. 2008, though Black still went on to win)
14.h4 (or 14.Qxf7 Bd7, followed by ...c7-c6 again) 14...c6 15.Bg5+ f6 16.Nf4 Qe8 17.Nh5 Qe7
18.Nxg7 Qxg7 19.Qc3 Bg4 20.Bxg8 Qxg8 21.Bxf6+ Kc7 and White is running out of ammunition.

C22: 5.h4

The Hamppe-Allgaier Gambit, which is probably unsound as well, but it’s a great deal more
complicated.

102
5...g4 6.Ng5

The retreat 6.Ng1 is best answered by 6...Nf6! with ideas of ...d7-d5 or ...Nh5; e.g. 7.d3 (if 7.d4
then 7...Bb4 8.Bd3 d5 9.e5 Nh5 10.Nge2 Be7 and ...Bxh4; or 7.e5 Nh5 8.Qxg4 Ng3 9.Rh3 d6
10.Qxf4 Bxh3, S.Nicolenco-T.Van Bommel, corr. 2009) 7...d5 8.Bxf4 d4 9.Nb5 (or 9.Nce2 Nh5)
9...Bb4+ 10.c3 (or 10.Bd2 Bxd2+ 11.Qxd2 a6 12.Na3 Be6) 10...dxc3 11.bxc3 Ba5 12.Ne2 a6 13.Na3
Nh5 14.Qd2 Be6 with a clear advantage, since White’s pieces are treading on each others’ toes.

6...h6 7.Nxf7 Kxf7

103
8.d4

Hoping to follow with Bxf4 and get something going on the f-file.
a) 8.Qxg4? is no good because of 8...Nf6 9.Qxf4 Bd6! 10.Qf2 (or 10.Qe3 Be5 11.Bc4+ Kg7 12.Ne2
Ng4 13.Qf3 d5, J.Cueto Rodriguez-T.Miranda Rodriguez, Havana 2009) 10...Be5! (Bangiev &
Hergert) and White has nothing for the piece; e.g. 11.Nd5 (or 11.Bc4+ Kg7 12.d4 Nxd4 13.Be3 c5)
11...Kg7 12.Nxf6 Qxf6 and Black won, blackcrowman-tsmenace, ChessWorld.net 2004. I’d
completely forgotten all this by the time it came round again: 10...Kg7 11.d4 Nh5? (11...Ng4! 12.Qf3
h5 is still good for Black) 12.e5 Nxe5 13.dxe5 Bxe5 14.Qf5 Nf6??, C.Dorrington-J.Tait, Notts
League 2016, when 15.Bxh6+! Rxh6 16.Qxe5 d6 17.0-0-0 would have won for White. Instead, more
mutual mistakes led to a draw.
b) 8.Bc4+ is well met by 8...d5! (jettisoning small material to accelerate development)

9.Bxd5+ (after 9.Nxd5 Ne5 10.Bb3 f3! 11.0-0 Qxh4 12.d4 g3 13.Rxf3+ Nxf3+ 14.Qxf3+,
M.Moskwinski-P.Sujkowski, Warsaw 2009, and 14...Kg6!, White has nothing) 9...Kg7 10.d4 (or
10.b4?! f3 11.gxf3 Ne5 – Burgess) 10...Bd6! (now jettisoning a tempo to facilitate a light square
blockade; this is stronger than 10...f3, transposing to 9.Bc4+ in the main line) 11.Bxc6 bxc6 12.e5 (or
12.h5 Ne7 13.0-0 Rf8 14.e5 Bb4 15.Ne4 Nd5 16.Qd3 Qh4 17.Ng3 Qg5 and Black won, nii_sama-
tsmenace, ChessWorld.net 2011) 12...Bb4 13.Bxf4 Be6 14.Qd3 (or 14.0-0 Bxc3 15.bxc3 Qxh4
16.Qd3 Ne7 17.Bg3 Qh5 18.Rf6 Bf5 and Black won, A.Shevtsov-L.Rosen, corr. 2009) 14...Ne7 15.0-
0-0 (or 15.0-0 Bxc3 16.bxc3 Nf5 – Bologan) 15...Bxc3 16.Qxc3 (or 16.bxc3 Qd5) 16...Nd5 and
White resigned, G.Welling-V.Mikhalevski, Gibraltar 2008.

8...f3!

This is a real spoiler to White’s plans.

104
9.gxf3

Again 9.Bc4+ is met by 9...d5! 10.Bxd5+ Kg7, here followed by 11.gxf3 Nf6! 12.Be3 (similarly
12.Bf4 Bd6!, intending 13.e5 Nxd5 14.Nxd5 Ba3 15.bxa3 Qxd5 and ...Rf8 – Bologan; or if 12.fxg4
then 12...Qe8 13.Bxc6 bxc6 14.Qd3 Bxg4 15.Bf4 Bb4 16.Be5 Rf8 defends) 12...Bd6!.

GM Bologan’s strategy. Black is willing to return the piece in order to consolidate a lesser
advantage. For example:
a) 13.e5 Nxe5 (or 13...Nxd5 14.Nxd5 Be7) 14.dxe5 Bxe5 15.Be4 (or 15.0-0 Re8) 15...Nxe4
16.Qxd8 Rxd8 17.fxe4 Be6 or 17...Rf8, intending 18.0-0-0 Rf3.
b) 13.Qe2 Bg3+ 14.Kd2 Bf4! 15.Bxf4 Nxd4 16.Qf2 Nxf3+ 17.Kc1 Nh5! 18.Be3 Rf8, followed by
...c7-c6.
c) 13.fxg4 is met by 13...Nb4! 14.Bb3 (or 14.g5 Nfxd5 15.gxh6+ Kh7 16.Nxd5 Nxd5 17.exd5 Bf5)
14...Nxg4, when Stockfish shows some nifty defensive ploys: 15.Qe2 (not 15.Rg1? Qxh4+) 15...Bg3+
(bringing the bishop round to g5 as much as winning a pawn) 16.Kd2 Bxh4 17.Rag1 (or 17.Qc4 Rf8!
18.Qxb4 Nxe3) 17...Bg5 18.Ke1 (or 18.Nd1 Re8 19.Qf3 Qxd4+ 20.Kc1 Qf6) 18...Nxe3 19.Qxe3 a5!
(targeting the b3-bishop and allowing the a8-rook to appear for the defence) 20.Qf4 (or 20.a3 Rf8
21.axb4 axb4 22.Nd1 Ra6 and ...Rg6) 20...Qe7 21.a3 a4 22.Bc4 (if 22.Rh5 then 22...Ra6! again, or
22.Nxa4 Rf8!) 22...Nxc2+ 23.Kd1 Rf8 24.Qd2 Nxd4 25.Qxd4+ Qf6 and Black has survived with an
extra pawn (if 26.Qc5 Bd7 27.Qxc7? then 27...Qd4+ 28.Kc2 Rac8 wins).

9...d5!

105
 GM Bologan’s improvement. Hitherto Black played 9...Be7, which dates back to I.Gunsberg-
G.H.Mackenzie, London 1886 (1-0, 30).

10.Nxd5

Other moves offer White little:

a) 10.Be3 Bb4!, followed by ...Be6.
b) 10.exd5 is met by 10...Bd6! again, this time intending 11.dxc6 Bg3+ 12.Ke2 Nf6 and ...Re8+
(Bologan).
c) 10.fxg4 dxe4 11.g5 (or 11.Bc4+ Ke8!) 11...hxg5 12.Bxg5 Qxd4 13.Qxd4 Nxd4 14.0-0-0 Nf3
15.Nxe4 Be7 16.Bc4+ Kg7 17.Be3 Bf5 and Black seems quite secure.

10...Nf6 11.Bc4 Kg7 12.Nxf6

Here 12.Be3 is a recent attempt to revive the variation, requiring more ingenuity from Stockfish:
12...Na5 13.Bd3 Nxd5 14.exd5 and now – rather than 14...Qxd5 15.Be4, followed by d4-d5 with
potential threats on the long diagonal, as in D.Jabot-J.Van Mechelen, corr. 2018 – the engine comes
up with 14...b5!, enabling the knight to rejoin the action.
For example: 15.Qe2 (not 15.Bxb5? Qxd5 16.Qe2 gxf3, or 15.fxg4 Qxd5 16.Rf1 Nc4) 15...Nc4
(15...Bb4+ 16.c3 Bd6 17.0-0-0 Re8 may be even better) 16.0-0-0 (or 16.Bxc4 bxc4 17.Qxc4 Bd6)
16...Nxe3 17.Qxe3 Bd6 18.fxg4 Bd7 with a clear advantage to Black, since the three pawns are not
worth the piece, and White is unable to get at the black king; e.g. 19.g5 (or 19.Qe4 Qe8) 19...hxg5
20.hxg5 (or 20.Rdg1 g4) 20...Qe7.

12...Qxf6 13.e5 Bb4+ 14.c3 Nxe5!

106
 GM Bologan writes: “Black has a strong initiative. I saw this theme frequently while analysing this
position. Black returns the piece to destroy White’s steamroller.”
I’ve had the opportunity to test this out and it does seem to be good for Black: 15.dxe5 Qxe5+
16.Qe2 Qg3+ 17.Qf2 Re8+ 18.Kf1 Qxf2+ 19.Kxf2 Bc5+! (playing to fix the kingside pawns, after
which Black’s greater activity will count) 20.Kg3 Bd6+ 21.f4 h5 (it’s surprisingly difficult for White
to develop and challenge the open files effectively) 22.Rd1 Bd7 23.Bd2 Rad8 24.Re1 Bc5 25.b4 Be6!
26.bxc5 Bxc4 27.Rxe8 Rxe8 (Black now has a decisive grip on the position) 28.f5 Kf6 29.a3 Kxf5
30.Bg5 (there’s no hope for White in the opposite-coloured bishop endgame after 30.Re1 Rxe1
31.Bxe1 Ke4) 30...Bd5 31.Rd1 Bc6 32.Kf2 Re5 33.Rd4 Rxc5 34.c4 Ra5 (the start of an enjoyable
rook tour) 35.Bc1 Ra6 36.Rf4+ Ke6 37.Bb2 Rb6 38.Rf6+ Ke7 39.Bd4 Rb3 40.Rg6 Rxa3 41.Rg5 Rh3
42.Re5+ Kd7 43.Rxh5 g3+ 0-1 AndyAndyO-tsmenace, ChessWorld.net 2017.

C23: 5.g3

107
 This is a variation of the Quaade Gambit (2.f4 exf4 3.Nf3 g5 4.Nc3!?). Remarkably, 5.g3 was
virtually unknown in this position until the noughties – and that’s the 21st century, not 19th century –
when it was tried by IM Paul Littlewood.

5...Bg7!

The exclamation mark is for simplicity. How often are you going to face this?
Sharper positions arise after 5...g4 6.Nh4 f3 7.d4 d6 8.Be3 Nf6 and then 9.Bd3 Be7 10.a3 or 9.Qd2
d5 10.e5 Ne4 11.Nxe4 dxe4 12.0-0-0 Qd5 13.c4 Qa5 14.Qxa5 Nxa5 15.d5. I’ve played these out to
numerous draws as White at ChessWorld.net – and to varied success in countless blitz games with
colours reversed (i.e. 3.g3 f5!? 4.exf5 Nf6 5.g4 g6 etc).

6.d4

Not 6.gxf4? g4 7.Ng5 h6 8.Nxf7 Qh4+ and wins (Jakobetz & Somlai).

6...d6

108
7.d5

Even though it means surrendering the e5-square, the pressure on d4 makes this advance almost
obligatory:
a) 7.gxf4? g4 8.Rg1 Bf6 9.d5 gxf3 10.dxc6 Bh4+ was good for Black in M.Castiglione-M.Bagi,
Slovakian League 2017; and GM Shaw’s 8...gxf3! 9.Rxg7 Qh4+ 10.Rg3 Qxh2 11.Qxf3 Nxd4 is even
stronger.
b) 7.Bb5?! is well met by 7...a6! 8.Bxc6+ (not 8.Ba4? b5 9.Bb3 Bg4) 8...bxc6 9.0-0 (both 9.Qd3 g4
10.Nh4 Qf6 11.e5 dxe5 12.0-0 f3, R.Speckner-T.Oral, German League 2008; and 9.h3 h5 10.gxf4 g4
11.hxg4 hxg4 12.Rxh8 Bxh8 13.Ng5 f6 14.Nh3 Bg7 15.Nf2 f5, K.Zhdanenia-A.Thurrott, corr. 2013,
are good for Black too) 9...Bh3! 10.Rf2 (or 10.Re1 h6 11.e5 fxg3 12.exd6+ Kf8!) 10...fxg3 11.hxg3
h6 with a clear advantage (Shaw); e.g. 12.Qd3 Qd7 13.e5 Ne7 14.Ne4 Bf5 15.Qe2 Qe6 16.d5? cxd5
17.Nd4 Qxe5 18.Nxf5 Nxf5 19.Rxf5 Qxe4 and Black won, D.Plassmann-Chr.Richter, Emsdetten
2015.

7...Ne5 8.gxf4 gxf4 9.Bxf4 Bg4

109
10.Be2

Forcing the black king to move first doesn’t help White: 10.Bb5+ Kf8 11.Be2 Nxf3+ 12.Bxf3
Bxc3+ (here 12...Qf6 13.Bxd6+ cxd6 14.Bxg4 is equal) 13.bxc3 Qf6 14.0-0 (or 14.Bxg4 Qxf4 15.Qf3
Qxf3 16.Bxf3 Nf6; not 15.Qd4?? Qxg4 16.Qxh8 Re8! and Black wins) 14...Qxf4 (if 14...Bh3, Shaw’s
suggested sac 15.Qd2 Bxf1 16.Rxf1 looks okay for White) 15.Bxg4 Qxe4 16.Bf5 Qe5 17.Qd2 (not
yet 17.Be6? Nf6 18.Qf3 Rg8+ 19.Kh1 Rg6) 17...Re8 18.Be6 Re7 and Black has what chances are
going; e.g. 19.Rae1 Qg7+ 20.Kh1 h5! (to safeguard the queen) 21.Rg1 Qh6 22.Qd4 Nf6 23.Ref1 Ng4
24.Qxa7 Qe3 25.Qd4 (or 25.Qxe3 Nxe3 26.Rf3 Nc4 and ...Ne5) 25...Qxd4 26.cxd4 Rg8, when
White’s weak pawns are an ongoing issue, J.Claridge-H.Hartl, corr. 2014.

10...Nxf3+

Alternatively, 10...Bxf3 11.Bxf3 Qf6 (or 11...Ne7 12.Bg5 h6 13.Bh4 Qd7, as in tsmenace-
Topdriver, ChessWorld.net 2019) 12.Bxe5 Qxe5 13.Qe2 0-0-0 14.0-0-0 Nf6 15.Kb1 Kb8 16.Rhg1
Rhg8 17.Bg4 Bh8 18.Bh3 Rxg1 19.Rxg1 a6 led to a draw in Stockfish-LcZero, TCEC 20 Superfinal
2021.

11.Bxf3

110
11...Qf6

Now if 11...Bxc3+ 12.bxc3 Qf6, White has 13.Bxg4 Qxf4 14.Qd4! Qxg4 15.Qxh8 since the king
still occupies e8. Then 15...0-0-0 16.Qxh7 Re8 17.Kd2 Qg2+ will be a draw.

12.0-0 Bh3 13.Bg2

Sacrificing the exchange with 13.Qd2? is no good here because of 13...Bxf1 14.Rxf1 Qd4+ and the
queens come off.

13...Bxg2 14.Kxg2 Qg6+ 15.Kh3 Bxc3 16.bxc3 0-0-0 17.Qg4+

The only way to ensure the white king’s safety.

17...Rd7

Instead, 17...Qxg4+ 18.Kxg4 Nf6+ 19.Kf5 Nh5 20.Be3 Rhg8 21.Rg1 Rg6 22.Rg5 Ng7+ 23.Kf4
was level in tsmenace-AlphaZero, ChessWorld.net 2018 (no, not that AlphaZero); as was 17...Kb8
18.Rae1 Nf6 19.Qxg6 hxg6+ 20.Kg2 Rde8 21.Be3 Ng4 22.Rxf7 Rxh2+ 23.Kg3 Ne5 24.Kxh2 Nxf7
25.Kg3 Ne5 26.Bd4 Nc4 27.Kf4, Stockfish-Komodo, TCEC 13 Superfinal 2018.
Stockfish 13 claims a slight edge for Black after the text, presumably in view of White’s wrecked
structure. It may be negligible but I doubt a human would enjoy the defence. LcZero-KomodoDragon,
TCEC 20 Bonus 2021, continued 18.Bg5 f6 19.Bh4 Qe8 20.c4 Kb8 21.Rae1 h5 22.Qf5 b6 23.c5!
(computer tactics; 23.Bxf6 Nxf6 24.Qxf6 Rd8 and ...Qa4 is good for Black) 23...dxc5 24.Qe6 Re7
25.Qc6 Qc8+ 26.Kg2 Qg4+ 27.Kh1 Qxh4 28.d6 Reh7 29.Qe8+ Kb7 30.d7 Ne7 31.d8=Q Rxe8
32.Qxe8 Nc6 33.Qg6 Re7 34.Re3 a5 35.a4 and was now unclear. The game still ended in a draw.

111
 Chapter Four
Bishop’s Gambit
(and other King’s Gambits)

1.e4 e5 2.f4 exf4

Former King’s Gambiteer GM Mark Hebden once gave me a helpful tip: “2 f4 loses a pawn.” Well,
I guess I’d better take it then.
In online games you then see all sorts of intriguing continuations. I’ve faced variously 3.b3, 3.d3,
3.e5, 3.g3, 3.Qg4, 3.h4 and 3.Qh5.
Of those, 3.Qg4 is perhaps the only one worth mentioning, when 3...d5 4.Qxf4 Bd6! 5.e5 (or 5.Qe3
Qh4+ and ...Qxe4) 5...Qe7 6.d4 f6 regains the pawn, as in CelticKing-tsmenace, ChessWorld.net
2004. GM Hebden was right!
More sensibly, 3.Nf3 is the main move, and this is examined in detail in the next chapter. In this
one we’ll look at five other serious options for White:
A: 3.Nc3
B: 3.Bc4
The remaining three, in ascending order of occurrence in the databases, are:
a) 3.Be2 (Petroff’s Gambit) can be met by 3...f5!, in keeping with line B below. I’ve never had the
chance to play this one myself, but 3...f5 here will be no surprise to anyone because it’s now
recommended everywhere.
My own investigations have little to add to those of GMs Shaw and Bologan:

112
 a1) 4.e5 d6 5.Nf3 (or 5.exd6 Qh4+ 6.Kf1 Bxd6 7.Nf3 transposing; while after 5.d4 dxe5 6.dxe5,
J.Mieses-G.Maróczy, Vienna 1903, and 6...Qxd1+! 7.Bxd1 Nc6 8.Bxf4 Bd7 and ...0-0-0, Black will
soon target the white e-pawn; or if 9.e6 Bxe6 10.Bxc7 Nf6, White is left disorganized and trailing in
development) 5...dxe5 6.Nxe5 Qh4+ 7.Kf1 Bd6 8.Nf3, H.Bird-L.Weiss, Bradford 1888, and now
8...Qh6! (Shaw) is very good for Black; e.g. 9.d4 Nf6 10.Qe1 (or 10.c4 c6) 10...0-0 11.Bc4+ Kh8
12.Ne5 Nc6! (Bologan) and if 13.Nf7+ Rxf7 14.Bxf7 then 14...b6 with ideas of ...Ba6/Bb7, ...Rf8,
...Nxd4 and suchlike.
a2) 4.exf5 Qh4+ 5.Kf1 d5 6.Nf3 (the check 6.Bh5+ Kd8 hardly signifies; and 6.Nc3 c6 7.d4 Bd6
8.Nf3 Qf6 will win the f5-pawn in view of 9.g4 h5! – Shaw; or similarly 8.Bd3 Ne7 9.Qe2, F.Crowl-
C.Purdy, corr. 1945, and 9...Qf6! 10.g4 h5!, while 10.Qh5+ g6 11.fxg6 hxg6 12.Qf3 g5 is even worse
for White) 6...Qf6 7.c4!? at least tries to make a game of it, and 7...d4 8.d3 Ne7 9.Bxf4 Nxf5 10.Nbd2
Ne3+ 11.Bxe3 dxe3 12.Ne4 Qf4 13.Kg1 Be7 14.Nc3 Be6 15.Qb3 Nd7 16.Qxb7 Rb8 17.Qc6 Rb6
18.Qa8+ Bd8 19.b3 wasn’t fatal in Tucano 9-Seer 1, TCEC 20 League 4 2020. However, Black was
rated a mere 3197 there; Stockfish (at 400 Elo higher) prefers 8...Nc6! 9.Bxf4 Bd6 and claims a close
to decisive advantage.
b) 3.Qf3 (the Breyer Gambit) has a unique solution in 3...Nc6!, aiming for the unguarded c2-square:

113
 b1) 4.Ne2 d5 5.exd5 Nb4!, for example; and if 6.Nxf4?, as in G.Welling-J.Gustafsson, Dresden
2001, Stockfish gives 6...Qh4+! 7.g3 Nxc2+ 8.Kd1 Bg4 or 8.Kf2 Bc5+ as completely winning.
b2) 4.c3 runs into 4...Qh4+! too, followed this time by 5.g3 fxg3 6.hxg3 Qe7 7.d4 d5 8.e5 f6 and
White’s centre is collapsing; e.g. 9.Bd3 fxe5 10.Bxh7 exd4+ 11.Kf1 Ne5 12.Qf4 Nf6 13.cxd4 Nf7
and the pinned bishop will fall sooner or later, A.Kuindzhy-G.Timoscenko, World Seniors Ch.,
Rijeka 2011.
b3) 4.Qxf4 is well met by 4...Bd6! (most authors go for 4...d5 here, perhaps because it’s simpler;
GM Neil McDonald is an exception) 5.Qe3 (there’s no better square; 5...Nf6 would follow anyway)
5...Nf6 6.d4 (both 6.Be2 Qe7 7.Nc3 Bc5, E.Paoli-L.Prins, Venice 1949; and 6.Nf3 0-0 7.d3 Be7 8.c4
Ng4 9.Qd2 f5, Wl.Schmidt-Mat.Bartel, Radom rapid 2015, are already very good for Black) 6...0-0
7.Bd3, as in Ethereal 11-Nemorino 5, TCEC Cup 4 2019, when Stockfish again shows its superiority,
proposing 7...Bb4+! 8.c3 d5 (offering a piece to open the centre) 9.e5 (if 9.cxb4 dxe4 10.Bc2 Nxd4
11.Bb3 Ng4 12.Qg3 e3, Black has threats everywhere) 9...Ng4 10.Qd2 Ba5 (or just 10...f6!?) 11.Nf3
f6 12.h3 (here 12.exf6 Re8+ is almost as bad; while after 12.e6 Qd6! 13.Qc2 Qxe6+ 14.Kd1 f5 Black
is a safe pawn up) 12...Ngxe5! 13.dxe5 fxe5 and White is unlikely to survive.
c) 3.d4 is a nameless variation, dating back to the 17th century (or earlier). Even in those days
Black knew to play 3...Qh4+!.

114
 c1) 4.g3?? fxg3 5.Nf3 (unfortunately the king’s leap to g2 is no longer legal) 5...g2+ 6.Nxh4
gxh1=Q wins, as even occurred in the 2007 European Blitz Championship.
c2) 4.Kd2 Qf2+! 5.Kd3 (neither 5.Qe2 Qxd4+, nor 5.Ne2 Nf6 6.Nbc3 Bb4 is really any better)
5...Nc6 6.Qf3 (or 6.c3 d5 7.Qe2 dxe4+ 8.Qxe4+ Nge7 9.Nh3 Bxh3 10.gxh3 0-0-0 11.Qe2 Rxd4+! –
Bologan) 6...Qxd4+ won in J.Sartorius-K.Konuhov, corr. 2006. However, there was a nice mate in
six: 6...Nb4+! 7.Kc4 Qxc2+ 8.Qc3 b5+! 9.Kxb5 Rb8+ 10.Ka5 Nc6+ 11.Qxc6 Bb4 mate.
c3) 4.Ke2 d5! (if 4...Qe7 5.Kf2!, Black’s best option is 5...Qh4+ 6.Ke2 and then 6...d5 anyway)
5.Nf3 (or 5.exd5 Bg4+ 6.Nf3, transposing) 5...Bg4 6.exd5 (the only move; not 6.Qd3? Nf6 7.exd5
Bd6 8.c4 0-0 9.Nc3 Re8+ 10.Kd1 Qh5 11.Be2 Na6 12.a3 c6 13.c5 Bxc5! 14.dxc6 Rad8 and Black
soon won, G.Guerin-M.Hebden, Montpellier 2006; and I’ve had both 6.Nc3? dxe4 7.Nxe4 Nc6 and
6.e5? g5 in online games) 6...Bxf3+! (better than any of 6...Bd6 7.c4 b6, I.Nepomniachtchi-Ding
Liren, Chess.com blitz 2019, and then 8.Kd2 Qh6 9.Bd3; or 6...Nf6 7.c4 Bd6 8.Qe1, G.Quparadze-
A.Kashlinskaya, Titled Tuesday blitz 2020; or if 7...c6, A.Bove-C.Seyfried, Bergamo 2009, then
8.Nc3 Bb4 9.Bxf4 – in all of which White is perhaps only slightly worse) 7.gxf3 (not 7.Kxf3?? Qh5+
8.Kf2 Qxd1 9.Bb5+ c6, winning a piece) 7...Nd7 (aiming to castle quickly while the white king is still
flapping around in the middle) 8.Qe1 (or 8.c4 0-0-0) 8...Qf6 9.Kd1+ Ne7 10.Qe4 g5 11.h4 h6 and
Black’s kingside is secure, whereas the white centre is liable to be blown up at any time.

A: 3.Nc3

115
 There’s a tendency to name an opening after the first master to play it, even if they seem only to
have played it once – and lost. Thus 3.Qf3 above is attributed to Gyula Breyer and 3.Nc3 here to
James Mason. GM Shaw prefers to credit 3.Nc3 to the great GM Paul Keres, who contributed to its
development in numerous postal games in the 1930s. Whatever it’s called, 3.Nc3 is – how shall I put
it? – not very good.

3...Qh4+!

Stronger than 3...Nc6, which transposes to line C in Chapter Three.

4.Ke2

Chess is supposed to be fun, and this often gets a laugh from the opponent.

4...d5!

There are decent spoiler variations, such as 4...Qe7 (recommended by Shaw, Bologan and
Lokander) and 4...d6 5.Nf3 Bg4 6.Nd5 Qd8 (as GM Hebden once played against me), both asking
why the white king is sitting on e2. But there’s no real need when the main line is strong and forcing
for Black.

5.Nxd5 Bg4+ 6.Nf3 Nc6!

116
 Go on, take my rook.

7.Nxc7+

If the rook is declined, Black will just castle long and break with ...f7-f5:
a) 7.d4 0-0-0 8.c3 (or 8.Kd3 f5 9.Qe2? fxe4+ 10.Qxe4 Bxf3 11.Qxf3 Rxd5 0-1 F.Crowl-C.Purdy,
corr. 1936; even after 8.Bxf4 f5, E.Rutanen-T.Aalto, corr. 1981, and 9.Ke3! Qh5 10.h3 fxe4 11.Bc4
Bxf3 12.gxf3 Rxd5 13.fxe4 Qxd1 14.Raxd1 Rd8 White’s centre is not worth the missing piece) 8...f5
9.Qd3 (or 9.Qe1 Qxe1+ 10.Kxe1 fxe4 11.Ng5 Rxd5 12.Nf7 e3!, threatening 13...Nxd4! 14.cxd4
Bb4+) 9...Nf6 10.Nxf6, P.Keres-W.Kunerth, corr. 1935, and now 10...Qxf6!, intending 11.Bxf4 (or
11.e5 Nxe5!) 11...Nxd4+! 12.cxd4 Rxd4 gives Black a huge attack.
b) 7.c3 0-0-0 8.Qe1 Qh6! (naturally keeping the queens on) 9.d4 (or if 9.d3, V.Shikova-H.Kube,
Halle 1970, then 9...g5 and 10...f5) 9...f5 10.Kf2 (if 10.Bxf4 Qh5 or 10.Kd1 g5, then ...f5xe4 follows)
10...Bd6 (not now 10...g5?! 11.h4!) 11.Bc4 Bxf3 12.gxf3 Nf6 with strong play against the king in the
centre; e.g. 13.Nxf6 Qxf6 14.Bd5 Rhe8 15.h4 Bc5 16.Bxc6 Qxc6 17.dxc5 fxe4 18.Bxf4 exf3 19.Be3
Qe4.

7...Kd8

The king goes here, rather than d7, in order to avoid subsequent bishop checks.

8.Nxa8 Ne5!

117
9.Qe1

Not 9.d4? Nxf3 10.gxf3 Bxf3+ 11.Kxf3 Qh5+ 12.Kg2 Qxd1 13.Bd3 Qg4+ and Black wins. Notice
that wouldn’t be the case had the king gone to d7, as 13.Bh3+ or 13.Bb5+ would regain the queen and
win for White.
Instead, after 9.h3

9...Bh5! (not 9...Bxf3+? 10.gxf3 Qg3 11.d4 Qxf3+ 12.Ke1 Qg3+ and Black has to take the draw)
10.d4 Nxf3 11.gxf3 Bxf3+ 12.Kxf3 Qh5+ 13.Kg2 Qxd1, White can play 14.Bd3 (now there’s no
check on g4) 14...Qh5 15.Bxf4 with nominally equal material, but it will take time to organize and

118
extract the knight. Meanwhile Black can target the enemy king; e.g. 15...Ne7 16.Rhf1 Ng6 17.Bg3,
Y.Arkhipkin-J.Klovans, Riga 1974, and now 17...Nh4+! (Estrin & Glazkov) 18.Kh2 (or 18.Bxh4+
Qxh4 19.Rxf7 Qg5+ 20.Kh2 Qd2+ 21.Kh1 Qe3) 18...g5! 19.Nc7 Rg8 20.Nd5 g4 21.Nf4 Qh6,
threatening ...Ng6, when the human assessment is that Black has a clear advantage. Stockfish says it’s
winning.

9...Nxf3 10.Qxh4+ Nxh4+ 11.Ke1 g5!

Chess is a game of information. Black props up the f4-pawn and waits to see what White does
before deciding how best to continue; ...Kc8 or ...Nf6 is likely to follow.
Playing either first allows White to respond accordingly: 11...Kc8 12.d4 g5 13.Kf2 Kb8 14.g3 or
11...Nf6 12.e5 Nd5 (not 12...Ne4? 13.d3, J.Tait-A.Gardner, corr. 1994, as 13...Bb4+ fails to the
Jaenisch trick 14.c3 Nxc3 15.a3! Ba5 16.Bd2) 13.d4 Be7 14.Kf2 (Bangiev & Hergert) and at least the
white pieces come out. Whereas after 11...g5, development is a serious issue for White.

Building the big centre with 12.d4 now runs into 12...Nf6! 13.e5 Ne4; e.g. 14.Rg1 (or 14.Bd3 f5)
14...Kc8 15.Bd3 f5 16.exf6 Nxf6 17.Kf2 Bd6!, intending 18.g3 Nf3.
The little centre doesn’t solve White’s problems: 12.d3 Kc8! (not 12...Bg7? 13.c3! Kc8 14.g3! Bf3?
15.Rg1 Ng6 16.gxf4 gxf4 17.Bxf4 and White won, B.Hanison-D.Ter Haar, corr. 2000) 13.c3 (or
13.Bd2 Kb8 14.Bc3 f6 15.Kf2 Kxa8 16.g3 Ng6) 13...Kb8 14.g3 Ng6! and Black is clearly better.
Even 14...Nf3+!? is unpleasant: 15.Kf2 Bc5+ 16.d4 Nxd4! 17.b4 (not 17.cxd4?? Bxd4+ 18.Kg2 f3
mate) 17...Nf6! 18.bxc5 Nxe4+ 19.Kg1 Nf3+ 20.Kg2 Ne1+ and Black has at least a draw.
In a recent online game I tried an artificial plan of pushing the a-pawn, together with some obscure
manoeuvres on the queenside. They didn’t come to anything: 12.a4 Kc8 13.a5 Kb8 14.b3 Bg7 15.Ra4
Kxa8 16.g3 Ng6 17.gxf4 gxf4 18.a6 b6 19.Rg1 Bd7 20.Rb4 N8e7 21.Bb5 Bc8 22.d4 Bf6 23.Ke2 h5
24.c3 Bg4+ 25.Kd3 Bh4 26.Kc2 f3 27.Rf1 f6 28.Rc4 Kb8 29.Ba3 Rd8 30.h3 Bxh3 31.Rh1 f2 (yes, I

119
gave that pawn away on move two, didn’t I) 32.Rxh3 f1=Q 33.Rc8+ Nxc8 34.Bxf1 and White has
saved the immediate loss but is now just a piece down. I’ll give the rest because of the final move:
34...Rh8 35.Be2 Kc7 36.Bb2 Nf4 37.Rxh4 Nxe2 38.c4 Ne7 39.Rh2 Ng3 40.d5 Rh7 41.Bxf6 Nxe4
42.Bxe7 Rxe7 43.Rxh5 Nc5 44.b4 Nxa6 45.Kb3 Rg7 46.Ka4 Kc8 47.Rh8+ Kb7 48.Rh4 Nc7 49.Rd4
Rg3 50.d6 b5+ 51.cxb5 Na8! 0-1 tsmenace-ianl, ChessWorld.net 2020. “N-QR1, Resigns” – how
often has that happened in chess?

B: 3.Bc4

The venerable Bishop’s Gambit. As against 3.Be2 (note ‘a’ at the top), Black can now push the f-
pawn.

3...f5!?

Thomas Johansson said that Harro Dahlgrün said that this is called the Nordic Counter-Gambit. I
first became interested in it when investigating 2.Bc4 f5 (see Chapter Two), which might conceivably
transpose here after 3.f4!? exf4.
I find it strange how 3...f5 has been neglected. MegaBase supplies 75 games, but most of those are
from the 19th century. From 1925 to the present day there aren’t even two dozen. My own games add
up to more than double that (P52, W41, D7, L4). More significantly, a lot of the old, handed-down
theory is wrong.
Black will answer most moves with 4...Qh4+, which begs the question: why not the other way
round? The reason is that 3...Qh4+ 4.Kf1 f5 allows 5.Nf3, followed by e4-e5 and d4-d5 with
advantage.
After 3...f5, I’ve faced three moves most often:
B1: 4.exf5

120
 B2: 4.Nc3
B3: 4.Qe2
Others:
a) 4.Bxg8 Rxg8 (or 4...Qh4+ first – E.Schiller) 5.Qh5+ g6 6.Qxh7 Rg7 7.Qh8 (or 7.Qh6 fxe4
8.Qxf4 d5 with a clear advantage, Stan Evans-tsmenace, ChessWorld.net 2005) 7...fxe4 8.Ne2 f3!
9.0-0 Rf7 10.gxf3 exf3 11.Ng3 d5 and the f3-pawn is actually extra, C.Mayet-G.Neumann, Berlin
1866.
b) 4.e5 Qh4+ 5.Kf1 runs into 5...f3! 6.d4 (if 6.Bxg8? fxg2+ 7.Kxg2 Rxg8, Black is virtually
winning, A.Kaulla-A.Von Rothschild, Vienna 1887) 6...fxg2+ 7.Kxg2 b6! 8.Bd5 (or 8.Nf3 Bb7
9.Nc3 Qg4+ 10.Kf2 Be7 11.h4 Nc6) 8...Nc6 9.Nf3 Qh5 10.Ng5 Qxd1 11.Rxd1 Bb7 and Black is
clearly better (in view of 12.Nf7? Nge7).
c) 4.d3 Qh4+ 5.Kf1 fxe4 6.dxe4 (or 6.Nc3 e3) 6...Bc5 7.Qf3 d5! (the simpler 7...d6 8.Bxf4 Bg4
9.Qg3 Qxg3 10.Bxg3 Nf6 11.Nd2 Nc6 12.Ngf3 0-0-0 is good for Black too, P.Kalata-O.Pejko,
Slovakian League 2012; whereas 7...Nh6?!, C.Mayet-G.Neumann, Berlin 1865, and 8.g3!, not so
much) 8.g3 (not 8.exd5? Bg4 9.Qxf4 Ne7, or 8.Bxd5 Ne7 9.Bxf4? Nxd5 10.exd5 0-0 11.g3 Qf6)
8...Qh6 9.Bxd5 (or 9.exd5 Nf6 10.d6 Nc6 11.dxc7 Rf8) 9...Ne7 10.Nc3 (still not 10.Bxf4? g5)
10...Rf8 11.gxf4 g5 and White is in trouble down the f-file.
d) 4.Nh3 Qh4+ 5.Nf2 (Max Lange’s idea) 5...fxe4 6.0-0 (or if 6.Bxg8 Rxg8 7.Nc3, which
Anderssen also played, then 7...c6 and ...d7-d5) 6...Nf6 7.Nc3, A.Anderssen-C.Mayet, Berlin (match)
1855, and now 7...Bc5 (or even 7...f3!?) 8.d4 (or 8.Qe1 Ng4 9.Qxe4+ Kd8) 8...exd3 9.Qxd3 Nc6
looks very good for Black; e.g. 10.Nd5 (or 10.Nce4 Ng4) 10...d6 11.Nxc7+ (or 11.Bxf4 Nxd5 12.Bg3
Qd4) 11...Kd8 12.Nxa8 Ne5 13.Qd2 Nxc4 14.Qxf4 Qxf4 15.Bxf4 Bf5, followed by ...Kd7 and
...Rxa8.
e) 4.Qh5+ changes the play somewhat:

121
 4...g6 5.Qe2 (or 5.Qf3 fxe4 6.Qxe4+ transposing; not 6.Qxf4? Nf6 7.Nc3 Be7! 8.Nxe4 d5 9.Nxf6+
Bxf6 or 8.Nd5 Nxd5 9.Bxd5 Rf8 10.Qxe4 c6 11.Bb3 d5 12.Qe2 Rf6 13.Kd1 c5 and Black is clearly
better, Sellanraa-tsmenace, ChessWorld.net 2018) 5...fxe4 (note that Black can’t play as in B3 below
because f6 is now weak; e.g. 5...Qh4+?! 6.Kd1 fxe4 7.Qxe4+ Be7 8.Nf3 Qh5 9.Re1 Nc6 10.Nc3 Nf6?
11.Qxf4 d5? 12.Qxf6 and wins) 6.Qxe4+ (or 6.Nc3 Nf6 7.Nxe4 Nxe4 8.Qxe4+ Qe7 9.d3 Qxe4+
10.dxe4 Bg7) 6...Be7 7.Qd5? (Schulten played 7.Bxg8 Rxg8 in another game, where 8.Nc3 c6 9.d3
d5 10.Qxf4 minimizes Black’s edge) 7...Nh6 8.Qe5 Rf8 9.Qg7 Bh4+ 10.g3 fxg3 11.hxg3 Bxg3+
12.Kd1 Ng4 13.Rxh7 Qf6 and Black won, J.Schulten-B.Suhle, Berlin 1863.

B1: 4.exf5?!

I’m surprised by how often I get this. Already White needs to play precisely not to be losing out of
the opening.

4...Qh4+ 5.Kf1

5...c6

The simple choice, with which I’ve scored 10/10. Objectively, the discovery with 5...f3 may be
stronger, and then:
a) 6.d3?! is weak because after 6...fxg2+ 7.Kxg2 Ne7 Black has the dangerous idea of 8...b5!,
opening the long light diagonal; e.g. 8.Qe2, S.Cremon-V.Pasteur, corr. 1999, and 8...b5! 9.Bxb5
Bb7+ 10.Nf3 g5!? with strong play. Even if White prevents that with 8.Qf3, something like 8...Nbc6
9.Nc3 Ne5 10.Qf4 Qxf4 11.Bxf4 Nxc4 12.dxc4 d6 and ...Bxf5 will leave Black a pawn up for
nothing.
b) 6.d4 fxg2+ 7.Kxg2 is a better try, though there is no record of it ever being tested in practice.

122
The pertinent issues are White’s exposed king and weak f-pawn. Stockfish consequently gives Black a
clear advantage, but I’ve been unable to coax the engine into increasing that to “winning” in any
specific line. If anyone wants to try their hand, three possibilities I’ve looked at are:
b1) 7...Nf6 8.Nc3 (here 8.Qe2+?! Be7 only helps Black) 8...b5!? (anyway).
b2) 7...d5!? 8.Bxd5 Nf6, gaining time for development.
b3) 7...Nc6, keeping both ...b7-b5 and ...d7-d5 as options.

6.d4

Alternatively, 6.Nf3 Qf6 7.d4 d5 8.Qe2+ (not 8.Bd3? Bd6 9.Kf2 Bxf5 10.Re1+ Ne7 11.Kg1 0-0
12.Nc3 g5, kocoski_d-tsmenace, Yahoo.com blitz 2010) 8...Ne7 and 6.Qe2+ Ne7 7.Nf3 Qf6 8.d4 d5
both transpose below.

6...d5 7.Qe2+

Not 7.Bd3? Bd6 8.Qf3 Ne7 9.Qe2 0-0 10.Nf3 Qf6 11.g3 Bxf5 12.Kg2 Nd7, D.Ashworth-J.Tait,
Sheffield Plate 2013.

7...Ne7 8.Nf3

And not 8.Bd3? Qf6 9.Bxf4 Bxf5 10.Bxf5 Qxf5 11.Qf2 g5 12.Be5 Rg8, Reprimand-tsmenace,
ChessWorld.net 2015.

8...Qf6

This is the critical position after 5...c6. I’ve only had it in one game (via 6.Qe2+), which continued
9.Bd3? Bxf5 10.Bxf5 Qxf5 11.c3 Nd7 12.b4 g5, steveberkley-tsmenace, ChessWorld.net 2006.

123
 Note that Bd3 was the key mistake in all the 12-move extracts above. Nobody seems to have
considered leaving the bishop on c4 to be taken, yet Stockfish does just that with the instant 9.Bxf4!,
after which it’s no longer so simple:
a) 9...dxc4 10.g4! is the point, when the b1-knight rushing to e4 is very dangerous; e.g. 10...h5
11.Nbd2 hxg4 12.Ne4, when Black needs to bail out quickly: 12...gxf3 13.Nxf6+ gxf6 14.Qxf3 Bxf5
15.Re1 Na6 with three pieces for the queen so Black is not worse – but is not better either. Returning
the piece with 10...Bxf5 11.gxf5 Qxf5 12.Bxb8 Rxb8 13.Nbd2 Kd7 14.Re1 also seems okay for
White.
b) 9...Bxf5 (declining the piece) 10.Be5 Qh6 11.Nbd2! (still leaving the bishop en prise, since
11...dxc4 12.Nxc4 Be6 13.Nd6+ Kd8 14.Nxb7+ Kc8 15.Nd6+ is a draw) 11...Nd7 12.Bd3 (at last)
12...0-0-0 13.Re1 and again White seems okay.
Should anyone ever get close to this, I might go back and look at 5...f3 some more.

B2: 4.Nc3

Thomas Johansson is inclined to give this a dubious mark as well. Considering the next note below,
he may be right.

4...Qh4+

I’ve always put the check in, thinking that all other moves were inferior. However, Stockfish comes
out with one I’d never even made a candidate – 4...Nc6! – and claims an edge for Black.

It’s hard to see why until considering what White might do next. It turns out that the various
choices on move four are generally worse with the two knight moves inserted:
a) 5.Qe2 runs into 5...Nd4! 6.Qd3 Qh4+ 7.Kd1 Qg4+ 8.Nge2 Nxe2 9.Qxe2 Qxe2+ 10.Bxe2 fxe4,
when White will not regain both pawns; e.g. 11.Nxe4 d5 12.Nc3 c6 13.d4 Bd6 14.Rf1 g5.

124
 b) 5.exf5 is met by 5...Qe7+! 6.Kf1 (not 6.Nge2? f3, or 6.Qe2 Nd4 7.Qxe7+ Nxe7 8.Bd3 d5)
6...Nf6 7.d4 d6 8.Bxf4 Bxf5 9.Nf3 0-0-0 and White is stuck with the vulnerable king.
c) 5.Nf3, as in J.Flores Chico-M.Hamer, Gibraltar 2009, also puts White behind after 5...fxe4 6.Qe2
Be7 7.Qxe4 Nf6 8.Qxf4 d5 9.Bb3 0-0 (or even 9...Nh5!? 10.Qa4 Be6).
d) 5.d4 Qh4+ 6.Kf1 fxe4 7.Nxe4 Be7 transposes to 6...Be7 in the main line below.
e) 5.d3 may be best, even if 5...Qh4+ 6.Kf1 fxe4 7.dxe4 (or 7.Nxe4 Be7 8.g3 fxg3 9.Kg2 d6
10.hxg3 Qg4) 7...Bc5 8.Qe1 Qxe1+ 9.Kxe1 Nd4 10.Bd3 Ne6 11.Nge2 Bd6 12.Nd5 (or 12.Bc4 Nf6)
12...b6 13.Nexf4 Bb7 14.Rf1 0-0-0 still leaves Black with the chances.

5.Kf1 fxe4

6.Nxe4

White can also start with 6.Qe2, when 6...c6 7.Nxe4 is the main line.
Alternatively, Black can try 6...Be7 7.Nxe4 (not 7.Qxe4? Nf6, or 7.Nd5? c6! 8.Nc7+ Kd8 9.Nxa8
d5 with huge play for the rook; e.g. 10.Bb3 Nf6 11.d3 Bd6 12.Bd2 Bg4 13.Qe1 Qh5 14.Bc3 e3
15.Bd4 e2+ 16.Nxe2 Re8 17.Qa5+ b6 18.Qxa7 Bxe2+ 19.Kg1 Re7 and Black won, alapin-tsmenace,
ChessWorld.net 2015) 7...Nc6 (as in the next note), although here 8.b3! (not 8.c3?! Ne5 9.Bb3 d5!, or
8.Nf3?! Qh5 9.Neg5 Nh6) 8...Nd4 9.Qd3 Nf5 10.Bb2 seems okay for White; e.g. 10...c6 11.Bxg8
Rxg8 12.Nf3 Qh5 13.c4 d6 14.Re1 Kd8 15.Be5 Kc7 16.Bxf4 g5 17.Nxd6 gxf4 18.Nxf5 Bb4 19.N5d4
Bd7 with rough equality.

125
6...c6

The most natural, threatening a fork with ...d7-d5, but White can cause problems.
Here Stockfish prefers 6...Be7 from G.Maróczy-G.Marco, Vienna 1903, which it updates for Black:
7.d4?! (for 7.Qe2 see 6.Qe2 above) 7...Nc6! (Marco won with 7...Nh6 8.Nf3 Qh5 9.Bxf4 d5 10.Ng3?
Qf7, but White can certainly improve on that) 8.Nf3 (if 8.c3, then now 8...Nh6 9.Qe2 Rf8 and Black
is better; or 8.d5 f3! 9.gxf3 Ne5) 8...Qh5 9.Bxg8 (not 9.Bxf4? d5) 9...Rxg8 10.d5 b6! 11.dxc6 Ba6+
and ...0-0-0 with very strong play for the piece.

7.Qe2

If 7.Nf3 Qe7 8.Qe2, Black can decline the transposition (8...Kd8) in favour of 8...d5! 9.Nd6+ Kd8
10.Qxe7+ Bxe7 11.Nxc8 Kxc8; so White should opt for 8.Bxg8 Qxe4 (8...Rxg8?! 9.Kf2 is good for
White, J.D.Adair-J.Wilcox, corr. 1877) 9.Kf2 Qf5 10.Re1+ Kd8 11.Bb3 d5 12.d4 and c2-c4 with
compensation.

7...Kd8 8.Nf3 Qe7

Opposing queens on the file is preferable to the diagonal so that Black can cover f7 with ...Kxe7 if
necessary. After 8...Qh5? 9.Neg5! d5 10.Ne5 (Flear) 10...Qxe2+ (not 10...f3? 11.Nef7+ Kc7
12.Qe5+! Kb6 13.Qe3+ Kc7 14.Bd3 and wins) 11.Bxe2, Black has to waste time on 11...Ke8 12.d3
Nf6 (not 12...h6? 13.Bh5+) 13.Bxf4 and White is clearly better.

126
9.Neg5!

Escaping the fork in advance. Not 9.Bxg8?! Rxg8 10.h4 (or 10.d3 g5) 10...d5 11.Neg5, as in
hbgchess-tsmenace, ChessWorld.net 2008, since 11...h6! 12.Qxe7+ Kxe7 13.Nh3 Bxh3 14.Rxh3 Ke6
is good for Black.
IM Timothy Taylor’s suggested 9.d3!? d5 10.Bxf4! is more interesting; e.g. 10...dxe4 11.dxe4 Bg4
12.h3 Bxf3 13.Qxf3 Nd7 14.Rd1 Ke8 15.e5 Nc5? 16.Bg5! Qc7 17.Bd8! Rxd8 18.Rxd8+ Kxd8
19.Qxf8+ Kd7 20.Qxg7+ 1-0 Fritz 12-T.Taylor, test blitz game 2013. Stockfish 13 goes for 13...Nf6
in that sequence (e.g. 14.Rd1+ Kc8 15.e5 Nfd7) and hence 12.Rd1+ Kc8 13.h3 Bxf3 14.gxf3! for
White (e.g. 14...Nf6 15.Rg1), which may well be difficult for a human to defend.
From the other side Stockfish proposes 10...Bd7!, leaving the forked pieces alone for the moment;
e.g. 11.Ne5 Be8, or 11.Nfg5 h6 12.Qf2 Qe8, or 11.Neg5 dxc4 12.Qxe7+ Bxe7 13.Nf7+ Ke8 14.Nxh8
cxd3 15.cxd3 Nf6, intending ...Kf8-g8. If instead 11.Re1 (or 11.Qd2) 11...dxe4 12.Ng5 Nh6 13.Nxe4
(threatening Bg5), then 13...Kc8 14.Nd6+ Qxd6 15.Bxd6 Bxd6 and Black shouldn’t be worse with
three pieces for the queen.

9...d5

I’ve played this position (quite badly) a few times and it’s not as easy for Black as it seems. After
10.Bd3 (or 10.Qxe7+ Kxe7, intelektualac-tsmenace, ChessWorld.net 2007, and 11.Bd3 h6
transposing; not 10.Bb3?, osirysm-tsmenace, ChessWorld.net 2005, since 10...h6! 11.Qxe7+ Kxe7
12.Nh3 g5 is again good) 10...h6 (or 10...Nh6 11.Qxe7+ Bxe7 12.Bxh7) 11.Qxe7+ Kxe7 12.Nh7 Nf6
13.Nxf6 Kxf6 14.b3! White has sufficient play to hold the balance; e.g. 14...Bd6 (or 14...c5 15.Bb2+
d4 16.c3) 15.Bb2+ Kf7 16.Ne5+ Bxe5 17.Bxe5 b6 18.Bxf4 Ba6 and the “0.00” assessment appears.
I’ll be trying 4...Nc6! next time.

127
B3: 4.Qe2

Most books give this as best for White and are probably right to do so, though rarely for the right
reasons. In any case it’s by no means a refutation of 3...f5.

4...Qh4+

Not 4...fxe4? 5.Qh5+ g6 (or 5...Ke7?? 6.Qe5 mate) 6.Qe5+ Qe7 7.Qxh8 Nf6 8.Nc3 c6, as in
H.Pillsbury-F.Marshall, Vienna 1903, because of 9.d3! d5 (or 9...exd3+ 10.Kf1) 10.Bxf4 exd3+ (or
10...dxc4 11.dxe4) 11.Kf1 dxc4 12.Bg5 Qg7 13.Bxf6 Qxh8 14.Bxh8 dxc2 15.Nf3 and Black has
nowhere near enough for the rook.
Well, I wonder what Stockfish says this time. It wants to play 4...Nc6 again. This one has had a few
tests:

128
 a) 5.exf5+ Qe7 6.Nf3 (not 6.d3? Nd4 7.Qxe7+ Nxe7 and Black is already much better, M.Giroux-
F.Trani, corr. 2006) 6...d5 (or 6...Qxe2+ 7.Kxe2 Nge7 8.d3 d5 9.Bb3 Bxf5 10.Bxf4 0-0-0 – Morgado)
7.Bxd5 Qxe2+ 8.Kxe2 Nge7 9.Nc3 Bxf5 10.d3?! Nxd5 11.Nxd5 0-0-0 12.Nxf4 Bd6 with good play
against the king in the middle, P.Ryan-F.Trani, corr. 2006.
b) 5.Nf3 (preventing both ...Nd4 and ...Qh4+) 5...fxe4 6.Qxe4+ Qe7 7.d3 (or 7.Qxe7+ Ngxe7 8.Nc3
h6 9.d4 g5 10.h4 g4 11.Ne5 f3 and Black is fine, InANutshell-rooksac, ChessWorld.net 2007; and
7.Nc3 Qxe4+ 8.Nxe4 Nge7 9.Nc3 h6 transposes) 7...Qxe4+ 8.dxe4 Bd6 9.Nc3 a6 10.Ne2 b5 11.Bd5
Nge7 12.Bxf4 Nxd5 13.exd5 Bxf4 14.Nxf4 0-0 15.Ne6 (15.dxc6 Rxf4 and 15.g3 Nb4 are both equal
too) 15...dxe6 ½-½ D.Petrovic-E.Noire, corr. 2008, in view of 16.dxc6 e5 17.0-0 Bf5.

5.Kd1

Vacating this square for the king was the main point of Qe2 after all.
Nonetheless, White sometimes plays 5.Kf1 anyway;

129
 e.g. 5...fxe4 6.Qxe4+? (definitely a mistake; White should play 6.Nc3!, for which see 6.Qe2 in line
B2 above) 6...Be7 7.Nf3 (or 7.d4 Nf6) 7...Qh5 8.d4 (or 8.Bxg8 Rxg8 9.d4 d5 10.Qxf4 Rf8 11.Qe3
Bg4 12.Nbd2 Nc6 13.c3 0-0-0 14.b3 Rde8 0-1 RKoelbach-tsmenace, ChessWorld.net 2005; or
8.Qxf4 d5 9.Qxc7 Nc6 10.Bb5 Bd7 11.Nc3 Nf6 12.d4 0-0 13.Be3 Rac8 14.Qg3 Ne4 15.Qe1 Rxf3+ 0-
1 mizuzul-tsmenace, ChessWorld.net 2018) 8...Nf6 9.Qxf4 d5 10.Qxc7 Nc6 11.Bb5 0-0! 12.Bxc6
bxc6 13.Qe5 (or 13.Qxe7 Bh3!) 13...Qg6 14.Nbd2 (or 14.Qxe7 Ba6+ 15.Kg1 Rae8 16.Qxa7 Ng4
17.Qxa6 Rxf3 and wins) 14...Qxc2 15.Kg1 Ne4 16.Nxe4 dxe4 17.Bh6 Bf6 18.Rc1 Qe2 0-1 Pawn
cocktail-tsmenace, ChessWorld.net 2017. Pretty much all I did in those games was get my pieces out.

5...fxe4 6.Nc3!

Thomas Johansson identifies 6.Nc3 as “White’s best option at this point.” According to the
databases it was last played in 1894. After that: nothing – the 20th century is a complete void – until
the first of my own games in 2006.
The supposed main line runs 6.Qxe4+ Be7 7.Nf3 (instead, 7.Bxg8 Rxg8 8.Nf3 Qh5 transposes to
note ‘c’ below; Keres’ 7.d4 Nf6 8.Qxf4 Qxf4 9.Bxf4 d5 is not quite “gleich”, since White can’t castle
and Black has the initiative; e.g. 10.Bd3 0-0 11.Ne2 Ng4 12.Ke1 c5 13.c3 Nc6 14.h3 Nf6 15.Nd2
Bd7 16.Nf3 Rae8, seawasp-tsmenace, ChessWorld.net 2008) 7...Qh5 and now:

130
 a) 8.Qxf4 Nf6 (there’s no need to worry about c7; whereas 8...d6?! 9.Nc3 Nf6 10.d4 Bg4 is only
equal, tomrules99-tsmenace, ChessWorld.net 2008) 9.Qxc7 Nc6 gives Black very good play for the
pawn; e.g. 10.Re1 (or 10.d4 d5 11.Bb5 0-0! 12.Bxc6 Bg4!) 10...d5 11.Bb5 (or 11.Bb3 Bh3!) 11...0-0
12.Bxc6 (or 12.Rxe7 Nxe7 13.Qxe7 Bg4) 12...Bd8 13.Qe5 bxc6 14.Qxh5 Nxh5 15.d4 Bg4 and White
is struggling.
b) 8.Re1 Nc6 9.Bxg8 (if 9.b4 Nf6 10.Qe2 – W.C.Spencer, then just 10...a6) 9...Rxg8 10.Nc3 d6
11.Nd5 (11.d3 Bg4 is certainly not equal either) 11...Bf5 12.Qc4 (or 12.Nxf4 Qf7 13.Qd5 Bxc2+
14.Ke2 Bf5 and Black is a clear pawn up, JETHROTULL-tsmenace, ChessWorld.net 2006)
12...Bxc2+! 13.Ke2 (not 13.Qxc2? Qxd5 14.Qxh7 0-0-0, C.Mayet-G.Neumann, Berlin match 1866;
or 13.Kxc2? Qxd5 and wins, X.Hawkins-W.Shinkman, corr. 1875, in view of 14.Qxd5 Nb4+)
13...Bh4! (13...Bg6!? is good too) 14.Kf1+ (relatively best; after 14.d4 0-0-0 15.Nxf4 Qe8+ 16.Be3
Bxe1 17.Rxe1 Bf5, Black won easily in B.Suhle-Hirschfeld & Guretzky Cornitz, consultation game,
Berlin 1860; and 14.Nxc7+ Kd7 15.Nxa8 Re8+ 16.Kf1 Rxe1+ 17.Nxe1 Qd1 – E.Schiller, also wins in
view of 18.Qe2 Bd3! 0-1 paardesprong-tsmenace, ChessWorld.net 2005) 14...Bxe1 15.Nxc7+ (not
15.Nf6+? gxf6 16.Qxg8+ Kd7 17.Qxa8 Nd4!) 15...Kd8! 16.Nxa8 Re8 17.Qxc2 Bg3! 18.h3 (or 18.d4
Qxh2!) 18...Re7 and Stockfish declares a win for Black; e.g. 19.b3 Ne5 20.Ba3 Nxf3 21.Bxd6 Re6
22.Qc7+ Ke8 23.Qc8+ Kf7 24.Qxb7+ Kg6 25.Qxf3 Qxf3+ 26.gxf3 Rxd6 27.Nc7 (or 27.Ke2 Rc6 and
...Rc8) 27...Rxd2 with a winning endgame.
c) 8.Bxg8! Rxg8 9.Nc3 is White’s best continuation in this line, when 9...Nc6 10.Nd5 (rather than
10.Re1?, transposing to ‘b’ above) 10...Kd8 11.Nxf4 Qf7 12.Qxh7, O.Cordel-G.Neumann, Berlin
match 1867, or 11.Re1 Re8 12.Nxf4 Qf7 13.d4 is roughly equal.
Instead, I’ve tried 9...c6 10.Re1 (or 10.Qxf4 Rf8, intending 11.Qg3 d5 12.Qxg7?? Rxf3) 10...Qf7
11.d4 (both 11.Qxh7 g5 and 11.Ne5 d5 12.Qxh7 Bf5 13.Nxf7 Bxh7 14.Nd6+ Kd7 15.Nxb7 Na6
16.d3 g5 are better for Black) 11...g5 12.d5 d6 in a couple of games. The extra pawn may give Black
the edge, though I’ve not managed to convert it; e.g. 13.Bd2 Kd8 (first 13...c5 might improve)

131
14.Nd4 Bd7 15.g3! fxg3 16.hxg3 Re8 17.a4 Bf6 18.Ne6+ Kc8 19.Qf5 cxd5 20.Bxg5 Rxe6 21.Rxe6
Bxe6 22.Qxf6 Qxf6 23.Bxf6 Nd7 24.Be7 and the endgame was drawn, LieutenantColumbo-tsmenace,
ChessWorld.net 2009.

6...Kd8

Stepping away from potential issues on the e-file. This move – or else 6...c6 7.Nxe4 Kd8,
transposing – was played in most of the old games, and in mine too. But, as in line B2, White can
cause problems.
Seeing my difficulties below, FM Stefan Bücker proposed an alternative: 6...Be7 7.Nxe4 (not
7.Nd5? c6 8.Nc7+ Kd8 9.Nxa8 d5 and Black again has huge play for the rook; nor 7.Bxg8? Rxg8, as
even then 8.Nd5 c6 9.Nc7+ Kd8 10.Nxa8 d5 looks good for Black, as does 8.Qxe4 d6) 7...Nc6! (this
move is getting less and less unexpected all the time) 8.Nf3 Qh5 9.Neg5 Nh6 (in line B2 this was
good for Black) 10.Re1 (this makes the difference, but not to the extent of giving White an
advantage) 10...Kd8 11.h4 (a new move; Bücker gave 11.Nh3 Rf8 12.d4 d5 13.Bb5 Bxh3 14.Bxc6
Bd6 as equal) 11...Bxg5 12.hxg5 (not 12.Nxg5?! Qxh4) 12...Nf5 13.d3 (or 13.d4 d6 14.d5 Ng3
15.Qf2 Ne5 16.Bxf4 Nxc4) 13...Qg4 14.b4 d6 15.b5 Nce7 and Black is not worse; e.g. 16.Bd2 Re8
17.Qf2 Qg3 18.Qxg3 fxg3 19.Bf7 Rf8 20.g6 hxg6 (or 20...h6 21.Bc3) 21.Bb3 Ng8 22.Bxg8 Rxg8
23.Bg5+ Kd7 24.Nh4 Re8 25.Kd2 a6 26.Nxf5 gxf5 27.Rxe8 Kxe8 28.Re1+ Kf8 29.Re7 axb5
30.Rxc7 Be6 31.Be7+ Kg8 ½-½ sinjoor1-tsmenace, ChessWorld.net 2018.

7.Nxe4

Not 7.Bxg8?! Rxg8 and Black is ready to prop up the f4-pawn with ...g7-g5 when required.

7...c6 8.Nf3

132
 Not 8.g3? Qe7 9.Bxg8 Rxg8 10.gxf4 d5 11.Ng5 Bf5 and Black is better, G.Neumann-Knorre &
Redlich, consultation game, Berlin 1865.

8...Qe7

This is the exact same position as in line B2, except for the location of the white king, which is
certainly in White’s favour. And 8...Qh5? is even worse than before: 9.Neg5 Nh6 (or 9...d5 10.Ne5,
as in their 2nd matchgame) 10.d4 d5? 11.Bxf4! Bg4 12.Qe5! and White won quickly, G.Neumann-
S.Rosenthal, 4th matchgame, Paris 1869.

9.Neg5!

Note that piece sacrifices are no good here because the white king obstructs the a1-rook; e.g. 9.d4?
d5 10.Bxf4 dxc4 11.Re1 Bf5 12.Bg5 Nf6 13.Nxf6 gxf6 14.Bf4 Qxe2+ 15.Rxe2 Nd7 and Black won,
Brenny-bhoy-tsmenace, ChessWorld.net 2008.
Keres’ 9.Bxg8 Rxg8 10.d3 isn’t straightforwardly equal either, because of 10...g5! (Morgado),
though White should be okay after 11.h4! (not 11.Nfxg5? d5 0-1 Zeitgeist-tsmenace, ChessWorld.net
2009; or 11.Nexg5?! Rxg5 12.Qxe7+ Bxe7) 11...g4 12.Nd4 d5 13.Ng5 Qxe2+ 14.Nxe2 h6 15.Nf7+
Ke7 16.Ne5 f3 17.gxf3 gxf3 18.Nxf3 Bg4 19.Rg1.
Not 10.Re1? h6 11.Qd3, as in E.Kirby-M.Michael, corr. 1881, because of 11...Qf7! 12.Nd6 (or
12.c4 d5 13.cxd5 Qxd5) 12...Bxd6 13.Qxd6 Qf8 and Black is clearly better.

133
9...d5

The alternative 9...Nh6 is no solution: 10.d4 d5 11.Bd3 Qxe2+ 12.Bxe2! (stronger than 12.Kxe2
Bd6 13.Bxh7 Nd7 14.Bd3 Re8+ 15.Kd1 Nf8 16.h3 Ne6 17.Rf1 Nxg5 18.Nxg5 Rf8, which led to a
draw in tarby-tsmenace, ChessWorld.net 2015) 12...Bd6 13.Ne5 Bxe5 14.dxe5 Re8 15.Bxf4 Nd7
16.Bh5! (provoking a weakness) 16...g6 17.Nxh7! gxh5 18.Bxh6 Nxe5 19.Nf6 Re6 20.Bg5 Nf7
21.Bh4, when Black has problems despite the reduced material; e.g. 21...Kc7 22.Kd2 Nd6 23.Rae1
Kb6 24.Rhf1 Rxe1 25.Bxe1 Bf5 26.Kc1! Bg6 27.Nxd5+ cxd5 28.Rf6 Be4 29.Rxd6+ Kc5 30.Rh6
Bxg2 31.Bf2+ Kb5 32.b3 b6 33.a4+ Ka6 34.Rxh5 and the passed h-pawn gave White an edge despite
the opposite-coloured bishops, Raff-zahn-tsmenace, ChessWorld.net 2015.

134
 I’ve played this position a few times too and it’s even less easy for Black than its equivalent in line
B2:
a) 10.Qxe7+ Bxe7 11.Bd3 (not 11.Nf7+?! Ke8 12.Nxh8 dxc4 13.b3 Kf8 14.Bb2 Bg4) 11...Nh6
12.h3 (if 12.Nxh7 or 12.Bxh7 then 12...Bg4) 12...g6 (12...c5!? might improve) 13.Re1
(Th.Johansson) gives White a long-term initiative: 13...Bd6 14.b3 Re8 15.Bb2 Bf5 16.Rxe8+ Kxe8
17.Nxh7 Nd7 18.Bxf5 Nxf5 19.c4 dxc4 20.bxc4 Kf7 21.Nhg5+ Kf8 22.a4 Re8 23.a5 Ke7 24.Kc2
Kd8 25.Nf7+ Kc7 26.Re1 Rxe1 27.Nxe1, which continued into the minor piece endgame, draco69-
tsmenace, ChessWorld.net 2017.
b) 10.Bd3! Nh6 (not 10...h6?? 11.Qxe7+ Kxe7 12.Re1+ and wins) 11.Re1 Bg4 12.b3 (or 12.b4!?)
12...Bxf3 (or 12...Qxe2+ 13.Bxe2 Be7 14.d4 Nd7 15.h3 Bxf3 16.Ne6+ and 17.Bxf3) 13.Nxf3 Qxe2+
14.Bxe2 Be7 15.Bb2 Rg8 16.c4 and again White has a long-term initiative; e.g. 16...Bf6 17.Bxf6+
gxf6 18.Rc1 Rxg2 19.cxd5 Nf7 20.Nd4 Rxh2 21.dxc6 Nxc6 22.Nxc6+ bxc6 23.Rxc6 Ke7 24.Bc4+
Ne5 25.Re6+ Kd7 26.Rxf6 Nxc4 27.bxc4 with a favourable rook endgame for White due to the
connected passed pawns, AndyAndyO-tsmenace, ChessWorld.net 2017.
I only drew one of the last three games noted above, so Black should probably avoid the whole
thing. Fortunately, there are two profitable ways to deviate this time: 6...Be7 (which I’ve already
tried) or 4...Nc6 again. Either way Black seems to be fine.

135
 Chapter Five
Wagenbach Defence

1.e4 e5 2.f4 exf4 3.Nf3 h5!?

GM Tony Miles once wrote: “I’m not quite sure why, but for many years I have spent idle moments
trying to refute the King’s Gambit. Refute it not in a boring way, but in the same spirit as the gambit
itself.” With that in mind it’s hard to suppress a sigh when some book or other recommends a
“practical” defence such as 2...d5 (intending 3.exd5 exf4) or 2...Bc5. Sure, those are perfectly good –
not least because almost every defence to the King’s Gambit is good – and if you like the arising
positions, fine, go for them. Otherwise, while a simple solution may appeal to a professional who
faces 2.f4 once every three years (if that), for club and tournament chess where the gambit still
appears regularly, I have to ask: why would you opt for something “practical” (which often lets White
off the hook) when there are so many more interesting lines available?
In particular, Schallopp’s 3...Nf6 is very much worth investigating, as recommended by Martin
Lokander and Nikolaos Ntirlis in their own repertoire books and since subjected to scrutiny by GM
Matthias Wahls (whose analysis is available as a free download on his website). As for me, for the
past thirty (!) years I have mostly been playing 3...h5!?. This was invented by my Mansfield
clubmate, János Wagenbach. It started out as a joke – the King’s Gambit is such rubbish that I can
even play 3...h5 against it – but proved surprisingly difficult to dismiss. We played endless blitz
games over many months, with János defending the Black side and me trying to find an advantage as
White. In the end I took up 3...h5 myself.
Because, rather than being a random silly move, 3...h5 has a serious point: to follow up with ...h5-
h4 and ...g7-g5, supporting the snaffled pawn while preventing White from undermining the phalanx

136
with h2-h4. The obvious cost is the time spent, especially as Black often plays ...d7-d6 and ...c7-c6 as
well before finally developing a piece at move eight. Consequently, White will have a huge lead in
development which you’d think really ought to lead to something; but it turns out that the Wagenbach
formation (f4/g5/h4) is remarkably resilient against attempts to bash it flat – such as attacks on f7
with Bc4 and Ne5 (see lines C and E1), sacrifices on g5 (line E22 primarily), or a rapid central
breakthrough with e4-e5 (line E33).
White might therefore consider a pre-emptive advance of the h2-pawn on moves four or five (lines
A and D). Alternatively, as GM Miles also reportedly said: “If I was playing Mr. Wagenbach, I’d play
3.Nc3.”
A: 4.h4
B: 4.Nc3
C: 4.Bc4
D: 4.d4
E: 4.d4 g5 5.Bc4 h4
Several other moves have been tried:
a) 4.Qe2!? is examined in the notes to line B.
b) 4.g3?, making the gambit permanent, is unwarranted at this early stage: 4...fxg3 5.hxg3 d5 6.Nc3
dxe4 7.Nxe4 Qe7 8.Bd3? f5 and Black won, foxworth_2-tsmenace, ChessWorld.net 2008.
c) 4.d3 is too quiet: 4...g5 5.h4 (5.Bd2 h4 6.Bc3 Rh6 and 5.Qd2 Nc6 6.h4 g4 7.Ng1 Bh6 are no
more effective) 5...g4 6.Nd4 (the attempted justification of the small pawn move) 6...Qf6 7.e5!? (or
7.Nb5 Na6) 7...Qxe5+ 8.Qe2 Bg7 9.Nb5 Qxe2+ 10.Bxe2, P.Cumbers-Fidelity Master, Grantham (30
min game) 1998, and here 10...Kd8 11.Bxf4 d6 is just good for Black, who is a clear pawn up.
d) 4.Be2 doesn’t pose a challenge either, since it’s not clear what the bishop is doing there.
J.Simmelink-J.Elburg, corr. 1998, continued 4...g5 5.0-0 d6 6.d4 Bg7 7.Qd3 c6?! (7...h4 is more
appropriate, before 8.h4! interferes with Black’s set-up) 8.b4?! Nh6 9.b5 0-0 10.c3 f5!? (this whole
game is quite un-Wagenbachly) 11.e5 g4! 12.exd6 gxf3 13.Qxf3 Qxd6 14.Bxf4 Qg6 and White didn’t
have enough for his sacrifice.
e) 4.b3!? is an idea borrowed from Becker’s Defence (3...h6).

137
 That the h-pawn has gone one square further doesn’t harm Black’s chances: 4...d6 (4...g5 5.Bb2
Rh6!? is similar; and 4...d5!? 5.exd5 Nf6, G.McDonald-F.Eastwood, corr. 1995, seems fine too)
5.Bb2 h4 6.d4 Rh6!? (there are more sensible-looking options) 7.Qd2 g5 8.h3 Nf6 9.Bd3 Nh5 10.Nc3
Ng3 11.Rg1 Rh5!? (intending perhaps ...Bh6 and ...g5-g4, while pre-empting any Nd5 and Nxg5
tricks) 12.Nh2 Nc6 13.0-0-0 Bg7 and Black was not worse, D.Amour-J.Tait, corr. 2000.

A: 4.h4

White blocks the black h-pawn, thus pre-empting the standard Wagenbach formation (f4/g5/h4) and
hoping for a more typical King’s Knight Gambit. The question is which side benefits most from the
two h-pawn moves. It’s not enough to claim that White has two weaknesses (g3 and g4) while Black

138
only has one (g5) since the possibility of Ng5 may be more important. It is up to Black to ensure that
it isn’t. Given that all systems with ...g7-g5 are also ruled out, the logical continuation is to play in the
centre with ...d7-d5. Black then has to decide how this might best be achieved.

4...Nf6

I like this move, first played by Phil Briggs. Note that we’re not in a kind of Schallopp because the
f6-knight cannot go to h5 if hit. Instead, Black aims for ...d7-d5 in its most favourable version.
Alternatively:
a) 4...d5 immediately is also okay, even if it allows White more flexibility, while Black’s ...h7-h5
means castling short is far less attractive than in a true Modern (3...d5) position; e.g. 5.exd5 Nf6
(5...Bd6 6.d4 Bg4 7.c4 is perhaps slightly better for White) 6.Bc4 (6.c4 would be met by 6...c6)
6...Nxd5 (the spite check 6...Qe7+!? is not without merit; e.g. 7.Be2 Nxd5 8.0-0 Nc6 or 7.Qe2 Qxe2+
8.Kxe2 Bd6 9.Re1 0-0 10.Kf1 Bg4) 7.0-0 (instead, 7.Nc3 would transpose to the main line, while
7.Bxd5 Qxd5 8.Nc3 can be met by 8...Qe6+ 9.Kf2 Qg4) 7...Bg4?! (the ‘normal’ ...Be7, ...0-0 can be
discounted, but 7...Be6! should be fine despite inviting 8.Ng5) 8.d4 Nc6 9.Nc3 Nxc3 10.bxc3 Bd6
11.Qe1+ Qe7 12.Ng5! Qxe1 13.Rxe1+ Kd7 14.Nxf7 Rhf8 was P.Cumbers-J.Tait, Sheffield League
2002, where 15.Rb1 Rab8 16.Rb5!? (intending Rg5) looks rather good for White.
b) 4...Ne7!? is interesting, in the fashion of a Bonch-Osmolovsky variation (3...Ne7), to which it
can indeed transpose.

a) 5.Bc4 Ng6 6.Ng5 Ne5 7.Bb3 f6 8.Nh3 g5 9.hxg5 fxg5 10.d4 Nf7 11.g3 fxg3 12.Bxf7+ Kxf7
13.Nxg5+ was winning for White in one of Greco’s games from the 17th century, but Black can
easily improve; e.g. 8...f3!? 9.gxf3 Nbc6 10.f4 (or 10.d4? Nxd4) 10...d5!.
b) 5.d4 Ng6 is an old-fashioned Bonch (nowadays Black usually follows 3...Ne7 with ...d7-d5 and
...Nxd5), which may well be playable too:

139
 b1) 6.Bc4 Be7 7.0-0 (or 7.Nc3 c6) 7...d6 8.Nc3 Bg4!? (or 8...c6 again, as in ‘b2’) 9.Ne2 Nxh4
10.Nxh4 Bxh4 11.Bxf4 Qd7 12.Qd2 Nc6 13.Bg5 Bxg5 14.Qxg5 Bxe2!? (there was nothing wrong
with 14...0-0) 15.Bxf7+ Qxf7 16.Rxf7 Kxf7 17.Qf5+ Ke7 18.Qg5+ Kd7 19.Qxg7+? (Stockfish gives
19.d5 Ne7 20.e5 as “0.00”) 19...Ne7 20.Qg3 Rag8 and Black had a very strong attack in J.Elburg-
J.Tait, corr. 1998.
b2) 6.Nc3 c6 7.Bc4 Be7

8.0-0 (here 8.e5 d5 9.exd6 Qxd6, A.Bragin-Adishev, corr. 1990 is nothing to fear; nor is 8.Qe2 d6
9.Bd2 Bg4 10.0-0-0 Nd7, jalfano-S.Tatlow, corr. 1999) 8...d6 9.Nh2?! (9.Ne2 as in ‘b1’ is preferable)
9...Bxh4 10.Bxf4 Nxf4 11.Rxf4 Bf6? (11...Be6 is better for Black) 12.e5 dxe5 13.Qe2? (another
mistake; 13.dxe5 Qxd1+ 14.Rxd1 Bxe5 15.Rxf7 looks good for White) 13...Be6 14.dxe5 Bg5
15.Re4?! Nd7 16.Bxe6 fxe6 17.Rd4 Qb6 18.Nf3 Be7 19.Qc4 Rh6 20.Ne4 0-0-0 and Black still had
the pawn in B.Spassky-A.Tolush, Kislovodsk 1960, a fluctuating game which eventually ended in a
draw.

140
5.Nc3

The usual reply.

As might be expected, 5.d3?! is too passive: 5...d5 6.e5 (nothing improves: 6.exd5 Nxd5, or 6.Nc3
Bb4, or 6.Qe2 dxe4 7.dxe4 Ng4 8.Bxf4 Qf6, 2buddies-tsmenace, ChessWorld.net 2017) 6...Ng4
7.d4?! (merely creating a target; White should take on f4 straight away) 7...c5 8.Bxf4 Nc6 9.Bb5 Be6
10.c3 Qb6 had somehow turned into a very nice French for Black, theblizz-ard-tsmenace,
ChessWorld.net 2005.
The main alternative is 5.e5 Ne4 (heading for the hole on g3; targeting the centre with 5...Ng4! 6.d4
d6 is another option) 6.d4 Ng3 7.Rg1 (an improvement on 7.Rh2?! Nxf1 8.Kxf1 d5 9.Bxf4 Bg4
10.Qd3 Nc6 11.c3 Qd7 12.Nbd2 0-0-0 13.g3 Re8 14.Re1 Nd8 15.Kg1 Ne6 16.Qe3 Be7 and White
suffered on the light squares, P.Cumbers-P.Briggs, Nottingham 2002) 7...Nxf1 8.Rxf1 d5 (Stockfish
prefers to open things up with 8...d6! 9.Bxf4 dxe5, never mind Black’s lack of development) 9.Bxf4
Bg4 10.Bg5 Qd7 11.Qd3 Nc6 12.a3 Qe6, intending ...f7-f6, is fine for Black, especially after
13.Kf2?! Bxf3 14.gxf3 Be7 15.Bxe7 Qxe7 16.Rh1 0-0-0 17.Nd2 f6, clb108-tsmenace,
ChessWorld.net 2007.

5...d5 6.exd5

Now 6.e5?! Ne4 is just good for Black; e.g. 7.Ne2 (if 7.d3 Ng3 8.Rh2 then 8...d4, or 7.d4 Bb4
8.Bd2 Ng3 9.Rh2 Nxf1 10.Kxf1 Bg4 11.Bxf4 Qd7 12.Qd3 Nc6 13.g3 0-0-0, Verloc-tsmenace,
ChessWorld.net 2006) 7...Bg4 8.d3 Be7 9.dxe4 Bxf3 10.Qxd5 (not 10.gxf3? Bxh4+ 11.Kd2? dxe4+
and wins) 10...Bxh4+ 11.Kd1 Bxe4 12.Qxd8+ Bxd8 with an extra pawn even when White picks up
the one on f4.

6...Nxd5 7.Bc4

141
 After 7.Nxd5 Qxd5 8.d4 Bd6 White will have to work to regain the pawn; e.g. 9.c4 Qe6+ 10.Be2
(or 10.Kf2 c5) 10...Qf6 11.c5 Be7 12.Qd2 (or 12.Ne5 Nc6 13.Nxc6 Qxc6 14.0-0 Be6 15.Rxf4 0-0-0)
12...0-0 13.Qxf4 Nc6 14.Be3 Bg4 and White’s inferior structure means Black is better if anyone.

7...Qe7+

I now prefer this to 7...Nxc3 8.bxc3 Qe7+ (note that Black can’t just play ‘normal’ moves; e.g.
8...Bd6 9.d4 Bg4 10.Qd3 Nc6 11.0-0 0-0?? 12.Ng5 wins at once), which allows 9.Be2! (rather than
9.Kf1 Nd7 10.d4 Nb6 11.Bb3 Bg4 12.Bxf4 0-0-0 0-1 B.Ayoub-J.Tait, corr. 2002, as my opponent
withdrew from the tournament) 9...g6!? (trying to keep the pawn; 9...Bg4 10.0-0 Nd7 11.d4 0-0-0
12.Bxf4 Nb6 is probably safer and equal) 10.d4 Bh6, bobby fissure-tsmenace, ChessWorld.net 2009,
when 11.0-0 0-0 12.Ng5 Bxg5 13.hxg5 Qxg5? 14.Bxf4 gives White deadly play for the pawn, while
12...f6? runs into 13.Bc4+ Kg7 14.Rxf4 fxg5 15.Rxf8 Kxf8 (or 15...Qxf8 16.hxg5) 16.Ba3! Qxa3
17.Qf3+ Ke8 18.Re1+ and wins.

8.Qe2

After 8.Be2 Be6 or 8.Kf1 Nb6 9.Bb3 Nc6 10.d3 Be6 11.Bxf4 0-0-0, Black is certainly not worse.

8...Nb4

This doesn’t achieve much. Exchanging queens is simpler: i.e. 8...Qxe2+ 9.Nxe2 (not 9.Kxe2?!
Nxc3+ 10.dxc3 Bd6) 9...Be6 with an equal position. White has little hope of getting anything more
than the pawn back.

9.Kd1! Qxe2+ 10.Nxe2 Nd7 11.Nxf4 Bd6 12.Re1+ Kf8

142
 So White has regained the pawn and now has an extra one in the centre; on the other hand, Black
can complete development more easily. What does Stockfish say about it? “0.00” of course.
The game Aethelbald-tsmenace, ChessWorld.net 2010, continued 13.d3 Nb6 14.Bb3 Bg4 15.a3
Nc6 16.c3 Bxf3+ 17.gxf3 Re8 (or just 17...g6 18.Kc2 Kg7 19.Bd2 Rhe8 with equality) 18.Rxe8+
Kxe8 19.Bd2 Ne7 20.Kc2 Nf5 21.Re1+ Kf8 22.Ng2 g6 23.Rh1 Kg7 24.Ne3 Ng3 25.Rg1 Re8 26.a4
Bf4 27.Re1 Kf8 (avoiding the trick with Nf5+ etc) 28.a5 Nc8 29.d4 Ne7 30.Nc4 Bxd2 31.Kxd2 Nef5
32.Ne5 Re7 (under slight pressure, facing threats against the h4-pawn and the centre via ...c7-c5,
White finds a nice way to liquidate) 33.a6! bxa6 (not 33...b6?? 34.Nc6) 34.Bxf7 Rxf7 35.Nxf7 Kxf7
36.Ra1 Nxh4 37.Ke1 ½-½. Play might have ended 37...Ngf5 (not 37...Nxf3+? 38.Kf2) 38.Kf2 g5
39.Rxa6 Ke7 40.Rxa7 Kd7 41.b4 g4 42.fxg4 hxg4 43.b5 g3+ 44.Ke2 Kc8 45.Ra8+ Kb7 46.Rg8 Ng2
47.Kf3 Nge3 48.Rxg3!? Nxg3 49.Kxe3 Kb6 50.c4 with a draw, as tablebases confirm.

B: 4.Nc3

143
 This generally transposes to the main lines (i.e. after d2-d4 and Bc4), but White can keep it
independent, not least with an early Qe2.
White has even tried 4.Qe2!? at once, when 4...g5 5.Nc3 h4 6.d4 transposes below, but 5...Nc6! is
stronger via this move order. For example, 6.Qf2 (6.d4 runs into 6...g4!) 6...Bg7 7.d3 d6 8.Bd2 h4
9.0-0-0 Qf6 10.d4!? Nxd4 11.e5 dxe5 12.Ne4?! (12.Nxd4 exd4 13.Nb5 offers White more
compensation; e.g. 13...Qe7 14.Nxd4 Nf6 15.Bb5+ Kf8 16.Rhe1 Ne4 17.Bb4!) 12...Qe7 13.Qe1 a5
(preventing Bb4) 14.Nxd4 exd4 15.Nxg5 Rh5! 16.Nf3 Be6 17.Kb1 0-0-0 18.Bxa5 Qf6 19.Bd3 Ne7
20.Bd2 Nd5 and Black won, juliangon-tsmenace, ChessWorld.net 2016.

4.Nc3 g5

Instead:
a) 4...d6 gives White various options: 5.Bc4 h4 6.d4 g5 is a main line (E3); 5.d4 g5 is the same as
4...g5 5.d4 d6 below; 5.h4!? can be compared (perhaps favourably) with line A; and even 5.Qe2!?
Nc6 (or 5...g5 6.Qb5+) 6.d4 g5 7.h4 g4 8.Ng5 comes into consideration.
b) 4...h4 5.Qe2!? (otherwise 5.d4 g5 6.Qe2 transposes below; while 5.Bc4 is covered in line C)
5...a6 (5...g5 6.d4 is 4...g5 etc again; not 5...d6? 6.d4 g5 7.Qb5+ etc) 6.b3!? (simply 6.d4 g5 7.e5
looks very promising) 6...d6 7.Bb2 Nc6 8.Nd5 h3!? (this move is almost always wrong in the
Wagenbach, but Black is struggling anyway) 9.Nxf4 hxg2 10.Bxg2 Bg4 11.0-0-0 Ne5 12.h3 Bxf3
13.Bxf3 Nxf3 14.Qxf3 Rh4?! (14...Qd7 keeps Black in the game for the moment) 15.Rdf1 Qd7
16.Ng6 Rh6 17.Rhg1 Rxh3 18.Qg2 f6 19.e5 dxe5 20.Nxe5 Qe6 21.Ng4 Rh5 22.Re1 Re5 23.Nxe5 0-
0-0 24.Ng6 1-0 J.Tait-J.Wagenbach, corr. 1998.

5.d4

Others:

144
 a) 5.Qe2 returns to 4.Qe2 above.
b) 5.h4 g4 6.Ng5!? (not 6.Nd4?! Nc6 7.Nxc6 dxc6 8.d4 Bh6) 6...f6 7.d4 fxg5 8.hxg5 Qxg5 (or
8...Bg7 9.Bxf4) 9.Nd5 Qd8 10.Nxf4 may give White enough play for the piece but unlikely any more
than that.
c) 5.Bc4 h4 (not 5...d6? 6.h4) 6.Ne5!? (otherwise 6.d4 is the main line E3) 6...Nh6 7.Qh5 is
different from E1 (with d2-d4) in that 7...Qf6? fails to 8.Nf3. It’s necessary to throw in 7...d5!, after
which 8.Bxd5 (or 8.Nxd5 Bg7) 8...Qf6 9.d4 (or 9.Nf3? Be7) 9...Nc6 10.Bxc6+ (or 10.Nxc6 Bg4!
11.Bxf7+ Qxf7 12.Qxg5 bxc6 13.Qe5+? Kd7 14.d5 c5) 10...bxc6 11.Qf3 (or 11.Qe2 Bb7, intending
...0-0-0) 11...Ba6 12.g3 0-0-0 13.gxf4 Rxd4 seems no worse for Black.

The position after 5.d4 can also arise via 4.d4 g5 (line D) 5.Nc3, so it has some significance.

5...h4

Instead:
a) 5...d6 6.h4 (6.Bc4 h4 is main line E3 again) 6...g4 7.Ng1 produces a type of Fischer Defence
(3...d6), similar to in line D, except that Nc3 and ...d7-d6 have already been inserted which is
somewhat in White’s favour; e.g. 7...f3 (or 7...Bh6 8.Nge2 f3 9.Ng3 Bxc1 10.Qxc1, when I prefer
White) 8.Bg5 (or 8.Be3 Be7 9.Bf2) 8...Be7 9.Qd2 fxg2 10.Bxg2 and White certainly has enough for
the pawn as it is not clear how Black should develop.
b) 5...g4!? 6.Bxf4 gxf3 7.Qxf3 is a sort of Rosentreter where Black has defended with ...h7-h5!?.
White has scored very well from here; e.g. tsmenace-Reprimand, ChessWorld.net 2015, continued
7...d6 8.Bc4 Be6 9.Bxe6 fxe6 10.0-0 Qd7 11.e5 Nc6?! (on 11...d5, I intended 12.Bg5 Qg7 13.Qh3!
Qxg5 14.Nxd5 Qg4 15.Nxc7+ Kd7 16.Qxg4 hxg4 17.Rxf8 Kxc7 18.Raf1 with excellent
compensation even for two knights, since Black cannot move anything) 12.Bg5 Bg7 13.exd6 cxd6
14.Rae1 Bxd4+ 15.Kh1 Nge7 16.Qf7+ Kd8 17.Rxe6 Bxc3 18.bxc3 Re8 19.Rd1 Kc7 20.Rdxd6 Qxd6

145
21.Bf4 Qxf4 22.Qxf4+ Kc8 23.h4 and White’s queen and kingside pawns eventually overcame
Black’s assortment of pieces.

6.Qe2!?

Once again, 6.Bc4 is a main line (E3). 6.Qd3!? is similar to the text, except that after 6...Bg7 7.e5
the white queen’s absence from the e-file allows Black to play 7...d5! 8.exd6 cxd6 9.Qb5+ (or 9.Ne4
f6) 9...Nc6 10.Qxg5 Qxg5 11.Nxg5 Nxd4 12.Bc4 Ne6 13.Nxe6 Bxe6 14.Bxf4 0-0-0 with equality.

With 6.Qe2 White deters the natural ...d7-d6 (due to Qb5+ etc) while preparing queenside castling
(Bd2, 0-0-0) and/or a quick g2-g3. Black needs to respond purposefully:

146
 a) 6...Be7?! (too routine) 7.Bd2 (7.e5 Kf8 is less clear) 7...d6 8.0-0-0 c6 9.Qf2, planning Bd3 and
g2-g3, looks very dangerous.
b) 6...Bh6 7.Qb5!? (7.e5 Kf8 8.g3!? is interesting too, or just 7.g3!, since the bishop obstructs the
black rook on the h-file) 7...Kf8 8.h3 d6 9.Bd2 c6 10.Qb3 Nd7 11.0-0-0 Kg7 12.Bc4 Qe7 13.Rhe1 b5
14.Bd3 a6 15.Kb1, juliangon-tsmenace, ChessWorld.net 2014, works out well for White, who can
prepare a central breakthrough.
c) 6...Bg7!? attempts to interfere with White’s plans by attacking d4. This can be followed up by
...Nc6 and even ...g5-g4 (note that preventing the latter with h2-h3 rules out g2-g3). The game
remyrey-tsmenace, ChessWorld.net 2019, continued 7.e5 Kf8 (so the d7-pawn can advance) 8.Ne4
Bh6 (or 8...Rh5!?) 9.d5 (the very sharp 9.Bd2 g4! 10.Bb4+ Ne7 11.Nfg5 d5! 12.exd6 cxd6 13.Nxf7!
Kxf7 14.Nxd6+ Kg7 15.Qe5+ Kh7 16.Bd3+ Ng6 is inconclusive; engine analysis shows that White
has sufficient play for the piece, if no more than that) 9...g4 10.Nfg5 Bxg5 11.Qxg4 Bh6 12.Bxf4 Nc6
13.Bxh6+ Rxh6 14.dxc6 d5 15.Qg5 Qxg5 16.Nxg5 bxc6 17.0-0-0 Rh5 18.Nf3 Bg4 19.Be2 Bxf3
20.Bxf3 Rxe5 and was later drawn.

C: 4.Bc4

The main lines arise after 4.Bc4 h4 5.d4 g5 or 4.d4 g5 5.Bc4 h4. By each move order White has the
chance to deviate (and Black too, if less advisably). Here we’ll look at independent lines with 4.Bc4.

4...h4!

This move is essential, otherwise White will likely throw in h2-h4 next. For example: 4...g5? (or
similarly 4...d6? 5.d4 g5 6.h4 g4 7.Ng5) 4...g5? 5.h4! g4 6.Ng5! (6.Ne5 would be a Long Whip)
6...Nh6 7.d4 f6 8.Bxf4 fxg5 9.hxg5 Nf7 10.g6 and Black is probably losing by force: 10...Ng5 (by
transposition 10...Nd6 11.Bxd6 cxd6 12.Bf7+ Ke7 13.Nc3 Qa5 14.Qd2 Bh6 15.Qf2 was
K.Poschauko-F.Burgi, Vienna 1954; I don’t have any more of that game but had worked out a forced

147
win and got to play it out: 15...Bg7 16.e5 Kd8 17.exd6 Qg5 18.Ne4 g3 19.Nxg3 1-0 tsmenace-
JHDonner, ChessWorld.net 2010 – not actually Donner of course) 11.Qd2 (or 11.Nc3 first, with
which I’ve won three times) 11...Nxe4? (but after 11...Ne6 12.Nc3 White’s position is still far too
strong) 12.Bf7+ Ke7 13.Bg5+ Nxg5 14.Qxg5+ Kd6 15.Qc5 mate, H.Bird-Pfander, England 1903,
another game that arose by transposition.

5.Ne5!?

Going straight for the “weak” f7-square. Instead:

a) 5.d4 g5 is the main line again.
b) 5.0-0 d6 will transpose after 6.d4 g5.
c) 5.Nc3 d6 (for 5...g5 see line B) 6.d4 g5 is the same again. For 6.d3 g5 see note ‘g2’ below. The
forcing 6.Nd5?! g5 7.Nxg5 doesn’t work after 7...Qxg5! 8.Nxc7+ Kd8 9.Nxa8 Qxg2 10.Rf1 Qxh2,
because the Wagenbach pawn is too fast: 11.d4 Qg3+ 12.Rf2 h3 13.Bxf4 Qg1+ 14.Rf1 h2 15.Qd2
Qxf1+ 16.Bxf1 h1=Q and Black won, MarcShaw-tsmenace, ChessWorld.net 2012.
d) 5.h3 d6 (or 5...g5) 6.d4 g5 transposes again. Here 6.d3 g5 7.0-0 Bh6 8.Nxh4!? (objectively “?”)
once took me by surprise: 8...gxh4 9.Bxf4 Bxf4 (9...Be6! is correct, when White has nothing) 10.Rxf4
Be6?! 11.Bxe6 fxe6 12.Qg4 Kd7 13.Nc3 Nc6? (terrible defence; 13...Ne7 14.Nd5 c6 is still about
level) 14.Rf7+ Nge7? 15.Nd5! Ne5?! 16.Nf6+ Kc6 17.Qxe6 Nxf7 18.Qc4+ Kb6 19.Qb4+ Ka6
20.Qa4+ Kb6 ½-½ G.Murfet-J.Tait, Notts League 2000, and I gratefully accepted the draw, since
21.Nd7+ Qxd7 22.Qxd7 Rae8 23.a4! is just winning for White.
e) 5.Qe2 g5 (5...d6 6.Nc3 c6 7.d4 g5 is another main line – see E32 below) 6.Nc3 Nc6 (6...d6?!
would be met by 7.e5!) 7.h3 Bg7 8.Bb3!? (the start of an imaginative plan, but Black defends
strongly) 8...Nh6 9.Qb5 Rg8! 10.Qd5? (not 10.Qxg5? since 10...Qxg5 11.Nxg5 Bxc3 wins material;
10.Ne2 is probably best) 10...Bxc3! 11.bxc3 Qe7 12.0-0 b6! 13.d4 Bb7 (now it’s clear that the white

148
queen has just wasted a lot of time) 14.Qb5 Na5 15.Re1 0-0-0 and Black soon won on the kingside,
samurai-Aethelbald, ChessWorld.net 2010.
f) 5.b3!?

(suggested by CCGM Maurice Johnson) tries to improve on 4.b3 (see note ‘e’ at the beginning of
the chapter), seeing that Black has now spent two moves with the h-pawn. After 5...d6 6.Bb2 c6
(6...Nf6!? 7.Qe2 Nh5 8.Nc3 Bg4?! 9.e5! Be7 10.0-0-0 Nc6 11.d4 0-0 12.e6!? f5 13.Qf2 saw Black
struggling in J.Tait-J.Simmelink, corr. 1998) 7.Nc3 Rh6?! (there’s no need to prepare 7...g5 as the
discovery on the long diagonal is not a threat) 8.Qe2 g5 9.0-0-0 (a preliminary 9.h3! might be best, as
in line E31) 9...Bg7 10.Qf2 Bg4 11.g3? (the correct plan except for the h6-rook; instead, 11.d4 Nd7 is
unclear) 11...fxg3 12.hxg3? (12.Qe3 is necessary) 12...Rf6! 13.Be2? Bxf3 14.Bxf3 g4 and Black won,
sholders-tsmenace, Itsyourturn.com 2006.
g) 5.d3 g5 is just an inferior (for White) version of the main line:

149
 g1) 6.Ne5?! Nh6 7.Qh5 can be met by 7...Qf6 (as in line E1), when 8.Nf3 (not 8.Nxf7? Nxf7
9.Bxf7+ Ke7) 8...d5! 9.Bxd5 Be7 10.e5 (or 10.h3 Nd7) 10...Qg7 11.e6 Kd8! leaves Black with a clear
advantage.
g2) 6.Nc3 d6 7.0-0 c6 8.Nxg5 Qxg5 9.Rxf4?! (for 9.Bxf7+ Kxf7 10.Bxf4 Qg4 11.Be5+ Nf6
compare line E22) 9...f6 10.Qf3 Be7 11.Be3 Nh6 12.Rf1 Ng4 13.d4 Nxe3 14.Qxe3 h3 15.g3 Bg4 and
White had nothing much for the piece, J.Pinto Veira-J.Tait, Notts Championship 2019 (via 5.Nc3 d6
6.d3 g5).
g3) 6.Bd2 d6 (6...Bg7 7.Bc3 Bxc3+ 8.Nxc3 d6 9.Qe2 c6 10.0-0-0 Nh6 11.g3! is clearly dangerous
for Black, L.Wooldridge-J.Wagenbach, Knaresborough rapid 1999) 7.Bc3 Rh7 (7...Rh6 seems
preferable now) 8.h3 Be6 9.Nbd2 Be7?! (either 9...Nd7 or 9...Bxc4 10.Nxc4 d5 is better) 10.Qe2 c6
11.Bxe6 fxe6 12.0-0-0 Nd7 13.Nd4 Nf8 14.Nc4 Qd7 15.Nf3 Nh6 16.Kb1 0-0-0 17.Ba5 Re8 18.d4
Kb8 was very unclear when White defaulted in Mario Cruz-tsmenace, ChessWorld.net 2005.

5...Nh6

Fortunately, the obvious reply works for Black.

6.Qh5

Having said “A”... Anything else, such as 6.d4 d6 or 6.0-0 d6, and White has just wasted two tempi
with the knight (since Nd3 makes little sense).

150
6...Qe7!

It is this move, in conjunction with 7...g6, which safeguards Black’s position.

In contrast to line E1 (where d2-d4 and ...g7-g5 have been included), here 6...Qf6?! doesn’t defend
the rook in the corner so 7.Nxf7! regains the pawn at once. After 7...g6 8.Qe5+ Qxe5 9.Nxe5 g5
10.Ng6 Rh7 11.Nxf8 Kxf8 12.g3! hxg3 13.hxg3 fxg3 14.d4 Nc6 15.c3 d6 16.Kf1! Kg7 17.Bxg5 Bd7
18.Kg2, White was soon a pawn up and went on to win in J.Wagenbach-S.Tatlow, corr. 1999.

7.Nc3

151
 The best if not only move.
a) 7.Nxf7? is met by 7...g6 8.Qe5? Nxf7 9.Bxf7+ Kxf7 10.Qxh8 Qxe4+ 11.Kf1 f3 0-1 J.Elburg-
J.Tait, corr. 1997. A later opponent played it out: 12.gxf3 Qxf3+ 13.Kg1 h3 14.Qxh3 Bc5+ 0-1
wormhole-tsmenace, ChessWorld.net 2005.
b) 7.d4 g6 (7...d6! is stronger, since 8.Nxf7 Qxe4+ 9.Kf1 g6! 10.Ng5 Qxc2 is hopeless for White)
8.Qe2 d6 9.Nf3 g5 10.Nc3 c6 11.h3 Bg7 12.Bd2 Nd7 and, two tempi up on a main line, Black
eventually won in J.Willow-J.Wagenbach, Notts League 2014.
c) 7.d3!? g6 (not now 7...d6?! 8.Nxf7 g6 9.Qxg6 Nxf7 10.Bxf4, followed by 0-0 with strong play
for the piece) 8.Nxg6 fxg6 9.Qxg6+ Kd8 10.Bxf4 looks threatening, until you spot 10...Qh7! 11.Qh5
(or 11.Bg5+ Be7 12.Bxe7+ Qxe7, rudov-tsmenace, ChessWorld.net 2005) 11...Be7 12.Nc3 d6
13.Nd5 (or 13.Bg5 Nf5 14.Bxe7+ Nxe7, J.Elburg-J.Tait, corr. 1997) 13...Bg4 14.Qxh6 Qxh6 15.Bxh6
Rxh6 16.0-0 Bh5 17.Rf4 Be8 18.Raf1 Nd7 19.d4 c6 20.Nxe7 Kxe7, A.Beatty-J.Tait, corr. 1999.
Black won all these games.

7...g6 8.Nxg6 fxg6 9.Qxg6+ Kd8

My database has 15 games with this position, Black scoring a healthy 13 points. Nevertheless,
consolidating the extra piece is no trivial task, so we’ll examine my only loss.

10.d4

The most natural continuation.

White has also tried 10.Nd5!? Qg7 11.Nxf4, aiming to keep the queen hanging about for a bit.
Black therefore plays to force the exchange: 11...Nc6 (Stockfish prefers 11...Ng8! and ...Ne7 at once)
12.c3 Ng8 (12...Ne5 is another option, as in J.Wagenbach-J.Elburg, corr. 2000) 13.d4 Nge7 14.Qxg7
Bxg7 15.0-0 Bh6 16.Bd2?! (16.Nh3 is better, heading for g5 after the exchange of bishops) 16...h3 (to

152
give the c8-bishop a square; e.g. 17.g3 d6 18.Rf2 Bg4) 17.gxh3 Na5 18.Be2 d5 19.exd5 Bxh3
20.Nxh3 Bxd2, J.Wagenbach-J.Tait, corr. 2000, where Black emerged with a slight edge in view of
21.b4 Bxc3 22.Rab1 Bxd4+ 23.Kh1 Nxd5 24.bxa5 Nc3 25.Rbe1 Nxe2 26.Rxe2 Kd7.

10...Qg7 11.Qxg7 Bxg7

12.Bxf4!

Generating an initiative is more important than defending the d-pawn, whereas 12.Ne2 allows
Black to develop more easily. Even so, White is not without chances: 12...Re8 (12...Nc6 13.c3 Re8?!,
J.Wagenbach-J.Simmelink, corr. 1998, might be met by 14.0-0! since 14...Rxe4? 15.Nxf4, followed
by Nh5, only lands Black in trouble; or if 12...d6 13.Bxf4 Bg4 14.c3 Re8, J.Wagenbach-
D.Nightingale, corr. 1998, then 15.Bg5+ Kd7 16.Nf4! Rxe4+ 17.Kd2, threatening 18.h3) 13.e5?!
(even here 13.0-0! looks best, intending 13...Rxe4 14.c3, followed by Nxf4) 13...c5 14.c3 d6 15.Bxf4
dxe5 16.Bg5+ Kc7 17.dxc5 Be6 18.Bxe6 Rxe6 19.Bxh4 Nd7 20.0-0-0 (or 20.b4 Rf8) 20...Nxc5
21.Rhf1 Rf8 and Black went on to win, J.Wagenbach-J.Tait, corr. 1998.

12...Bxd4

Black might as well take it, since refusing the pawn is no sinecure; e.g. 12...d6 13.e5! Bd7 14.0-0-0
Nc6 (or 14...Nf5 15.g4! hxg3 16.hxg3 Rxh1 17.Rxh1 Nxd4 18.Bg5+ Kc8 19.Bf6) 15.Rhf1 Kc8
16.exd6 cxd6 17.Ne4 Nf5 18.Nxd6+ Nxd6 19.Bxd6 Bxd4 20.Rf7, still with play for the piece.

153
13.Nd5

In Chess Mail I gave 13.Bg5+ Ke8 14.Nd5 Be5 as good for Black, which was very much an
overestimation. After 15.0-0 Nc6 16.Nf6+! Bxf6 17.Bxf6 (not 17.Rxf6?? Ne5 18.Bxh6 Nxc4 and
Black won, R.Hearne-S.Wells, Nottingham rapid 2002) 17...Rf8 (or 17...Rh7 18.e5) 18.Bg7 Rxf1+
19.Rxf1 Ng4 20.Rf8+ Ke7 21.Rg8, White has sufficient play to draw; e.g. 21...d6 22.Bf8+ Kf6
23.Bg7+ Kg5 24.Be5+ Kh6 25.Rh8+ Kg6 26.Rg8+ with perpetual check.

13...d6 14.0-0-0 Bg7 15.Bg5+ Kd7 16.Rhf1 c6 17.Nf4

If 17.Nf6+ Kc7 18.Bf4, then 18...b5! 19.Rxd6 Kb7 20.Be6 Nf7! 21.Bxc8+ (or 21.Bxf7 Rf8)
21...Kxc8 22.Re6 Bxf6 23.Rxf6 Nd8 defends.

17...Kc7

Here 17...b5 was probably simpler: 18.Be6+ (or 18.Ne6 Be5) 18...Kc7 19.Bxc8 Kxc8 20.Ng6 Re8
21.Rxd6 Ng4 and rough equality ensues; e.g. 22.Rf7 Be5 23.Nxe5 Nxe5 24.Rh7 Nbd7 25.Bxh4 Kc7
26.Rd1 Re6 and so forth. The text would consolidate completely, were it not for my opponent’s reply.

18.e5! d5

Not 18...Bxe5? 19.Ng6 and White is better, in view of 19...Rh7? 20.Nxe5 dxe5? 21.Bd8 mate.

19.Bxd5

154
19...Ng4?!

Targeting the e5-pawn was overambitious, especially as Black can’t take it. It was better to accept
the piece: 19...cxd5 20.Nxd5+ Kc6 21.Bf6 (or 21.Nb4+ Kc7 22.Nd5+ Kc6, repeating) 21...Nf5
22.Rf3 Nd7 23.Rc3+ Nc5 24.b4 b6 25.Bxg7 Nxg7 26.bxc5 bxc5 27.Nf4 Rb8 28.Rd6+ Kb7 29.Rxc5
Ka8 with a likely draw, given that the e-pawn isn’t dangerous.

20.Be6! Ne3

Not 20...Nxe5? 21.Bxc8 Kxc8 22.Ne6 and the g7-bishop can’t be saved since 22...Bh6 (or 22...Rg8
23.Nxg7 Rxg7 24.Rf8+) 23.Bxh6 Rxh6 24.Rd8 is mate again.

21.Bxc8 Kxc8?

A serious mistake; I can’t see now why I didn’t just take the rook: 21...Nxf1 22.Bh3 (or 22.Bf5
Bxe5) 22...Bh6! 23.Bxh6 Rxh6 24.Rxf1 Nd7 and Black is still probably okay.

22.Ng6 Bh6 23.Nxh8 Bxg5 24.Rf8+ Kc7 25.Nf7 Nxd1+ 26.Nxg5 Ne3 27.Ne6+ Kd7

Not 27...Kb6? 28.Rf3 and White wins at once, since Rb3+ and Nc7 follows if the e3-knight moves
away.

28.Nc5+ Ke7

Or 28...Kc7? 29.Rf7+ Kb6 30.b4 (threatening Rxb7 mate) 30...Kb5 31.Rxb7+ Kc4 32.e6 and wins.

29.Rh8 b6 30.Na6! Nxa6 31.Rxa8

155
 The active rook and pawns beat Black’s two knights quite easily. AndyAndyO-tsmenace,
ChessWorld.net 2013, concluded 31...Ke6 32.Rxa7 Nb4 33.a4 Kxe5 34.Re7+ Kf4 35.Rb7 Nexc2
36.Rxb6 c5 37.a5 Nd4 38.a6 Ndc6 39.Kd2 Ke4 40.Kc3 Kd5 41.Kb3 Na5+ 42.Ka4 Nc4 43.Rh6 Nxa6
44.Rxa6 Nxb2+ 45.Kb3 Nd1 46.Rh6 c4+ (or 46...Ne3 47.Rxh4 Nxg2 48.Rg4 Ne3 49.Rg5+ and 50.h4
etc with a fairly simple win – the tablebases declare mate in 24) 47.Kb4 c3 48.Kb3 Ne3 49.Rh5+ Kd4
50.Rxh4+ Kd3 51.Rh3 c2 52.Rxe3+ 1-0.

D: 4.d4 g5

Other moves have been played here, but none is really worth considering. Each gives Black a
different defensive set-up with an unhelpful ...h7-h5.

156
5.h4

This is the line GM John Shaw suggests for White in his big book. It has the advantage of
preventing Black from setting up the typical formation with f4/g5/h4 and takes the game into more
orthodox-looking King’s Gambits.
Instead, 5.Bc4 h4 is another route to the main line (E), while 5.Nc3 was examined in line B. White
has also tried:
a) 5.c3 h4 6.Qb3 (intending 6...d6 7.Qb5+) 6...Bg7 7.Rg1? (contemplating g2-g3) 7...Qe7 8.Bd3 f5
9.e5? d6 10.Bxf5 was some over-the-board whimsy in J.Tait-J.Wagenbach, Notts Championship
2000, which Stockfish stamps flat with 10...Bxf5! 11.Qxb7 Nf6 12.Qxa8 0-0 13.Qxa7 dxe5, declaring
a win for Black.
b) 5.d5!? was a casual idea by Paul Cumbers, hoping for something like 5...h4 6.Qd4 Rh7? 7.Qe5+!
Be7 (or 7...Qe7 8.Qxg5) 8.Nxg5 Rh5 (or 8...f6 9.Qf5) 9.Bxf4 f6 10.Qf5 fxg5 11.Qg6+ Kf8 12.Be5
Bf6 13.Bxf6 Qxf6 14.Qxh5 Qxb2 15.Be2 and wins. Simply 6...Rh6 avoids all that.
c) 5.Qd3 can be met by 5...Bh6!? (for 5...h4 6.Nc3 see line B) 6.Qb3 (now if 6.Nc3 then 6...g4!
7.Ng1 Nc6 8.Nge2 Nb4 or 8.Nd5 Nf6 9.Nxf4 d5) 6...Qe7 (if 6...Nc6 7.d5 Nce7, afms-remyrey,
ChessWorld.net 2019, then 8.d6!? is unclear) 7.Nc3 c6 8.Bd2 g4 (White is too fast against routine
play; e.g. 8...d6 9.0-0-0 h4 10.h3 Nd7 11.Re1 and so forth) 9.Ne5 d6 10.Nc4 b5 11.Nxb5! cxb5
12.Qxb5+ Bd7 13.Qb7 Nf6 14.e5 (or 14.Qxa8 0-0 15.Qb7 Nxe4 16.0-0-0 Nxd2 17.Nxd2 Nc6 with
compensation) 14...0-0 15.0-0-0 dxe5 16.dxe5 Qe6 17.exf6 Bc6 18.Qe7 Nd7 19.Re1 Qd5 20.Ne5
Qxa2 21.Qa3 Qxa3 22.bxa3 Rac8 led to a draw in afms-tsmenace, ChessWorld.net 2019.
d) 5.g3!? g4 (5...d5!? is also logical) 6.Nh4 f3 7.Nc3 is a reasonable-looking Quaade. Black could
therefore play it as such with 7...Nc6 8.Bf4 (or 8.Be3 Bb4) 8...d6 9.Qd2 Bg7 10.Bg5, as in
M.Hrubcik-V.Fejfar, corr. 2015, and now 10...Qd7, intending 11.Nf5 f6 12.Bh4 Qf7. Alternatively,
7...Be7 8.Bf4 (or 8.Nf5 Bf6 9.e5 d5) 8...Bxh4 9.gxh4 Qxh4+ 10.Bg3 Qg5 11.Nd5 d6! is not
unthinkable; or if 7.Bf4, blackcrowman-tsmenace, Chess-World.net 2005, then 7...Nf6 8.Nc3 d5.

5...g4

157
6.Ne5

Transposing to a Kieseritzky Long Whip (some of the games below did in fact arrive via this route)
where White has played 6.d4!? instead of the standard 6.Bc4. This means Black is able to drive the
e5-knight away without having to defend f7 first. Nevertheless, White’s strong centre needs to be
taken seriously.
There are two other knight moves too:
a) 6.Ng1 can be compared with the Fischer Defence (swap 3...h5 for 3...d6). In the Wagenbach the
h-pawn is quite useful on h5, defending g4, and the d7-pawn may yet advance two squares. As in the
main line below, pushing on with 6...f3 looks like the best option.

158
 And now:
a1) 7.Be3 Be7 8.Bf2 can be met by 8...d5.
a2) 7.Bg5 is a strong response in the Fischer, and 7...Be7 8.Qd2 fxg2 9.Bxg2 d6 is a possible
transposition there, except that Black has no reason to oblige. Instead, 7...f6 8.Be3 d5! or 8.Bf4 d5
seems perfectly okay.
a3) 7.gxf3 (the usual move) 7...Be7 (or even 7...d5!? 8.exd5 Qxd5, as in P.Cumbers-J.Tait,
Sheffield League 2003) 8.Be3 (not 8.fxg4? Bxh4+ 9.Kd2 hxg4 10.Qxg4?? Qg5+ and wins; while
after 8.Bg5 Bxg5 9.hxg5 Qxg5 10.Qd2 Qxd2+ 11.Nxd2 d6 12.fxg4 Bxg4 Black kept the pawn, Fritz
3-J.Tait, 30 min game, Farnsfield 1995 – back in the day when you still had a chance against the
engines) 8...Bxh4+ 9.Kd2 Bg5 10.f4 (10.Qe2 can be met by 10...Nc6; and 10.fxg4?! Bxe3+ 11.Kxe3
d5 is just asking for trouble) 10...Be7 11.Nc3 Bb4 (Stockfish goes for 11...d5!? 12.Nxd5 f5) 12.Qe2
(or 12.Nge2 d5) 12...Bxc3+ (here too Stockfish wants to push the d-pawn: 12...d5! 13.exd5 Kf8)
13.bxc3 d5 14.Bf2 dxe4 15.Qxe4+ Qe7 16.Qd3 Qf6? 17.Re1+ Kf8 18.d5 Na6 19.Qe4? (19.Ne2 was
far more testing) 19...Bf5 20.Qe5 ½-½ Jo.Nelson-J.Tait, Sheffield League 2003, since 20...Qxe5
21.Rxe5 (not 21.fxe5?! Ne7) 21...Bg6 is quite equal.
b) 6.Ng5!? is an unusual Allgaier

159
 where Black has delayed attacking the intruding knight and White has refrained from assailing f7;
i.e. 3...g5 4.h4 g4 5.Ng5 h5?! 6.d4!?, which was seen as early as the 18th century: 6...f6 7.Bxf4 fxg5
8.hxg5 d5!? (8...d6 is simpler, as in J.Weiss-Von Gomperz, Vienna 1906) 9.e5? Bf5 10.Bd3 Ne7 and
Black was already winning, Cotter-H.Von Brühl, London 1788.
That’s not the final word though. Fred Galvin has drawn attention to 8.Bxg5! Be7 9.Qd2, aiming
for rapid development which Black will struggle to match.

I had in fact faced this once before: silverclaret-tsmenace, ChessWorld.net 2008, saw 9...d6 10.Nc3
Be6 11.0-0-0 Nd7 12.Bd3 Nf8 13.e5? d5 14.Na4 b6 15.b3 c6 16.Rhf1 Kd7 17.Rf2 Kc7 18.Rdf1 Nd7
and Black consolidated. However, White has several improvements: in the game itself 13.d5! Bf7
14.e5 looks close to decisive; and while 12...Bxg5 13.hxg5 Qe7 would avoid that, White can swerve

160
in turn with 11.Bd3! or 10.Bc4! (Galvin), keeping the option of castling short. For example: 10.Bc4
Bxg5 11.hxg5 Qe7 12.0-0 Be6 13.d5 Bc8 14.Qc3!? (or 14.e5 dxe5 15.Na3! – Galvin) 14...Qe5
15.Qxe5+ (unfortunately for Black, the exchange of queens hardly lessens the pressure) 15...dxe5
16.Nc3 Rh7 17.d6 Rg7 18.Nb5 Na6 19.dxc7 Nxc7 20.Nxc7+ Rxc7 21.Bxg8 and having regained the
piece White was clearly better, tsmenace-remyrey, ChessWorld.net 2019.
Black’s main problem is in finding a way to castle long safely – or even at all. Hence I decided that,
should I ever face this line again, I might try 9...Nf6 and 10...Nh7!? with the idea of castling short.
Despite the airy kingside, the closed g- and h-files offer cover for Black who still has an extra piece.
I was later very surprised to discover a game from the Top Chess Engine Championships with this
exact defence: 9...Nf6 10.Nc3 Nh7!? 11.Bf4 (here 11.Bxe7 Qxe7 12.0-0-0 d6 13.e5 0-0 14.Bc4+ Kh8
or 12.Bc4 d5!? 13.Bxd5 c6 14.Bb3 g3!? is another unresolved mess) 11...d6 12.0-0-0 0-0 13.Bc4+
Kh8 14.Bh6 Nc6 (obviously Black doesn’t mind returning the exchange) 15.Nd5 Bd7 16.Kb1 Be8
17.a3!? (don’t ask me why) 17...Rg8 18.e5 Rg6 19.Bf4 Bf8 20.Bg5 Nxg5 21.hxg5 Qxg5 22.Nxc7
Rd8 23.Nxe8 Rxe8 24.Bf7 Rd8 25.Bxg6 Qxg6 26.exd6 Bxd6 27.Qd3 Kg7 led to a draw in Komodo-
Stockfish, TCEC 13 Superfinal 2018.
All in all, 6.Ng5!? is a very interesting positional piece sacrifice, whose soundness is evidenced by
the fact that Stockfish only managed to draw. And it won the return game, which saw 9...Rh7 10.Nc3
Rg7 11.Bxe7 Qxe7 12.e5 Qe6 13.Bd3 Nc6 14.Nb5 d5 15.0-0 Qh6 16.Qf2 Re7 17.c4 with a decisive
attack, Stockfish-Komodo, TCEC 13 Superfinal 2018.

6...d6

7.Nd3

Sacrificing with 7.Nxf7?! is over-optimistic, even if it succeeded in J.Kvicala-NN, Prague 1891 (via

161
3...g5): 7...Kxf7 8.Bc4+ Be6 9.Bxe6+ Kxe6 10.Bxf4 Bg7 11.c3 Nf6? (if Black runs away with
11...Kd7, White has nothing) 12.Nd2 Nbd7? (blocking the king’s path) 13.Qb3+ Ke7 14.e5 dxe5
15.dxe5 Ne8? 16.Bg5+ Ndf6 17.0-0-0 and White won.

7...f3!

Shaw correctly identifies this move as best: “if Black is to stand any chance of justifying his
opening play.” Personally I think Black has every chance of that.
Shaw actually gives 7...Be7 8.Bxf4 Bxh4+ 9.g3 as his main line, writing: “White has great
compensation: a lead in development, central domination and plenty of targets on the kingside.”

Then:
a) 9...Bg5 was played in our first recorded 3...h5 game. D.Newhouse-J.Wagenbach, Burnley (rapid)
1992, continued 10.Nc3?! (by transposition an even earlier game saw 10.Be2?! Nc6 11.Qd2 Bf6
12.e5 dxe5 13.dxe5, J.Louma-B.Prikryl, Dustoj 1937, when 13...Bg7 seems playable; but 10.Qd2! at
once looks very good for White) 10...c6? (10...Nc6 isn’t so bad) 11.Qd2 Bxf4 12.Nxf4 Qg5 13.0-0-0
h4 14.gxh4 Rxh4 15.Rxh4 Qxh4 16.Bg2 and with Rh1 coming next White was very much better.
b) 9...Be7 10.Nc3 (not 10.Bg2?, S.Tidman-J.Wagenbach, Birmingham rapid 1998, which just
encourages 10...h4) 10...Nc6 11.Be3 (not 11.d5? Nd4) 11...h4 12.Nf4 leaves Black with probably
insolvable problems over development.
c) 9...Bf6 gives White a tempo to push the e-pawn: 10.e5! (better than 10.c3 h4 as in J.Tait-
J.Elburg, corr. 1998) 10...Be7 (10...dxe5 11.dxe5 is no improvement) 11.Nc3 Nc6 12.Qd2 Nxd4 13.0-
0-0 Ne6 (or 13...Nf3 14.Qe2 d5 15.Ne1! c6 16.Nxf3 gxf3 17.Qxf3 Be6 18.Ne4! with excellent play –
Shaw) 14.exd6 cxd6 15.Nd5 h4 16.Kb1 Nf6 17.Nxe7 Qxe7 18.Re1 Rg8 (or 18...h3 19.Nf2!) 19.Rxh4
1-0 M.Mueller Alves-J.Simmelink, corr. 2007.

162
8.Be3

Ignoring the f-pawn looks like the most promising plan for White.
a) 8.Bf4, with the same idea, is inferior because Black can exploit the unprotected d4-pawn. For
example:
a1) 8...Qf6!? 9.Nc3?! (neither 9.c3 Bh6, nor 9.e5 Qf5 seems to get White anywhere) 9...Qxd4
10.gxf3 Be6 11.Nd5 Bxd5 12.c3 Qg7 13.exd5 Nd7 14.Qa4 a6 15.0-0-0 0-0-0 and White’s
compensation for the pawn is nebulous, even if I eventually managed to grovel a draw, tsmenace-
samurai, ChessWorld.net 2013.
a2) 8...Nc6 9.c3 Nce7! may be even better; e.g. 10.gxf3 Ng6 11.Bg3 (or 11.Be3 Bh6) 11...Be7
12.Nd2 Bxh4 13.Bxh4 Nxh4 14.Qe2 Qe7 15.0-0-0 Nxf3 16.Nxf3 gxf3 17.Qd2 Be6 18.Bh3 0-0-0 and
Black consolidated, Decus Tutamen-tsmenace, ChessWorld.net 2020.
b) 8.gxf3 “should be met by 8...Be7” (Shaw) – rather than 8...gxf3 9.Qxf3 Bg4 10.Qf2,
O.Wieczorek-A.Iwanow, Wroclaw 2005 – though 8...Be7 is not at all a novelty.

163
 b1) 9.Bg5 Bxg5 10.hxg5 Qxg5 11.Qd2 (or 11.fxg4? Nc6 12.c3 Bxg4, K.Matthews-F.Eastwood,
corr. 1995) 11...Qxd2+ 12.Nxd2 Nc6 13.c3 gxf3 14.Nxf3 Bg4 15.Kf2 Kd7? (both weak and pointless
since 15...0-0-0 and ...Rce8 is just as fast) 16.Bh3? (missing the chance for 16.d5! Nce7 17.Nfe5+!
dxe5 18.Nxe5+ Kc8 19.Nxg4 Ng6 20.Ne3, when White has the edge) 16...Nf6 17.Rae1 Rae8 and
Black consolidated the extra pawn, D.Flynn-J.Tait, Notts League 2010.
b2) 9.c3!? (suggested by MNb on the ChessPublishing forum)

9...Bxh4+ 10.Kd2 gxf3 (10...g3!? 11.Kc2 Bf6 and ...h5-h4 is more belligerent) 11.Kc2 Bg4 12.Nd2
Qe7 13.Nxf3 Qxe4 14.Bg2 (or 14.Rxh4 Qxf3) 14...Be7 15.Qf1 Nd7 16.Nfe5 Qe2+ 17.Qxe2 Bxe2
18.Bxb7 Rb8 19.Bf3 dxe5 20.Bxe2 exd4 21.cxd4 Bf6 22.Bxh5 Bxd4 led to a draw in Raffzahn-
tsmenace, ChessWorld.net 2014.

164
 b3) 9.Be3 Bxh4+ 10.Kd2 Bg5 11.f4 Bh6!? (Stockfish prefers 11...Bf6; both improve on 11...Be7
12.Nc3 f5? 13.Bg2 c6 14.Qe2 Kd7 15.Rae1 Kc7? 16.Kc1 fxe4 17.Nxe4 Qf8 18.d5! Bd7 19.f5! with a
crushing position for White, JHDonner-tsmenace, ChessWorld.net 2010) 12.Kc1 (not 12.Rxh5? Nf6
and ...Nxe4+; while if 12.Nc3 then 12...Bg7 and ...Nc6 targets the d4-pawn) 12...h4 13.Nc3 h3
14.Qe2 Nf6 (having secured the connected passed pawns in enemy territory, Black starts to get pieces
out) 15.f5!? (trying to make something happen) 15...Nc6 16.Bxh6 Rxh6 17.Qe3 Rh8 18.Nf4 Ne7
19.Bd3 c6 20.Nfe2? (clearing the way for the queen is too slow; 20.e5!? was more to the point)
20...d5 21.e5 Ne4! 22.Bxe4 dxe4 23.Nxe4 Nxf5 24.Qf4 Qh4 25.Nf6+ Ke7 26.Nxg4 Nxd4 27.Qxd4
Qxg4 and Black went on to win, tsmenace-Magellen, ChessWorld.net 2010.

8...Nc6!

A notable idea by Joop Simmelink, not least because it gets a piece out at last! I think it was
Maróczy who first talked about returning a gambit pawn “at the right time”, and here Black does just
that in order to facilitate development.
Taking the g2-pawn makes it a permanent gambit, but White should be quite content since Black
has no counterplay and no way to develop smoothly. For example:
a) 8...Nf6 9.Nc3 fxg2 10.Bxg2 was W.Mead-B.Pritchett, Brighton 1885, where 10...Qe7? 11.Bg5
led to a quick crush: 11...c6 12.Rf1 (or 12.0-0!) 12...Bg7 13.Qe2 0-0 14.e5 dxe5 15.dxe5 Qc7 16.Bxf6
Nd7 17.Qe3 Re8 18.Bxg7 Kxg7 19.Qg5+ 1-0.
b) 8...Be7 9.Bf2 fxg2? (9...c5! would make a game of it) 10.Bxg2 c6 11.Nc3 Nd7 12.Qe2 Nf8
13.Bg3 Ng6 14.Qf2 Bd7 15.0-0-0 and White is clearly better, Shaolin Fist-tsmenace, ChessWorld.net
2006.
c) 8...d5!? 9.gxf3 Be7 10.Bf2 “gave White the upper hand” (Shaw), which looks to be true after
10...dxe4?! 11.fxe4 Nf6, S.Istomin-D.Raleus, corr. 2001, and now 12.Nc3!. But 10...Nc6 improves

165
there, so White should just take the pawn with 10.exd5.

9.gxf3

Reinforcing the d-pawn with 9.c3 takes the square from the b1-knight and reduces White’s activity,
so that 9...fxg2 becomes a more serious possibility, or perhaps 9...Bh6!? first.
If 9.d5 then certainly 9...fxg2, since Black no longer has any problems with development; e.g.
10.Bxg2 Ne5 11.Nxe5 dxe5 12.Qd2 Ne7 13.Bg5 Qd6 14.Na3 Ng6 15.Rf1 Bd7 16.Rf6 (or 16.0-0-0
Bg7, when ...0-0 is perfectly safe) 16...Qb4 17.c3 Qa4 18.Nc2 Bg7 19.d6!? Be6 20.0-0-0 Qd7!
21.Ne3 (or 21.dxc7 Qxd2+ 22.Rxd2 Bxf6 23.Bxf6 0-0) 21...Qxd6 22.Qf2 Qb6 23.Nd5 Bxf6!
24.Nxf6+ (or 24.Nxb6 Bxg5+ 25.hxg5 axb6) 24...Kf8 and Black won in benny1501-tsmenace,
ChessWorld.net 2015.

9...Bh6

Black contests the dark squares, utilizing the fact that the white d-pawn is slightly vulnerable. For
instance, the desirable 10.Qd2 now loses to 10...Nxd4.

10.f4

Closing the c1-h6 diagonal and increasing White’s central pawn presence at the cost of resolving
the tension on the kingside.
White doesn’t want to exchange bishops here, as Black can then develop smoothly; e.g. 10.Bxh6
Nxh6 11.c3 Qe7 12.Be2 gxf3 13.Bxf3 Bg4 14.Nd2 0-0-0 15.Qe2 f5 and suchlike.
The only logical alternative is to use the knight, 10.Nf4, when praxis has seen 10...Bxf4 11.Bxf4
Qf6 12.Be3 Nge7 (the safe option; keeping the pawn with 12...gxf3!? is quite risky: 13.Qd2 Bg4
14.Nc3 0-0-0 15.0-0-0 Qg6 16.d5 Ne5 17.Qd4 and the black king is rather vulnerable on the

166
queenside with so many enemy pieces pointing in that direction, though Black drew in F.Irmer-
J.Simmelink, corr. 2008) 13.Nc3 Qxf3 14.Qxf3 gxf3, followed by ...Bg4, and Black is at least equal,
Reprimand-AndyAndyO, ChessWorld.net 2014.

10...d5!

With the dark square issues resolved, Black switches play to the other colour complex, and this is
the most effective way to do so.

11.Nc3

Obviously not 11.e5? Nge7 and Black is in full control.

11...dxe4 12.Nc5

Nor 12.Nxe4? Qe7 13.Bg2 Bf5.

167
12...f5?!

This was AndyAndyO-tsmenace, ChessWorld.net 2013. It’s become almost a fixed rule for me now
never to play ...f7-f5 in the Wagenbach. There may be no obvious short-term refutation, but the
squares it leaves weak behind can cause Black endless difficulties.
Here 12...Nf6 is more solid; e.g. 13.d5 (before Black plays ...Ne7/Nb4 and ...Nd5) 13...Ne7 (or
13...Nb4!? 14.Bb5+ Kf8) 14.Bd4 (or 14.N5xe4 Nxe4 15.Nxe4 Bg7 16.Bc4 0-0) 14...Nf5 15.N5xe4
Nxe4 16.Nxe4 (not 16.Bxh8? Neg3) 16...Qe7 17.Bg2 Nxd4 18.Qxd4 0-0 and Black seems safe
enough.

13.d5 Ne5

Making use of the ‘pin’ on the c1-h6 diagonal. 13...Nce7 14.Qd4 Ng6 15.0-0-0 gives White decent
play for the pawn(s).

14.Bb5+

Or 14.fxe5 Bxe3 15.Bb5+ Kf7, which comes to the same thing.

14...Kf7 15.fxe5 Bxe3 16.N5xe4!

The sacrifice is perfectly good. This is just as well since by now White has nothing else.

16...fxe4?

Stockfish prefers to decline the piece and sacrifice one of its own: 16...Kg7 17.Qd3 f4 18.e6 c6
19.Bc4 b5 20.Bb3 Bxe6! 21.dxe6 Qxd3 22.cxd3 Ne7, which it assesses as “0.00”, presumably

168
judging that the kingside pawns will recoup the material. I doubt I even considered this.

17.Qe2 Bd4 18.Qxe4 Bxc3+ 19.bxc3 Kg7 20.0-0-0

White has two big pawns for the knight, with strong threats against the black king after moves like
e5-e6, Rhf1 and Bd3; meanwhile all of Black’s pieces are on the back rank.

20...Rh6 21.e6!

Shutting Black’s queenside out of the game.

21...Nf6 22.e7?

Letting me off. I had thought I was okay anyway, but Stockfish disagrees: 22.Qf5! and it’s hard to
see how to secure Black’s position. For example: 22...a6 (or 22...c6 23.dxc6 Qc7 24.Bd3 Kh8 25.Qg5
Ng8 26.Rhe1 Bxe6 27.Bc4, intending 27...Bxc4? 28.Rd7 etc) 23.Be2 b5 24.Rhf1 Qe7 (or 24...Bb7
25.Bd3! Qxd5 26.Qf4 Qc5 27.Bf5, intending Qg5+) 25.Qe5! Bb7 (or 25...Bd7 26.Bd3 Be8 27.Qg5+
Bg6 28.Rf5 Raxf8 29.Rdf1 and Black can do nothing while White prepares to push the two c-pawns;
if 29...c6? then 30.dxc6 Qxe6 31.c7 wins) 26.d6! cxd6 27.Rxd6 Kg8 28.Qg5+ Qg7 29.e7 Re8 30.Rd8
Bf3 (or 30...Bc6 31.Qc5) 31.Bxf3 gxf3 32.Rxf3 Kh7 33.Rxf6 Qxf6 34.Qxf6 Rxf6 35.Rxe8 Re6
36.Kb2 with a winning rook endgame. Conclusion: Never play ...f7-f5 in the Wagenbach!

22...Qe8!

Throwing the queen just to slow the e-pawn down is a nice move to make.

23.Bxe8 Nxe4 24.Bc6 Nd6 25.e8=Q Nxe8 26.Bxe8 Bf5

169
 Black has survived and even has a nominally superior pawn structure, if no actual winning chances.
The game concluded 27.Rhe1 Rd8 28.c4 Rh8 29.Re7+ Kh6 30.Ba4 Rh7 31.Re5 Rf7 32.c5 Rf6
33.Bb3 Rd7 34.d6 cxd6 35.Rxf5! Rxf5 36.Be6 Rff7 37.Bxd7 ½-½, since 37...Rxd7 38.Rxd6+ Rxd6
39.cxd6 g3 leads to a drawn queen endgame.

170
The Main Line
E: 4.d4 g5 5.Bc4 h4

or 4.Bc4 h4 5.d4 g5

Finally, the main line – the basic tabiya of the Wagenbach, where Black has completed the desired
formation (f4/g5/h4) around the extra pawn. In return White has a big centre, two pieces out to none,
and now “only” needs to decide how to prosecute the initiative.
E1: 6.Ne5
E2: 6.0-0
E3: 6.Nc3
Instead, 6.Qe2 d6 7.Nc3 is covered in line E3; as is 6.h3 d6 7.Nc3, while 7.0-0 c6 8.c3 transposes to
E2. Several other moves have been seen as well:
a) 6.Qd3 plans a “trick” (after ...d7-d6) with Qb3 and Qb5+ to pick up the g5-pawn, but Black can
just allow it: 6...d6! 7.Qb3 (for 7.Nc3 see the notes to E32) 7...Qe7 (7...Nh6!? may be even stronger,
and if 8.Qb5+, as in S.Young-D.Toye, corr. 1996, then 8...Nc6 9.d5 a6 10.Qa4 h3!) 8.Qb5+?! Nc6
9.Bd5 (but if 9.Nc3 then 9...Rh5; and taking the g5-pawn is even worse: 9.Nxg5 Bh6 10.Bxf7+ Kd8
11.Bxf4 a6, or 9.Qxg5 Nf6! 10.Nc3 h3!, intending 11.g3?! fxg3 12.hxg3 Bh6 13.Qb5 Bxc1 14.Rxc1
Bd7 15.Qxb7?! Rb8 16.Qxc7 Rc8 17.Qb7 Na5 and wins) 9...Bd7 10.Qxb7 (or 10.Nxg5 Nd8)
10...Rb8 11.Qa6 (or 11.Qxc7 Rc8 12.Qb7 g4) 11...Rb6 12.Qe2 g4 13.Bxc6 Rxc6 14.Ng1 Bg7 15.c3
Bxd4 16.Bxf4 Bb6 and Black is clearly better, D.Sakellarakis-D.Koetsier, corr. 1999.
b) 6.c3 has the same idea and can be met similarly:

171
 6...d6! (not 6...Be7? 7.Ne5! Nh6 8.Qh5 Kf8 9.Bxf7 Rh7, BABA 2004-tsmenace, ChessWorld.net
2005, as 10.g3! is then very strong) 7.Qb3 (7.0-0 is line E21) 7...Qe7 8.Qb5+?! Nc6 9.Nxg5 (if 9.Bd5
then 9...Bh6 10.Bxc6+ Kf8!, or 9.Qxg5 Nf6 10.0-0 h3 11.Qxf4 hxg2 12.Rf2 Ng4 13.Rxg2 Bh6
14.Ng5 f6) 9...f6 10.Nf3 (or 10.Nf7 a6, followed by ...Rh7) 10...Qxe4+ 11.Kf2 a6 12.Qb3 Nge7
13.Re1 Qf5 and Black is at least equal.
c) 6.g3? seeks to blow up Black’s formation at once. I analysed this with Paul Cumbers. Then
6...hxg3 7.Ne5 can lead to an interesting draw: 7...Rxh2 8.Rxh2 gxh2 9.Bxf7+ Ke7 10.Qh5 Nf6
11.Ng6+ Kxf7 12.Ne5+ Ke6 13.Qf7+ Kd6 14.Nc4+ Kc6 15.Ne5+ Kd6 (15...Kb6?? 16.Qb3+ leads to
mate) 16.Nc4+ etc. Black can improve by throwing in 7...d5!, when 8.Bxd5 Be6 9.Bxb7 Rxh2 is
pretty much winning.
d) 6.Kf2? tries to prepare g2-g3 by covering the h1-rook first. The white king is not entirely happy
with this plan. After 6...d6 7.g3, rather than get involved in 7...hxg3+ 8.hxg3 Rxh1 9.Qxh1 g4
10.Bxf7+!? Kxf7 11.Qh5+ Kg7 12.Ng5 Nh6, C.Davies-Mi.Lane, corr. 1998, Black can play 7...g4
8.Nxh4 Rxh4 9.gxh4 Qxh4+ and ...f4-f3; or else 7...Nf6!?, intending 8.Re1? Ng4+ 9.Kg1 hxg3
10.hxg3 Rh3 and wins (Wagenbach).
e) 6.e5?, as in E.Efendiyev-J.Simmelink, corr. 1999, is premature and can be met by 6...d5! 7.exd6
Bxd6 with a definite advantage. Black is a pawn up and can develop easily, and the king will be quite
safe on f8.
f) 6.Bxf7+?? Kxf7 7.Ne5+ Kg7 8.Bxf4 is no good at all, as long as Black avoids 8...gxf4?? 9.Qg4+
and mates. After 8...d6 9.0-0 gxf4 Black was just two pieces up in DanielP4R-tsmenace,
ChessWorld.net 2005.

E1: 6.Ne5

172
 This move has been played 82 times to date, with Black scoring 77% (+56 =14 -11); and those
eleven losses included a default, an input error, and a 3 minute blitz game blunder. In other words, the
direct attack on f7 is nothing to be afraid of – with the proviso that Black is sufficiently accurate over
the next few moves.

6...Nh6

Not 6...Rh7?! 7.Bxf7+! Rxf7 8.Nxf7 Kxf7 9.g3 and Black’s position is already difficult.

7.Qh5

The consistent continuation; otherwise 7...d6 drives the knight away again at once.
The line-opening 7.g3?! is dubious: 7...d6 8.Nf3 (after 8.gxf4 dxe5 9.fxg5 Bg4 White has very little
for the piece; while 8.Nd3? Bg4 9.Qd2, D.Flynn-J.Wagenbach, Notts League 2006, and 9...Nc6 is just
good for Black) 8...hxg3 9.hxg3 g4 (even 9...fxg3!? 10.Bxg5 g2 11.Rg1 Be7 is probably about equal)
10.Nh4 (10.Bxf4? gxf3 11.Qxf3 is no good; e.g. 11...Bg4 12.Qf2 Nc6 13.Bxh6 f5!) 10...fxg3 (10...f3?
11.Nc3 looks like decent compensation) 11.Ng2 Nc6 12.c3 Qe7 13.Qe2 Bd7 14.Bf4 0-0-0 and Black
is better.

7...Qf6

173
8.Nc3

On reaching this position, White usually realizes that 8.Nxf7 doesn’t win material and gets the
other knight out instead, which is indeed the only good move.
a) 8.Nxf7? Nxf7 9.Bxf7+ fails to 9...Ke7! and White loses material: 10.Qg6 Qxf7 (Stockfish
actually prefers 10...Nc6! 11.c3 Qxf7 or 11.Qxf6+ Kxf6, followed by ...Nxd4) 11.Qxg5+ Qf6 12.Bxf4
(or 12.Qxf6+ Kxf6 13.Bxf4 d6, R.Fulford-J.Wagenbach, Leighton Buzzard rapid 1995) 12...Qxg5
13.Bxg5+ Ke8 and Black is clearly better, P.Cumbers-J.Tait, Grantham (rapid) 1993.
b) 8.Nf3? can be met by 8...d5! 9.Bxd5 Be7 and the white queen is in trouble; e.g. 10.e5 (if 10.h3
then 10...Nf5, or 10.0-0 Qg7) 10...Qg7 11.e6 Kf8.
c) 8.g3? is even less effective now: 8...d6 9.gxf4 dxe5 (after 9...gxf4 10.Nf3 Be7? 11.e5 dxe5
12.Qxe5 Qxe5+ 13.Nxe5 White came out okay in P.Cumbers-J.Wagenbach, Grantham rapid 1998)
10.fxg5 (neither 10.fxe5 Qb6, nor 10.dxe5 Qc6 is any good for White) 10...Ng8! 11.Bxf7+ Qxf7
12.Qxh8 Nc6 and Black is clearly better; e.g. 13.Rf1 Qg6 14.Qxh4 Nxd4 15.Rf2 Nh6! 16.gxh6 Be7
17.Qg3 Qxe4+ 18.Be3 Nxc2+ 19.Rxc2 Bh4.

8...d6!

János’s discovery; Black has no need to fear the knight coming to d5. Previously we looked to
prevent it with either 8...Bb4 or 8...c6, allowing White a greater initiative.

174
9.Nd5?

The natural move, but not the best.

It’s still bad to take the f-pawn: 9.Nxf7? Nxf7 10.Bxf7+ Kd8 (obviously not 10...Ke7?? here)
11.Qg6 (or if 11.Nd5 Qg7 12.Qg6 c6 13.Nc3, G.Allenby-J.Wagenbach, corr. 1998, then 13...Ke7!
pretty much wins) 11...Qxg6 12.Bxg6 Nc6 13.Ne2 Bg7 14.c3 Ke7 (intending ...Kf6, when Black will
regain the pawn and consolidate) 15.e5 (or 15.Rf1 Kf6 16.Bf5 Bxf5 17.exf5 Rae8, T.Walker-J.Tait,
Notts League 2001) 15...dxe5 16.d5 Nd8 17.c4 Kf6 18.Bb1 Bf5 19.0-0 Nf7 20.b3 Bxb1 21.Rxb1 c6
22.dxc6 bxc6 23.Bd2 Rad8 and Black went on to win, Deepmind_5-tsmenace, ChessWorld.net 2006.
White should retreat and take action before ...Bg4 traps the queen: 9.Nf3! Rg8 10.e5! (not 10.Nd5?!
Qd8 which transposes to the main line, while 10.h3 loses the option of breaking up the kingside,
allowing Black to consolidate more easily; e.g. 10...c6 11.e5 Qg6 12.Qxg6 Rxg6 13.exd6 Bf5 14.0-0
Bxd6 15.Re1+ Kd8 16.Ne4 Bxe4 17.Rxe4 Nd7 18.Bd3 Kc7 19.Bd2 Rgg8, Magellen-tsmenace,
ChessWorld.net 2010) 10...Qg6 11.Qxg6 Rxg6 12.Nd5 Kd8 13.g3! (Black now needs to be precise)
13...Be6? (insufficiently forcing; both 13...Bg4 14.Be2 Bxf3 15.Bxf3 Nc6 16.Be4 Nxd4 and 13...b5!?
14.Bd3 Bb7 15.Be4 dxe5 improve for Black, who is not worse) 14.gxf4 c6 15.Ne3 gxf4 16.Nxh4 Rg5
(or 16...d5 17.Nxg6 fxg6 18.Ng2 dxc4 19.Nxf4, when the rook and soon-to-be two pawns look
superior to Black’s uncoordinated minor pieces) 17.Nf3 fxe3 18.Bxe6 Rh5 19.Bb3 Nd7 20.Bxe3 dxe5
21.dxe5 Ng4 22.Bd4 Ndxe5 23.Nxe5 Nxe5 24.0-0 Bd6 25.Rae1 Kc7 26.Bxf7 Rg5+ 27.Kh1 Rh8
28.Re2 and now a pawn up with the two bishops, White went on to win, AndyAndyO-samurai,
ChessWorld.net 2013.

9...Qd8!

175
10.Nf3

The only move. All sacrifices lose:

a) 10.Nxf7? Nxf7 11.Qg6 Rh6, D.Jones-J.Wagenbach, Bolton (rapid) 2001.
b) 10.Nxc7+? Qxc7 11.Nxf7 Nxf7 12.Bxf7+ Qxf7 13.Qxh8 Bg4, R.Harskin-J.Wagenbach,
Nottingham 1996.
c) 10.Nb6? dxe5 11.Nxa8 Bb4+! 12.c3 Bg4 13.Bxf7+ Kf8 (all as I gave in Chess Mail) 14.Qg6
Nxf7 15.cxb4 Qxd4 0-1 bridlingtonbrummie-tsmenace, ChessWorld.net 2009.
d) 10.Bxf4? gxf4 11.Nb6 dxe5 (not 11...axb6?? 12.Bxf7+ Ke7 13.Qg5 mate) 12.Qxe5+ Qe7
13.Nxa8 Qxe5 14.dxe5 c6 and the white knight will be trapped later, stormytlc-tsmenace,
ChessWorld.net 2009.

10...Rg8 11.h3

White has to answer the threat of ...Bg4 and there is no e4-e5 this time. Stockfish actually suggests
11.Nxg5!? Rxg5 12.Qxh4, though 12...Nd7 13.Nxf4 (or 13.Bxf4 Rxg2) 13...Rg8 is clearly good for
Black.

11...c6

Another option is 11...Nd7, as in P.Adams-J.Wagenbach, corr. 1998, intending a quick ...Nf6, or if

12.e5 then 12...c6 13.Nxf4 gxf4 14.Bxf4 Nb6 15.Bd3 (or 15.Bxh6 Nxc4) 15...Nd5 16.Bd2 Be6 17.c4
dxe5!? 18.cxd5 Qxd5 19.Bxh6 Rh8 is at least a little better for Black.

12.Nxf4!

176
 Fighting for the initiative. After 12.Nc3 Nd7, the threat of ...Nf6 again forces White’s other knight
to scurry backwards as well.

12...gxf4 13.Bxf4 Qf6 14.Bg5

Swapping the queens off at this stage, such as with 14.Qxh4 Qxh4+ 15.Nxh4, would be an
admission that White’s opening has failed.

14...Rxg5!

Returning an exchange is a big improvement on 14...Qg6 15.Qxh4, where White is close to full
mobilization and can look forward to good play against the enemy king; e.g. 15...Nd7 16.0-0-0 f6,
GoBoSox-tsmenace, ChessWorld.net 2005, and now 17.Bf4! Nf7 18.Rde1 Be7 19.Qf2 Nb6 20.Bb3
Bd7 (20...Qxg2? 21.Qe3 looks very greedy) 21.a4 0-0-0 22.a5 Na8 23.e5 with two pawns and an
ongoing initiative for the piece.

15.Nxg5

After 15.Qxg5 Qxg5 16.Nxg5 Be7 17.Nf3 Nd7 Black’s extra bishop and knight are stronger than
White’s rook and pawn.

15...Qxd4 16.Nxf7

If 16.Bxf7+ Kd8 17.Qxh4 Qe3+ 18.Kd1 Kc7, it’s White’s king who has the greater problems.

16...Qxe4+

177
17.Kd1?

The last hope was 17.Be2 Nxf7 18.0-0 with threats on the e-file and against f7, but Black can afford
to give a piece back to evacuate the king: 18...Be6 19.Bg4 Nd7 20.Rae1 Qd4+ 21.Kh2 0-0-0 22.Bxe6
Nfe5 23.c3 Qd3 24.Qxh4 Kc7 and Black has all the chances in the middlegame before White can
even think about utilizing the kingside pawns.

17...Qd4+ 18.Kc1?

Black should win anyway after 18.Ke1 Kd7 19.Nxh6 Qe3+ 20.Qe2 Bxh6.

18...Qf4+ 19.Kb1 Nxf7

Now with three pieces for a boxed-in rook Black won quickly in Redddy-Coroticus,
ChessWorld.net 2006.

E2: 6.0-0

Castling is the most natural and popular move. White completes kingside development and is then
ready to take a decisive initiative. Perhaps. Stockfish indicates that White may already have gone
wrong.

178
6...d6

The standard reply; Black opens the diagonal for the c8-bishop, takes the e5-square from the white
knight, and can now close the centre in response to e4-e5.
The only “master” game to feature the Wagenbach saw 6...Nh6!?, which is not entirely bad, even if
it’s not clear that the knight wants to be there just yet. N.McDonald-R.Dive, 4th Café Baroque,
London 1994, continued 7.Nxg5!? (more speculation than an outright refutation; 7.Nc3 d6 8.e5 looks
more critical) 7...Qxg5 8.Bxf4 Qg4?? (this is no good without ...d7-d6 already in; 8...Qg6 was correct
– and equal according to Stockfish, though it might be less easy for a human to prove that over the
board) 9.Bxf7+! Kd8 10.Qxg4 Nxg4 11.h3 Nh6 12.Be5 Nxf7 13.Bxh8 Ke8 14.Bf6 (now White will
have rook and three pawns for two undeveloped pieces at a minimum; in fact the game didn’t last
much longer) 14...b6 15.Nc3 Bb7 16.Nd5 Bd6 17.Bxh4 Bxd5 18.exd5 Na6 19.Rf5 1-0.

179
 E21: 7.c3
E22: 7.Nc3
In line E22 we’ll focus particularly on piece sacrifices at f7 and/or g5 with 7...c6 8.Bxf7+ or
8.Nxg5. The immediate 7.Nxg5? and 7.Bxf7+? are examined there too; as we’ll see, they’re not much
cop for White.
Instead:
a) 7.h3 c6 transposes below after 8.c3 (except for 8.Nc3, which is considered in E31); similarly
7...Be7, followed by 8...c6. Fianchettoing the bishop with 7...Bg7 seems less appropriate as it leaves
both g5 and d6 a little vulnerable. After 8.c3 (for 8.Nc3 see E31 again) 8...c6 9.Nbd2 Nh6,
DraigMawr-tsmenace, ChessWorld.net 2005, White might play 10.Bb3 with ideas of Nc4 and Nxg5;
e.g. 10...Nd7 11.Nc4 Nb6 12.Nxg5! Nxc4 (not 12...Qxg5? 13.Bxf4 and 14.Nxd6+) 13.Bxf4 b5 14.a4
with strong play for the piece.
b) 7.g3? is no good at all here: 7...hxg3 8.hxg3 Bh3 9.Nxg5? (White’s position is terrible anyway)
9...Qxg5 10.Bxf4 Qh5 11.Be2 Bg4 12.Bxg4? Qh2 mate, H.Siebelhoff-R.Hendriks, Haarlem 1999.
c) 7.e5? is very much premature: 7...d5 8.Bd3 Be6 (vs. c2-c4 which might revive White’s position)
9.c3 (or 9.c4 dxc4 10.Be4 c6) 9...Nd7 10.b3 Nh6 11.a4 Be7 12.c4 c6 13.Ba3?! Ng4 14.Bxe7 Qxe7
15.Qd2 Ne3 16.Re1? g4 0-1 sumbadger-tsmenace, ChessWorld.net 2009.
d) 7.Qd3 c6 8.Qb3 is no more effective than in line E21 below: 8...Nh6 9.Nc3 Bg7!? (9...Be7 is
more usual) 10.Rd1 Nd7 11.Be2 Nf8 12.d5 Ng6 13.dxc6 bxc6 14.Qb4 Bf8 15.Bd2 Be7 and despite
having “wasted” two moves with this bishop, Black still has a very good position, G.Allenby-
J.Wagenbach, corr. 1999.

E21: 7.c3

180
 I faced this move in my very first game with the Wagenbach (see note ‘a’ below). White
overprotects the d-pawn in Hanstein fashion and clears the path for the queen to b3. On the other
hand, it’s not easy to see how to proceed from there, since f7 is easily defended.

7...c6

Usual, if not essential here. 7...Be7 or 7...Nh6 can be played first.

8.h3

Another frequently seen move, preventing variously ...Bg4, ...h4-h3, and ...g5-g4 (at least for the
time being). While this has merit if part of a cohesive plan (as we’ll see in line E31), here it does little
more than create a weakness where White has just castled. In response, the idea of ...Be7
(overprotecting g5), ...Nd7-f8-g6 (overprotecting f4), and then ...g5-g4 has proved to be very
dangerous.
For reasons of thematicity, 8.Nxg5 and 8.Bxf7+ are covered in the notes to line E22 below. Others:
a) 8.Qb3 Nh6 9.d5 (instead, 9.h3 Be7 transposes below; whereas after 9.Nxg5? Qxg5 10.Bxf4 Qg6
11.Bxh6 Qxh6 12.Bxf7+ Kd8 White had nothing much for his sacrifice, J.Van der Kooij-J.Elburg,
corr. 1998) 9...Be7 10.e5? Nd7?! (the dubious mark is because 10...dxe5! 11.Nxe5 Bc5+ 12.Kh1 Nf5
and ...Ng3+ wins at once) 11.dxc6 bxc6 12.Bxf7+? Nxf7 13.e6 Qb6+ (and here 13...Nfe5! 14.exd7+
Qxd7 is stronger) 14.Nd4 Qxb3 15.axb3 Nfe5 16.exd7+ Bxd7 and Black was a pawn up for less than
nothing, J.Lyth-J.Tait, Rhyl 1992.
b) 8.b4 Be7 9.a4 Nh6 10.Ra2 (after 10.Qb3 h3! 11.d5? hxg2 12.Kxg2 g4 13.Nd4 f3+ 14.Kg1 Bf6
15.Ra2 Be5 Black had a decisive attack in Steve Cullip-tsmenace, ChessWorld.net 2009) 10...Bg4
(this was an early game; 10...Nd7 and ...Nf8-g6 is more standard now) 11.h3 Bh5 12.Qe1 0-0 13.Raf2
d5!? (13...Bxf3 might be more prudent) 14.Bd3 Bg6 15.Nxh4!? (very tempting once Black has

181
castled short) 15...gxh4 16.Bxf4 Bg5 17.Nd2 Nd7 18.Qe3 Bxf4 19.Rxf4, N.Lazonby-J.Wagenbach,
Notts League 1995, and now 19...dxe4 20.Nxe4 Re8 looks okay for Black, even if White has
reasonable practical chances.

8...Be7

Given that White isn’t threatening anything Black has more options with move order. At first we
tended to develop the g8-knight more quickly. I actually played 7...Nh6 in Lyth-Tait above, and
8...Nh6 has been seen here too: 9.b4 Be7 10.b5 Nd7 11.a4 Nf8 12.Qe2 Ng6 (Black is now ready to
push the g-pawn, which White’s next only encourages) 13.Nbd2? g4! 14.hxg4 Nxg4 15.Ne1 h3
16.bxc6 bxc6 17.Ba3?! Qa5! (heading for the kingside along the fifth rank) 18.gxh3 Rxh3 19.Rf3
Rh2 0-1 J.Simmelink-J.Tait, corr. 1998.

9.Qb3

The only move to have been tried so far. Forcing the black knight to the edge seems quite logical,
except that it was going there anyway.

9...Nh6 10.a4 Nd7 11.a5 Nf8

It’s always nice to have a plan. Stockfish considers Black to be clearly better.

12.Nbd2

Playing to undermine the queenside with 12.d5!? is more purposeful, but it leaves a huge hole:
12...Ng6 (it’s even worth inserting 12...a6!?, since 13.e5 Ng6 14.e6 0-0 is not a problem) 13.Nd4 Ne5
(13...c5! looks even stronger) 14.a6 Qb6 15.Nd2 Qxb3 16.Bxb3 c5 17.axb7 Bxb7 and White has very
little for the pawn, pdchessvamp-tsmenace, ChessWorld.net 2010.

182
12...Ng6

Now Black is ready to play ...g5-g4 whenever it seems most appropriate. It’s not easy to suggest a
defence for White.

13.Re1

Or 13.Nh2 Qc7 14.Ndf3 Bd7 15.Bd2 Rg8 16.Rae1 Rg7 17.a6 b6 18.d5 c5 19.Bb5 0-0-0 20.Bxd7+
Qxd7 21.c4 Bf6 22.Kh1 Ne5 23.Nxe5 Bxe5 24.Bc3 f6 25.Qc2 g4! (at last) 26.hxg4 Nxg4 27.Nxg4
Rxg4 28.Rf3 Rdg8 29.Bxe5 fxe5 and Black won quickly, bobagem-tsmenace, ChessWorld.net 2010.

13...Qc7

14.Qa2

If White wants to push the b-pawn it’s better to do so before playing Qb3 rather than waste time
moving the queen again now.
Instead, 14.Nh2 prevents ...g5-g4 for the moment but does nothing positive, so Black can continue
to build up with 14...Bd7 and ...Rg8-g7, as in the previous note.
Another game from the early days – before we’d worked out definite plans – saw 14.Nf1 0-0!?
(14...g4! 15.hxg4 Bxg4 16.N1h2 Bd7 is more to the point; or else 14...Rg8 15.N1h2 Rg7 and so forth)
15.a6 b5 16.Bd3 Qb6 17.Kh1 Bxa6!? (a bit greedy; 17...Be6 or 17...Kg7 would prevent what follows)
18.e5 d5 (not 18...dxe5?? 19.Bxg6) 19.Qc2 Kg7 20.e6 f5 21.b3 Bb7 22.Ba3 Rae8? (dropping the g5-
pawn; 22...Qc7 was still okay) 23.Bxe7 Rxe7 24.Nxg5 and despite the mess White was clearly better,
F.Holt-J.Wagenbach, corr. 1995.

183
14...g4!

Now seems a particularly good moment. Very likely Black is already winning.

15.hxg4 Bxg4 16.a6 b6 17.Bf1 0-0-0 18.b4 Rdg8 19.Qb3

I’ve actually reached this position twice. Hari1210-tsmenace, ChessWorld.net 2008, continued
19.Kh1 h3 20.gxh3 Nh4 21.Nh2 Bh5 22.b5 Rg3 23.bxc6 Rhg8 24.Bb5 Rxh3 25.Bf1 and White
resigned before 25...Rhg3 26.Bb5 Ng4 finished things off.

19...h3 20.gxh3 Be6 21.d5 Ne5+ 22.Kf2

Or 22.Kh1 Bxh3! 23.Bxh3+ Nhg4 24.Kg2 Rxh3 25.Kxh3 Qd7 and mates.

22...cxd5 23.exd5 Bf5 24.Bg2

Or 24.Ne4 Nhg4+! 25.hxg4 Nxg4+ 26.Ke2 (or 26.Kg1 Ne5+ 27.Kf2 Bh4+) 26...Bxe4 27.Bxf4 Re8
and my notes merely append the “with attack” symbol. Stockfish now gives 28.Kd2 as “best”, when
28...Bxf3 just leaves White a piece down.

24...b5! 0-1

White has no reasonable defence to the threat of ...Qb6+ and so resigned. Reprimand-tsmenace,
ChessWorld.net 2012. I notice that Stockfish prefers 24...Nd3+ 25.Kf1 Rxg2! 26.Kxg2 Bxh3+!
27.Kxh3 Rg8!, but you only need to find one win.

E22: 7.Nc3

This makes more sense than c2-c3 (E21), as there’s no urgency for White to reinforce the d4-pawn

184
here.
In this section we’ll mostly concentrate on attempts to bash the Wagenbach flat by sacrificing on g5
– and often f7 as well. If that’s the intention then preparing it with 7.Nc3 is the best way forwards.
Even so, I think the sacrifice forces White into a difficult fight for equality at best. Stronger options
are given in the notes to move eight.
The immediate 7.Nxg5? Qxg5 is just good for Black:

a) 8.Bxf7+ Kxf7 (there is no cause to refuse the piece – especially as 7.Bxf7+? Kxf7 8.Nxg5+ Qxg5
comes to the same thing) 9.Bxf4 (not 9.Rxf4+? Nf6 and White has nothing) 9...Qg4 (9...Qh5?!
10.Bxd6+ Nf6 11.Qxh5+ Rxh5 12.Bxf8 Kxf8 13.Rxf6+ Kg7 14.Rf4 Nc6 15.c3 isn’t quite as bad,
J.Toscano-J.Tait, Scarborough 1992) 10.Qxg4 (10.Be5+ Nf6 11.Qxg4 Bxg4 transposes) 10...Bxg4
11.Be5+ (or 11.Bxd6+?! Nf6 12.Bxf8 Rxf8 13.e5 Nbd7 and Black is winning, TomOne-tsmenace,
ChessWorld.net 2008) 11...Nf6 12.Bxf6 (or 12.Rxf6+?! Ke7 13.Rxd6 Rg8) 12...Bg7 13.Bxh4+ Kg6
14.Bf2 Nc6 15.c3 Bh6 16.Na3 Rae8 17.h3?! Be2 18.Rfe1 Rxe4 19.Nb5 Bd2 and wins, A.Hall-
J.Wagenbach, corr. 1995.
b) 8.Bxf4 Qg4 (more forcing than 8...Qg6 or 8...Qg7, though Stockfish prefers the latter move)
9.Qd2 (as we’ve seen, 9.Bxf7+ Kxf7 is nothing to fear) 9...h3 10.Bxf7+? (but if 10.g3 then 10...Be6
11.d5 Bc8 and ...Nd7 is one way to consolidate) 10...Kxf7 11.Be5+ Kg6 12.Bxh8 Bh6 13.Qf2 0-1
e4forme-tsmenace, ChessWorld.net 2008, in view of 13...Be3! etc.
c) 8.Rxf4 is perhaps the best try:

185
 8...Nf6 (for 8...f6 9.Nc3 c6 see 9.Rxf4 in the main line) 9.Nc3 (not 9.e5?! dxe5 10.Re4 Qg7
11.Rxe5+ Be7 12.Qe2 Nc6 and Black won, D.Zhichkin-S.Tatlow, corr. 1997) 9...Be7 (not 9...c6? due
to 10.Rf5! as we’ll see below) 10.Rf5 (after 10.Rxf6 Qxf6 11.Nd5 Qg7 12.Nxc7+ Kd8 13.Nxa8 Bh3
14.Bf1 Bxg2! 15.Bxg2 h3, Black comes out on top) 10...Qg4 (10...Qg8!? 11.Rg5 Qf8 is another
option) 11.Qxg4 Nxg4 12.Rxf7 Nc6 and Black is for preference.
There are also the sacrifices after 7.c3 c6, which we held over from line E21 (note that, if desired,
playing 7...Be7 first would rule them out since 8.Nxg5? Bxg5 9.Bxf4 Nh6 10.Qh5 Rg8 leaves White
with nothing): 8.Bxf7+ (both 8.Nxg5 Qxg5 9.Bxf7+ Kxf7 10.Bxf4 Qg4 and 9.Bxf4 Qg4 10.Qxg4
Bxg4 11.Bxf7+ Kxf7 transpose; not 9.Rxf4? here in view of 9...Nf6 10.e5 Nd5!) 8...Kxf7 9.Nxg5+
Qxg5 10.Bxf4 Qg4! 11.Qxg4 Bxg4 12.Be5+ Nf6 (12...Kg6!? 13.Bxh8 Nd7 looks even better)
13.Bxf6 Bg7 14.Be5+ Kg6 15.Bxd6 h3 16.Rf2 Nd7 17.Nd2 Rae8 18.Re1 and White was worse but
managed to hang on for a draw, G.Grasso-F.Obertin, corr. 1999.

7.Nc3

186
7...c6

Directed against e4-e5 which can now be answered by ...d6-d5. Otherwise Black has to be wary of
White pushing the e-pawn (or indeed sacrificing on g5) more favourably than on the previous move –
in particular transposing to favourable variations in line E33.
a) 7...h3? was played in the first recorded game to reach the Wagenbach tabiya after Black’s 5th
move (albeit via 2...h5!?). A.Moll-E.Mendelssohn, Amsterdam 1988, was a slaughter: 8.g3 g4? 9.Ne1
(9.Bxf4 gxf3 10.Qxf3 is very strong too) 9...f3 10.Nxf3! gxf3 11.Qxf3 f6 12.Bxg8 Rxg8 13.Qh5+
Kd7 14.Qf7+ Be7 15.Nd5 1-0.
b) 7...Be7 doesn’t actually prevent 8.Nxg5!? (instead, 8.e5 Nc6! is okay for Black) 8...Bxg5 9.Bxf4,
which looks to give White decent compensation; e.g. 9...Nh6 10.Qc1! Bxf4 (10...Rg8 11.Nd5 Be6
12.Bxg5 Qxg5 13.Nf6+ is why the queen went to c1 rather than d2) 11.Qxf4 Be6 12.d5 Bg4 13.h3
Bxh3 (14.e5 follows if the bishop retreats) 14.gxh3 Nd7 15.Bb5 and White has regained the material
with a slight edge.
c) 7...Bg7 8.Nxg5!? (8.e5! is even better) 8...Qxg5 9.Bxf4 Qd8? (9...Qg6 is correct, as in TIGRIS-
brawn, ChessWorld.net 2005, when Black may yet defend) 10.Be3 and Black is already in big
trouble; e.g. 10...d5 (or 10...f6 11.Bxg8 Rxg8 12.Qh5+ Kf8 13.e5 Rh8 14.Qg6 Qe8 15.Rxf6+! and
wins – Konikowski) 11.Bxd5 f6 12.Bxg8 h3 13.e5 hxg2 14.Rf2 Bh3 15.exf6 Bxf6 16.Qf3 1-0
A.Krzyzanowski-E.Bortnik, corr. 2000.
d) 7...Be6 feels wrong, since the exchange of light-squared bishops doesn’t really help Black:
8.Bxe6 fxe6 9.Qd3 c6? (unfortunately, Qb5+ was not the only threat) 10.e5! Rh6 11.Ne4 dxe5
12.Nexg5 and White had a huge advantage, D.Koetsier-D.Sakellarakis, corr. 1999.
e) 7...Nc6!? makes a strange impression but may be okay; e.g. 8.Nxg5!? (for 8.e5?! dxe5 see 8.0-0
in line E33; while 8.h3 can be met by 8...Bh6!? and ...Nge7, or 9.Bd2? g4! 10.hxg4 Bxg4 11.e5 h3

187
and White’s position fell apart, Alexvjv-remyrey, ChessWorld.net 2019) 8...Qxg5 9.Bxf7+ (or 9.Rxf4
f6 10.Nd5 Qg7) 9...Kxf7 10.Rxf4+ (or 10.Bxf4 Qh5!? 11.Bxd6+ Nf6 12.Qxh5+ Rxh5 13.Bxf8 Nxd4)
10...Nf6 11.Nd5 Qg7 12.Nxf6 (or 12.Rxf6+ Ke8) 12...Ke7 13.e5 and White has a definite initiative
for the piece, though nothing conclusive according to Stockfish.
f) 7...Bg4 8.Qd3 (for 8.e5 Nc6! see 7...Bg4 in line E33) 8...Nd7 is an “orthodox” alternative; e.g.
9.h3 Bh5 10.Bd2 Be7 11.Nd5 Nb6 12.Nxe7 Qxe7 13.Bb3 Bxf3!? 14.Qxf3 0-0-0 15.Rfe1 Kb8 16.a4
a6 with equal chances, Alexvjv-afms, ChessWorld.net 2019; White has two bishops, Black an extra
pawn.

8.Bxf7+

This is fun to bash out if nothing else. The next few moves are forced.
White more often plays 8.Nxg5 first, when 8...Qxg5 9.Bxf7+ (or 9.Bxf4 Qg4 10.Bxf7+ Kxf7)
9...Kxf7 just transposes. By taking on f7 at once White avoids other options for Black, such as ...Kd8
(after Bxf7+) or ...Qg6 (after Bxf4). On the other hand, I think accepting the transposition is best for
Black, so it’s all swings and roundabouts.
Let’s look at the independent lines nonetheless; i.e. 8.Nxg5 Qxg5 and now:

188
 a) 9.Bxf7+ Kd8? 10.Bxf4 Qg7 11.Bb3 Bg4 12.Qd2 Nd7 13.Rae1 Be7 14.e5 d5 15.Nxd5 cxd5
16.Bxd5 gave White superb play with the central pawn mass, M.Vujadinovic-J.Simmelink, corr.
1998.
b) 9.Bxf4 Qg4! (note that 9...Qg6 doesn’t prevent 10.Bxf7+! Qxf7 11.Bxd6 Nf6 12.Bxf8 Rxf8
13.e5 and White has quite enough, B.Wood-J.Wagenbach, MCCU U-175, Bolton 1995) 10.Bxf7+
Kd8? 11.Qd2 h3 12.g3 Qg7 13.d5 (13.e5 looks good too) 13...Nh6 14.Be6! Qg6? 15.dxc6! Qxe6
16.Bg5+ and White won quickly, G.Ruggeri Laderchi-J.Elburg, corr. 1999.
c) 9.Rxf4!? f6 (not 9...Nf6? since 10.Rf5 Qg6 11.Rg5 Qh7 12.Bd3! is good for White –
Th.Johansson) 10.e5!? h3 (a useful inclusion in these positions; 10...Qg6 at once is also possible)
11.g3 Qg6 12.Qe2 (not 12.Ne4? d5; or if 12.exd6 Bxd6 13.Rf2 then 13...f5 holds the light squares)
12...dxe5 13.dxe5 f5 14.e6 Na6 is a typical mess. Andy- AndyO-Panza, ChessWorld.net 2013,
continued 15.Bxa6 bxa6 16.Be3 Qxe6 17.Re1 Rh7! 18.Qf3 Re7 19.Rxf5 Qxf5 20.Qxc6+ Qd7
21.Qg6+ Rf7 22.Qe4+ Ne7 23.Qxa8 Bg7 24.Ne4 Bd4 25.Nf2 Bxb2 26.Bc5 Kf8 27.Qe4 Qc7 28.Nd3
Bb7 29.Qe6 Kg8 30.Re2 Qxc5+! and White resigned in view of 31.Nxc5 Bd4+ 32.Rf2 Bxf2+ 33.Kf1
Bxc5+. The engines don’t throw out any instant improvements for White in all this.

189
 Apart from sacrificing, White has tried numerous other moves, including:
a) 8.h3 is covered in the notes to line E31 (see 8.0-0 there).
b) 8.e5? is mistimed since Black can safely push past: 8...d5 9.Bd3 Nh6 and was already clearly
better in A.Pready-J.Wagenbach, Birmingham 1998.
c) 8.d5?! is premature if White cannot follow up with e4-e5, so Black should fight for that square:
8...Nd7 9.dxc6 bxc6 10.Kh1 (or 10.Qd4 Rh7 11.Bd3 Be7 12.Bd2 Ne5 13.Nxe5 dxe5 14.Qxd8+ Bxd8,
Brynjar-tsmenace, ChessWorld.net 2009) 10...Be7 11.Nd4 Bb7 12.Nf5 Ne5 13.Be2 Nh6 14.Nxe7
Qxe7 15.h3 0-0-0 16.b4 d5 17.Qd4 c5! 18.bxc5 dxe4 19.Qa4 Nf5 and Black won, jl777-tsmenace,
ChessWorld.net 2010.
d) 8.a4!? is more testing. White can consider undermining the queenside with the a- and b-pawns,
as well as building a big position in the centre. For example: 8...Bg4 9.Qd3 (the direct 9.b4 Nd7 10.b5
c5 11.e5 dxe5 12.dxe5 Bg7 works out well for Black; while 9.a5 allows b2-b4 to be answered by
...a7-a6, e.g. 9...Nd7 10.Qd3 Be7 11.Bd2 Nh6 12.b4 a6 13.h3 Bh5 14.Rae1 Qc7 15.Na4 Rg8 16.Bb3
g4 17.hxg4 Rxg4, followed by ...0-0-0 with good play, mantyk-tsmenace, ChessWorld.net 2008)
9...Be7 (in hindsight 9...Nd7 and 10...Bg7 might be preferable) 10.Bd2 Nh6 11.Rae1 Nd7 12.b4 Bh5
13.b5, Jo.Nelson-J.Wagenbach, Derby 1995, and now 13...g4! is very sharp indeed, as shown by the
following engine-derived analysis: 14.bxc6 bxc6 15.Ne2 gxf3 16.Nxf4 fxg2 17.Rf2 (or 17.Kxg2 h3+!
18.Kh1 Bg4! 19.e5 d5 20.e6 dxc4 21.exf7+ Nxf7 22.Qg6 Nf8! and 23.Qxc6+ Bd7 or 23.Qxg4 Qd7)
17...Ng4! 18.Nxh5 Nxf2 19.Qf3 Rh7 20.Qf5 Nf8 21.e5 d5 22.Nf6+ Bxf6 23.exf6+ Ne4 24.Rxe4+
dxe4 25.Qxe4+ Ne6 26.Bxe6 Qxf6 27.Qxc6+ Kf8 28.Qxa8+ Kg7 29.Qxg2+ Kh8 30.Qa8+ (or 30.Bg4
Qxd4+ 31.Kf1 Rg7 32.h3 f5) 30...Kg7 31.Qg2+ Kh8 with a draw.

8...Kxf7 9.Nxg5+ Qxg5

190
10.Bxf4

Not 10.Rxf4+? Nf6 11.e5 dxe5 12.Ne4 Qxf4 13.Bxf4 exf4 and Black has far too much for the
queen.

10...Qg4

Once again 10...Qh5?! is inferior: 11.Bxd6+ Nf6 12.Qxh5+ Rxh5 13.Bxf8 with three good pawns
for the piece, R.Eames-P.Briggs, Sheffield 1997.

11.Be5+

Now 11.Bxd6+? Nf6 is just good for Black; e.g. 12.Bxf8 (or 12.Rxf6+?! Kxf6 13.Be5+ Kg6
14.Bxh8 Qxd1+ 15.Rxd1 Bg4 16.Rf1 Nd7, H.Groffen-J.Simmelink, corr. 2000) 12...Qxd1 13.Raxd1
Rxf8 14.e5 Kg6 15.exf6 Bf5 16.d5 Rxf6, S.Tatlow-J.Wagenbach, corr. 1996.

11...Nf6

191
12.Bxf6

This time 12.Rxf6+?! is met by 12...Ke7! 13.Rxd6 (for 13.Bxd6+ Kxf6 see the previous note)
13...Qxd1+, followed by ...Rg8, and Black is clearly better, Alexvjv-tsmenace, ChessWorld.net 2019.

12...Bg7?!

Black’s choice in every game so far, for no obvious reason. Removing the queens with 12...Qxd1!
13.Raxd1 and then 13...Rh5 (or 13...Rh7) 14.Bxh4+ Kg8 looks like an obvious improvement –
certainly Stockfish thinks so.

192
13.Bxg7+

The strongest continuation. Praxis has also seen:

a) 13.Be5+ Kg8 14.Bxd6 h3 (to give the c8-bishop the g4-square once the queens come off)
15.Qxg4 Bxg4 16.e5 hxg2 17.Rf2 Nd7 18.Re1 Be6 19.Ne4 Rh4 20.Be7 Rg4 21.h3 Rg6 22.Rxg2
Rxg2+ 23.Kxg2 Bd5 and Black has a definite edge, even if I failed to convert it in barri-tsmenace,
ChessWorld.net 2009.
b) 13.Bxh4+ Ke8 14.Bg3 Qxd1 15.Raxd1 Rh6 16.Bf4 Rg6 kept things under control for Black, who
went on to win in Sh.Khan-J.Simmelink, corr. 2008.
c) 13.e5 h3?! (here 13...Qxd1 14.Raxd1 Bxf6 15.Rxf6+ Ke7 16.Rxd6 Bg4 17.Rf1 Nd7 looks like an
improvement) 14.Qxg4 Bxg4 15.Ne4 dxe5 16.Bxe5+ Kg6 17.Bxg7 Kxg7 18.Rf4 Be6 19.Nc5 Bd5
20.Rg4+ Kh6 21.Re1 Na6 22.Re3 Kh5 23.Rg7 Bxg2 24.Reg3 ½-½ N.Rutter-J.Wagenbach, corr.
2002, since perpetual check follows.
d) 13.Qd3!? Bxf6 14.e5 improves on note ‘c’ since ...Bf5 isn’t possible, while 14...Nd7 15.Ne4
dxe5 16.dxe5 h3 17.g3 Nxe5 18.Rxf6+ Kg7 19.Qd4 Re8 or 19...Rh5 leads to a draw according to
Stockfish.

13...Kxg7 14.Qd2

Keeping the queens on is far more promising than 14.Qxg4+? Bxg4 15.Rf4 Bh5 16.Rxh4 Nd7
17.Rf1 Bg6 18.Rhf4 Rae8, which led to another win for Black in Y.Yarmolyuk-J.Simmelink, corr.
2007.

14...h3 15.g3 Qg6 16.Rf4 Rf8 17.Rxf8 Kxf8 18.Rf1+ Kg7

193
19.Qf4

The immediate 19.e5?! would be met by 19...Bg4! 20.Rf6 Qh5 21.Rxd6 Nd7, followed by ...Rf8.

19...Nd7 20.e5 b6

The only way to develop the queenside.

21.Qh4

Here 21.e6 Qxe6 (or 21...Ba6 22.Qf7+) 22.Qg5+ forces a draw.

21...Ba6 22.Qe7+ Kg8 23.Rf2 Rf8

The safe solution. 23...Nf8 24.Rf6 Qe8 allows the game to continue, albeit with few winning
chances for either side.

24.Qxd7 Rxf2 25.Kxf2 Qxc2+ 26.Ke1 Qxh2

The Wagenbach h-pawn is now so dangerous that White has nothing better than give endless
checks, which they proceeded to do.

27.Qe6+ Kh8 28.Qf6+ Kh7 29.Qf7+ Kh8 30.Qf8+ Kh7 31.Qf7+ Kh8 32.Qf8+ Kh7 33.Qf7+ Kh8
½-½

A draw by threefold repetition, PARIS-tsmenace, ChessWorld.net 2007. Stockfish maintains a

steady “0.00” after 12...Bg7, which might be regarded as success for either side. Then again, if given
another chance, I would definitely try for more with 12...Qxd1!.

194
E3: 6.Nc3

This natural move is stronger than castling (E2), as it leaves more options open. White may yet
castle short, or sacrifice if it seems good to do so, or castle long and try for g2-g3, or aim for a central
breakthrough, perhaps after h2-h3 and castling long again.
Sometimes White opts to play 6.h3 first (or even earlier). After 6...d6 this almost always transposes
elsewhere. For instance, 7.Nc3, 7.Qe2 c6 8.Nc3 and 7.Qd3 c6 8.Nc3 are all covered in E31, as is 7.0-
0 c6 8.Nc3 (for other continuations after 7.0-0 see E2 above).
Finally, 7.b3 c6 8.Bb2 can be met by 8...Nf6! and ...Nh5, since 9.Nxg5?? loses to 9...Qa5+ etc.

6.Nc3 d6

Other moves are clearly inferior:

a) 6...Bb4?! is too important a piece to give away on c3.

195
 For example: 7.Qd3 Qe7 8.Bd2 Bxc3 9.Bxc3 d6, J.Cadman-S.Tatlow, corr. 1996, and now 10.0-0-0
is very good for White, or even 10.e5!? since 10...dxe5? 11.Nxe5 f6? loses quickly: 12.0-0-0 fxe5
13.Qg6+ Kd8 14.dxe5+ Bd7 15.e6.
b) 6...Nc6?! 7.d5 Na5 puts the knight at risk for no substantive positional gain. White can choose
between 8.Be2 Bg7 9.Qd3 (threatening b2-b4) 9...c5 10.e5 g4 11.Bxf4 with a huge attack, or 8.Qd4!
f6 9.Be2 (threatening variously Nxg5, e4-e5 and b2-b4). Either way White is probably winning.

E31: 7.h3
E32: 7.Qe2
E33: 7.e5

196
 Instead, 7.0-0 returns to line E22 above; 7.Bd2 and 7.Qd3 are considered in E32. Finally, 7.b3!?
Bg7 (either 7...Bg4 or 7...c6 is fine too) 8.h3 Nc6 transposes to M.Moruzzi-J.Wagenbach, Notts
League 2010, where 9.Bb2 g4! 10.hxg4 Bxg4 was at least equal for Black.

E31: 7.h3

The engines have generally been very keen on inserting h2-h3 at some point, seemingly concerned
about ...g5-g4 or ...h4-h3, neither of which Black is usually intending to play any time soon.
Meanwhile White creates a hole on g3 and gives up any real possibility of breaking up the kingside
with the g-pawn. The one major point in favour of h2-h3 is that the c8-bishop now has no easy
development, which means that White will have a lot of time to castle long and arrange a central
break with d4-d5 and e4-e5.

7...c6

The standard move, but with White already committed to h2-h3 Black has a more profitable range
of options:
a) 7...Be7 8.Qe2 c6 9.Bd2 Nd7 transposes to the main line.
b) 7...Bh6 8.Qe2 (8.Qd3 is possible too) 8...Ne7 transposes to E32.
c) 7...Bg7 has more point to it now that e4-e5 isn’t an immediate threat; Black can play against the
d4-pawn with ...Nc6 (whereas ...c7-c6 should transpose to variations with ...Bg7 in the main line). For
example: 8.Qd3 (or 8.0-0 Nc6 9.Nd5 Be6, when 10.Nxg5? Bxd4+ 11.Kh1 Qxg5 12.Nxc7+ Kd7 is
just good for Black, I.Gilbert-J.Wagenbach, Blackpool 1997) 8...Nc6 9.a3 (not 9.Bd2? Nb4 10.Bxf7+
Kf8!) 9...Nh6 (after 9...g4? 10.hxg4 Bxg4 11.Bxf4 Bxf3 12.gxf3 Nxd4 13.0-0-0, White had more than
enough for the pawn, juliangon-Reprimand, ChessWorld.net 2014) 10.Bd2 0-0 11.Nd5 Ne7
12.Nxe7+ Qxe7 13.0-0-0 c6 and Black was fine, samurai-JHDonner, ChessWorld.net 2010.

197
8.Qe2

The usual continuation; White makes ready to castle long after Bd2 and then advance in the centre.
Instead:
a) 8.d5?! is premature: 8...Bg7 9.Qe2 Nd7 10.0-0?! Ne5 and Black has already consolidated,
B.Oldham-J.Wagenbach, Derby (rapid) 1995.
b) 8.Bd2 will just transpose below after 9.Qe2.
c) 8.Qd3 is slightly less effective since the queen is better placed behind the e-pawn here; e.g.
8...Nh6 9.Bd2 Bg7 (rather than 9...a5!? 10.a3 b5 11.Ba2 Ba6, when 12.b4! axb4 13.axb4 Nd7 14.0-0
Bg7 15.Rfe1 0-0 16.e5! was good for White in H.Davies-J.Wagenbach, Scunthorpe rapid 1995)
10.d5?! (now 10.0-0-0 a5!?, or 10.0-0 Nd7) 10...c5! with advantage, as 11.e5 Bxe5 or 11.Bb5+ Kf8
doesn’t achieve anything.
d) 8.a4!? is not without merit: 8...Bh6 (8...Nh6! seems preferable, waiting to see what White is up
to before committing the bishop) 9.e5? (here 9.b3 and Ba3 was more testing) 9...d5 10.Bd3 Be6 11.a5
Ne7 12.Bd2 Nf5 13.Bxf5 Bxf5 14.Na4 Nd7 15.b4 0-0 and Black is clearly better, D.Flynn-J.Tait,
Fiveways vs. Bunkers Hill 2002.
e) 8.0-0 has more significance...

because the position can arise (and has arisen) via several move orders. From the other side, the
inclusion of h2-h3 creates a target for counterplay via ...Be7, ...Nd7-f8-g6 (reinforcing the f4-pawn)
and ...g5-g4. Black should probably start with 8...Nd7 (in the games below I generally played 8...Be7
first, but then 9.Nxg5!? Bxg5 10.Bxf4 is quite difficult to defend; and 8...Bg7 9.Qe1!? Nh6 10.Ne2
Nd7 11.c3 Nb6 12.Bd3 Be6 13.a4 Nd7 14.a5 a6 15.Qd1 Rg8 16.Kh1 Qe7 17.Nxg5!? Qxg5 18.Nxf4
Qe7 19.Qa4 Bf8 20.Bd2 also gave White good play in Stockfish-Leelenstein, CCC 9 Final 2019,
though Black managed to hold on)

198
 A few examples in chronological order:
a) 9.Nxg5 Qxg5 10.Rxf4 (or 10.Bxf4 Qg6) 10...Nh6 11.Rxf7 Qg6 12.Rf3 Nb6 (or 12...Be7) 13.Bd3
Qg7 is nothing much to worry about.
b) 9.Qd3 Be7 10.Bb3 Nf8 11.e5?! d5 12.Ne2 Nh6 (after 12...Ng6?! 13.c4! White found some play
in B.Hanison-J.Wagenbach, corr. 1999) 13.c4 Bf5 is good for Black.
c) 9.Qe2 Be7 10.Bd2 Nh6?! (10...Nf8 is more accurate, intending 11.d5 Ng6 12.e5 dxe5 13.Nxe5
Nxe5 14.Qxe5 Rh6) 11.Rae1 Nf8 12.Bb3?! (missing the moment for 12.d5!) 12...Ng6 13.Nh2 Bd7
14.Kh1 Qc7 15.d5 Ne5 16.dxc6 bxc6 and Black was now clearly better, jeweler-tsmenace,
ChessWorld.net 2008.
d) 9.a4 Be7 10.b3 (or 10.a5 Nf8) 10...Nf8 11.Qd2 Ng6 12.Ne2 Nh6 13.Ba3 Rg8 14.a5 Qc7 15.d5?!
(fixing the centre can only help Black) 15...c5 16.a6 b6 17.b4 Bf6 18.Rab1 Ne5 19.Nxe5 Bxe5
20.bxc5 dxc5 21.Rbd1 Rg6 22.Qd3 Bd7 23.Nc3 0-0-0 24.Bc1 Qd6 25.Nb5 Bxb5 26.Bxb5 Rdg8
27.Kh1 g4! (at last) 28.Rd2 gxh3 29.Qxh3+ f5 with a decisive attack, WaterDragon-tsmenace,
ChessWorld.net 2011.
e) 9.b4! seems the best try:

199
 9...Be7 (via this move order 9...Nh6 and ...Bg7 might be preferred, that is if ...Be7 hasn’t been
played already) 10.b5 Nf8?! (and this is too slow; 10...Nh6 is still probably correct) 11.Re1?! (here
11.bxc6! bxc6 12.d5 is far more testing) 11...Ng6 (Black is now fine) 12.bxc6 (or 12.e5 d5 13.Bd3
Be6 14.Bxg6 fxg6 15.bxc6 bxc6 16.Qd3 Bf5 17.Qa6 Qc8, ibrachess-tsmenace, ChessWorld.net
2009) 12...bxc6 13.Bd3 g4! 14.hxg4 Bxg4 15.e5 Nxe5!? 16.Bxf4 (if 16.dxe5 then 16...Qb6+ 17.Kh2
h3 18.Qd2 hxg2+ 19.Kxg2 Bh3+ draws) 16...Bxf3 17.gxf3 Nxd3 18.Qxd3 d5 19.Rab1 Kf8 20.Rb7
Rh5 21.Qe3 Bd6 22.Kh1 Bxf4 23.Qxf4 Qf6 soon led to a drawn rook endgame, alapin-tsmenace,
ChessWorld.net 2014.

8...Nd7

Knights before bishops; it seems more flexible since the bishop might yet go to g7.
Stockfish, on the other hand, preferred to play 8...Bg7 straight away. Leelenstein-Stockfish, CCC 9
Final 2019, continued 9.Bd2 a5!? (not 9...b5? 10.Nxb5! cxb5 11.Bd5; instead, 9...Nd7 transposes to
9...Bg7 below) 10.0-0-0 (similarly 10.Rf1 b5 11.Bd3 Na6, rather than 10...Nd7?! 11.0-0-0 b5, jkettle-
tsmenace, ChessWorld.net 2008, when Stockfish suggests 12.Bxf7+!? Kxf7 13.e5 as promising for
White) 10...b5 11.Bd3 Na6 12.Rhe1 Kf8! 13.Qf2 Rh5! (these are all classic Wagenbach moves!)
14.Kb1 Nb4 15.Be2 d5 16.a3 dxe4 17.Nxe4 Nd5 18.Bd3 b4 19.a4 Bf5 20.Bc1 Ra7! 21.Qf1 Bxe4
22.Rxe4 Re7 and Stockfish emerged with the advantage. It later won by promoting the h-pawn in
complete justification of its third move.

200
9.Bd2

Proceeding as planned. The only other move to have been tried is 9.d5?!, when Romerm-tsmenace,
ChessWorld.net 2013, continued 9...c5!? (both 9...Bg7 and 9...Ne5 are possible too) 10.e5 (otherwise
10...f6 might well follow) 10...Nxe5 11.Nxe5 dxe5 12.Qxe5+ Qe7 13.Qxe7+ Bxe7 14.Nb5 Bd8
15.Bd2 Rh6 16.0-0-0 a6 17.Nc3 b5 18.Be2 f5 19.Bf3 Ra7 and Black consolidated.

9...Be7

I have generally opted for this in my own games, but other people might consider Black’s
subsequent play (pushing the queenside pawns as well) to be too ridiculous even for the Wagenbach
(and they could be right). In that case this is the last moment to choose something else, such as
9...Bg7 and ...Nh6. For example: 10.0-0-0 Nh6 11.e5?! (but if 11.Rhe1 then 11...b5 12.Bb3 Nb6; or
11.a3 Nb6 12.Ba2 0-0, when 13.Nxh4!? gxh4 14.Bxf4 Be6 15.Qh5 Kh7 defends) 11...d5 12.e6 Nb6
13.exf7+ Kxf7 14.Bd3 Nf5 15.Qf2 Ng3 16.Rhe1 Bf5 17.Bxf5 Nxf5 and Black has pretty much
consolidated, Romerm-samurai, ChessWorld.net 2013.

201
10.0-0-0

And again. It still seems too soon for 10.d5?!, since White isn’t ready to follow with e4-e5.
Nonetheless, 10...Bf6 11.0-0-0 (rather than 11.0-0?! Ne5, as in Bopeep-tsmenace, ChessWorld.net
2017) 11...Ne5 12.dxc6 bxc6 13.Na4! may be just about okay for White.
Prophylactic moves are more challenging:
a) 10.a3!? (spending a tempo – of which White has plenty – to pre-empt Black’s plan in the main
line) 10...Nh6 (either 10...a5 or 10...Kf8 comes into consideration; 10...b5?! makes little sense before
the white king has gone long) 11.0-0-0 Nb6 12.Ba2 Qc7 (not 12...0-0? 13.Nxh4! gxh4 14.Qh5 Kg7
15.g3! and the black king is in big trouble) 13.d5 Nd7 (back again, since 13...f6 14.dxc6 bxc6
15.Nxh4! gxh4 16.Qh5+ Kd8 17.Bxf4 is also too dangerous) 14.Rhf1!? Ne5 15.Nxe5 dxe5 16.Be1
g4?! (16...Kf8 may yet be okay) 17.Rd3 Bg5 18.Kb1 Bd7?! 19.Bc4 f3 20.gxf3 g3 21.d6 Qd8 22.Qg2
Qc8?! 23.Rh1 Be6 24.Bd2! was very good for White, who won shortly in alapin-tsmenace,
ChessWorld.net 2011.
b) 10.Bb3!? (retreating in advance of ...b7-b5) 10...Kf8 (not yet 10...b5?! in view of 11.d5 c5 12.e5,
while 10...a5 can be met by 11.a4) 11.0-0-0 b5!? (not obligatory; 11...Kg7 is an option) 12.d5
(regaining the pawn; other pushes seem okay for Black: 12.e5 d5 13.e6 fxe6 14.Qxe6 Nb8, or 12.g3!?
hxg3 13.h4 g4 14.Ng5 f3) 12...c5 13.Nxb5 f6 (clamping down on the dark squares, to be followed by
...Nh6-f7-e5 if allowed; instead White elects to give up a piece for three pawns) 14.Bc3 Ne5 15.Nxe5
fxe5 16.Bxe5!? dxe5 17.d6 Bf6 18.Nc7 Rb8 19.Qc4 Nh6 20.Qxc5 Bd7 21.Qxa7 Nf7 22.a4 Qc8
23.Qc5 Kg7 24.a5 Qb7 25.Rhe1 Qc6 and Black managed to defend successfully in juliangon-
tsmenace, ChessWorld.net 2021.

10...b5

202
 Disrupting White’s plans requires radical measures. After this further pawn move the position
becomes incredibly random.
Note that 10...Nf8? is too slow in view of 11.d5! Ng6 12.e5! and White is winning already.

11.Bd3!

I think this is the best retreat. If 11.Bb3, Black continues the assault with 11...a5 and then:

a) 12.e5 d5 13.e6 fxe6 14.Qxe6 Ndf6 15.Qe5?! (15.Qxc6+ Bd7 16.Qb7 Rb8 17.Qa7 Ra8 etc is a
draw) 15...a4 16.Bxd5 cxd5 17.Nxg5 Qd7 18.Bxf4? b4 19.Nb1 Qf5 and Black was on top, jkettle-
tsmenace, ChessWorld.net 2008.
b) 12.d5 a4 13.dxc6 axb3 14.cxd7+ Qxd7 15.axb3 (or 15.cxb3 b4 16.Na4 Qb7, intending ...Rh6-e6)
15...b4 16.Na4 Rxa4 17.bxa4 Qxa4 18.Kb1 Ba6 19.Qf2 Bc4 20.b3 Bxb3 21.cxb3 Qxb3+ and draws.
c) 12.a3 b4 13.Nb1!? (after 13.Na4 bxa3 14.bxa3 Rh6 Black is at least equal) 13...Qb6 14.e5?!
(White should regain the pawn with 14.axb4 axb4 15.Qc4 Nh6 16.Qxb4, even if 16...d5 17.Qxb6
Nxb6 or 16...Qxb4 17.Bxb4 g4 keeps an edge for Black) 14...d5 15.e6, docjan-tsmenace,
ChessWorld.net 2010, and now 15...Nf8! 16.exf7+ Kxf7 would have been very good for Black; e.g.
17.Qe5 Ng6 18.Nxg5+ Bxg5 19.Qxg5 Nf6, threatening ...Rh5.

11...Kf8!?

The king moves off the e-file in advance of the white e-pawn advancing. The danger can be seen
after, for example, 11...a6? 12.e5 d5 (or 12...dxe5 13.Ne4 exd4 14.Nexg5) 13.e6! Nb6 (or 13...fxe6?
14.Qxe6, threatening Bg6+ and mates) 14.exf7+ Kxf7 15.g3! fxg3 (or 15...hxg3 16.h4) 16.Ne5+ Kg7
17.Rhf1 Bf6 18.Rxf6! Qxf6 (or 18...Nxf6 19.Bxg5) 19.Rf1 Qe7 20.Rf7+ Qxf7 21.Nxf7 Kxf7 22.Qf3+
and White is probably winning.
Stockfish prefers 11...b4 12.Na4 Nb6 13.Nxb6 Qxb6 and assesses this as roughly equal. In conde7-

203
tsmenace, ChessWorld.net 2010, I opted for 13...axb6?!, when simply 14.Kb1 is good for White.

12.e5

The most direct and dangerous continuation. I have also faced:

a) 12.Nb1 a6!? (enabling 13.d5 c5) 13.Rhf1 Kg7 14.a3 Rh6 15.Qe1 a5!? 16.d5 Ne5 17.Bc3 (or
17.dxc6 Nxd3+) 17...b4!? 18.Nxe5 bxc3 19.Nxc6 cxb2+ 20.Kxb2 Bf6+ 21.Ka2 Qe8 22.Qf2 Ne7
23.Nd4, samurai-tsmenace, ChessWorld.net 2010, and now 23...Qa4!? might give Black a little
something.
b) 12.Rhe1 b4 13.Na4 Nb6 14.Nxb6 Qxb6 15.Kb1 a5 16.Bc4 Rh6 17.c3 Be6 18.Bxe6 Rxe6 19.Qc4
(threatening d4-d5) 19...Qb7 20.Qd3 Kg7 21.c4 d5 and Black is certainly not worse, verdi-tsmenace,
ChessWorld.net 2015.

204
12...d5?!

This natural and long-prepared response is far more risky than it seems (or at least seemed to me).
Instead, 12...dxe5 13.dxe5 (now 13.Ne4?! exd4 is safe enough) 13...Nc5 14.Be1 Qc7 is an unclear
mess; e.g. 15.Bf2 Rh6 16.Bxc5 Bxc5 17.Ne4 Be3+ 18.Kb1 g4! 19.Nfg5 Qxe5 20.hxg4 Nf6 21.Nf3
Bxg4 22.Qxe3! fxe3 23.Nxe5 Bxd1 24.Nxf6 Kg7 25.Neg4 Bxg4 26.Nxg4 Re6 and Black should hold
this endgame.

13.e6! fxe6 14.Rhf1

Since Bg6 is no longer check, 14.Qxe6 can be met safely by 14...Nb6, when 15.Qxc6? Bd7 16.Qb7
Rh6! and ...Bc8 traps the queen.

14...Bf6?

A serious mistake. The king needed to move on again; i.e. 14...Kg7, intending 15.Rde1 b4
(15...Bf6? transposes below) 16.Na4 e5!? 17.dxe5 Nf8, or if 15.Qxe6 then 15...Rh6.

205
15.Rde1

I had thought Black was okay, as neither 15.Qxe6 Ne5 16.Nxg5 Nxd3+ 17.cxd3 Bxg5 18.Qxc6
Bf5, nor 15.Nxg5 Bxg5 16.Bxf4 Bxf4+ 17.Rxf4+ Ke7 succeeds for White. Stockfish has other ideas:
15.g3!! hxg3 16.h4 g4 (or 16...gxh4 17.Bxf4) 17.Ne5 f3 18.Rxf3! gxf3 19.Qxf3, which it assesses at
“+4.24”; i.e. White wins. I’m really getting to hate Stockfish.

15...Kg7 16.Qxe6?

Making things easier for Black. This time Stockfish prefers a quiet move, 16.Kb1, preparing to
sacrifice something: Nxg5, Nxh4, or g2-g3 again. It’s not clear how Black defends.

16...Nc5

Not 16...Ne5? because 17.Qxe5! Bxe5 18.Rxe5 gives White a huge attack for the queen.

17.Qxc6 Nxd3+ 18.cxd3 Bf5 19.Qxd5 Bxd3 20.Qb7+

This was 0404it-tsmenace, ChessWorld.net 2010, which continued 20...Ne7?! 21.Nxg5! Bxg5
22.Rxf4 Rb8 23.Qxa7 Ra8 24.Qb7 Rb8 25.Qf3 Ng6 26.Rg4! Bxd2+ 27.Kxd2 Bc4 28.Re5, when
White has three pawns and an unrelenting initiative for the piece. Stockfish resolutely claims “0.00”,
but the defence proved too difficult for me.
The safer human course was 20...Kf8, when 21.Nxg5 (or 21.Rf2 Rh7) 21...Bxg5 22.Bxf4 Bxf1
23.Rxf1 Bxf4+ 24.Rxf4+ Nf6 25.Rxf6+ Qxf6 26.Qxa8+ Kg7 27.Qxa7+ Kh6 should be a draw. But
I’d be deviating at move 11 or 12 in another game.

E32: 7.Qe2

206
 This move can be played with or without a subsequent h2-h3. In fact, it seems more accurate to
start with Qe2 in either case because it maintains more options. Another dangerous plan is to follow
up with Bd2, 0-0-0 and a rapid g2-g3, while White can also revert to e4-e5 (as in line E33), so Black
needs to react with all of these in mind.
White has two other moves with the same sort of aims:
a) 7.Bd2 is less threatening (no immediate e4-e5) and therefore gives Black more leeway: 7...Bg4
(or 7...Nc6!?, targeting the d4-pawn at once) 8.Qe2 (or 8.0-0?! Bg7 9.Ne2 Nc6 10.c3 Nh6 11.Nxg5?
Qxg5 12.Bxf4 Qe7 13.Bxh6 Bxh6 14.Bxf7+ Kd7 and Black was winning, A.Ruffle-J.Wagenbach,
Nottingham 2008) 8...Bg7 9.e5!?, I.Gregory-J.Wagenbach, Derbys League 2005, and now 9...Nc6!
10.exd6+ Kf8 11.dxc7 Qxc7 12.Nd5 Qd6 13.0-0-0 Re8 looks at least equal for Black, while Stockfish
goes for 13...a5!? 14.Rhe1 b5! 15.Nb6 Nxd4 16.Qe4 Rb8 17.Bxa5 bxc4 18.Nxd4 Nf6 and says Black
is clearly better.
b) 7.Qd3 can be criticized on similar grounds: 7...Bg7 (or 7...Nc6!? again, when 8.Bd2?? Nb4 and
8.d5? Ne5 show the white queen to be misplaced) 8.Bd2 Bg4 (or 8...g4!?, intending 9.Ng1 h3!
10.gxh3 Qh4+ 11.Kd1 Qf2) 9.0-0-0 Nc6 10.Bb5 (10.e5!? dxe5 11.Ne4 Bxf3 12.gxf3 Qxd4 13.Qb3 0-
0-0 14.Bxf4 Qb6 may be critical) 10...a6 11.Bxc6+ bxc6 12.Rde1 Bxf3 13.gxf3, K.Forman-
J.Wagenbach, Scarborough 2002, and now 13...Rh6, followed by ...Rg6, looks fine for Black.
After 7.Qe2, on the other hand, Black has to exercise considerable caution.
For a start, 7...c6

207
 is probably a mistake because 0-0-0 and g2-g3 comes very quickly: 8.Bd2! Bg7 9.0-0-0 (not 9.d5?!
b5 10.Bd3 b4 11.Nd1 c5 12.0-0?! Bg4 13.c3 a5 14.Bb5+ Kf8 15.Qd3 Bxf3 16.Qxf3 Be5 and Black
was clearly better, M.Alcock-J.Wagenbach, Notts League 1993) 9...Nh6 (if 9...Bg4, J.Lyth-
J.Wagenbach, Chorley rapid 1998, then 10.g3! is even stronger since the g4-bishop blocks the g-
pawn) 10.g3! hxg3 11.hxg3 g4 12.Nh4 f3 13.Qe3 (threatening Nf5) 13...Qe7 14.e5 d5 15.Bxd5! looks
close to winning for White.
With other moves, such as 7...Bg7 (intending 8.Bd2?! g4!) or 7...Be7 (also planning 8.Bd2 g4!?),
Black runs the risk of landing in an inferior 7.e5 line (E33) after 8.e5!.
One exception is 7...Bh6!?, since 8.e5 is not so critical then; and otherwise Black might follow with
...Ne7, enabling ...d6-d5 (after e4-e5) in a different way:

208
 a) 8.Bd2 can now be answered by 8...g4! 9.Ng1 Nc6, targeting d4 once again; e.g. 10.Qf2 (or 10.0-
0-0 Nxd4 11.Qf2 Nc6) 10...g3! 11.hxg3 hxg3 12.Qf1 Nxd4 13.0-0-0 Bg4 and Black seems to be
okay.
b) 8.g3!? prompts the immediate 8...Nc6, intending 9.gxf4 g4 10.Ng1 Nxd4, or 9.Qf2 hxg3 10.hxg3
g4 11.Nh4 f3 12.Bxh6 Rxh6 13.Qd2 Qf6 14.0-0-0 (if 14.Nd5 then 14...Qxd4!, or 14.Nb5 Kd8)
14...Nce7 15.e5 Qg7 16.Nb5 Kd8 and Black looks to have everything covered.
c) 8.h3 Ne7!? 9.Bd2 Nd7 10.0-0-0 Nb6 11.Bd3 (or 11.Bb3 Ng6 12.Rhe1 0-0) 11...Nc6 12.a3 Bg7
13.Qf2 0-0!?, as in tsmenace-Reprimand, ChessWorld.net 2016, showed a very original Wagenbach
set-up for Black in view of its “orthodox” piece deployment. I was unable to break it down: 14.Kb1
Qf6! 15.Nb5 Bd7 16.Rhe1 Rac8 17.e5 dxe5 18.dxe5 Qh6 19.b3?! Rfe8 20.Bc3 Nd5 21.Bb2 Ne3
22.Rxe3 fxe3 23.Qxe3 g4 24.Ng5 Re7 25.Bh7+ Kf8 26.Be4 Bxe5 27.Bc1 Qg7 28.Qc5 Ke8 29.hxg4
Bxg4 30.Rh1 Rd8! and Black’s extra exchange eventually told.
d) 8.e5 has generally arisen via 7.e5 Bh6 8.Qe2, but we’ll examine it here: 8...Kf8! 9.0-0 (not yet
9.e6?! Bxe6 10.Bxe6 fxe6 11.Qxe6? Qe7 and Black is already better, J.Shemilt-J.Wagenbach,
Sheffield rapid 1999; if 9.Bd2 then 9...g4! 10.Ng1 Nc6 11.0-0-0 Bf5 is fine for Black, M.Tarmak-
J.Simmelink, corr. 2009; while 9.h3 can be met by 9...Ne7, intending ...d6-d5 or 10.d5?! Nf5 again
with advantage, JHDonner-samurai, ChessWorld.net 2010) 9...Bg4 10.e6!? Nc6 11.exf7 Nf6 12.b3
d5 13.Ba3+ Kxf7 (13...Kg7!? is possible too) 14.Nxd5! Nxd5 15.Qe4 Be6 16.Rfe1 Qd7 17.Bb5 Nc3
18.Bxc6 Nxe4 19.Ne5+ Kg7 20.Bxd7 Bxd7 21.Nxd7 Rhe8 led to a draw in AndyAndyO-tsmenace,
ChessWorld.net 2014.
Although 7.Qe2 Bh6!? is largely uncharted territory, having only appeared directly in two games to
date, it certainly deserves more outings.

E33: 7.e5

209
 Decidedly critical. White makes sure of opening the centre by pushing on before Black is ready to
push past with ...d6-d5.

7...Nc6

The best response seems to be to apply maximum pressure to the advanced white pawn. Black has
tried numerous other moves and ideas:
a) 7...c6 is less than useful after 8.exd6 Bxd6 9.Ne4 Be7 (if 9...f6!? 10.Qe2 Kf8 11.Bd2 b5 12.Bb3
a5, tsmenace-juliangon, ChessWorld.net 2016, then 13.0-0-0 and g2-g3 is very good for White) 10.0-
0 Bf5 11.Nfxg5! (or 11.Nexg5!) 11...Bxg5 12.Nxg5 Qxg5 13.Bxf4 1-0 D.Nightingale-J.Wagenbach,
corr. 1999. Resignation would be premature in a proper game; still, the best Black can do is 13...Qg4
14.Qxg4 Bxg4 15.Rae1+ Be6 16.Bxe6 fxe6 17.Rxe6+ Kf7 18.Rd6 Na6 19.d5! (Nightingale)
19...cxd5 20.Be5+ Ke7 21.Rxa6 Rh5 22.Bf6+ Nxf6 23.Raxf6 Rh7 with a difficult ending a pawn
down.
b) 7...Bg4 is rather too routine: 8.Qe2! (not 8.0-0?! Nc6! 9.exd6 Bxd6 10.Ne4 Bxf3 11.Qxf3 f5!?
12.Qb3 fxe4 13.Bxg8 Qf6 14.Qxb7 Qxd4+ 15.Kh1 Rxg8 16.Qxa8+ Nd8 and Black is better with the
powerful pawn mass, alapin-Raffzahn, ChessWorld.net 2014) 8...Be7 (after 8...Bxf3?! 9.Qxf3 c6?!
10.Bd2 d5 11.Bd3 Qd7 12.0-0-0 Qe6 13.g3, Black’s kingside was falling apart, P.Swanson-J.Tait,
North Midlands Cup 1994) 9.e6!? f5 10.h3 Bh5?! 11.Qd3 Nh6 12.d5 and Black’s position is cut in
half, J.Tait-D.Nightingale, corr. 1998.
c) 7...Bg7 is insufficiently forcing:

8.0-0 (or 8.exd6 cxd6 9.0-0 f6 10.Bxg8 Rxg8 11.Nxh4! Qd7, J.Carleton-J.Wagenbach, corr. 2003,
when 12.Qh5+ Kd8 13.Ng6 is good for White; not 8.Qe2?! due to 8...Nc6!, transposing to the main
line) 8...Bg4 (8...h3 9.g3 doesn’t help Black) 9.exd6 cxd6 10.Qe1+! (suggested by Micawber at
ChessPublishing; both 10.Re1+? Kf8, G.Rosser-J.Wagenbach, corr. 2002, and 10.Qd3 Bxf3 11.Qxf3?

210
Bxd4+ 12.Kh1 Nc6, F.Meis-R.Hendriks, Hoogeveen 2002, were good for Black) 10...Kd7 (forced,
since 10...Kf8? 11.Nxg5 Qxg5 12.Bxf4 wins quickly) 11.Bb5+ Nc6 (or 11...Kc8? 12.Nxg5! Qxg5
13.Bxf4, followed by Ne4) 12.d5 is good for White.
d) 7...Be7 has brought Black some success, but White’s play can be strengthened: 8.Qe2! Kf8
9.Bd2 (9.0-0 can be met by 9...dxe5! 10.Nxe5 Qxd4+ 11.Kh1 Nc6 12.Nxf7 Bg4; while 9.e6!? fxe6
10.Bxe6 Bxe6 11.Qxe6 Qd7 also seems acceptable, e.g. 12.Qb3 Nc6 13.d5 Ne5 14.Nd4 h3 15.Ne6+
Ke8 16.g3 fxg3 17.Qxb7? Qc8! 18.Qxc8+ Rxc8 19.hxg3 g4 and Black was clearly better in
Will_Take_Your_Queen-tsmenace, ChessWorld.net 2019) 9...Nc6 10.0-0-0 (after 10.exd6 cxd6 11.d5
Black can return the pawn with 11...Ne5 12.Nxe5 dxe5 13.Qxe5 Bf6, securing the kingside with
tempo, and shouldn’t be worse; again 10.e6!? fxe6 11.d5 exd5 12.Nxd5 Nf6! 13.Nxg5 Nxd5 14.Ne6+
Bxe6 15.Qxe6 Qe8! is quite unclear) 10...Bg4 (fighting now for both e5 and d4; 10...g4 11.exd6 cxd6
12.Ne1 Nxd4 13.Qf2 Bf6 14.Bxf4 was good for White in Raff-zahn-samurai, ChessWorld.net 2014)
11.Be1 (to defend the d4-pawn; note that 11.Nd5!? can be met by 11...b5!, intending 12.Bxb5?! Nxd4
or 12.Bb3 a5 13.Qxb5 a4 14.Bxa4? h3! or 14.Bc4 Bxf3 15.Qxc6 Bxd1 16.Rxd1 Qe8 with equality)
11...Nh6 (nor does 11...Na5 12.Bd3 d5 solve Black’s problems after 13.Qf1! and g2-g3)

d1) 12.h3?! (an inferior move on principle: White gives up on g2-g3 and creates a hole on that
square which the h6-knight is pleased to occupy) 12...Bh5 13.Bf2 Nf5 14.Rhe1 Rh6 (a typical
Wagenbach move, reinforcing Black’s defences along the rank) 15.a3 (instead, 15.Bd5!? Ng3
16.Bxg3 hxg3 17.a3 dxe5 18.dxe5 Qe8 19.Qe4 Rd8 20.h4!? Bxf3 21.Qxf3 Rxh4 22.Kb1 Kg7 led to a
draw in Raffzahn-tsmenace, ChessWorld.net 2017) 15...Ng3 16.Bxg3 hxg3 17.e6 Na5!? 18.Ba2 c6! (a
lovely idea; Black is prepared to give up the knight in order to consolidate) 19.exf7 d5 20.b4 Bxf7
21.Ne5 (or 21.bxa5 Bxa3+ 22.Kd2 Bb4) 21...Bh5 22.Ng4 b5! 23.Qe5 Bxg4 24.hxg4 Bd6! 25.Qf5+
Qf6 26.bxa5 Bxa3+ 27.Kd2 Qxf5 28.gxf5 Bb4 29.Kd3 Bxa5 and Black has all the chances with three
solid pawns which shut in the extra piece, juliangon-AndyAndyO, ChessWorld.net 2016.
d2) 12.Nd5! is much stronger now that the d4-pawn is covered. Since there’s no ...b7-b5 trick,

211
White can take time to build up with Bc3, Qe4, Rhf1, say, before sacrificing something somewhere;
e.g. 12...Nf5 13.Bc3 Rh6 14.Qe4 Qd7 15.Rhf1 Re8 16.Kb1!, threatening 17.Nxf4 with a very strong
attack, and I can’t find a satisfactory defence (16...Nd8!? may be the best try). Whereas the immediate
16.Nxf4?! gxf4 17.Qxf4 would be met by 17...Rg6 18.h3 (or 18.Kb1 h3) 18...Bxf3 19.Qxf3 Bg5+
20.Kb1 Ne3 and Black survives.
e) 7...Bh6!? has similar aims to 7...Be7, and overprotecting the f4-pawn can prove useful. This time
White does best to castle: 8.0-0! (maintaining White’s options; for 8.Qe2 Kf8! see E32; inserting
8.exd6 cxd6 only helps Black; and 8.g3!? doesn’t quite work after 8...g4 9.Nxh4 dxe5 10.0-0 Nc6 or
9.Bxf4 Bxf4 10.gxf4 gxf3 11.Qxf3 Bf5 12.0-0-0 Nd7) 8...h3!? (very unorthodox for the Wagenbach,
but it’s hard to find a good move: 8...Nc6 9.e6! fxe6 10.d5 is much the same; 8...Bg4 is met by 9.Re1!
dxe5?! 10.Bxf7+ Kf8 11.Bb3! with a big advantage, C.Dubois-J.Johnsrud, corr. 2001, or 9...Kf8
10.b3 and Ba3, seeing that Black’s bishop is missing from e7; instead, 8...Ne7!? was a one-move,
untested suggestion in my Chess Mail article – should anyone wish to investigate, 9.Ne4 Ng6 10.b3
g4 11.Ba3 gxf3 12.exd6 may be critical) 9.g3 Nc6 10.e6! (Black has drawn against other moves; e.g.
10.Qe2 Nge7 11.exd6 cxd6 12.d5 Ne5 13.Nxe5 dxe5 14.Qxe5 0-0 15.gxf4 Ng6 16.Qe2 gxf4 and so
on, RickF-tsmenace, ChessWorld.net 2015) 10...fxe6 11.d5! (opening up against Black’s king in the
middle)

11...exd5 (without going into too much detail this seemed relatively best; e.g. 11...Na5 12.Bd3; or
11...Nce7 12.dxe6 Nf6 13.Qd4; or 11...Ne5 12.Nxe5 dxe5 13.Qh5+; or 11...Nb4!? 12.dxe6 Nf6 13.e7!
Kxe7 14.Nxg5! Bxg5 15.Qe2+ Kf8 16.Bxf4 Bxf4 17.Rxf4 with a huge attack) 12.Nxd5 Nge7 13.Re1
Kf8 14.gxf4 Nxd5 (or 14...Qd7 15.Nf6! Qf5 16.fxg5 Qc5+ 17.Kh1 Qxc4 18.gxh6 Bf5 19.Bg5,
AndyAndyO-Reprimand, ChessWorld.net 2016) 15.fxg5 Nb6 16.Rf1! (improving on 16.Nd4 Nxc4
17.Qf3+ Kg8 which isn’t a cast-iron win) 16...Ke8 (now 16...Nxc4 17.Ne5+ Ke7 18.Rf7+ wins)
17.Nd4 Nxc4 18.Nxc6! bxc6 19.Qh5+ Kd7 20.Rf7+ Ke6 21.gxh6? (a slip; 21.Qg6+! Kd5 22.b3! is
decisive) 21...Rg8+ 22.Rg7 Qe8 23.Qxh3+ Kd5 24.Qd3+ Kc5 25.b3 Bh3! 26.Qxh3 Qe5 27.Qf1 Ne3
28.Bxe3+ Qxe3+ 29.Qf2 Qxf2+ 30.Kxf2 Raf8+ 31.Ke3 Rh8? (capitulating; Stockfish somehow

212
defends with 31...Re8+ 32.Kd2 Rgf8) 32.h7 Rf6 33.Rh1 Rh6 34.Rxc7 R6xh7 35.Rxh7 Rxh7 36.h4
and White soon won, AndyAndyO-tsmenace, ChessWorld.net 2016.

8.Qe2!

The only testing move. After anything else Black may already be better; e.g. 8.exd6 Bxd6 9.Ne4 f6,
or 8.0-0 dxe5 9.Re1 Be7 10.Nxe5 Nxe5 11.Rxe5 Kf8.

8...Bg7!

Offering the pawn back with check to eliminate the wedge. Instead:
a) 8...Be7? 9.exd6 cxd6 10.Nxg5 Nxd4 11.Qe4 is just good for White with the black pawn phalanx
destroyed.
b) 8...d5? 9.Bxd5 g4 is well met by 10.Qc4! Na5 11.Bxf7+ Ke7 12.Qd3 Kxf7 13.e6+ Kg7 14.Bxf4
with a crushing position, R.Zeier-I.Foote, corr. 2008.
c) 8...dxe5?! runs into 9.d5! Nce7 (if 9...Nb4 10.a3 Nxc2+ 11.Qxc2 Bd6, the three pawns aren’t
worth the piece) 10.Nxe5 Bg7, taquin-tsmenace, ChessWorld.net 2008, and now 11.d6! Qxd6
12.Nxf7 Bxc3+ 13.bxc3 Qf6 14.0-0 (or just 14.Bb2) 14...Rh7 15.Nxg5 Qxg5 16.Bxf4 (Bücker) is
again very good for White.

213
9.exd6+

Refusing the pawn doesn’t get White anywhere:

a) 9.0-0 dxe5 10.d5 (or 10.dxe5 Nd4) 10...Nce7 11.Nxg5 Nh6.
b) 9.Bd2 Bg4 10.Bb5 (or 10.exd6+?! Kf8) 10...d5 or 10...Kf8.
c) 9.h3 dxe5 (9...Bf5 and 9...Kf8 are okay too) 10.dxe5 (or 10.Bd2 g4 11.hxg4 Bxg4) 10...Bf5 11.e6
(or 11.Ne4 Bxe4 12.Qxe4 Qe7 13.Bb5 Rh6) 11...fxe6 (or 11...Nh6!? 12.exf7+ Kf8) 12.Bxe6 Bxe6
13.Qxe6+ Qe7 14.Qxe7+ Ngxe7 15.Nxg5 Nd4 16.0-0 Nxc2 17.Rb1 Bd4+ 18.Kh1 Be3 19.Ne6 Kd7
20.Nxf4 Rae8 with equality.

9...Kf8 10.dxc7

The same applies here:

a) 10.d5 Nd4 (or 10...Na5) 11.Nxd4 Bxd4 12.Nb5 Qf6 13.Nxc7 (or 13.dxc7 Bxb2 14.Bxb2 Qxb2
15.Rd1 a6 16.Nd4 Qb4+ 17.Kf2 Qc5) 13...Qxd6 14.Nxa8 (or 14.Nb5 Qe5) 14...Nf6 with a powerful
initiative; e.g. 15.Qd3 Bc5 16.h3 Kg7 17.Kd1 b5! 18.Bb3 Ng4 19.Rf1 Ne5.
b) 10.Ne4 Nxd4 (or 10...Bg4) 11.Nxd4 Bxd4 12.Bd2 Bxb2 (or 12...cxd6 13.0-0-0 Qb6) 13.Rb1 Be5
14.dxc7 Qxc7 15.Nxg5 Nf6 16.Bxf7 (or 16.Nxf7 Rh5) 16...Bg4 17.Qc4 Rc8 18.Qxc7 Rxc7 19.h3 f3!
with equality.

10...Qxc7

In return for returning the pawn Black now has active play with threats at d4, c4 and on the e-file.

11.Nd5

214
 Not 11.Nxg5? Nxd4 and Black wins material; while if 11.Nb5 Qd8 12.Nxg5 (or 12.d5 a6)
12...Qxg5 13.Bxf4 then 13...Qh5 or even 13...Qxf4!? 14.Rf1 Qf6.

11...Qd8 12.b3!

Seeking to cause embarrassment on the a3-f8 diagonal. This is virtually the only move, given that
12.Nxg5 Qxg5 13.Bxf4 Qg6 doesn’t lead to anything.

12...Nce7!

An effective reply. In this quite obscure position, Black seems to be fine:

a) 13.Nxe7 Nxe7 14.c3 Bf6 15.0-0 Ng6 and ...Kg7 secures the kingside. Although the queenside
majority and d4-pawn look like assets, it’s hard to see how White can utilize them. Stockfish is
steadfast in its “0.00” assessment.
b) 13.Nxg5 Nxd5 14.Ba3+ Nge7 15.0-0-0 Qa5!? (15...Bf6 16.Ne4 Bg7 is a tacit draw offer) 16.Bc5
b6 17.Bxd5 bxc5 18.Bxa8 cxd4 and Black has excellent compensation for the exchange; e.g. 19.Nf3
Qxa2 20.Qc4 Bf5 21.Rhe1 d3 22.Be4 Bb2+ 23.Kd2 dxc2 24.Qxc2 Qa5+ 25.Ke2 Bc3 26.Rd8+
(returning the rook for a level endgame) 26...Qxd8 27.Qxc3 Rh5 28.Bxf5 Nxf5 29.Qb4+ ½-½
remyrey-afms, ChessWorld.net 2019.
c) 13.Ba3 b5! (a crucial resource) 14.Nxg5! (both 14.0-0-0? bxc4 and 14.Nxe7? Nxe7 15.Qxe7+
Qxe7 16.Bxe7+ Kxe7 are good for Black) 14...bxc4 15.Qxc4 Ba6! 16.Nxe7 (not 16.Qxa6?? Qxd5)
16...Qa5+ 17.Bb4 Bxc4 18.Ng6+ Ke8 19.Bxa5 Rh5! (skewering minors along the fifth rank) 20.Nxf4
(or 20.Ne4 Be6 21.Nxf4 Rxa5) 20...Rxg5 21.Bd2 Bd5 22.0-0-0 Be4 23.Rde1 f5, leif41no-tsmenace,
ChessWorld.net 2021, and the best to be said for White is that 24.Rhf1 is probably okay with three
pawns for the piece.
So that’s the Wagenbach Defence. Over the years it’s given us perhaps thousands of hours of

215
entertainment in both games and analysis. And success too: my own results – OTB (long/rapidplay)
and CC (mail/online) – with 3...h5 are P316, W234, D49, L33. That’s an 81.8% score as Black. May
you have such joy with your own openings.

216
 Chapter Six
Three Knights Game
(and Other Third Moves)

1.e4 e5 2.Nf3 Nc6

The starting position for Romantic Chess Openings, named as such in the title of the fourth world
correspondence champion Vladimir Zagorovsky’s 1982 book. From here the Ruy Lopez, Two
Knights, and Scotch each have their own chapter(s). This one will examine all the rest.
A: 3.c3
B: 3.d4
C: 3.Nc3
White has a few lesser options that get more outings than you might expect:
a) 3.d3 can be met by 3...f5, which is either a Latvian Gambit (1.e4 e5 2.Nf3 f5 3.d3 Nc6) or a
Lisitsin Gambit Deferred (1.Nf3 f5 2.d3 Nc6 3.e4 e5) – unless White plays 4.d4!? in which case it’s a
reversed Vienna, where Black, now having the first move, is obviously fine; e.g. 4...fxe4 5.Nxe5 Nf6
6.Be2 (or 6.Bc4 d5 7.Bb5 Bd7 or 6.Nc3 Bb4) 6...d5 7.0-0 Bd6 8.f4 exf3 9.Bxf3 0-0 and White has
equalized. Well done.
The position after 3...f5 actually brings up 600 games in the databases, but only 33 via this precise
route so it’s hardly worth spending a lot of time on. In any case Black is equal after, for example,
4.exf5?! d5, or 4.Be2 Nf6, or 4.Nc3 Nf6 5.exf5 d5 6.Nh4 (or 6.d4 exd4) 6...Be7, or 4.g3 fxe4 5.dxe4
Nf6 6.Bg2 Bc5.
b) 3.g3 f5!? is more provocative.

217
 Here White’s extra g2-g3 on a Vienna after 4.d4 is not so useful; e.g. 4...fxe4 5.Nxe5 Nf6 6.Bg2 (or
6.Nc3 Bb4) 6...d5 (or 6...Be7 7.0-0 0-0 8.c4 d6) 7.c4 (not 7.Nc3 Bd6 8.Bg5 0-0 9.Nxd5?? Bxe5
10.dxe5 Qxd5 11.Qxd5+ Nxd5 12.Bxe4 Be6 and Black is winning, S.Manush-A.Jain, Ahmedabad
2019) 7...Bd6 8.Bf4 (or 8.cxd5 Nxe5) 8...0-0 9.Nc3 dxc4 10.0-0 Qe8 11.Nxe4 Nxe4 12.Bxe4 Bxe5
13.dxe5 Be6 is about equal.
However, 4.exf5 is now more testing since White can defend the pawn with either Nh4 or Bh3. All
the same, I think Black should be okay: 4...e4 (not 4...d6?! 5.d4!, which is a bad Vienna with or
without g2-g3 inserted) 5.Nh4 Be7 (not 5...Nf6? 6.d3 d5 7.Bg2 and White is clearly better,
B.Predojevic-Ju.Sarkar, PNWCC blitz 2020) 6.d3 (or 6.Qh5+ Kf8 7.Ng6+ hxg6 8.Qxh8 Nd4 9.Kd1
d5 with good compensation) 6...Bxh4 7.Qh5+ (or 7.gxh4 Qxh4 8.Nc3 exd3 9.Bxd3 d6 10.Nd5 Nf6!
11.Nxc7+ Kd8 12.Ne6+ Bxe6 13.fxe6 Re8) 7...Kf8 8.Qxh4 (or 8.gxh4 Nd4) 8...Qxh4 9.gxh4 exd3
10.Bxd3 d6 11.Nc3 Nge7 12.f6 gxf6 13.Bh6+ Kf7 14.Rg1 Rg8.
Hitherto I’ve generally played 3...Nf6. The issue there is that 4.Nc3 makes it a Glek Four Knights,
and I’m now avoiding the Four Knights (see line C).
c) 3.Be2 renders 3...f5?! suspect, as after 4.d4! (rather than 4.d3 Nf6 5.Nbd2 d6 6.0-0 Be7 7.c3 0-0
and Black is fine, D.M.Adams-J.Tait, Sheffield League 2019) 4...fxe4 (or 4...exd4 5.exf5 Bc5 6.0-0
Nf6 7.Nbd2 and Nb3) 5.Nxe5 White’s extra Be2 is significant in the reversed Vienna; e.g. 5...Nf6
6.0-0 d6 7.Nxc6 bxc6 8.f3! with a definite advantage, R.Kreutzkamp-T.Michalczak, German
Bundesliga 1996.
This time Black should resort to 3...Nf6 and then:

218
 c1) 4.Nc3 is a feeble Four Knights: e.g. 4...d5 5.exd5 Nxd5 6.0-0 (6.Bb5!? gives Black the white
pieces again, this time in a reversed Scotch Four Knights) 6...Nxc3 7.bxc3 e4 8.Nd4 (or 8.Ne1 Bd6
9.f3 Bf5) 8...Nxd4 9.cxd4 Qxd4 10.c3 Qd5 11.d3 exd3 12.Re1 Be7 13.Bxd3 Be6 14.Be4 Qxd1
15.Rxd1 Rb8 16.Bf4 Bd6 17.Bxd6 cxd6 18.Rxd6 Ke7 19.Rd3 Rhc8 and Black has the better chances
in the endgame, N.Short-A.Volokitin, Wijk aan Zee 2009.
c2) 4.d3 d5 5.Nbd2 (or 5.exd5 Nxd5 6.0-0 Be7) 5...Bc5 is a reversed Philidor. GM Baadur Jobava
has won several games with this as White, though objectively Black already stands well; e.g. 6.0-0 0-
0 7.c3 a5 8.a4 Re8 9.Qc2 h6 10.h3 Be6 11.Re1 d4 12.Nf1 Nd7 13.Ng3 Bf8 14.cxd4 exd4 15.Nd2 Nc5
with a clear advantage, A.Grischuk-F.Caruana, Paris GCT (blitz) 2017.
c3) 4.d4!? exd4 5.e5 was played by IM Michael Basman a few times. Here 5...Ng4 is in keeping
with Chapter Eight; e.g. 6.0-0 Be7 (rather than 6...d6 7.Nxd4 Ngxe5 8.f4 Nxd4 9.Qxd4 Nd7?! 10.Re1
Qf6, S.Grayland-J.Tait, corr. 1998, when 11.Qe3+ is good for White according to Stockfish) 7.Bf4 (or
7.Nxd4 Ngxe5) 7...d6 8.Bb5 (or 8.exd6 Bxd6 9.Bxd6 Qxd6 10.Na3 a6) 8...dxe5 9.Bxc6+ bxc6
10.Nxe5 Nxe5 11.Bxe5 c5! and Black is better.

A: 3.c3

Ponziani’s Opening, preparing to take the centre with d2-d4. Black can pre-empt that by getting in
first.

3...d5

219
 In case anyone was expecting ...f7-f5 here, when White has played c2-c3 I prefer to push the d-
pawn. The resulting positions are a lot more interesting than those after 3...f5 4.d4, which is another
reversed Vienna.
A1: 4.Bb5
A2: 4.Qa4
There are other options:
a) 4.d4?! dxe4 5.Nxe5 (not 5.Ng5? exd4 6.Bc4 Ne5) 5...Nxe5 6.dxe5 Qxd1+ 7.Kxd1 is no good for
White; e.g. 7...Bf5 8.Be3 0-0-0+ 9.Nd2 Ne7, M.Bertel-M.Sadowski, corr. 2007.
b) 4.Bd3?! Nf6 5.Qe2 (or 5.exd5 Qxd5 6.Qe2 Be6) 5...dxe4 6.Bxe4 Nxe4 7.Qxe4 f5 8.Qe2 e4 9.d3
Qxd3 10.Qxd3 exd3 11.Bf4, G.A.Szabo-M.Csikos, Hungarian League 1999, and 11...Be6 12.Bxc7
Bd5 isn’t much better.
c) 4.d3 is another reversed Philidor, which is hardly a serious concern; e.g. 4...Nf6 5.Qc2 (for
5.Nbd2 a5 6.Be2 Bc5 7.0-0 0-0 see 3.Be2 Nf6 4.d3 earlier on) 5...a5 6.Be2 h6 7.0-0 Bd6 (since
7...Bc5 might now be met by 8.d4!?, as in B.Markevich-J.Morgado, corr. 1976) 8.Nbd2 (or 8.b3 0-0
9.Nbd2 Re8 10.a3 Ne7 11.Rb1 c6 12.b4 b5 13.Re1 Ng6, craig4000-tsmenace, ChessWorld.net 2018)
8...0-0 9.Rd1 Re8 10.Nf1 Be6 11.Ng3 Qc8 12.h3 a4 13.Be3 Ne7 14.Re1 c5 and Black stood well,
Zhang Zhong-E.Bareev, Wijk aan Zee 2004.
d) 4.exd5 Qxd5 5.d4! proposes a Göring Gambit after 5...exd4 or 5...Bg4 6.Be2 exd4 7.cxd4, which
Black may as well accept – see line B. Other moves are worse for White: 5.Qb3?! (not 5.Be2? e4,
Adamantin-tsmenace, ChessWorld.net 2004; or 5.d3? Bf5 6.Be2 0-0-0, braindeadcz-tsmenace,
ChessWorld.net 2018) 5.Qb3?! Nf6 6.Bb5 (or 6.Qxd5 Nxd5 7.Bc4 Nf4) 6...Qe4+ 7.Be2 Be6! 8.c4 (or
8.Qxb7 Rb8 9.Qxc7 Bd7, threatening ...Nd5) 8...Nd4 9.Nxd4 exd4 10.f3 Qg6 11.Qxb7 Rd8 12.0-0 d3
13.Bd1 Bc5+ 14.Kh1 0-0 15.Nc3 Nh5 with a decisive attack, mizuzul-tsmenace, ChessWorld.net
2018.

220
A1: 4.Bb5?!

Perhaps “?” or even “??” is warranted, because Black now gets a very powerful position by force.

4...dxe4! 5.Nxe5 Qg5!

This is similar to 7...Qg5 in Chapter Eleven, except that the queen is there sooner which makes it
much better for Black.

6.Qa4

The forcing line; 6.Bxc6+ bxc6 7.Qa4 Qxg2 comes to the same thing.
Instead, 6.d4 Qxg2 7.Rf1 is well met by 7...Bd6! 8.Nxc6 (not 8.Qa4 Ne7 9.Nxc6 bxc6 10.Bxc6+
Nxc6 11.Qxc6+ Ke7! and wins; e.g. 12.Qxa8 Ba6 13.Qxh8 Qxf1+ 14.Kd2 Qe2 mate – T&H; or
8.Qh5 g6 9.Qh4 Ne7! 10.Qf6 Rf8 11.Bh6 Ng8! – Bologan) 8...Bd7 9.Nxa7 (if 9.Qa4 then 9...a6!
10.d5 Rd8 or 10.Na3 Rc8 – T&H) 9...c6 10.Bc4 (or 10.Nxc6 bxc6 11.Be2 Ne7 with ideas of ...0-0,
...Bxh2 and/or ...f7-f5 as appropriate) 10...Rxa7 11.d5 c5 12.Na3 Rxa3 13.bxa3 b5 14.Be2 Nf6
15.Be3 0-0 16.Qd2 Qxh2 17.0-0-0 Ra8 18.Rg1 Rxa3 with two pawns and a strong initiative for the
exchange, Croatia Kings-tsmenace, ChessWorld.net 2009. Indeed, looking at it with Stockfish now,
I’m not sure how I failed to win.

6...Qxg2 7.Bxc6+

Here 7.Rf1 Bh3 8.Bxc6+ bxc6 9.Qxc6+ Kd8 transposes, as neither 8.Nxc6 bxc6 9.Bxc6+ Kd8
10.Qc4 Qxf1+ 11.Qxf1 Bxf1 12.Bxa8 Bd3, A.Mietkiewicz-M.Bergstrom, corr. 1985; nor 8.d4 exd3
9.Nd2 Ne7, S.Wolfl-F.Colombo Berra, corr. 1992, is any better.

221
7...bxc6 8.Qxc6+ Kd8!

9.Rf1

There are no improvements: 9.Nxf7+? Ke7 10.Qxc7+ Ke8 11.Rf1 Bh3 wins. 9.Qxa8? Qxh1+
10.Ke2 Bd6 11.Nxf7+ Kd7 wins again, since 12.Nxh8 Ba6+ leads to mate. 9.Qd5+ Bd6 10.Nxf7+
Ke7 11.Qg5+ Qxg5 12.Nxg5 Bb7 is terrible for White, who has no light squares or good
development. The same applies after 9.Kd1 Qxh1+ 10.Kc2 Nh6 11.Qxa8 Bd6; e.g. 12.Nc6+ Kd7
13.c4 Ng4 14.Nc3 Qf3 15.Kb3 Qd3 16.a3?! Qxc4+! 17.Kxc4 Ba6+ 18.Kb3 Rxa8 19.Nb4 Bb7 and
White’s position is hopeless, stormytlc-tsmenace, ChessWorld.net 2010.

9...Bh3 10.Qxa8+ Ke7 11.Kd1

Or 11.Nc6+ Kd6.

11...Qxf1+ 12.Kc2 Bf5

The theoretical move. Interestingly, Stockfish goes for 12...Nh6!? instead; e.g. 13.Qc6 f6 14.Qxc7+
Ke6 15.d4 fxe5 16.Qxe5+ Kf7 17.Bxh6 Qd3+ 18.Kc1 gxh6 19.Qh5+ Kg7 20.Qe5+ Kg6 21.Qxh8
Qf1+ 22.Kc2 Qxf2+ 23.Nd2 e3 24.Rg1+ Kh5 25.Qe5+ Bf5+ 26.Kb3 exd2 27.Rd1 Kg6. It seems no
one has ever tried this.

222
 In this very complicated position theory claims a win for Black:
e1) 13.Qd5 Nh6 14.Nc6+ (or 14.c4 Kf6! 15.Nc3 Bd6 16.Nxe4+ Bxe4+ 17.Qxe4 Bxe5) 14...Kf6
15.Qd4+ Kg6 16.Ne5+ Kh5 and, with the king quite safe, Black won quickly in P.Soderberg-
P.W.Smith, corr. 1985.
e2) 13.Na3 f6! (Emms) 14.Nc6+ Kf7 15.Nd4 Qd3+! 16.Kb3 Bg4 17.Nac2 (or if 17.Qc6 Bd1+
18.Nac2, I.S.Campbell-E.Efendiyev, corr. 2001, then 18...Bd6! 19.Qd5+ Kg6 20.Qf5+ Kh6 21.Qxe4
Bxc2+! – Bologan) 17...c5 18.Qd5+ Kg6 19.f3 Bxf3 (simpler than 19...exf3 20.Ne1 Ne7 21.Qb7 Qf1,
as in M.Schäfer-J.Fleck, German League 2002) 20.Qf5+ Kh6 21.Qh3+ Bh5 22.Nf5+ Kg6 23.Qg3+
(or 23.Qxd3 exd3 24.Nce3 Kg5 25.Kc4 Bg6 26.Ng3 f5 27.h4+ Kxh4 28.Ngxf5+ Kg5 29.Ng3 h5,
Rol.Markus-M.Szalai, corr. 2002) 23...Qxg3 (simpler again than 23...Kf7 24.Qc7+ Be7 25.Nce3 g6
26.Nxe7 Nxe7, B.Radoor-Ka.Martins, corr. 2005) 24.Nxg3 Bd1 25.c4 f5 26.Kc3 Bd6 “and Black
should win” (Bologan).
e3) 13.Qc6!? (Stockfish’s best defence) 13...Qe2! (instead, 13...Nf6 14.b3! e3+ 15.d3 Qxf2+
16.Nd2, or 13...e3+ 14.d3 Qe2+ 15.Nd2 exf2 16.b3! is rather less clear) 14.Qxc7+ (or 14.c4 e3+
15.Kc3 exd2) 14...Kf6 15.Nd7+ Kg5 16.Nxf8 Nf6 17.h4+ (or 17.Qg3+ Kh5 18.Qe3 Qd3+! 19.Qxd3
exd3+ 20.Kb3 Rxf8, intending ...Kg4-h3xh2, ...h7-h5 etc) 17...Kh5 18.Qf4 Qg4 19.d3 exd3+ 20.Kb3
Rxf8 21.f3 Be6+ 22.Ka3 Qxf4 23.Bxf4 Nd5 24.Bd2 Rc8 25.b3 Kxh4 26.c4 Ne7 27.Nc3 Kg3
28.Rg1+ Kxf3 29.Rxg7 h5 30.Rh7 Kg4 and again Black should win.

A2: 4.Qa4

I can see why the Ponziani appeals to some players – already the game has entered its own unique
territory.

4...f6

223
 Steinitz’s move, which is indeed very Steinitzian. Black defends the centre e-pawn in the most solid
way possible, unconcerned about compromising the kingside.
A21: 5.d3
A22: 5.Bb5
Instead, 5.exd5 Qxd5 6.Bb5 gives Black an extra option in 6...e4! (otherwise 6...Ne7 returns to the
main line) 7.c4 (for 7.Nd4 Ne7 8.0-0 Bd7 see 7.0-0 e4 8.Nd4 Bd7 below) 7...Qd7, intending 8.Bxc6
bxc6 9.Nd4 c5 (Bologan).
Note that 6.Bc4? is bad because of 6...Qe4+ 7.Kf1 (or 7.Kd1 Qg6) 7...Be6 8.Bb5 Qxa4 9.Bxa4
Bc4+ and White will have great trouble in developing, R.Bryant-M.Hebden, Jersey 2017; while 6.d4
Bd7 7.Bb5? (here 7.Qc4 Qxc4 8.Bxc4 exd4 9.Nxd4 Nxd4 10.cxd4 0-0-0 restricts Black to an edge
based on the weak d-pawn) 7...a6 8.c4 Qe4+ 9.Be3 Rd8 10.Nbd2 axb5 11.cxb5 Qg4 12.bxc6 Bxc6
13.Qb3 e4 14.h3, M.Reznicek-K.Kuzmicz, Karvina 2010, and 14...Qxg2 15.0-0-0 Qg6 and ...Qf7 is
very good for Black.

A21: 5.d3

Another Philidorish position. The white queen looks oddly placed but has done a job in
encouraging ...f7-f6 and might easily drop back to c2. Another idea is to follow up with Nbd2 and d3-
d4. The task for Black is how to negate the apparent weakening and even play for a win. To that end I
like:

5...Ne7

More usually 5...Be6 is played before ...Ne7. I’ve switched them round to enable a different option
against the d3-d4 plan.

224
6.Be2

Now 6.Nbd2 can be met by 6...a6!?, intending 7.d4 exd4 8.Nxd4 Bd7 9.Be2 dxe4 10.Nxe4 Nxd4
11.Qxd4 Bc6 12.Qxd8+ Rxd8 13.f3 h5 14.h4 Ng6 15.Be3 Bxe4 16.fxe4 Bd6 17.0-0-0 Bf4 and Black
is at least equal, Stoofvlees II-LcZero, TCEC 18 Testing 2020.

6...Be6 7.0-0 g5!?

Utilizing the f6-pawn. Now Black is ready for ...Bg7 and ...0-0, while set for a kingside attack at
some stage. The position is quite concrete yet relatively unexplored, so here’s a summary of practice
so far:

a) 8.Be3 Bg7 9.Nbd2 0-0 10.h3 (if 10.Rfe1 Qd7 11.Nf1, L.Neagu-C.Voiculescu, corr. 2011, then
11...Ng6 with an edge) 10...h5 11.Rfe1 Qd7 12.Bd1 Rad8 13.Bb3 b6 14.Rad1 Na5 15.Qxd7 Bxd7
16.Bc2 c5 17.d4 exd4 18.cxd4 c4 and Black is no worse, P.Tonisson-D.Babic, corr. 2016.
b) 8.b4 Ng6 (or simply 8...a6!) 9.exd5 Qxd5 10.b5 Nce7 11.Nfd2, A.Sitnikov-P.Tsvetkov,
Chelyabinsk 2020, and now 11...g4 looks fine, intending ...f6-f5 and ...Bg7, while preventing Bf3.
c) 8.Nfd2!? d4 9.Nb3 h5 10.Nc5 Bc8 (Black shouldn’t be shy about retreating if necessary) 11.Qd1
Ng6!? 12.Bxh5 Bxc5 13.Bxg6+ Ke7 14.Qb3 Qg8 15.Qxg8 Rxg8 with a lot of activity for the pawn,
juliangon-tsmenace, ChessWorld.net 2018.
d) 8.Nbd2 Bg7 9.Nb3!? Bd7!? (rejecting 9...0-0 on account of 10.Nc5 Bc8 11.h3 b6? 12.exd5 Qxd5
13.d4, threatening Bc4) 10.Qa3 (now if 10.Nc5 then 10...Nd4 11.Qd1 Nxe2+ 12.Qxe2 Bc8 and ...b7-
b6) 10...0-0 11.Nc5 Bc8 12.exd5!? (otherwise ...b7-b6 will push the knight back to b3) 12...Nxd5
13.h3 b6 14.Ne4 h6 15.Ng3 f5 16.Nh5!? (quite a knight tour) 16...Bh8 17.Bd1 Bb7 18.Re1 Qd7
19.Ba4 Qf7 20.Ng3 g4!? (since it seemed time to start something) 21.Nh4 Nce7 22.c4 (or 22.Bxh6
Qf6!) 22...f4!? 23.Ne4 Nf6 24.hxg4 Nxg4 25.Bd7?! (possibly 25.Bd1 was better) 25...Ng6 26.Bxg4
(or 26.Nxg6 Qxd7!) 26...Nxh4 27.d4 h5! 28.Ng5 Qg6 29.Be6+ Kg7 30.Rxe5 Kh6! 31.Bf7 Rxf7

225
32.Nxf7+ Qxf7 33.Qe7 Qxe7 34.Rxe7 Bxg2 35.Bxf4+ Kg6 and Black emerged with a big advantage,
bobby47-tsmenace, ChessWorld.net 2021.

A22: 5.Bb5

The consistent and intended move.

5...Ne7

Supporting the other knight without obstructing the queen.

6.exd5

White’s only plausible plan is a d2-d4 break, so 6.d3?! now makes little sense; e.g. 6...Be6 7.Be3
(or 7.0-0 a6 8.Bxc6+ Nxc6 9.Be3 Be7 10.Nbd2 0-0 11.d4 dxe4 12.Nxe4 f5 13.Neg5 Bd5 14.dxe5,
S.Tolstoy-A.Romashkevich, corr. 1900, and then 14...h6) 7...a6 8.Bxc6+ Nxc6 9.Nbd2 dxe4 10.dxe4
Qd3 and Black is clearly better, P.Utkin-M.Chigorin, corr. 1899.
Note that exchanging on d5 is necessary since 6.0-0?! allows 6...dxe4 7.Qxe4 Bf5 8.Qa4 (not
8.Bxc6+? bxc6 9.Qa4 Bd3 with a big advantage, N.MacLeod-W.Pollock, New York 1889) 8...a6
9.Bxc6+ Nxc6 10.d4 Qd7 11.dxe5 Nxe5 12.Qxd7+ Nxd7 13.Bf4 0-0-0 and Black has the edge with
the unopposed light-squared bishop.

6...Qxd5

226
7.0-0

White’s most popular move, preparing d2-d4.

The alternative is to push the pawn immediately: 7.d4 e4 8.c4 (this pending pawn sacrifice is the
point of White’s play; not 8.Nfd2?! Qg5! 9.Kf1 a6 10.Na3? axb5! 11.Qxa8 Qg4 12.Nc2 Nd5 13.Ne3
Nxe3+ 14.fxe3 Bd6 15.Nxe4 Qd1+ 16.Kf2 Qc2+ 17.Nd2 0-0 with a huge attack, R.Ovetchkin-
A.Hamatgaleev, Pskov 1998; nor 8.Bc4?! Qh5 9.Nfd2 Bd7 10.Qc2 f5 11.Nf1?! g5 12.b4 Qg6
13.Nbd2 0-0-0 and Black is clearly better, O.Roething-F.Marshall, Sylvan Beach 1904; instead,
8.Ng1!? a6 9.Bf1 showed a nice sense of humour in L.Webb-T.Pähtz, Senden 2014, though after
9...Bd7 10.Qb3 and 10...0-0-0 Black’s lead in development must be worth a little something)
8...Qd7!? (Zagorovsky’s suggested improvement on Keres’ 8...Qd8; Stockfish actually prefers the
latter, and if 9.Nfd2 Kf7! 10.Nxe4 Nxd4 11.c5 Nxb5 12.Qxb5 Qd7 13.Qe2, as in G.Mercky-
T.Boychuk, corr. 2006, then 13...Ng6, followed by ...Be7, ...Re8 and so on)

227
 a) 9.d5?? exf3 10.gxf3 (or 10.dxc6 bxc6 11.Ba6 Bxa6 12.Qxa6 fxg2 – Zagorovsky) 10...a6 11.dxc6
bxc6 0-1 pickle47-tsmenace, ChessWorld.net 2009.
b) 9.Nfd2 a6 10.d5 (here 10.Nc3 Qxd4 11.0-0 Bd7 12.Ndxe4 Rc8! 13.Bxc6 Bxc6 14.Qd1 Qxd1
15.Rxd1 Kf7 is just a slight edge with the bishops; whereas 11.Ndxe4?! axb5! 12.Qxa8 b4 13.Be3
Qe5 14.0-0 bxc3 15.Nxc3 Qa5! 16.Nb5 Qxa8 17.Nxc7+ Kf7 18.Nxa8 – T&H, and 18...Nf5 looks
better for Black) 10...axb5! 11.Qxa8 Nd4 (or 11...Nb4 12.0-0 Nc2 transposing; not 12.Na3? Nd3+
13.Kf1 Qf5 14.f3 exf3 15.Nxf3 Ne5 and wins – Zagorovsky) 12.0-0 Nc2 13.Nc3 (or 13.Nxe4 Nxa1
14.Nbc3 etc; not 13.a4? Nxa1 14.axb5 Nc2 15.Nxe4 Nf5 16.Qa4 Ncd4 and Black consolidated,
M.Doran-J.Tait, Midland Jamboree 1993) 13...Nxa1 14.Ndxe4 may yet be okay for White; e.g.
14...Nf5 (or 14...Kf7 15.cxb5 Nxd5 16.Rd1 c6 17.bxc6 bxc6 18.Qa4) 15.cxb5 Nc2 16.Qa4 Ncd4
17.Qc4 Bd6 18.Rd1 0-0 (or 18...Be5 19.f4) 19.Nxd6 Qxd6 20.Rxd4 Nxd4 21.Qxd4 with two pawns
for the exchange.
c) 9.Ng1 a6!

228
 10.Ne2 (not 10.d5? axb5 11.Qxa8 Nb4 12.Na3 Qg4 13.g3 Nd3+ 14.Kf1 Qd1+ 15.Kg2 Ne1+ –
Zagorovsky; or 13.Bd2 Nd3+ 14.Kf1 Nxf2 15.Kxf2 Nxd5 with a huge attack, P.Soderberg-
E.G.Svensson, corr. 1984) 10...Rb8 (Zagorovsky; if 10...axb5!? 11.Qxa8 Nb4 – T&H, then 12.0-0
Nc2 13.Nbc3 Nxa1 14.cxb5 Ng6 15.Qa4 f5 16.Bd2 Bd6 17.Rxa1 0-0 looks roughly equal) 11.Bxc6
(or 11.0-0!? axb5 12.cxb5, when 12...Nd8 13.Qa7 Nec6 14.bxc6 Nxc6 15.Qa4 b5 16.Qb3 is an
obscure transposition to 14.Qb3 below) 11...Nxc6 12.0-0 b5 13.cxb5 axb5 14.Qb3 and if Black has an
edge, it isn’t much of one; e.g. 14...Ba6 (or 14...f5 15.Bf4) or 15.Re1 (or 15.Nbc3 f5 16.a4) 15...f5
16.d5 Ne5 17.Nd4.

7...e4

229
 Pre-empting d2-d4.

8.c4!?

“A speculative move which we believe needs further testing” (T&H). Which it has had since their
book came out in 2009.
Instead:
a) 8.Nd4 Bd7 9.Nxc6 (here 9.d3!? Nxd4 10.Bxd7+ Qxd7 11.Qxd7+ Kxd7 12.cxd4 exd3 13.Rd1
Rd8 14.Rxd3 Kc8 15.Nc3 Nd5 gave Black a slight edge due to the IQP in afms-tsmenace,
ChessWorld.net 2010, though not one I was able to convert, so simply 9...exd3 10.Bxd3 0-0-0 might
improve; and 9.Re1 Nxd4 10.Bxd7+ Qxd7 11.Qxd4 Qxd4 12.cxd4 f5 13.d3 exd3 14.Nc3 0-0-0
15.Bg5 Rd7 16.Rad1 h6 17.Bxe7 Bxe7 18.Rxd3 Bf6 was good for Black in K.Child-J.Tait, Notts
League 2014, where 9...0-0-0! looks even stronger) 9...Nxc6! 10.d4 (or 10.d3 exd3 etc; if 10.Qb3,
J.Niewold-M.Pintor, corr. 2005, then 10...Qf5! 11.d4 exd3 is even better, e.g. 12.Re1+ Be7 13.Rd1 0-
0-0; while 10.Re1 0-0-0! 11.Rxe4? a6 12.Bxa6 Nb4 just wins – Steinitz) 10...exd3 11.Rd1,
P.Bersamina-A.Diermair, Tromsø Olympiad 2014, and now 11...0-0-0 12.Rxd3 Qf5 is very good for
Black, intending 13.Be3 a6 14.Bxa6 Nb8! etc.
b) 8.Ne1

can be met by 8...Bd7 and ...a7-a6. For example: 9.f3 (or if 9.d3 a6 10.Na3, I.Poulsen-S.Bjerke,
Odense 2013, then 10...Ng6 11.dxe4 Qh5 12.Nf3 Bd6 13.Be2 Nce5 14.Nxe5 Qxe5 15.f4 Bxa4
16.fxe5 Nxe5 with an edge) 9...exf3 10.Nxf3 a6 11.Be2 0-0-0 12.d4 g5!? with some initiative. B.Pott-
P.Speisser, corr. 2007, continued 13.Na3 Re8 14.Nd2 Ng6 15.Bf3 Qe6 16.Qc4 Qe3+ 17.Kh1 Qe7
18.b4 Nh4 19.Bh5 Be6 20.d5 Ne5 21.Qe4 Bf7 22.Bxf7 Qxf7 23.b5 Nd7 24.Qd3 Qg6! 25.Qh3
(25.Qxg6? hxg6 opens the h-file, with threats of ...Nf5-g3+) 25...Re2 26.bxa6 bxa6 27.Nac4 Bc5
28.Nb3 Bf2 29.Qg4 Rhe8 30.Be3 R2xe3 31.Nxe3 Bxe3 32.Qe2 Nf5 33.Qxa6+ Kd8 34.Qa8+ Ke7

230
35.d6+ Nxd6 36.Qf3 Kf8 and Black went on to win.
c) 8.Bxc6+ Qxc6! (if 8...Nxc6 9.Re1 f5 10.d4 Bd7, as in J.Showalter-H.Pillsbury, Buffalo 1894,
then 11.Qb3 Qxb3 12.axb3 a6 13.Nfd2 0-0-0 14.f3 looks okay for White) 9.Qxc6+ bxc6 and White
will suffer on the light squares; e.g. 10.Nd4 c5 11.Nb5 (or 11.Nb3 Nd5! 12.Re1 Kf7, intending
13.Rxe4 Bf5 14.Re1 Bd3) 11...Kd7 12.c4 (or 12.Rd1 Kc6 13.N5a3 Bf5, M.Mujunen-O.Rajala, corr.
2018) 12...Ng6 13.N1c3 a6 14.Na3 Bb7 15.Rd1 Bd6 16.d3 exd3 17.Rxd3 Rad8 18.Be3 Kc8 and even
now White has not equalized, D.Nenneman-Y.Del Toro Montoya, corr. 2015.

8...Qd7!?

I now think this is best, which is rather satisfying because it parallels Zagorovsky’s move in the
7.d4 line above. Instead:
a) 8...Qd8?! 9.Bxc6+ bxc6 (or 9...Nxc6 10.Re1 f5 11.d3 – T&H) 10.Ne1 Bf5 11.Nc3 Qd7 12.c5
Nd5?! 13.Nxe4 Be7 14.Ng3 Bg6 15.d4 was the stem game, J.Showalter-H.Pillsbury, New York 1897,
which White won.
b) 8...Qd6 is well met by 9.c5! Qxc5 10.d4 Qd6 (after 10...Qb6 11.d5 a6 12.Nc3 exf3 13.dxc6 axb5
14.Qxa8 Qxc6 15.Re1 Kf7 16.Bf4 Be6 17.Rac1 Ng6 18.Bg3 Bd6 19.Qa5, White should be okay,
K.De Smet-T.Llorach Gracia, corr. 2010) 11.d5! (here 11.Nfd2 Bd7! 12.Nc3 0-0-0 13.Rd1 Bg4
14.Ndxe4 Qb4! 15.f3 Be6 16.Qxb4 Nxb4 17.Ba4 Bc4 18.Bb3 Bxb3 19.axb3 a6 was better for Black
in view of the damaged white pawn structure, J.K.Andersen-V.Silin, corr. 2011) 11...exf3 12.dxc6
bxc6 13.Bc4 Nd5 14.Re1+ Be6, as in rooksac-tsmenace, ChessWorld.net 2018. Although I won that
game, Stockfish is unimpressed and claims equality; e.g. 15.Nc3 Kf7 16.Qd1 fxg2 17.Qf3 Re8
18.Rxe6 Rxe6 19.Nxd5 cxd5 20.Bxd5 Qd7 21.Bf4 Bd6 22.Bxd6 Qxd6 23.Re1 Rhe8 24.Qh5+ Kf8
25.Bxe6 Rxe6 26.Rxe6 Qxe6 27.Qc5+ with a drawn queen ending.
c) 8...Qh5 9.Re1 Bg4

231
 10.Nd4 (not 10.Bxc6+?! Nxc6 11.Rxe4+ Be7) 10...0-0-0! 11.Nxc6 Nxc6 12.Bxc6 bxc6 13.Qxc6
Bd6 14.Qa6+ Kb8 15.Qb5+ Qxb5 16.cxb5 Rhe8 is more interesting, in that Black’s development and
light square control offer more than enough for the pawn. Nonetheless, all three games to reach this
position ended in draws; e.g. 17.Nc3 Bf5 18.h3 Bg6 19.Nd5 Re5 20.Nc3 f5 21.b3 Ree8 22.Bb2 Be5
23.Na4 Rxd2 24.Bc3 Bxc3 25.Nxc3 Bh5 26.Re3 Rc2 27.Rae1 Rd8 28.g4! fxg4 29.hxg4 Bxg4 30.a4
Bf3 31.Nxe4 and so on, A.Serner-R.Carvaga, corr. 2013.

9.Ne1

Not 9.Bxc6? Nxc6 10.Re1 f5 11.d3 Nb4! and wins (T&H).

9...a6 10.Nc3 b6!

232
 The post-T&H improvement on 10...Rb8 11.Bxc6 Nxc6 12.Nxe4 Be7 13.d3 0-0 with mutual
chances and an eventual draw in V.Demian-W.Van Vugt, corr. 2004. After the text, Black intends
simply ...Bb7, so White has to make a decision about the b5-bishop:
a) 11.d4 exd3 12.Bxc6 (or 12.Nxd3 Bb7 13.Bxc6 Bxc6 14.Qc2 0-0-0) 12...Qxc6 13.Qxc6+ Nxc6 is
better for Black; e.g. 14.Bf4 (if 14.Nxd3 then 14...Bf5, or 14.Nd5 Bd6) 14...Bg4 15.Nd5 (or 15.Bxc7
Be2; and not 15.f3? Bc5+ 16.Kh1 Bf5) 15...0-0-0 16.Bxc7 Rxd5 17.cxd5 Kxc7 18.Rc1 Be2 19.Rxc6+
Kb7 20.Nf3 Bxf1 21.Kxf1 Bc5, followed by ...Rd8 and ...Rxd5.
b) 11.Bxc6 Qxc6 12.Qxc6+ (12.d4 exd3 is note ‘a’) 12...Nxc6 13.Nxe4 Bf5 14.d3 (or 14.Ng3 Bg6)
14...0-0-0 gives Black an excellent position for the pawn. White is worse if the d-pawn is let go, but
defending it ties the pieces down enormously, allowing Black to probe at will; e.g. 15.Be3 Bg6 16.c5
Kb7 17.cxb6 cxb6 18.Nc3 Nd4 19.Rc1 Bb4 20.Kh1 Nf5 21.a3 Be7 22.Bf4 Rd7 23.Ne4 Nd4 24.Be3
Re8 25.Rc4 Nc6 and White continues to suffer, C.Glanville-J.Potrata, corr. 2013.
c) 11.c5!? Bb7 12.Be2 has been seen in four games. In all four Black played 12...Nd4, which
doesn’t lead to very much: 13.Qxd7+ Kxd7 14.Bd1 Nef5 (or 14...bxc5 15.d3) 15.d3 exd3 16.cxb6
cxb6 17.Nxd3 Bc6 with only a tiny edge in J.Arufe Vazquez-V.Mondino, corr. 2016 (before Black
defaulted two moves later). Instead, 12...b5!? might improve; e.g. 13.Qxe4 0-0-0 14.b4 Nd4 15.Qb1
(the only move; not 15.Qd3? Nec6) 15...Nxe2+ 16.Nxe2 h5 and Black has an ongoing initiative.

B: 3.d4 exd4 4.c3

The Göring Gambit. I used to play this myself, when I was 10, and once won a brilliancy prize with
it – a copy of The Games of Anatoly Karpov by Kevin O’Connell & Jimmy Adams. You can surmise
what influence that book has had on my chess.
The two main moves here are 4.Nxd4 and 4.Bc4 – the former is the subject of Chapter Seven; the
latter transposes after 4...Nf6 to line A in Chapter Eight.

233
 A fourth option, 4.Bb5?!, can be met by 4...Bc5 5.0-0 (or 5.c3 dxc3 6.Nxc3 Nge7 7.Ng5 f6! 8.Nf7
Bxf2+ 9.Kxf2 Kxf7 – Lokander) 5...a6 (a useful insertion; White can hardly take the knight) 6.Ba4
(or 6.Bc4 d6) 6...Nge7, when White is scratching around for compensation: 7.c3 (after 7.Ng5 f6!
8.Nf7 Kxf7 9.Qh5+ g6 10.Qxc5 b5 11.Bb3+ Kg7 12.a4 d6 13.Qa3 b4 14.Qa2, the queen says
everything about White’s position, G.Donzellotti-A.Kashapov, corr. 2011; while 7.b4 Bxb4 8.Nxd4
0-0 9.Bb2, hoping for a Bxg7 trick, is neutralized by 9...Bc5 10.Nxc6 dxc6 11.Qh5 Bd4 12.c3,
E.Schiller-I.V.Ivanov, Los Angeles 1995, and 12...Ba7 13.Rd1 Qe8 with a clear advantage) 7...dxc3
8.Ng5!? d6 9.Qh5, K.Skold-H.Kokkoris, Varna Olympiad 1962, and now 9...g6 10.Qh4 h6 11.Nxc3
(or 11.Bb3 Ne5) 11...Bd4! 12.Nf3 Bg7 should consolidate.

4...d5

Again the d-pawn goes forward. If I recollect correctly 4...dxc3 5.Bc4 cxb2 6.Bxb2 was played in
my childhood game. It’s not easy to defend as Black, and there may not be any great advantage even
if you succeed. Anyway, I like the positions after 4...d5, and your opponent can easily get deflated,
which is always a good thing.

5.exd5

Other moves:
a) 5.e5? dxc3 (accepted, now that lines are closed) 6.Nxc3 d4! 7.Nb5 (or 7.Ne4 Bb4+ 8.Bd2 Qd5
9.Bd3 Bf5 10.Qe2 Nge7, A.Ermeni-J.Mesples, Cappelle online blitz 2021) 7...Bg4 8.Bc4 Bb4+ 9.Bd2
Bxd2+ 10.Qxd2 Bxf3 11.gxf3, M.Wangler-I.Natarov, KCF University Cup (rapid) 2021, and now
11...a6 is near enough winning.
b) 5.Bb5?! (or 5.Nxd4 first) 5...dxe4 6.Nxd4 Bd7 7.Qa4 (or 7.Nxc6 bxc6 8.Bc4 Bd6 9.Qd4 Nf6,
V.Valkov-E.Ovod, Albena 2012) 7...Nxd4 8.Qxd4 Bxb5 9.Qe5+ Qe7 10.Qxb5+ c6 11.Qa4 f5 and
Black is just a pawn up, A.Deev-E.Egorov, Pavlodar 2008.

234
 c) 5.Bd3!? has a couple of Ukrainian IM adherents so shouldn’t be discounted. The position bears
distinct similarities to a Chigorin Queen’s Gambit (not coincidentally, as it happens), in which vein
Black might play 5...Bg4!.

Then 6.cxd4 Bxf3 7.gxf3 dxe4 (not 7...Nxd4?! 8.Qa4+ Nc6 9.Ba6!) 8.Bxe4 Nf6 9.Bxc6+ bxc6 is
like a main line, where White’s d-pawn is at least as weak as Black’s doubled c-pawns. If instead
6.Qb3 (L.Sobolevsky), then 6...Bxf3 7.gxf3 Nge7! looks thematic; e.g. 8.Bf4 (or 8.Qxb7 Rb8 9.Qa6
g6 10.exd5 Qxd5 11.Be4 Qe6) 8...g6 9.Qxb7 Rb8 10.Qxc7 Qxc7 11.Bxc7 Rxb2 with the advantage.
Even after 6.0-0 Nf6 7.Re1 (or 7.e5 Bxf3 8.gxf3 Nd7) 7...Be7 8.e5 Nd7 9.Bb5 (not 9.Bf4?! 0-0
10.cxd4 Nxd4 11.Bxh7+ Kxh7 12.Qxd4 Bxf3 13.gxf3 Nc5 14.Nc3 c6 and Black is better,
E.Bogdanov-T.Lirindzakis, Anogia 2019) 9...dxc3 10.Nxc3 a6 11.Bxc6 bxc6, J.Llorente Alvarez-
H.Daurelle, corr. 2010, and 12.Qa4! Bxf3 13.gxf3 Nb6 14.Qxc6+ Qd7 15.Qxd7+ Kxd7 16.Rd1 c6, I
prefer Black’s chances.

5...Qxd5 6.cxd4

Not 6.Nxd4?! Nf6, when White is behind in development and with a stupid pawn on c3.
There’s no point in trying to be clever either: 6.Be2 Bf5 7.cxd4 Bxb1! 8.Rxb1 Bb4+ 9.Bd2 Bxd2+
10.Qxd2 0-0-0 11.0-0 and the d-pawn is a weakness, A.Zacik-K.Motuz, Tatranske Zruby 2008.
Indeed, Bologan suggests taking it straight away: 11...Nxd4! 12.Nxd4 Qxd4 13.Qc2 Nf6 and Black
cannot be worse.

6...Bg4

235
7.Be2

White does have a significant alternative here: 7.Nc3 Bb4 (the sharp 7...Bxf3 8.Nxd5! Bxd1
9.Nxc7+ Kd7 10.Nxa8 Bh5 11.d5 Nd4 12.Bd3 is close to a forced draw; e.g. 12...Bg6! 13.Bxg6 hxg6
14.Kd1! Nh6 15.Be3 Nhf5 16.Rf1! Nxe3+ 17.fxe3 Nf5 18.Rf3 Bd6 19.e4 Rxh2 20.Rc1 Rh1+ 21.Kd2
Bb4+ 22.Kc2 Nd4+ 23.Kd3 Rxc1 24.Rxf7+ Ke8 25.Rxb7 – Müller & Voigt, and now 25...Rd1+
26.Kc4 Bd6 27.Rxa7 Rc1+ 28.Kd3 Rd1+ ½-½ J.Llorente Alvarez-K.Herzog, corr. 2009) 8.a3!?
(instead of 8.Be2 as in the main line) 8...Bxf3 9.axb4 (or 9.gxf3 Qe6+ 10.Be2 Bxc3+ 11.bxc3 Nge7,
when White’s structural wreckage balances the two bishops) 9...Qe6+ (or 9...Bxd1 10.Nxd5 0-0-0
11.Bc4 Bg4 12.b5 Nxd4 13.Rxa7, H.Warzecha-K.Tauscher, corr. 2012) 10.Ne2 Bd5 11.Be3 Nf6
12.Nf4 Qe7 13.b5 Nb4 14.Be2 0-0 15.0-0 with equality, H.Warzecha-E.Hebel, corr. 2008.

7...Bb4+ 8.Nc3

In truth I’ve generally reached this position via entirely different routes, such as 1.Nf3 Nc6!? 2.d4
d5 3.c4 Bg4 (this is another Chigorin) 4.e3 e5!? 5.cxd5 Qxd5 6.Nc3 Bb4 7.Be2 exd4 8.exd4. Yes, it
really is the same thing.

8...Bxf3 9.Bxf3 Qc4!

236
 Capablanca’s solution. The queen is a decided irritant, both preventing castling and reinforcing the
attack on c3.

10.Bxc6+

Other moves:
a) 10.d5? Nd4 makes things very much worse for White.
b) 10.Bd2?! 0-0-0 11.Bxc6 Qxc6 12.Qg4+ Kb8 13.0-0 (13.Qxg7? Re8+ is far too greedy) 13...Nf6
sticks White with the weak d-pawn, D.Doncevic-D.M.Pantovic, European Ch., Skopje 2019.
c) 10.Be3 0-0-0 (instead, 10...Bxc3+ 11.bxc3 Qxc3+ 12.Kf1 Qc4+ may repeat after 13.Be2 Qd5
14.Bf3 Qc4+ or 13.Kg1 Nge7 14.Rc1 Qxa2 15.Ra1 Qc4 16.Rc1 ½-½ F.Marshall-J.R.Capablanca,
Lake Hopatcong 1926) 11.Qb3 (not 11.Be2? Bxc3+ 12.Kf1 Qb4 13.bxc3 Qxc3 14.Rc1 Qa3 and
Black has safely captured the pawn, S.Petrik-T.Petenyi, Slovakian League 2008; nor 11.Bxc6??, as in
T.Beerdsen-V.Kovalev, Titled Arena blitz 2021, even if White won with it twice, since simply
11...Bxc3+! 12.bxc3 Qxc3+ and ...Qxc6 should reverse the result) 11...Qxb3 12.axb3 Nge7 is note
‘d’.
d) 10.Qb3 Qxb3 11.axb3 (here 11.Bxc6+ bxc6 12.axb3 is 11.Qb3 in the main line) 11...Nge7
12.Be3 (or 12.0-0 0-0-0, intending 13.Be3 a6 or 13.Ra4?! a5) 12...0-0-0 13.0-0 a6 is also perfectly
satisfactory for Black, who will naturally target the isolated d-pawn. For example: 14.Ra4 (or 14.Rfd1
Kb8 15.Ne4? Nf5 16.d5 Ne5 17.Bf4 Nxf3+ 18.gxf3 Rd7 19.Rac1 Rhd8 20.Nc5 Bxc5 21.Rxc5 Ne7
and Black won, A.Polivanov-S.Bogdanovich, Donetsk 2011) 14...a5 (or 14...Bd6 – Bologan) 15.Rd1
Kb8 16.g3 Rd7 17.d5 Ne5 18.Bg2 Rhd8 19.Rda1?! (one might almost wonder whether the engine
mistakenly picked up the wrong rook) 19...Nxd5! 20.Nxd5 Rxd5 21.Bxd5 Rxd5 22.Rc1 Rd3 23.Bf4
f6 24.Raa1 Rxb3 25.Rc2 b5, when Black’s knight and pawn majority were worth a lot more than
White’s rook, Xiphos-LcZero, TCEC Cup 2 2019.

237
10...bxc6!

Leaving the black queen in situ is more important than structure. Now in order to castle, White has
to take the queens off.

11.Qe2+

Or 11.Qb3 Qxb3 12.axb3, hoping to make something of the a-file. Then 12...a5 (another idea is
12...Ne7 13.Ra6 Kd7 and ...Rhb8, as in J.Moreno Ruiz-V.Gallego Jimenez, Madrid 1992) 13.Bd2 0-
0-0 14.Ne2 Ne7 15.Bxb4 axb4 16.Kd2 Nd5 looks equal; whereas after 15.0-0-0?! Rhe8 16.Rhe1 Nf5
17.Bc3, M.Herbold-P.Haba, Bayerisch Eisenstein 2013, and 17...Re4 I prefer Black, who can switch
files at will.
The same idea featured in one of my games: 13.0-0 0-0-0 14.Be3 Ne7 15.Rfd1 Rhe8 16.g3 Nf5
17.Rd3 h5 18.Ne2?! c5 19.Rad1 Rd7 20.Kf1? (20.dxc5 was necessary) 20...Rde7! (switching files)
21.Nc3, P.Dodd-J.Tait, corr. 2003, where Stockfish gives 21...Rxe3! as the cleanest win.

11...Qxe2+ 12.Kxe2 0-0-0 13.Be3 Ne7

Although objectively equal (“0.00”) I like this position for Black, whose doubled c-pawns are no
weaker than the white d-pawn, which can potentially be attacked by the knight, bishop, both rooks,
and even both c-pawns. It also helps when White has gotten here via 1.d4 or 1.Nf3 and may not know
the Göring Gambit exists, let alone what to do once there.
A few sample variations:
a) 14.Rhd1 Rhe8 15.a3?! (very obliging, since the bishop wanted to go to a5 anyway; instead,
15.Rac1 is note ‘b’; while 15.Rd3 Nf5 16.Rad1 c5 17.dxc5 Rxd3 18.Rxd3 Nxe3 19.Rxe3 Rxe3+
20.Kxe3 Bxc5+ 21.Ke2 Kd7 was soon drawn in B.Spassky-I.Bondarevsky, Sochi 1964) 15...Ba5

238
16.Kf3 Bb6 17.Na4 Nf5 18.Nc5 Rd5 19.Rac1?! Red8 20.b4 Nxd4+ 21.Bxd4 Rxd4 22.Rxd4?! Rxd4
23.Rc3 Bxc5 24.bxc5 Kd7 and Black won the rook endgame, A.Miles-J.Nunn, Islington 1970.
b) 14.Rac1 Nf5 15.Rhd1 Rhe8 16.Na4!? (offering to trade weaknesses; 16.Kf3 Rd6 17.g3 Ba5
18.Rd3 Bb6 19.Rcd1 Red8 20.d5?! Ne7 21.Bf4 R6d7 22.d6 Ng6 23.Ne4 Nxf4 24.Kxf4?! cxd6
25.Rc1 d5, V.Narandzic-D.Blagojevic, Montenegrin Ch., Tivat 2000, is another way White can go
wrong) 16...Rd6 (or 16...Nxd4+ 17.Kf1 Nb5 18.Rxd8+ Rxd8 19.Rxc6 Bd2) 17.Rc4 Ba5 18.Rc5 Nxe3
19.fxe3 Bb6 20.Rc4 Rg6 was soon drawn, S.Calvi-I.Rutkowski, Frankfurt 2004.
c) 14.Kd3!? (given as best by Flear and Bologan) 14...Nf5 (otherwise 14...c5 15.Kc4 cxd4 16.Bxd4
Nc6 has produced 13 quick draws in the databases) 15.Kc4 Ba5 16.b4 Bb6 17.Rad1 Rhe8 18.Rd3
Re6 19.g4 Nd6+ 20.Kb3 Ne4 21.Rhd1 Nxc3 22.Kxc3 Rd5 was also drawn, this time in 89 moves of a
fluctuating game, M.Herbold-S.Milliet, Gibraltar 2017.

C: 3.Nc3

The Three Knights, heading for the Four Knights. Sigh. I face the Four Knights so rarely that, on
the occasional occasion it occurs, I’ve forgotten the intricacies of its quite extensive theory and end
up trying to fake it against a decent opponent who’s aiming to win on technique.
Yes, I have looked at 3...f5. It loses by force; e.g. 4.d4! fxe4 5.Nxe5 Nf6 6.Bc4! d5 7.Nxd5! Nxd5
8.Qh5+ g6 9.Nxg6 hxg6 10.Qxg6+! Kd7 11.Bxd5 Qe8 12.Bf7 Qe7 13.Bg5 Ne5 14.Qf5+ 1-0
G.Breyer-Z.Balla, Bad Pistyan 1912.
Searching for another solution, I noticed an apparent discrepancy.

3...Bc5!?

In New Ideas in the Four Knights GM John Nunn calls this move “dubious”, adding that “the
impression that this line is very bad for Black (...) is probably true”. Yet a trawl through the databases

239
shows GM Oleg Romanishin playing 3...Bc5 twenty times (or more) for a plus score (+6 =10 -4) as
Black.
That’s not to say Nunn is necessarily wrong. To quote from Wikipedia, GM Romanishin “has a
reputation for the use of rare, offbeat and sometimes, long since discarded systems. Only by means of
deep research and accurate preparation has he been able to employ these openings as weapons to
sidestep known theory and fight for the full point.”
A fine fellow, in other words, whose example we should surely follow.

4.Nxe5

The only dangerous reply, playing for a typical fork trick.

a) 4.Na4 Bd6!? is not at all silly. The point is seen in comparison with 4...Be7, when 5.d4! gives
White some initiative; e.g. 5...exd4 6.Nxd4 Nxd4 7.Qxd4 Bf6 8.Qd3 Ne7 9.Be3 0-0 10.0-0-0,
Glaurung 2-Rodent III, TCEC 16 Testing 2019. Whereas 4...Bd6 5.d4 exd4 6.Nxd4 is far less
threatening; e.g. 6...Nf6 7.f3 (or 7.Nf5 Bb4+! 8.c3 Bf8, intending ...d7-d5, or 9.Qe2 d6 and ...g7-g6)
7...0-0 8.Nf5 Bb4+! 9.c3 Be7 10.Nxe7+ (not 10.Bc4? d5!) 10...Qxe7 11.c4 d5! 12.cxd5 Nxe4!
13.fxe4 Qh4+ 14.Kd2 Qf4+ 15.Kd3 (or 15.Kc3 Qe5+) 15...Nb4+ 16.Kd4 Nc6+ with a draw.
b) 4.Be2 d6 (not 4...a6? 5.Nxe5!) 5.Na4 isn’t very threatening either; e.g. 5...Nge7 6.0-0 Ng6 7.c3
0-0 (or 7...a5!?) 8.d4 Bb6 9.d5 Nce7 10.b4 Bd7 11.Nb2?! (White might as well take the bishop)
11...f5 12.exf5 Nxf5 13.Nc4, L.Gutman-O.Romanishin, Bad Endbach 1995, and now 13...e4! 14.Ng5
e3 (Romanishin) is good for Black, especially with Stockfish in charge: 15.Nxb6 Qxg5! 16.Nxd7 Nf4
17.Bf3 (or 17.g3 Nh3+ 18.Kg2 Nxf2 19.Rxf2 Nh4+) 17...e2 18.Bxf4 exd1=Q 19.Bxg5 Qd3 20.Nxf8
Rxf8 and White doesn’t have enough for the queen.
c) 4.g3 d6 5.Bg2 transposes to 5.Nf3 in line A of Chapter Three, where Black plays 5...f5!?, and
5.Na4 can be answered similarly: 5...f5!? (or 5...Bb6 6.Nxb6 axb6 7.Bg2 f5, and if 8.Nh4?!,
P.Ornstein-A.Flaata, Gausdal 2000, then 8...fxe4 9.Bxe4 Nf6 10.Bg2 Bg4 11.f3 Be6 leaves the h4-
knight stranded) 6.Nxc5 (or 6.d3 fxe4 7.dxe4 Bb6) 6...dxc5 7.exf5 (or 7.Bb5 fxe4 8.Nxe5 Qd5)
7...Bxf5 8.Bg2 Nf6 9.0-0 0-0 and Black’s central control is more relevant than the isolated e-pawn.
d) 4.Bb5 is a Classical (3...Bc5) Ruy with the anodyne 4.Nc3, which allows Black to switch to a
good version of Bird’s Defence with 4...Nd4!.

240
 For example: 5.Nxe5 (here 5.Nxd4 does nothing after 5...Bxd4, e.g. 6.Ne2 Bb6 7.d4 c6 8.Ba4 Nf6
9.Bg5, W.Simmons-W.Jonas, Australian Ch., Melbourne 1922, and now 9...h6! with an edge;
similarly 5.Ba4 c6 6.0-0 d6 7.Nxd4 Bxd4 8.Qf3 Nf6, F.Serra Fontoya-J.Vives Esmerat, Catalan Team
Ch. 2014; while 5.Bc4 d6 6.Nxd4 Bxd4 7.d3 Qh4 8.Qf3 Nf6 was drawn in a few more moves,
S.Tartakower-M.Vidmar, Bled 1931) 5...Qg5 (5...Qe7 is also good) 6.Ng4, L.Steiner-
F.Sämisch,.Dresden 1926, and now 6...Nh6! 7.h3 (not 7.Nxh6? Qxg2 8.Rf1 Nf3+ 9.Ke2 c6 and wins;
while 7.Ne3?! Nxb5 8.Nxb5 Bxe3 9.Nxc7+ Kd8 10.Nxa8 Qxg2 11.Rf1 Bd4 is good for Black after
any of 12.c3 Qxe4+, 12.d3 Ng4, 12.f3 d5, or 12.Qe2 d6) 7...Nxg4 8.hxg4 Nxb5 9.Nxb5 Bxf2+
10.Kxf2 Qxb5 and Black is at least equal.
e) 4.Bc4 d6!? (avoiding 4...Nf6 with a Giuoco Pianissimo, beloved of beginners everywhere)

241
 5.Na4 (going after the c5-bishop; if White instead plays 5.d3, then Black reciprocates with 5...Na5!;
if White holds back on both, Black has useful moves with the rooks’ pawns, e.g. 5.a3 a6 6.d3 h6 7.h3
Nf6 8.0-0 and now 8...g5!? 9.g4 Nxg4 10.hxg4 Bxg4 11.Be3, as in M.Kurbanov-V.Demakov, Irkutsk
2011, by transposition, and 11...Nd4 12.Bxd4 exd4 13.Nb1 Qd7 with a strong attack) 5...Qf6!? (or
5...h6 6.Nxc5 dxc5 7.d3 Nge7 8.Be3 Qd6 9.Nd2 b6 10.0-0 g5!? with unbalanced play, E.Can-
Mi.Adams, Turkish League 2010) 6.Nxc5 (or if 6.c3 Nge7 7.h3, Y.Kruppa-R.Dautov, Frunze 1988,
then 7...h6) 6...dxc5 7.d3 h6 8.Be3 b6 9.h3 Nge7 10.Qe2 Be6 and Black stands quite well; e.g.
11.Nd2 Bxc4 12.dxc4?! Nd4 13.Qd1 Qg6 14.0-0 Rd8 15.c3 0-0 16.Nb1?! (16.cxd4 cxd4 regains the
piece) 16...Ne6 17.Qg4 Rd3 18.Qxg6 Nxg6 and Black dominates, P.Teclaf-Mat.Bartel, Polish Ch.,
Warsaw 2020.

4...Nxe5

The following sequence is effectively forced. Not 4...Bxf2+? 5.Kxf2 Nxe5 6.d4 and White is
clearly better.

5.d4 Bd6 6.dxe5

The Halloween Gambit-esque 6.f4? is unsound; e.g. 6...Ng6 7.e5 (or 7.Bc4 c6 8.0-0 Bc7 9.f5 d5!
10.exd5 N6e7) 7...Bb4 8.Bd3 (not 8.f5? Bxc3+ 9.bxc3 Qh4+ with another check to pick up the f5-
pawn), St.Kristjansson-Gy.Magnusson, Dos Hermanas (blitz) 2003, when Stockfish comes up with
8...f5!? 9.Bxf5 N8e7 10.Bd3 (or 10.Bh3 0-0 11.0-0 d5) 10...d6 11.0-0 dxe5 12.dxe5 Bf5 and Black
consolidates.

6...Bxe5

The critical position which dates back to the Paulsen-Anderssen match of 1877. White has tried 17
moves here, many of which have some merit. IM Andrey Obodchuk, the only author to take 3...Bc5

242
seriously, identifies the following three as the most testing:
C1: 7.Qd3
C2: 7.f4
C3: 7.Bd3
Of the rest:
a) 7.Bc4 is the most natural and certainly the m``ost frequent choice – and yet hardly threatens. So
little in fact that Obodchuk doesn’t bother with it at all. Black responds with 7...Nf6, unconcerned
about any Bg5 pin since the e5-bishop already defends the knight and a further ...c7-c6 solves almost
all problems.

For example:
a1) 8.Bg5 0-0 9.Nd5?! c6 10.f4?! Bxb2 11.Nxf6+ Bxf6 12.Bxf6 Qxf6 13.0-0 d5 and Black is
clearly better, bobby fissure-tsmenace, ChessWorld.net 2019.
a2) 8.Qe2 0-0 9.0-0 Re8 10.Bg5 c6 11.f4? Bxc3 12.bxc3 d5 with a huge advantage, craig4000-
tsmenace, ChessWorld.net 2019.
a3) 8.Qd3 0-0 9.Ne2 (not 9.Bd2?! c6 10.0-0?! b5! 11.Bb3 a5 12.a3 Ba6 13.f4 Bc7 14.Nd1 d5 and
White is in big trouble, Earl of Norfolk-tsmenace, ChessWorld.net 2019) 9...c6 10.Bg5, as in
C.H.Alexander-C.Dörner, Hilversum 1947, can be answered by 10...Bxb2 quite safely; e.g. 11.Rb1
Be5 12.f4 Qa5+ 13.c3 (or 13.Kf1 d5) 13...Nxe4! 14.Qxe4 Bxc3+ 15.Kf1 d5 16.Qf3 dxc4 17.Qxc3
Qxc3 18.Nxc3 f6 19.Bh4 b5 with a pawn mass for the knight.
a4) 8.0-0 0-0 9.Re1 (or 9.Bg5 c6 10.Re1 d6 11.Qd2 Re8 12.Nd1?! h6 13.Bxf6 Qxf6 14.c3 Bf4
15.Qc2 d5 16.Bd3 Bf5 17.Ne3 Bxe4 and Black won, R.Tomczak-O.Romanishin, German League
1992) 9...Re8 10.Bg5 (here 10.f4 Bxc3 11.bxc3 Rxe4 12.Rxe4 Nxe4 13.Bxf7+ Kxf7 14.Qd5+ Kf8
15.Qxe4 Qe8 is level) 10...c6 11.f4? Bxc3 (11...Qb6+! 12.Kh1 Bxc3 13.bxc3 Nxe4 is even better)

243
12.bxc3 Qb6+ 13.Qd4 Nxe4 and Black was a pawn up with ...d7-d5 coming, J.Willow-J.Tait, Notts
Championship 2017. (Jonah still managed to draw, aided by me later missing mate in four.)
b) 7.Be2!? is trickier than it looks.

By developing small, White leaves the d-file open, planning 7...Nf6 8.Nb5! c6? (as in line C3)
9.Nd6+! with a clear advantage, A.Khakpoor-M.Lodhi, Asian Seniors Ch., Larestan 2015; nor does
8...a6 9.f4!, M.Samusenko-P.Dumakov, Kazan 2017, fully equalize. Probably Black should prefer
7...d6 here, intending 8.f4 Bxc3+ 9.bxc3 Qh4+ 10.g3 Qh3 as per line C2. Other arrangements of f2-f4
are not to be feared either; e.g. 8.0-0 Nf6 9.Bg5 c6 10.Bh4 0-0 11.f4 Bxc3 12.bxc3 Qb6+ 13.Bf2 Qa5;
or 8.Qd3 Nf6 9.Nd1 Nd7 10.f4 Bf6, followed by ...0-0 and ...Re8, when Black seems okay.
c) 7.g3 has a 100% record for White. I’m not sure why. After 7...Nf6 8.Bg2 0-0 9.0-0 (or 9.Nd5
Nxd5 10.exd5 c5 11.0-0 d6, Rodent III-Glaurung 2, TCEC 16 Testing 2019) 9...Re8 (or 9...Bxc3
10.bxc3 d6) 10.Re1 d6 11.f4 (or 11.h3 Bxc3 12.bxc3 Bd7 13.Qd4 Bc6 and ...Nd7, P.Potapov-
T.Nyback, European Ch., Rijeka 2010) 11...Bxc3 12.bxc3, M.Bagi-M.Jurcik, Slovakian League 2016,
and 12...Be6 Black looks fine to me.
d) 7.Be3 plans to block the black bishop with Bd4, aiming for a simple space advantage.

244
 The bishop should therefore be dumped: 7...Bxc3+ 8.bxc3 d6 9.Qd4 (or if 9.Bc4 Qe7 10.Qd3 Nf6
11.f3 0-0 12.0-0, R.Sehgal-E.Hu, Seattle 2014, then 12...Be6) 9...Nf6 10.e5!? (or 10.0-0-0 0-0 11.Bg5
Qe7, since Bxf6 etc is not a problem given White’s own doubled pawns) 10...Ng4 11.exd6 0-0!
12.dxc7 Qxc7, when Black has more than adequate compensation; e.g. 13.0-0-0?! (13.Qc5 is safer)
13...Nxe3 14.Qxe3 Be6 15.Kb2 Qa5 16.a3 Rae8! with a strong initiative, boqsa-tsmenace,
ChessWorld.net 2020.
e) 7.Bd2 is “devoid of ambition” (Obodchuk). In this case 7...Nf6 8.Bd3 d5! seems more than
satisfactory; e.g. 9.exd5 (or 9.Qe2 0-0 10.0-0-0 Re8) 9...0-0 (not 9...Nxd5? 10.0-0 Nxc3?, as in
D.Moldovan-N.Mitkov, World Junior Ch., Mamaia 1991, because of 11.Qh5! with a big advantage –
Obodchuk) 10.0-0 Bg4 11.f3 (or 11.Qb1!? h6) 11...Bh5 12.Bg5 h6 and Black is fine.
f) 7.Nd5 threatens f2-f4, so Black should hit the knight at once: 7...c6 8.Ne3 Nf6 9.Bd3 (not
9.Bc4?! 0-0 10.Qd3? Re8 11.Qb3 d5! 12.exd5 cxd5 13.Bb5, L.Paulsen-A.Anderssen, 7th matchgame,
Leipzig 1877, since 13...d4! 14.Bxe8 Qxe8 is terrible for White; and 9.Nc4 doesn’t trouble Black
after 9...0-0 10.Nxe5 Qa5+ and 11...Qxe5) 9...d5 10.exd5 cxd5 11.0-0 0-0 12.c3 Re8 and Black stood
well in A.Naiditsch-Z.Efimenko, 6th matchgame, Mukachevo 2010, if not well enough given that
White only needed to draw.
g) 7.Ne2 also intends f2-f4, so Black should give the bishop a retreat: 7...c6 8.f4 Bc7 9.Ng3 (or
9.Nc3 Nh6!?) 9...d6 (or 9...d5!? 10.exd5 Nf6 11.dxc6 Qxd1+ 12.Kxd1 Bg4+ 13.Ke1 0-0-0 with
compensation) 10.Be3 Nf6 11.Bd3, G.Gossip-D.Baird, New York 1889, and now Stockfish likes
11...h5!? 12.h3 h4 13.Ne2 0-0 14.Nc3 Re8 for Black.
h) 7.Qf3 Nf6 8.Bf4 (or 8.Bg5 c6) 8...d6 9.Bxe5 dxe5 10.Qg3 0-0 11.Qxe5 Nxe4 12.Nxe4 Re8
13.Qf4 Qd6 14.Qxd6 Rxe4+ 15.Be2 cxd6 16.f3 Re5 is an engine line, played four times between
them. Clearly they consider the d6-pawn to give White a little something, but it’s Black’s only
weakness and all four games were drawn; e.g. 17.Rd1 Bd7 18.Kf2 (not 18.Rxd6?? Bb5 19.Rd2 Rae8)
18...Rc8 19.c4 Rce8 20.Rd2 Kf8 21.Rhd1 R8e6 22.f4 Rf6 23.g3 Ra5 24.a3 Ke7 25.Rd5 Rxd5

245
26.Rxd5 b6 and so on, Stockfish-Houdini 6, CCC 2.Blitz Battle 2018.
i) 7.Qh5 innocently attacks e5 and f7 (after Bc4). Both are easily defended: 7...d6 (there’s no need
to give up the bishop just yet, even if 7...Bxc3+ 8.bxc3 Qe7 is okay too) 8.Bc4, V.Tilicheev-
I.Babikov, Moscow 2012, and now 8...Qd7, followed by ...Nf6, drives the white queen away with
equality.

C1: 7.Qd3

An ostensibly aggressive option, intending Bd2 and to castle long. Black responds by castling short
as quickly as possible.

7...Nf6 8.Bd2

Using the queen to maintain the pawn structure after 8.f4?! Bxc3+ 9.Qxc3 is not an improvement
on 7.f4 (line B) because the e-pawn needs securing: 9...0-0 10.e5 (or 10.Bd3? Nxe4) 10...Ne4 11.Qf3
d5 “with good counterplay for Black” (Obodchuk); e.g. 12.Be3 f5 13.0-0-0 Be6 14.Rg1 c5 15.g4!?
fxg4 16.Rxg4 Qa5!? (16...Bxg4 17.Qxg4 Qe8 18.Bg2 is unclear) 17.a3 b5 and Black’s attack is at
least as strong.
Inserting 8.Bg5 h6 9.Bd2 doesn’t enhance White’s chances either; e.g. 9...0-0 10.0-0-0 Re8 11.f3
c6 12.Be3?! d5 13.exd5 Bxc3 14.Qxc3 Nxd5 15.Qa3, O.Brendel-C.Pieper Emden, German League
1992, and now 15...a5, followed by ...b7-b5 is attractive, since grabbing the piece with 16.c4 is very
dangerous after 16...Bf5! 17.cxd5 cxd5 18.Rd2 Rc8+ 19.Kd1 Qc7.
Black might even leave the bishop where it is and just castle: 8...0-0, intending 9.0-0-0 (or 9.f4
Bxc3+ 10.bxc3 Qe7 11.e5 d6) 9...Bxc3 10.bxc3 (not 10.Qxc3? Nxe4) 10...Qe7 with a perfectly sound
position.

246
8...0-0 9.0-0-0

Still not 9.f4?! Bxc3 10.Bxc3 d5 11.e5 Ne4 and Black is fine.

9...Re8

Threatening ...Bxc3 and thus deterring f2-f4 since the e-pawn must again be defended.

10.f3

The alternative 10.Re1 is met the same way: 10...c6, preparing a freeing ...d7-d5.

Now 11.f4 Bxc3 12.Bxc3 can be answered by 12...d5 with equality. Note that after 13.exd5 Nxd5
14.Rxe8+ Qxe8 15.Bd2 Nf6 16.Qe3 Bf5 17.Qxe8+ Rxe8 the two bishops do not offer White
anything, since the knight is quite as good as the dark-squared bishop. Indeed, R.Adamson-
L.Altounian, US Ch., San Diego 2004, eventually reduced to those two pieces and Black won the
endgame.
White might find a more useful move than f2-f4 in this variation. For instance: 11.g3 Bxc3 (more
accurate than 11...d5 12.exd5 Bxc3, when White has the option of 13.Rxe8+ Qxe8 14.Qxc3 Nxd5
15.Qd4) 12.Bxc3 d5 13.exd5 (or 13.Bg2 Nxe4) 13...Nxd5 14.Rxe8+ Qxe8 was assessed by
Romanishin himself (in 1992) as slightly better for White.
That didn’t stop him playing it as Black: 15.Bd2 Be6 16.Qd4, and now Obodchuk suggests
16...b6!? (improving on 16...Qd7 17.c4 Nb6 18.Qxd7 Nxd7 19.Bc3 Rd8 20.Be2 Nc5 21.Rd1 Rxd1+
22.Kxd1, when Black had to endure a difficult defence to draw, O.Korneev-O.Romanishin, Saragossa
1996), which looks okay to me; e.g. 17.Bd3 (if 17.Bg2 then 17...Rd8 – Obodchuk; or 17.c4?! c5)
17...Qd7 18.Re1 Bf5 19.c4 Nb4! 20.Qxd7 Nxd3+ 21.Qxd3 Bxd3 22.b3 Kf8 23.Bf4 Rd8 and the
opposite-coloured bishops increase the likelihood of a draw.

247
 Another possibility is 11.h3, after which 11...Bxc3 12.Bxc3 d5 13.exd5 Nxd5 14.Rxe8+ Qxe8
15.Bd2, with ideas of c2-c4 and g2-g4, looks more promising as Black’s minor pieces lack good
squares.
But there is no obligation to follow through with ...Bxc3 just yet. Black can find things to do
instead on the queenside: 11...a5!? (not 11...h5?!, Be.Larsen-E.Torre, Bad Homburg 1998, which asks
for 12.g4! – Mikhalevski) 12.g4 (or 12.f4 Bxc3 13.Bxc3 d5 again; not 12.g3 b5 13.Bg2? Ba6,
threatening ...b5-b4) 12...Qc7 13.Bg2 b5 14.f4 (what else?) 14...Bxf4 15.Nxb5 Bxd2+ 16.Qxd2 cxb5
17.e5 Ne4! 18.Bxe4 Bb7 and Black is fine.

10...c6

Seemingly preparing ...d7-d5 once more, but it often has other purposes here.

11.g4

The logical corollary to f2-f3. The more restrained 11.g3 encourages 11...b5 12.Bg2 b4, and if
13.Ne2?!, M.Golubev-E.Torre, Leuven 2003, then 13...Qb6 (or 13...a5!?, threatening ...Ba6) 14.Be3
Qa5 15.Kb1 d5 with a strong initiative.

11...Qb6

Romanishin’s improvement on 11...d5 12.exd5 Bxc3 13.Bxc3 Nxd5 14.Qd4 Qg5+ 15.Bd2 Qf6
16.Qxf6 Nxf6, V.Bologan-O.Romanishin, Nikolaev Zonal 1995, where White has a good version of
the previous endgames because Black’s minor pieces again lack squares. In this line 12...cxd5!? is a
better try, intending 13.Be3 (or 13.g5 d4) 13...Be6 14.g5 Nh5 15.Nxd5 Rc8 with compensation.
Another game saw 11...Qc7!? 12.Kb1 b5 13.Be3 (or 13.g5 Nh5 14.Ne2 a5 15.Qe3 d5 16.f4 Bxf4
17.Nxf4 Rxe4 18.Nxd5 cxd5 19.Qf2 with a mess) 13...b4 14.Ne2 a5 (or 14...d5!? 15.f4 Bxb2) 15.g5

248
Nh5 16.Qd2 d5! 17.f4 d4 18.Nxd4 Bxf4 (or 18...Bxd4 19.Bxd4 Bg4) 19.Be2 Bxe3 20.Qxe3 Nf4 and
Black was fine, Ma.Pavlov-S.Bogdanovich, Odessa 2011.

12.Be3

Now if 12.Re1 then 12...d5! 13.exd5 Nxd5 makes sense, since 14.Nxd5?? is unthinkable and
14.Kb1 Bd7 15.Na4 Qc7 16.Nc5 Rad8 17.Nxd7 Rxd7 gave Black good play in J.Pálkövi-
O.Romanishin, Balatonbereny 1996, continuing 18.Qb3 Rde7 (or 18...Red8 – Obodchuk) 19.Bc4 Nb6
20.f4 Nxc4 21.Qxc4 Bd6 22.Rxe7 Qxe7 23.Bc3 Qe4 24.Qxe4 Rxe4 25.Rd1 Bf8 26.Rd8? (26.Rd4 is
equal) 26...Rxf4 27.a3 f5! 28.g5 (or 28.Bb4 fxg4) 28...b6 and Black won.

12...Qa5 13.Bd4

After 13.g5 Nh5 14.Kb1 d5 15.exd5 Bxc3 16.bxc3 cxd5 “Black is not at all worse” (Obodchuk). In
his notes to the game above, Bologan assessed 13.Bd4 as good for White. Subsequent praxis has
shown that to be erroneous.

13...d6

More ambitious than 13...d5 14.Bxe5 dxe4 15.Qd4 (or 15.Nxe4 Qxe5) 15...Qxe5 16.Qxe5 Rxe5
17.Rd8+ Re8 18.Rxe8+ Nxe8 19.Nxe4 with equality (Obodchuk).

Here White has tried:

a) 14.a3 Be6 15.Ne2 ½-½ B.Szuk-Pa.Petran, Hungarian League 2004, just as the game was getting
going.
b) 14.Qe3 Bxd4 (not really necessary; 14...Be6 15.a3 b5 is more enterprising) 15.Rxd4 d5 16.g5
Nd7 17.Qf4 Nb6 18.Bd3 Be6 19.exd5 Nxd5 20.Nxd5 and now 20...Bxd5 is still equal, whereas
20...cxd5? landed Black with a weak pawn and bad bishop for no compensation, D.Majer-J.Hamer,

249
corr. 2008.
c) 14.Kb1?! Be6 15.h4 Rad8 16.Qd2 c5?! (either 16...b5 17.Nd5 Qxd2 18.Nxf6+ Bxf6 19.Rxd2
Bxd4 20.Rxd4 d5 – Obodchuk; or 16...Bxd4 17.Qxd4 d5 18.e5 Nd7, intending ...c6-c5 or ...b7-b5, is
good for Black) 17.Bxe5 dxe5 18.Qe1 Rxd1+ 19.Qxd1 Rd8 20.Qc1 b5, T.Willemze-O.Romanishin,
Hoogeveen 2009, where Stockfish proclaims 21.Bxb5 Rb8 22.Qd1! Bxa2+ (or 22...a6 23.Qd6!)
23.Nxa2 Qxb5 24.Qc1 to be roughly equal.

C2: 7.f4

A very direct response. White gives naught for structure, preferring to prioritize space and the
bishop pair.

7...Bxc3+ 8.bxc3

Nunn’s rejection of 3...Bc5 is supported by 8...Nf6?! 9.e5 Ne4 10.Qd5! Qh4+? (10...f5 is forced)
11.g3 Nxg3 12.hxg3 Qxg3+ 13.Kd1 and White won, A.Martorelli-F.Bellia, Italian Ch., Chianciano
1985. So we shouldn’t play that.

8...d6

250
9.Bd3

The usual move, defending the e4-pawn and preparing to castle short with a secure advantage. So
Black needs to disrupt that plan a little.
Instead:
a) 9.Be2 Qh4+ 10.g3 Qh3 11.Bf1 transposes to the main line as White has nothing better: 11.Bf3
Nf6 12.e5 dxe5 13.fxe5 Nd7 14.Bg4 Qg2 15.Bf3 Qh3 is a draw; and so is 11.Ba3 Nf6 12.e5 0-0
13.exf6 Qg2 14.Kd2 Re8 15.Rg1 (or 15.Qf1 Rxe2+ 16.Qxe2 Qd5+ 17.Qd3 Qg2+) 15...Qe4 16.Qe1
Qa4 17.Qc1 (not 17.Bb2?? Bf5 18.Rc1 Qxa2 19.Ba1 Qd5+ and wins) 17...Qe4 18.Qe1 Qa4 etc.
b) 9.Bc4 Qh4+ 10.g3 Qh3?, however, does not transpose because of 11.Qd5! Nh6 12.f5! 0-0 and
now, rather than 13.Bf1 Qh5 14.Be2 Qh3 15.f6 Ng4 16.Bf1 c6! 17.Qg5 Qh6 18.Qxh6 gxh6 19.Bg2
Nxf6 20.0-0 Ng4 21.h3 Ne5 22.Bxh6 ½-½ alexkhesin-tsmenace, ChessWorld.net 2021, simply
13.Bf4! sets serious problems for Black whose queen is shut out of the game; e.g. 13...Ng4 14.0-0-0
Nf2 (or 14...Nf6 15.Qd4 Re8 16.Rhe1 Bd7 17.Kb2! Bc6 18.e5 dxe5 19.Rxe5) 15.Rhe1 Nxd1
16.Qxd1 Re8 17.Bf1 Qxh2 18.Bb5 Re7 19.Rh1 Qf2 20.Bd3 Qc5 21.Qh5 h6 22.Bxh6 Qxc3 23.f6
Qa1+ 24.Kd2 Qxf6 25.Bg5 Qf2+ 26.Kc1 f6 27.Bxf6! Qxf6 28.Qh8+ Kf7 29.Rf1 with a clear
advantage. The old adage “long line, wrong line” no longer holds with a modern engine threshing off
the chaff.
So the queen must retreat and 10...Qf6 is probably the best option (if 10...Qe7 11.0-0 Be6,
D.Thielen-C.Schlingensiepen, Goch 1994, then 12.Bb5+!? Bd7 13.Bxd7+ Qxd7 14.Qd4 Nf6 15.e5
looks like an edge for White). 10...Qf6

251
 For example:
b1) 11.0-0 Qxc3 12.Qd5 Qxa1 13.Qxf7+ Kd8 14.Qf8+ Kd7 15.Qf5+ is a draw.
b2) 11.Qd3 Bh3 12.Rg1 (if 12.Be3 Qe7 13.Bd4 then 13...f5! equalizes already, E.Kapatos-
Art.Kovacs, corr. 2015) 12...h5 13.Ba3 0-0-0 14.0-0-0 Bg4 15.Rd2 Nh6 (or 15...Ne7) 16.Bd5 Kb8
17.Kb1 Bc8 18.Ka1 Rhe8 saw White unable to break in, Laser-Ethereal 11, CCC 2 Blitz Battle 2018.
b3) 11.Qd4 Bh3 (or 11...Ne7 12.Qxf6 gxf6 13.e5?! fxe5 14.fxe5 dxe5 15.0-0 Be6 16.Bxe6 fxe6
17.Bg5 Rg8 18.Bf6 e4 19.Rae1 Rg4 and Black is not worse, M.Lagarde-E.Bacrot, French Ch., Pau
2012) 12.Kf2 (or 12.Bb5+ c6 13.Bd3 Qxd4 14.cxd4 d5) 12...Nh6 13.Be2 Ng4+ 14.Bxg4 Bxg4
15.Qxf6 gxf6 16.f5 Bh3! 17.Rg1 0-0-0 18.Bh6 Rhg8 with equality, since the black bishop cannot be
trapped, Laser--Fire 7, CCC 2 Blitz Battle 2018.
c) 9.Qd4 (“mit klarem Raumübergewicht für Weiß” – Keres) is the old refutation of 3...Bc5, based
on 9...Qh4+?! (pointless before the f1-bishop has moved) 10.g3 Qe7 11.Bg2 and 0-0 with advantage,
L.Paulsen-A.Anderssen, 3rd matchgame, Leipzig 1877.
134 years later, IM Obodchuk suggested a solution which “suffices for equality”: 9...Nf6 10.e5 c5!
(the surprising key)

252
 11.Bb5+ (if 11.Qxd6 Qxd6 12.exd6 then 12...Kd8! and ...Re8 is good) 11...Bd7 12.Qd3 dxe5
13.fxe5 0-0 14.0-0 (not 14.exf6? Re8+ 15.Kf2 Qxf6+ and wins) 14...Bxb5 15.Qxb5 Nd5 16.Qxc5 (or
16.Qxb7 Nxc3 17.Be3 Qb6 – Obodchuk) 16...Rc8 17.Qxa7 Nxc3 18.Qe3 Qd4!? 19.Qxd4 Ne2+
20.Kh1 Nxd4 21.Rb1 Rxc2 and draws. Even after 11.Qf2 Qe7 12.Ba3 0-0 13.0-0-0 dxe5 14.Qxc5 (or
14.Bxc5 Ne4) 14...Qxc5 15.Bxc5 Bg4 16.Bxf8 Bxd1 17.fxe5 Rxf8 18.exf6 Bg4 19.fxg7 Rc8, White’s
advantage is minimal.
Note that Black can play the same way against other moves; e.g. 10.Bb2 0-0 11.0-0-0 c5!; or
10.Bd3 c5! 11.Qe3 0-0 12.0-0 Re8 13.Qf3, J.Abdalla-H.Van Riemsdijk, Americana 2013, and now
13...Bg4! 14.Qg3 c4, intending 15.Bxc4 Nxe4 16.Qxg4? Qb6+ etc.
d) 9.Qf3 didn’t give White anything after 9...Nf6 10.Bd3 Bd7 11.0-0 Bc6 12.Bb2 Qe7 13.Rae1 0-0-
0 14.Qe3 Kb8 15.c4 Rhe8 in G.Djurovic-S.Kozarcanin, Velika Gorica 2006, but 12.Re1 improves a
little, so reversing Black’s moves might be more precise; i.e. 9...Bd7 10.Qg3 (or 10.Bd3 Qh4+ 11.g3
Qh3) 10...Nf6. For example: 11.Bd3 (not 11.Qxg7? Rg8 12.Qh6 Qe7) 11...0-0 12.e5 (or 12.0-0 Bc6)
12...dxe5 13.fxe5 Qe7 14.0-0 Ng4 15.Bf4 Rae8 16.Rae1 f6! 17.exf6 Qc5+ 18.Kh1 Nf2+ 19.Kg1
Ng4+ with a draw.

9...Qh4+!

Inducing a weakness.

10.g3 Qh3

253
 And then sitting on it.

11.Bf1

Driving the intruder away. If instead 11.Qf3?!, as in R.Feist-He.Luther, Bechhofen 1994, then
11...Bg4 12.Qf2 Nf6 is more than okay for Black, since 13.f5? (which would be an issue after
12...Ne7) now fails to 13...0-0 14.Bf1 Qh5 15.h3 Rae8 and so on.

11...Qe6 12.Bg2 Bd7

Bringing the bishop quickly to the long diagonal and preparing to castle long. I prefer this as being
more active than castling short with 12...Nf6 13.0-0 0-0, D.Pedzich-O.Romanishin, Koszalin 1997; or
12...Ne7 13.0-0 0-0, N.Bujukliev-D.Dimitrievski, Stip 2002. In particular, Black can now apply a lot
of pressure on the white e-pawn.

13.0-0 Bc6 14.Re1

Defending the pawn for the moment. If 14.f5 Qe5 15.Qd4 Nf6 16.Bf4 Qxd4+ 17.cxd4 Bxe4
18.Rfe1 d5 19.Bg5 0-0-0 20.Bxf6 gxf6 21.Bxe4 dxe4 22.Rxe4 Rd5, Black is at least equal in the
double rook endgame, S.Shapovalov-G.Obukhov, corr. 2015.

14...0-0-0 15.Qd4

Forking a7 and g7. Fortunately, the former doesn’t actually matter.

15...Nf6!

254
 Attacking e4 again and with ...Re8 (either rook) to follow. White can only defend the pawn by
making a big concession.
a) 16.Qxa7?! Nxe4 17.Bxe4? (since White’s idea doesn’t work, this just leaves the light squares
fatally weak) 17...Bxe4 18.Rb1 (threatening Rb4 or Rxb7; here 18.Qa8+ Kd7 19.Qa4+ Bc6 20.Rxe6
Bxa4 21.Re2 Rde8 is very good for Black) 18...Qg4! 19.Qf2 (if 19.Rxe4 then 19...Qd1+ and ...Qxc2+
picks up a rook) 19...Rde8 20.Rb4 Bc6 21.Be3 Qf5 22.Rd4 Re7 23.a4 Rhe8 24.c4 h5! 25.h4 Qh3 and
wins, Grandpa_Steve-tsmenace, ChessWorld.net 2019.
b) 16.a4?! Rde8 17.Qxa7 Nxe4 18.a5 f5?! (Stockfish suggests 18...h5! as good for Black) 19.Qd4?!
(and here 19.Be3 Kd7 20.Qd4 as roughly equal) 19...Bb5?! (again 19...h5!) 20.Qa7 h5 (finally)
21.Rb1 Ba6 22.Rb4! Kd7 23.Bxe4 fxe4 24.Qe3 d5 25.c4! Bxc4 26.Rxb7 Qc6 27.Ra7 h4 28.f5 Kc8
29.c3 led to a draw in tripoduk-jatait47, Chess.com 2017, as each side completely controls a colour
complex.
c) 16.f5! (releasing the c1-bishop even at the cost of a huge hole) 16...Qe5

255
 17.Bf4 (exchanging queens would make things much worse, while after 17.Bg5?! Rhe8 18.Rad1 h6
19.Bxf6 Qxf6 20.Qxa7, B.Bogoevski-R.Navarro Segura, Merida 2001, and 20...Re5 Black is clearly
better despite the missing pawn) 17...Qa5 18.Bg5 (or 18.Bh6!? gxh6 19.Qxf6, which led to a draw in
Mis.Pap-Mo.Nikolov, Hofheim 2014) 18...Qe5 (or 18...Rhe8!? 19.Bxf6 gxf6 20.Qxf6 Rd7 and Black
is no worse nor no better) 19.Qxa7 (otherwise 19.Bf4 Qa5 repeats) 19...Rde8 and the position is equal
(“0.00”) as White’s extra pawn doesn’t signify.

C3: 7.Bd3

This quiet-looking move may well be best because it’s flexible. White develops a piece,
overprotects the e-pawn, prepares to castle, and doesn’t rule anything out. In particular, White can

256
aim to achieve f2-f4 in a more favourable situation, perhaps after Bg5, or Ne2 or Nb5, or simply
against an inferior response from Black.

7...Nf6

Black replies in like manner, keeping options with the d-pawn.

This was GM Romanishin’s second try, after his sharper choice got smashed: 7...Qh4?! (preventing
both 0-0 and f2-f4, but only temporarily) 8.Nd5! Ne7?! (but 8...c6 9.g3 Qd8 10.f4 and 8...Nf6 9.g3
Qh3 10.Bf4 are good for White too) 9.g3 Qh3 10.Nxe7 Kxe7 11.f4! Qg2 (or 11...Bd6 12.Qf3) 12.Rf1
Bd4 13.Qh5 d5 14.Bd2 Rd8 15.f5 Bxb2 16.Rd1 dxe4 17.Bc4 1-0 I.Glek-O.Romanishin, Biel 1996.
Other moves show the efficacy of White’s set-up: 7...Bxc3+ 8.bxc3 d6 9.0-0 and f2-f4 is what
White hoped to achieve in line C2; while 7...Ne7 8.Bd2 0-0 9.Qh5 Ng6 10.g3 c6 11.f4 Bd4 12.0-0-0,
N.V.Pedersen-S.Drazic, Skanderborg 2010; and 7...d6 8.0-0 Nf6 9.Ne2 Nd7 10.f4 Bf6 11.c3, Houdini
6-Ethereal 11, CCC 2 Blitz Battle 2018, are similarly passive.

8.0-0

Instead:
a) 8.f4 can now be met by 8...Bxc3+ 9.bxc3 d5! 10.e5 Bg4 11.Qd2 Ne4, and if 12.Qe3 (not
12.Bxe4?! dxe4 13.Ba3 Qxd2+ 14.Kxd2 0-0-0+ 15.Ke3 h5 and Black is better, S.Kuipers-
O.Romanishin, Hoogeveen 2008) 12...Qh4+ 13.g3 Qh3 14.Bf1 Qh5 15.Bg2 0-0 16.Ba3 Rfe8 17.c4,
Sa.Robin-C.Mania, Chamalieres 2009, then 17...Nd6 18.cxd5 Nc4 19.Qc3 Nxe5! equalizes, since
White can’t take the knight: 20.fxe5?? Rxe5+ 21.Kf1 Qf5+ 22.Kg1 Re2 23.Bc5 Rxg2+! 24.Kxg2
Bh3+ 25.Kg1 Qxd5 26.Kf2 Re8 27.Be3 Qg2+ and wins.
b) 8.Bd2 also allows 8...d5!, as seen via 7.Bd2 earlier.

257
 c) 8.Bg5 asks for 8...h6 9.Bh4, when ...g7-g5 can be thrown in as required; e.g. 9...c6 10.0-0 (or
10.Qd2 Qc7, intending 11.0-0-0?? Bf4) 10...g5 11.Bg3 d6 with a solid set-up on the dark squares.
d) 8.Ne2 threatens f2-f4, so Black must make room for the bishop: 8...c6 9.f4 Bc7 10.e5 (otherwise
...d7-d5 would likely follow; e.g. 10.0-0 d5 11.e5 Ne4 12.Bxe4 dxe4 13.Qxd8+ Bxd8 14.Ng3 Bb6+
15.Kh1 0-0 16.Nxe4 Bf5 with compensation) 10...Nd5 11.Bd2 (if 11.c4 Nb4 12.Be4, J.Hector-
N.Mitkov, Metz 1992, then 12...d5 looks fine)

11...d6 12.c4 Ne7 13.Qc2 dxe5 14.0-0-0 Ba5!? (not 14...Bg4?! 15.Bc3 Qc8 16.fxe5 and White was
better in T.Shaked-O.Romanishin, Linares 1997) 15.Rhe1 (any of 15.fxe5 Ng6, or 15.Be1 Bxe1
16.Rhxe1 Qa5, or 15.b4 Bc7 16.Bc3 Qd7 17.fxe5 Qg4 or 16.Rhe1 a5 is quite unclear) 15...Bxd2+
16.Qxd2 Qc7 17.Qc3 (or 17.fxe5 0-0) 17...Be6 18.fxe5 Rd8 19.Nd4 0-0 20.Nxe6 ½-½ J.De Waard-
B.Fister, corr. 2014.
Another possibility is 11...Qh4+!? 12.g3 Qe7 13.c4 (if 13.Be4 then 13...d6, or 13.a3 d6 14.c4 Nb6
15.exd6 Bxd6 16.Qc2 Bg4) 13...Nb4 14.Be4 d5 15.cxd5 Nxd5 16.Qb3 a5 17.Bxd5 cxd5 18.Qxd5 (or
18.Nc3 Be6 19.Qb5+ Qd7) 18...0-0 with compensation on the light squares.
e) 8.Nb5!? has the same intention, and 8...a6 9.f4 Bxf4 (or 9...Bxb2 10.Nxc7+ Qxc7 11.Bxb2 Qxf4
12.Qf3) 10.Bxf4 axb5 11.0-0 d5 (or if 11...Qe7, Y.Khegay-A.Malaikanova, Cholpon Ata 2019, then
12.Qf3 d6 13.e5) 12.Bg5 (or 12.exd5 Qxd5 13.Qe1+ Be6 14.Qb4) 12...dxe4 13.Bxb5+ Bd7 14.Bxf6
gxf6 15.Qxd7+ Qxd7 16.Bxd7+ Kxd7 17.Rxf6 gives White a pleasant endgame due to Black’s
weaker pawns (K&S). Although Stockfish considers 17...Rhf8 to be close to equal, it wouldn’t be fun
for a human to defend, while its suggested alternative of 13...Ke7!? just looks insane.
So Black might prefer 8...c6 9.f4 Bb8, when 10.e5 cxb5 11.exf6 Qxf6 12.Qh5 (or 12.0-0 0-0)
12...h6 13.0-0 0-0 transposes to 9.Nb5 in the main line.

8...0-0

258
9.f4

The main line based on early praxis, but White has more appealing options:
a) 9.Bg5 “is strong now, as ...h7-h6 and ...g7-g5 is unattractive for Black. White is comfortably
better here” (IM Christof Sielecki). For example: 9...c6 10.Bh4!? (or 10.f4 Bxc3 11.bxc3 d5 12.e5
Qb6+ 13.Kh1 Ne4, J.Kovar-Gü.Schulz, corr. 2003, and 14.Qh5!?) 10...Qb6 11.Bxf6 Bxf6 12.e5 Bd8
13.Ne4 Qxb2 14.c3! d5 15.exd6 – K&S.
However, 9...Qe8!?, breaking the pin, doesn’t look so bad;

e.g. 10.Bh4!? (not 10.f4? Bxc3 11.bxc3 Nxe4, or 10.Nd5?! Nxd5 11.exd5 f6; while 10.Qe1 Qe6

259
11.f4 Bxc3 12.bxc3 d5 13.Bxf6 Qxf6 14.e5 Qb6+ 15.Kh1 f5 16.h4 Be6 led to a draw in Stoofvlees-
Mystery, Chess.com Engine Tuning 2020) 10...Bxc3 11.bxc3 Qe6 12.Qe2 (or 12.Re1 d6) 12...Re8
13.c4 d6 14.Rae1 Qg4 15.Qxg4 (or 15.Bxf6 Qxe2 16.Rxe2 gxf6 17.f4 Bd7) 15...Bxg4 16.f4 Be6
17.h3 h6 18.g4 b6 19.Re2 Rab8 20.Bg3 Bc8! 21.e5 dxe5 22.Rxe5 Bb7 23.Rfe1 Kf8 and Black was
fine, lysol-tsmenace, ChessWorld.net 2019.
b) 9.Nb5 has the usual threat, so 9...c6 (if 9...d5 10.f4 Bg4 11.Be2 Bxe2 12.Qxe2 Bd6 13.e5 Bc5+
14.Be3, “White still maintains the initiative” – K&S; and 11.Qd2!? may be even stronger) 10.f4 Bb8
11.e5 (not 11.Nc3?! d5 12.e5 Ng4 13.Qf3 a6! 14.Qg3 f6 15.Qh4, M.Stankovic-B.Lajthajm, Bosnian
Senior Team Ch. 2015, when 15...Nh6! is at least equal) 11...cxb5 12.exf6 Qxf6 13.Qh5 h6 (rather
than 13...g6?! 14.Qh6 Qg7 15.f5 – K&S, which looks much better for White)

14.Bd2 (either 14.Bxb5 a6 15.Bd3 Ba7+ 16.Kh1 d6, or 14.Qxb5 a6 15.Qd5 Qe6 seems okay for
Black) 14...d5 15.Qxd5 (if 15.g4 Bc7 16.g5 then 16...Qxb2 and 17...g6; or 15.Rae1 Bd7 16.g4 Bc7
17.g5 Qc6 18.f5 Bxf5 19.Bxf5 g6) 15...a6 16.Kh1 Qc6 17.Qh5 Be6 18.Rf3 (or 18.Bc3 Bc4 19.Qg4
f6) 18...Bc4 19.Rg3 (threatening Bc3; Black has only one defence) 19...Bxd3 20.cxd3 b4! 21.Bxb4
Bxf4 22.Rg4 g5! 23.Rxf4 (or 23.Bxf8 Rxf8 24.Rf1 f5) 23...gxf4 24.Bxf8 Rxf8 25.Qg4+ Qg6 26.Qxf4
Rc8 with sufficient activity to hold the endgame: 27.Rf1 Rc2 28.Qb8+ Kg7 29.Qxb7 Qf5! 30.Re1
Qxd3 31.h3 ½-½ ST_KAOURAS-tsmenace, ChessWorld.net 2020.
c) 9.Ne2 c6 10.f4, A.Guzman-J.P.Casanovas, Europe-Echecs (blitz) 2021, can be answered by
10...Bc7 11.e5 Ng4 12.Qe1 d5, intending ...f7-f6; e.g. 13.Kh1!? (or 13.h3 Nh6 14.g4 f6) 13...Re8
14.Bd2 f6 15.Qh4 Nh6 16.exf6 (or 16.f5 Bxe5 17.Bxh6 gxh6 18.Qxh6 Re7) 16...Bf5 17.Bxf5 Nxf5
18.f7+ Kxf7 19.Qxh7 Qf6 with compensation.
d) 9.h3 Re8 10.Re1 (or 10.Nb5 c6 11.f4 Bb8 12.Nc3 d5 13.e5 Nd7) 10...c6 11.f4 Bd4+ went very
wrong for White: 12.Kh2? (12.Kf1 is still equal) 12...d5! 13.exd5?? Ng4+! 0-1 D.Leroy-B.Lajthajm,
Novi Sad 2015, since 14.hxg4 Qh4 is mate.

260
9...Bd4+

The direct 9...Bxc3 10.bxc3 d5 is very logical, and 11.e5 Ne4 12.f5 (Obodchuk’s 12.c4 looks
better) 12...Qe7 13.Bxe4 (or 13.f6 Qxe5 14.fxg7 Re8) 13...dxe4 14.Qd4 Re8 15.e6 fxe6 16.f6 gxf6
17.Rxf6 Rd8! 18.Qe5 Rd1+ 19.Rf1 (not 19.Kf2?! Rd5) 19...Rxf1+ 20.Kxf1 Qg7 21.Qxe4 Bd7 was
soon drawn in D.Sermek-O.Romanishin, Bled 1998. On the other hand, given the option, Black might
just as well keep the bishop.

10.Kh1 d6 11.Qf3

And from my own praxis: 11.Ne2 Bb6 12.Qe1 Nd7? (going after the d3-bishop is very much the
wrong plan; simply 12...Re8 13.Ng3 Ng4 14.h3 Qh4 or 13.Qh4 d5 14.e5 Ne4 is equal) 13.Ng3 (here
13.f5! Ne5 14.f6 was more to the point) 13...Nc5 14.Be3 (still 14.f5!) 14...Nxd3 15.cxd3 Bd7 16.f5 f6
17.Nh5 Qe8 18.Nf4 Qe5 19.Bxb6 axb6 20.Qf2 Rf7 21.Rfe1 Qb5 22.Re3 Qc5 23.Qe2 c6 24.a3 Re8
25.Re1 Rfe7 26.Qh5? (26.g4) 26...Qe5 27.Qh4 Qxb2 and Black was clearly better, J.Willow-J.Tait,
Notts League 2019. (Jonah won this one.)

11...Re8 12.Bd2 Bd7 13.Rae1

Here IM Obodchuk writes: “White’s position looks very pretty, but (...) it is not clear why Black
should stand worse.” Quite so.
For example: 13...Bc6! (not 13...h5?! 14.Qg3 c5, V.Lilov-M.Vasilev, Bulgarian Ch., Plovdiv 2008,
since 15.e5! is good for White) 14.Qg3 g6 (or 14...Nh5) 15.f5 Bxc3 16.Bxc3 Nxe4 17.Rxe4 Rxe4
(not 17...Bxe4?? 18.fxg6 fxg6, D.Campora-N.Mitkov, Elgoibar 1995, when 19.Qf4! wins at once)
18.Bxe4 Bxe4 19.Qe3 gxf5! 20.Qg3+ (not 20.Qh6?! f6 21.Bxf6 Qf8) 20...Kf8 21.Qg7+ Ke8 22.Bf6
Qc8 23.Qg8+ Kd7 24.Qxf7+ Kc6 25.Qc4+ with a draw.
which is either a Latvian Gambit (1.e4 e5 2.Nf3 f5 3.d3 Nc6)

261
 or a Lisitsin Gambit Deferred (1.Nf3 f5 2.d3 Nc6 3.e4 e5) – unless White plays 4.d4!? in which
case it’s a reversed Vienna, where Black, now having the first move, is obviously fine;
In truth I’ve generally reached this position via entirely different routes, such as 1.Nf3 Nc6!? 2.d4
d5 3.c4 Bg4 (this is another Chigorin) 4.e3 e5!? 5.cxd5 Qxd5 6.Nc3 Bb4 7.Be2 exd4 8.exd4. Yes, it
really is the same thing.

262
 Chapter Seven
Scotch Game

1.e4 e5 2.Nf3 Nc6 3.d4 exd4 4.Nxd4 Qh4!?

I’ve never enjoyed facing the Scotch. Checking my results, I see I’ve only ever lost to it three times
over the board, but the games always feel uncomfortable, like playing away with my opponent as the
referee. I much prefer to play at home, even on a ramshackle non-league pitch, especially on that kind
of pitch. So how might something similar be arranged in the Scotch? With the first world champion
Wilhelm Steinitz’s favourite 4...Qh4 of course!
As the diagram shows, White’s fourth carelessly removed the knight from f3, allowing the black
queen to rush in and attack the e4-pawn, which is surprisingly difficult to defend. To derive anything
from the opening White is almost obliged to give the pawn up, obtaining in return a strong initiative.
The pertinent question then is: how strong is strong? We’ll see when we get to line C below. GM Lev
Gutman wrote a 270-page book entirely on 4...Qh4, so it does have some pedigree.
At this precise moment White has three main options:
A: 5.Qd3
B: 5.Nb5
C: 5.Nc3
Other moves are less testing:
a) 5.Nxc6? (played 216 times, including by two GMs) makes no sense to me. After 5...Qxe4+
6.Be2 (or 6.Be3 dxc6, or 6.Qe2 Qxe2+ 7.Bxe2 dxc6) 6...dxc6! 7.0-0 Be6 8.Nc3 (or 8.Re1 Qd5, or
8.Nd2 Qd4) 8...Qh4 9.g3 Qd8! (B.Zuckerman), the most White can hope for is that Black’s extra

263
doubled c-pawn won’t be enough to win.
b) 5.Be2? is an incorrect move order because of 5...Bc5! 6.Be3 (or 6.c3 Qxe4) 6...Bxd4 7.Bxd4
Qxe4 8.Bxg7 (otherwise 8.Be3 Qxg2 9.Bf3 Qh3 or 8.Bc3 Qxg2 9.Bf3 Qg5 sees White two pawns
down) 8...Qxg2 9.Bxh8 Qxh1+ 10.Bf1 Qe4+ 11.Qe2 Qxe2+ 12.Bxe2, A.Mühlbauer-W.Schwarm,
Bayern 1999, when 12...d6, followed by ...Be6 and ...0-0-0, leaves White with very little for the
pawn.
c) 5.Nf5?! wastes time to no real purpose.

After 5...Qxe4+ 6.Ne3 Nf6 (or 6...Bb4+ 7.Nd2 Qe7!?) 7.Nc3 (or if 7.Bd3, L.Paulsen-J.Minckwitz,
Leipzig 1879, then 7...Qe7 and ...d7-d5) 7...Qe5 8.Bd3 d5 9.0-0 Be6, White would have to work hard
to show any compensation; e.g. 10.Re1 (or 10.f4 Qd4) 10...Qd6 11.Nf5 Qd7 12.Bb5 a6 13.Ba4 b5
14.Nxb5 axb5 15.Bxb5 Rb8 16.a4 Ne4 17.Nd4 Rxb5 18.axb5 Nd8 and even here Black is better with
two pieces for rook and pawn.
d) 5.Be3 can be answered by 5...Nf6! (ignoring the e4-pawn for the moment, as after 5...Qxe4
6.Nc3 Bb4 7.Nb5, and if 7...Bxc3+ 8.Nxc3 Qe5 then 9.Bd3!, Black will quickly have issues on the e-
file; e.g. 9...Nf6 10.0-0 0-0 11.Re1 d6 12.Bxa7 and White regains the pawn anyway with an edge)

264
 Now ...Ng4 becomes an option as well as taking the e-pawn:
d1) 6.Nd2 Ng4 7.Nf5 Nxe3 8.Nxe3 Bc5 and if 9.Nf3? Qxe4 10.Bd3, D.Kononenko-V.Stoockalov,
Alushta 2005, then 10...Qb4+ 11.c3 Qxb2 12.0-0 Bxe3 should win.
d2) 6.Bd3 Ng4 7.Nf5 Nxe3 8.Nxe3 Bc5 and if 9.Qd2, E.Cosma-M.Duminica, Bucharest 1991, then
9...Ne5 10.Nc3 c6 11.0-0-0 Nxd3+ 12.cxd3 0-0 and Black is not worse.
d3) 6.Be2, as in A.Aaberg-J.Hector, Copenhagen 1991, prevents ...Ng4, but instead 6...Nxd4
7.Bxd4 Nxe4 8.0-0 d5 is fine for Black.
d4) 6.Nc3 Bb4 7.Nb5 (there is nothing better) 7...Ba5! can arise via multiple routes. Then 8.Bd3? is
a mistake in view of 8...a6! 9.Nd4 (or 9.g3 Qh3 10.Nd4 d5! – Gutman) 9...Nxe4 10.Bxe4 Qxe4 11.0-0
Bxc3 12.bxc3 0-0 and Black is clearly better, Wo.Schmid-M.Holzhaeuer, German League 1990.
Despite the tempo loss, 8.Bd2 is relatively best, as in W.Boschmann-M.Steinhart, Ditzingen 2015;
e.g. 8...a6 9.g3 Qg4 10.Qxg4 Nxg4 11.Na3, or 8...Nxe4 9.Qe2 Kd8 10.Nxe4 Bxd2+ 11.Kxd2 Re8
12.Nbc3 f5 13.Re1 fxe4 14.Kc1, when White should regain the e4-pawn and equalize.
d5) 6.Nb5 is also met by 6...Bb4+ and ...Ba5. The only independent line is 7.Nd2 (instead, 7.N1c3
Ba5 is ‘d4’ above; while for 7.c3 Ba5 see 6.c3 in line B) 7...Ba5 (threatening ...Nxe4) 8.g3! (note that
8.c3 is met by 8...Ng4!, rather than 8...a6) 8...Qxe4 9.Nxc7+ Kd8 10.c3 Qxh1 11.Nxa8, MimiSiku-
Melton, playchess.de 2005, when 11...Qxh2 is equal; e.g. 12.Nf3 (if 12.Nc4 then 12...Re8!, or 12.Qf3
Ne5! 13.Qf4 Qh5) 12...Qh5 13.Ng5 Qxd1+ 14.Rxd1 Ke7 15.b4 h6! 16.Nxf7 Kxf7 17.bxa5 d5.
e) 5.Nf3!? offers no more than reasonable compensation for the pawn: 5...Qxe4+ 6.Be2 (not
6.Be3?! Nb4! 7.Na3? Nf6 8.Ng5 Qe7 9.Nb5 Nfd5 10.a3 Nxe3 11.fxe3 Na6 12.h4? h6 13.Nf3 c6
14.Nbd4 Qxe3+ 15.Be2 d5 and Black won, A.C.Vasquez-W.Steinitz, 5th matchgame, Havana 1888;
even 7.Bd3 Nxd3+ 8.cxd3 Bb4+ 9.Nc3 Bxc3+ 10.bxc3 Qg6 11.0-0 Nf6 leaves White struggling to
justify the pawn sacrifice) 6...Bb4+ (I like this patzer’s check; 6...Qe7!? is another sound option, first
played in I.Orchard-E.E.Gilbert, corr. 1875, where Black intends ...Qd8, ...Be7, ...Nf6, ...0-0, ...d7-d6,

265
and to sit tight with the extra pawn) 7.c3 (anything else and Black can swap the bishop off; e.g. 7.Bd2
Nf6 8.0-0 Bxd2 9.Qxd2 0-0 10.Nc3 Qe7 11.Rfe1 d6 12.Bc4 Qd8, G.Kunert-J.Gelfenboim, German
League 1999; or 7.Nbd2 Bxd2+ 8.Bxd2 Nf6 9.0-0 0-0 10.Bc3 Nd5 11.Bd3 Nxc3 12.bxc3 Qf4,
Anishchenko-A.Vaisman, USSR 1963; or 7.Nc3 Bxc3+ 8.bxc3 Nf6 9.0-0 0-0 10.Bd3 Qg4 11.h3 Qh5
12.Rb1 a6, Em.Lasker-N.Manson, San Francisco simul 1902 – Black won at least two of those
games) 7...Be7! (having encouraged c2-c3, the bishop retreats to its safest post) 8.0-0 Nf6 and then:

e1) 9.Re1 Qd5 10.Nd4 Nxd4 11.cxd4 and now 11...Qa5!, followed by ...d7-d5, solves Black’s
problems. Not 11...0-0? 12.Bf3 Qd6 13.b3! Bd8 14.Ba3 Qb6 15.Bc5 Qa6 16.Bxf8 Kxf8 17.Nc3 and
White was winning in C.Golmayo Zupide-W.Steinitz, 7th matchgame, Havana 1883.
e2) 9.Bb5 Qd5 10.Qe2 is well met by 10...Nd8! and ...Ne6 (Freeborough & Ranken).
e3) 9.Nbd2 Qf5 10.Nh4 (or if 10.Re1 0-0 11.Nf1 then 11...d6 12.Ng3 Qd7) 10...Qe6 11.Re1 d5
12.Ndf3 Qd6 was good for Black at this point in M.Vivo-A.Gori, Milan 1992.
e4) 9.c4 is logical, clearing the c3-square again. No one would likely answer 6...Nf6 7.0-0 Be7 with
8.c4, but it’s not a bad move. Here White developed a very strong initiative after 9...d6 (Black might
prefer 9...0-0 first, when 10.Nc3 Qf5 11.Nd5?! allows 11...Nxd5 12.cxd5 Nb4 with advantage, while
11.Nh4 Qe5 12.Nf3 Qf5 just repeats) 10.Nc3 Qf5 11.Nd5 Bd8? (11...Qd7 was correct and roughly
equal) 12.Bd3 Qh5 13.Re1+ Kf8 14.Nf4 Qg4 15.h3 Qd7 16.Bc2 and won quite quickly in H.Knoll-
G.Kreischer, Aschach 2009.
Well, if you send your queen out to grab a pawn early doors, you should accept some suffering in
consequence. So let that last game serve as a cautionary lesson: observe due measure; moderation is
best in all things.

A: 5.Qd3

Given that the f2-pawn is pinned, 5.Nc3 would get pinned, and 5.Nd2 or 5.Bd3 or 5.Qf3 blunders a

266
piece, this is the only way to defend the e-pawn.

5...Nf6

Now White has to consider the problem anew and the best they can hope for is equality.

6.Nc3

Allowing the pin on the knight after all.

Instead:
a) 6.Nd2? avoids the pin but runs into 6...Ng4! 7.g3 (not 7.N2f3? Qxf2+ 8.Kd1 Nxd4 9.Nxd4 Qf6)
7...Qf6 8.f3 (not 8.N2f3? Nxd4 9.Nxd4 Nxf2, nor 8.N4f3? Nb4 9.Qb3 Qb6 – Gutman) 8...Nb4 (or
8...Qxd4 9.fxg4 Nb4) 9.Qb3 (not 9.Qc3? Qxd4 10.Qxd4 Nxc2+ 11.Kd1 Nxd4 12.fxg4 d5, D.Aldama
Degurnay-B.Lugo, Matanzas 1992) 9...Qxd4 10.fxg4 Bc5 and Black has a huge advantage,
M.Stankovic-A.Dimovska, Yugoslav League 2000. If White wants to play Nd2 it is necessary to
exchange on c6 first:
b) 6.Nxc6 dxc6 7.Nd2 (alternatively, 7.Nc3 Bb4 transposes to the main line; not 7.e5? due to
7...Ng4! again, intending 8.g3 Qh5, or 8.Qg3 Qxg3 9.hxg3 Bf5, or 8.Be3 Nxe3 9.Qxe3 Qb4+) 7...Bc5
(7...Ng4 is answered the same way; or if 7...Bd6 then 8.g3 Qh5 9.Qe2! Bg4 10.f3) 8.Qg3! (not 8.g3?!
Qh5 9.Nb3? Bg4 10.Nxc5 Rd8 and Black is winning, P.Journoud-J.Arnous de Riviere, 12th
matchgame, Paris 1859; in this line 9.Be2 is relatively best, e.g. 9...Bg4 10.f3 Bh3 11.Nb3 Bb6
12.Be3 Rd8 13.Qc3 Bg2 14.Bxb6 axb6 15.Rg1 Qxh2 16.Qe3 Qxg3+ 17.Qf2 Qxf2+ 18.Kxf2 Bh3
19.Rxg7 h5, when Black has an extra passed h-pawn as well as the box of pawns) 8...Qxg3 9.hxg3
(not 9.fxg3? 0-0 10.Be2 Re8 11.Bf3 Bf5 and Black won, mizuzul-tsmenace, ChessWorld.net 2019)
9...Be6 and the game is quite equal; e.g. 10.Be2 0-0-0 11.f3 (not 11.f4? h5! 12.e5 Ng4 13.Ne4 Bb6
and Black is clearly better, as I’ve had reversed via 2.Nc3 Nf6 3.a3 d5 etc) 11...Bd6 12.Nf1, followed

267
by Bf4 or Be3.
c) 6.Nb5 is properly met by 6...Bb4+ (rather than 6...Kd8?! 7.N1c3 Bb4 8.a3! or 7...Nb4 8.Qe2 Bc5
9.g3 Qh5 10.Qxh5 Nxh5 11.Na3, when White has the edge if anything) 7.c3 (for 7.N1c3 or 7.Nd2
Ba5 8.g3 Qxe4+ see 7.Nb5 in the main line; not 7.Bd2?! in view of 7...Nxe4! 8.g3 Qe7 9.Nxc7+ Kd8
10.Nd5 Qe5 11.Nxb4 Nxd2+ 12.Kxd2 Nxb4 and Black is better, e.g. 13.Qb3 Nxc2! 14.Kxc2 Qe4+
15.Bd3 Qxh1, S.Kota-S.Horka, corr. 2002) 7...Ba5 8.b4?! (here 8.Nd2 d6 9.Qg3 keeps it level) 8...a6
9.N5a3? Bb6 10.Be3 Bxe3 11.Qxe3 Qxe4 and Black is close to winning, D.Pibernik-S.Neelmeier,
corr. 2015.

6...Bb4

If you want to avoid note ‘c’ below, there is always 6...Ng4!? (G.Palm), although 7.g3 Qf6 8.Nf5
Bb4 9.Bh3 Nge5 10.Qe2 is no more than equal; e.g. 10...g6 (or if 10...Bxc3+ 11.bxc3 d6, L.Nemeth-
M.Kossarski, corr. 2004, then 12.f4! g6 13.0-0) 11.f4 gxf5 12.fxe5 Qxe5 13.0-0 Bxc3 14.bxc3 d5.

7.Nxc6

Necessary preparation for Bd2.

a) 7.Bd2? at once is a mistake because the bishop is there to be taken after 7...Bxc3 8.Bxc3 Qxe4+
9.Qxe4+ Nxe4 10.Nxc6 Nxc3!, when even regaining the pawn with 11.Nxa7 Rxa7 12.bxc3 gives
Black a very good endgame (dating back to Bilguer 1912).
b) 7.Nb5 is answered by 7...Ba5 again, when the e4-pawn remains an issue for White: 8.g3 (the
only move; not 8.Bd2?? Nb4 and White can’t defend c2, D.Boros-P.Ponkratov, Titled Tuesday blitz
2020; and 8.Qe2?! 0-0 9.Bd2 a6 10.g3 Qg4 11.Qxg4 Nxg4 12.Na3 b5 looks dreadful too, A.Sustarsic-
R.Lorbek, corr. 1992) 8...Qxe4+ 9.Qxe4+ Nxe4 10.Nxc7+ Bxc7 11.Nxe4 and Black is certainly not
worse; e.g. 11...0-0 12.Nc3 Nb4 13.Kd1 d5, R.Rodriguez Perez-M.Vujadinovic, corr. 2000.

268
 c) 7.Nf5! has been played 20 times with 17 wins for Black. Nevertheless, it’s the best move:
7...Qxe4+ 8.Qxe4+ Nxe4 9.Nxg7+! Kf8 (not 9...Kd8?, G.Borgardts-G.Kieninger, Cologne 1933, due
to 10.a3! – Gutman; Black can try 9...Ke7, but 10.a3! Bxc3+ 11.bxc3 d5 12.f3 Nxc3 13.Nh5 or
10...Ba5 11.b4 Nxc3 12.bxa5 d5 13.Nh5 is still only equal) 10.Bh6! Nxc3 (not 10...Rg8? 11.Ne6+
Ke7 12.Nxc7 Nxc3 13.Nxa8 Ne4+ 14.c3 – Gutman) 11.Nh5+ Ke7 (not 11...Ke8 12.Bg7! Nd5+?
13.c3 Rg8 14.Rd1 Nxc3 15.bxc3 Be7 16.Bd3 and White is better, I.Cavatorta-R.Antoniacci, EU Ch.,
Arvier 2007) 12.Bg7! Ne4+ (not 12...Rg8?! 13.Bf6+ Kf8 14.bxc3 Be7 15.Bxe7+ Kxe7 16.0-0-0 d6
17.Re1+ Be6 18.f4 with the initiative) 13.c3 Bc5! (and not 13...Re8? 14.cxb4 Nxb4 15.f3! with a
clear advantage) 14.Bxh8 Nxf2 15.Rg1 (inserting 15.b4 Bb6 or 15.Bf6+ Kf8 makes no difference)
15...Ng4 16.Rh1 Nf2 is a draw by repetition.

7...dxc6 8.Bd2

The alternative 8.e5?! is bad because of 8...Ng4! 9.Qe2 (not 9.g3? Qe7 10.f4 Qc5! 11.Qf3 Bxc3+
12.bxc3 Be6 and Black is winning, J.McAloon-Z.Markov, Segovia 1990; and 9.Qg3 doesn’t work
here due to 9...Qxg3 10.hxg3 Bf5 11.Bf4 0-0-0) 9...Nxh2 (threatening ...Nf3+) 10.g4 (the only try;
not 10.g3? Qd4! and wins, P.Horatschek-P.Leisebein, corr. 2002) 10...Bxg4 (Stockfish is keen on
10...g5!? for some reason) 11.Rxh2 Qxh2 12.Qxg4 Qxe5+ 13.Qe2 Qxe2+ 14.Kxe2 0-0-0 and Black is
at least slightly better with rook and three pawns for the two pieces.

8...Bxc3 9.Bxc3 Nxe4!

Forcing White to play accurately in order to equalize. Not 9...Qxe4+? 10.Qxe4+ Nxe4 11.Bxg7 and
White is better with the two bishops and superior structure; while 9...0-0 10.e5 Re8 11.Be2 Nd5 12.g3
is just equal, R.Duff-J.Spaight, corr. 1886.

Here White has tried:

a) 10.g3? only helps Black: 10...Qe7 11.Bxg7? Rg8 12.0-0-0 Nxf2 13.Qd4 Nxd1 0-1 S.Gorts-

269
I.Werner, Charleroi 2008.
b) 10.0-0-0? allows 10...0-0 11.Qd4 Qh6+ 12.Kb1 Bf5 (or just 12...Nxc3+ 13.Qxc3 Be6) 13.Bb4
Rfe8 and Black is a pawn up, M.Pacholsky-J.Krajnak, Slovakian League 2006.
c) 10.Qe3 can be met by 10...Bf5! 11.Bxg7 (or 11.Bd3 0-0 12.Bxe4 Rfe8) 11...Rg8 12.Bd4
(Gutman prefers 12.0-0-0 Rxg7 13.f3 Qg5 14.Re1 Qxe3+ 15.Rxe3 Kf8 16.fxe4 Re8, P.Grott-
P.Leisebein, corr. 1999, followed by 17.Ra3 Bxe4 18.Rxa7 c5 19.Rg1 “and White is close to
equalizing”, though 19...Bc6 is not quite equal yet) 12...0-0-0 13.0-0-0 c5! 14.Bxc5, U.Vetter-
P.Leisebein, corr. 2001, and then 14...Rxd1+ 15.Kxd1 Qd8+ 16.Kc1 Qd5 17.Bd4 Rd8 with the
initiative; e.g. 18.f3 Nc5 19.Qe5 Qxd4 20.Qxf5+ Kb8 21.Bd3 Na4 22.Qb5 a6 23.Qb3 Rd6 24.Re1
Qf4+ 25.Kd1 Nc5 26.Qc3 Nxd3 27.cxd3 Rc6 28.Qd2 Qxh2 and most of those were “only moves” for
White.
d) 10.Qd4! forces 10...Qe7 (not now 10...Bf5? 11.Bd3 0-0?? 12.Qxg7 mate, while 11...Qf6 12.Qxf6
gxf6 13.Bxe4 Bxe4 14.0-0-0 is good for White despite the missing pawn and opposite-coloured
bishops) 11.0-0-0 Nxc3 (and not 11...Qg5+?? 12.f4 Qxf4+ 13.Bd2 Qg4? 14.Qd8+! Kxd8 15.Bg5+
Ke8 16.Rd8 mate, L.Maczuski-I.Kolisch, Paris 1864; the last mistake hardly signifies since 13...Qh4
14.Re1 0-0 15.Rxe4 or 14...f5 15.Qxg7 wins anyway) 12.Qxg7 (otherwise 12.Qxc3?! 0-0 keeps the
pawn safely) 12...Nxa2+ 13.Kb1 Rf8 14.Kxa2 Bd7 15.Qxh7 (not 15.Bc4?! 0-0-0 and White has
merely made the bishop a target, T.Brüske-P.Leisebein, corr. 2004) 15...0-0-0 16.Kb1, which is
“0.00” according to Stockfish. All the same, Black’s game looks easier to play; e.g. 16...Qf6 17.Be2
Be6 18.Qe4 Qxf2 19.Bg4 Rxd1+ 20.Rxd1 Re8 21.Bxe6+?! Rxe6 22.Qh7?! b6 and Black is now
clearly better with a safer king and strong queenside majority, D.Junghänel-R.Zajontz, corr. 1993.

B: 5.Nb5

Once considered White’s best move, this immediate knight sally is now regarded as less accurate
(than 7.Nb5 in line C) because it allows Black the possibility of defending the c7-pawn with the

270
bishop.

5...Bb4+!

I have never played 5...Bc5 and am not going to look at it here.

6.Bd2

The strongest way to block the check, preventing ...Ba5. Instead:

a) 6.N1c3 Ba5! transposes to 5.Nc3 Bb4 6.Nb5 in line C.
b) 6.Nd2? runs into 6...Qxe4+! (there’s no need for 6...Ba5 here) 7.Be2 (not 7.Qe2? Qxe2+ 8.Bxe2
Ba5 9.0-0 a6 10.Na3 Bb6 11.Nf3 d5, Blueaxe-tsmenace, ChessWorld.net 2019) 7...Qxg2! 8.Bf3 Qh3
9.Nxc7+ (or 9.Bg4 Qh4) 9...Kd8 10.Nxa8? (White has to play 10.c3 Kxc7 11.cxb4 Nf6 and is only a
pawn down) 10...Nf6 11.c3 (or if 11.a3 Re8+ 12.Be2 Qg2 13.Rf1 Nd4 14.axb4, A.Manukyan-
M.Carrabeo, corr. 2000, then 14...d5 intending ...Bg4 wins) 11...Re8+ 12.Be2 Ne5! 13.Qa4 Nd3+
14.Kd1 Nxf2+ 15.Kc2 (or 15.Ke1 Qd3 and mates) 15...Rxe2 16.Qxb4 Qd3+ 17.Kb3, R.Brajovic-
D.Perovic, Yugoslavia 1978, and now 17...Nxh1 wins (Gutman).
c) 6.c3 Ba5 (not 6...Qxe4+? 7.Be3 Ba5 8.Nd2 Qg6, as in S.Lalic-J.Tait, Hereford 1990, because of
9.Nc4 a6 10.Nd4 Nge7 11.h4! and White is clearly better)

271
 c1) 7.Be3 Nf6! (stronger than 7...a6?! 8.Nd4, intending 8...Qxe4 9.Nd2 and Nc4 – Gutman) 8.b4 (or
8.Nd2 Ng4! 9.Nf3 Nxe3 10.Nxh4 Nxd1 11.Rxd1 d6 12.Bc4 ½-½ K.Jergler-B.Besner, Oberbayern
1996, where Black would be justified in playing on) 8...a6 9.Nd6+! (not 9.Nd4? Bb6, or 9.N5a3? Ng4
10.g3 Nxe3 11.Qf3 Qh6 12.fxe3 Bb6 13.Nc4 Ba7 with a big advantage, R.Theissl Pokorna-
V.Zhidkov, Bratislava 1995) 9...cxd6 10.bxa5 and White has just about enough compensation in view
of the wreckage on the d-file.
c2) 7.Bd3 a6 8.N5a3 Bb6 9.g3 (not 9.0-0? Ne5 10.h3? d5 11.exd5 Bxh3! and Black won, J.Brychta-
P.Surnak, Plzen 1997; nor 9.Qe2?! Ne5 10.Be3 Bxe3 11.Qxe3, S.Macak-F.Tropp, Slovakian League
2002, due to 11...Qf4! with an edge) 9...Qh3 10.Bf1 Qe6 11.Bg2 is an inferior version of note ‘c3’
below, since White has not played Nd2 yet. Black can even consider 11...d5!? 12.Qxd5 Nf6
13.Qxe6+ Bxe6 with ideas of ...0-0-0, ...Ne5, ...Ng4.
c3) 7.Nd2 a6 8.Na3 Bb6 (securing the bishop before Nac4 comes) 9.g3 looks like the best try for
White, aiming for a useful space advantage; e.g. 9...Qe7 10.Bg2 Nf6 11.0-0 0-0 12.Ndc4 Ba7 13.Re1
(not yet 13.Be3?!, I.Radulov-J.Trapl, Decin 1976, because of 13...Nxe4! 14.Bxa7 Rxa7 15.Re1 f5,
intending 16.f3?! Qc5+) 13...d6 14.Be3 Bxe3 (or 14...Bb8 15.Qc2) 15.Nxe3 Re8 16.f4.
Black should therefore prefer 9...Qf6! 10.Qe2 (if 10.f4 then 10...Qe7! 11.Qe2 d6 12.Ndc4 Nf6!
13.Nxb6 cxb6 14.Bg2 0-0 15.0-0 Bf5, targeting the e4-pawn) 10...d6 11.Nac4 (or 11.Ndc4 Ba7
12.Be3 Bxe3 13.Nxe3 Nge7 14.Bg2 h5) 11...Ba7 12.Bg2 Nge7 13.0-0 h5! 14.Ne3?! (14.Nf3 Bg4 is
equal) 14...Ne5 15.h4? Bg4 16.f3 Bd7 17.Kh2 g5! with a huge attack, G.Harutjunyan-I.Kovalenko,
Alushta 2007.

6...Qxe4+

Of course! The exclamation mark is there because I actually played 6...Qd8? in Jo.Nelson-J.Tait,
Sheffield League 2009. I can’t explain that move, never mind the subsequent draw.

272
7.Be2

Not 7.Qe2? Qxe2+ 8.Bxe2 Kd8 and Black is a pawn up for practically nothing.

7...Kd8

Everything else loses:

a) 7...Bxd2+?? 8.Nxd2! Qe5 (or 8...Qf4 9.g3) 9.Nc4 (Steinitz).
b) 7...Qxg2?? 8.Bf3 Qh3 9.Nxc7+ Kd8 10.Bxb4 Nxb4 11.Nxa8 Nf6 12.Kd2 (Gutman).
c) 7...Qe5?? 8.f4! Qxf4 (or 8...Qxb2 9.N1c3 Nd4 10.Rb1 Qxc2 11.Rxb4, N.Shavtvaladze-
A.Kostouros, Ikaros 2003) 9.c3 Qh4+ 10.g3 Qe4 11.Nxc7+ Kd8 12.Rf1 Kxc7 13.cxb4 and White is
winning according to Stockfish; e.g. 13...Nf6 14.Nc3 Qe6 15.b5 Ne5 16.Bh6! gxh6 17.Rxf6 Qe8
18.Qd6+ Kd8 19.Nd5. There’s no arguing with the engine in this sort of position.

8.0-0

273
8...Bxd2

The most efficacious move, eliminating both the potentially dangerous white bishop and any issues
with Black’s own bishop and queen, bearing in mind that loose pieces can drop off.

9.Nxd2

Developing with tempo. After the alternative 9.Qxd2, the white knights are slightly discomfited:
9...a6 10.N1c3 (if 10.N5c3 then 10...Qd4! 11.Qg5+ Qf6, or 10.Bf3 Qb4! 11.Qg5+ Nge7 – Gutman)
10...Qh4 (or 10...Qe5) 11.Na3 Qd4! (Gutman) 12.Qg5+ (neither 12.Bd3 Nf6 13.Nc4 d6, nor 12.Qe1
Nf6 13.Rd1 Qc5 gets White anywhere) 12...Qf6 13.Qd2 (both 13.f4 d6 and 13.Qe3 Nge7 are better
for Black) 13...Qd4 repeats if desired. Otherwise 13...h6, 13...Nge7, 13...d6, and 13...b5 are ways to
continue; Black is not worse.

9...Qf4

274
 The black queen goes to all sorts of strange squares after 4...Qh4 and this is one of the strangest.
When a queen is attacked the natural course is to bring her swiftly to safety. Here she hangs about
directly in front of the opposing army – and yet is quite unharmed there, off the e-file and off the light
squares, Almost the only way she can be attacked is by g2-g3, when she can retreat to the very safe
h6-square.
Stockfish wants to play 9...Qf5!? and claims equality. It seems no human has ever tried or even
considered that move.

10.g3

White drives the insolent intruder away.

Instead:
a) 10.a4!? (planning to attack the queen with the a1-rook) 10...a6! (so that the white knight will be
in the way on the third rank) 11.g3 Qh6 (as in the main line; not 11...Qf6 12.Nc3 Nh6?,
G.MacDonnell-J.Wisker, London match 1874, because 13.Nd5! is very strong, e.g. 13...Qxb2? 14.c3
Re8 15.Bc4 Ne5 16.Ra2 wins – Gutman) 12.Nc3 Nf6 looks quite solid for Black; e.g. 13.Bf3 (or
13.Nc4 d6) 13...Re8 14.Nde4 Nxe4 15.Bxe4 d6 16.Re1 f5 17.Bg2 Rxe1+ 18.Qxe1 Be6.
b) 10.c4 Nf6 11.Nf3 (for 11.g3 Qh6 see 11.c4 in the main line) 11...Re8, and if 12.Re1 (12.g3 Qh6
transposes again) 12...d6 13.c5, then 13...d5 seems safe enough. Not 11...Ng4? 12.g3 Qf6 and Black
has just wasted time to bring the queen to a more exposed square. The game Vienna CC-City of
London CC, corr. 1872, continued 13.Nc3 Re8 14.Nd2 Nh6 15.Nde4 Qg6 16.Bh5 Qf5 17.Bf3 b6
18.Bg2 Bb7 19.Qd2 f6 20.Rad1 Nf7 21.Rfe1 and White had a huge position for the pawn.
c) 10.Nb3 Nge7 (or 10...Nf6) 11.g3 Qf6 (this makes more sense now the d2-knight has gone to the
queenside) 12.c3?! h5! 13.Qd2 h4 14.Rad1 Qh6 15.Nc5 Qxd2 16.Rxd2 b6 17.Ne4 hxg3 18.fxg3 f6
and Black soon consolidated, C.Golmayo Zupide-W.Steinitz, 5th matchgame, Havana 1883.

275
10...Qh6

The best placement for the queen for the time being, comfortably out of the way yet governing a
key rank and diagonal, while leaving the f6-square for the knight.

11.Nc4

The main move, as well as the only one examined by theory. In practice, White has tried numerous
others:
a) 11.Bc4?! (the beginning of a bad plan) 11...Nf6 12.Re1 (not 12.Bxf7?? Ne5 13.Bb3 Qxd2! and
wins) 12...d6 13.Bxf7? (still bad; 13.Nf3 is still fairly level) 13...Rf8 14.Be6 (giving away the key
light square defender, but if 14.Bb3 then 14...Ng4 15.Nf3 a6 16.Nc3 Nxf2! follows) 14...Bxe6
15.Rxe6 Qh3 16.Re1? a6 and Black was winning in P.Ster-Ma.Rohde, corr. 2001, in view of 17.Nc3
Ng4 18.Nf3 Nd4.
b) 11.Re1, as in W.Pommerel Brouwer-A.Curado, corr. 2003, asks for 11...a6 12.Nc3 d6, followed
by ...Nf6, ...Bd7, ...Re8 and so on.
c) 11.Ne4, M.Pila Diez-G.Gomez Cid, Oropesa del Mar 1999, can be answered by 11...Nf6 12.Bf3
Re8 13.Re1 Qg6 (not 13...d6?? 14.Nxc7!) 14.Nbc3 (or 14.Nxf6 Rxe1+ 15.Qxe1 Qxf6) 14...d6 and so
forth.
d) 11.Nb3 Nf6 (Steinitz, who thought the queen belonged on f6, would no doubt approve of
11...Qf6!? 12.Re1 h5 13.Nc5 Qe5!? 14.b4 Nxb4 15.Nb3 h4 16.Bc4 Qf6 17.Qd2 Nc6 with a random
“0.00” position which Black won in MarcShaw-AndyAndyO, ChessWorld.net 2017) 12.Bf3 (if
12.N5d4 then 12...Nxd4 13.Qxd4 d6, or 12.c4 Re8) 12...d6 13.Re1 Bd7 14.c4 Re8 15.Rxe8+ Nxe8
16.N5d4 Ne5!? (16...Rb8 was possible too) 17.Bxb7 Rb8 18.Na5 c5! 19.Nac6+ Nxc6 20.Nxc6+ Bxc6
21.Bxc6 Rxb2 22.Rb1 Qd2 23.Qxd2 Rxd2 24.Rb7 Nc7 25.Rb8+ ½-½ MarcShaw-tsmenace,
ChessWorld.net 2017, in view of 25...Ke7 26.Rb7 Kd8 27.Rb8+, repeating.

276
 e) 11.c4 Nf6 12.f4 (here 12.c5 seems premature after 12...a6 13.Nc3 Re8 14.Re1 Rb8 15.Nf3 b5!
16.cxb6 Rxb6, G.Villumsen-H.Lykke, corr. 2000; the same applies to 12.Nf3 Re8 13.c5, as in
S.Voitsekhovsky-S.Tikhomirov, Tula 2001, and 13...b6!, while if 13.h4 – Steinitz, then 13...Ne4
14.Bd3 f5 looks fine) 12...Re8 13.Bf3 a6 14.Nc3 Nd4 15.Bg2 (or 15.c5 Nxf3+ 16.Nxf3 Qh5) 15...Nf5
16.Qf3 Re3 17.Qd1 Re6 18.Qf3 Ne3 (otherwise 18...Re3 repeats) 19.Rfe1 Nxg2 20.Qxg2 d6 and
Black is at least equal, tripoduk-jatait47, Chess.com 2020.

11...Nge7

With threats looming at c6 and d6, it’s prudent to support the c6-knight and leave the queen’s sight
clear along the rank. Whereas after 11...Nf6?! 12.Bf3 Re8 13.Nd4 d5 (or 13...Ne5 14.Nxe5 Rxe5
15.Re1 – Gutman) 14.Nxc6+ bxc6, Genius-Hiarcs 6, computer game 1997, and then 15.Na5! Bd7
16.c4, Black’s position is very shaky.

12.Bf3

IM George Botterill’s improvement on 12.Qd3 a6 13.Nd4, after which 13...Nxd4 (or 13...Qg6
14.Nxc6+ Nxc6 15.Qd2 Qf6 16.c3 d6 17.Rad1 Be6, C.Golmayo Zupide-W.Steinitz, 3rd matchgame,
Havana 1883) 14.Qxd4 Nc6 15.Qc3 Re8 16.Bf3 (or 16.Rfe1 d6 17.Bf3 Bd7 18.Bd5 Qg5) 16...Qf6
17.Qxf6+ gxf6 18.Bd5 Re7 19.Rfe1 d6 (or 19...Ne5 20.Ne3 c6 21.f4 cxd5 22.fxe5 Rxe5 23.Rad1)
20.Rxe7 Kxe7 21.Re1+ Kf8 22.Be4 Be6 23.Ne3 Kg7 gave Black a minimal edge, J.Mieses-
M.Chigorin, Hastings 1895.

277
12...d6!?

A suggested counter-improvement on 12...a6 13.Nd4 Rb8?! (here 13...Nxd4 14.Qxd4 Qf6 is

somewhat better – B&H) 14.Re1 b5??, G.Botterill-M.Staples, Manchester 1974, when 15.Bxc6!
Nxc6 16.Nxc6+ Qxc6 17.Ne5 Qf6 18.Qh5 wins for White; e.g. 18...Rf8 (not 18...g6? 19.Nxf7+! Qxf7
20.Qg5+) 19.Ng4 g6 20.Qc5 Qd6 21.Qe3, followed by Rad1.
I’m less keen on 12...Qe6 13.Qd3 (13.Ne3 is another option) 13...Qf5 (Gutman) 14.Qb3 (obviously
not 14.Qxf5? Nxf5) 14...Qc5, as in M.Bergmann-J.Simmelink, corr. 1999, as 15.Rfe1 looks good for
White; e.g. 15...a6 16.Bxc6 Nxc6 17.Qc3 axb5 18.Qxg7 Qf8 19.Qf6+ Ne7 20.Rxe7 Qxe7 21.Qxh8+
Qe8 22.Qf6+ Qe7 23.Qc3 and Black can’t capture the knight because 23...bxc4? 24.Re1 wins.

278
 This position has received very few tests since B&H first mentioned it in 1977, but I have defended
it twice:
a) 13.Nd4 Bd7 14.Nxc6+ Nxc6 15.Bxc6? (15.b4 is better and probably level) 15...bxc6 16.Qf3 Qf6
17.Qxf6+ gxf6 was good for Black in K.Amann-C.Callow, corr. 2004.
b) 13.Re1 can be met by 13...Qf6! (Gutman), followed by ...a7-a6 and/or ...h7-h5, with shades of
MarcShaw-AndyAndyO above.
c) 13.Nbxd6 cxd6 14.Nxd6 Kc7! (a required resource) 15.Nxf7 Qf6 16.Nxh8 Bh3 (thus far B&H)

17.Bg2 Bxg2 18.Kxg2 Rxh8 19.c3 leaves White with rook and two pawns for the two knights,
which Gutman assesses as “unclear”. However, it’s Black who needs to be careful lest the knights

279
become bystanders. I tried 19...g5!? (to increase Black’s influence on the kingside) 20.Qf3 (or 20.f4
Nf5 21.Qe2 h5) 20...Qxf3+ 21.Kxf3 Ne5+ 22.Kg2 (or 22.Ke4 Nc4, or 22.Ke2 g4) 22...Nd3 (not now
22...g4? 23.Rfe1) 23.b3 Ng6 24.Rad1 Rd8 25.h3 (or 25.Kg1 Nge5 26.f4 gxf4 27.gxf4 Ng6 28.f5
Nge5) 25...h5 26.Kf3 Nde5+ 27.Ke4 Re8 28.Kf5 Rf8+ 29.Kxg5 Ne7 30.Kxh5 with a draw in
scepticus-tsmenace, ChessWorld.net 2018, since the white king cannot escape the checks: 30...Rf5+
31.Kh6 Rf6+ 32.Kg7 Rf7+ etc.
d) 13.Qd3 a6 14.Ncxd6 (or 14.Nd4 Nxd4 15.Qxd4 Qf6) 14...cxd6 15.Nxd6 Kc7! (again) 16.Nxf7
Qf6 17.Nxh8 is similar, except that this time Black has 17...Bf5! 18.Qe3 Rxh8 (thus far Gutman)
19.Rfe1 Rd8 20.Bg2 g5!? (again this move) 21.Rac1 (or 21.Qc5 Bg6) 21...h5 22.f4 g4 23.c3 Rd3
24.Qe2 h4 25.Rcd1 Rxd1 26.Rxd1 h3 27.Bh1 Nc8 28.Bxc6 Qxc6 29.Qe5+ Kb6 30.Rd2 Qc5+
31.Qxc5+ Kxc5, reaching a level endgame, aperturaf-tsmenace, ChessWorld.net 2018.

C: 5.Nc3 Bb4

Pinning the knight. No points for anything else.

6.Be2!

The strength – even existence – of this move was discovered after B&H’s book was published in
1977. Rather than attack c7 straight away, White defends a check on e4 in advance, prepares to castle,
threatens Nf5, and waits to see how Black responds.
Instead:
a) 6.Qd3 Nf6 returns to line A.
b) 6.Be3 should be answered by 6...Nf6!, for which see 5.Be3 (note ‘d4’) at the top.
c) 6.Nf3?! Qxe4+ 7.Be2 appears in a bracket (7.Nc3) to 5.Nf3 (note ‘e’) at the top.

280
 d) 6.Nf5? Qxe4+ 7.Ne3 (and 7.Qe2 Qxe2+ 8.Bxe2 g6 is just silly) 7...Bxc3+ 8.bxc3 d6 9.Bd3 Qh4
leaves White with very little for the pawn.
e) 6.Nxc6?! is somewhat better than on the previous move, if only because after 6...Qxe4+ 7.Be2
(both 7.Be3?! dxc6 and 7.Qe2?! Qxe2+ 8.Bxe2 dxc6 are good for Black) 7...dxc6 8.0-0 Qh4 9.g3 Qd8
(or 9...Qe7 10.Bg4; not 9...Qh3?, S.Santaella Amate-A.Garcia Ruiz, Malaga 2005, as 10.Re1 Be7
11.Ne4 gives White very good play for the pawn) 10.Qxd8+ Kxd8 11.Ne4, it is more difficult for
Black to consolidate.
Another option is 7...Bxc3+ 8.bxc3 Qxc6 9.0-0 Nf6 10.Bf3 d5 11.Re1+ Be6 12.Ba3, as in N.Cinar-
C.Foisor, Mangalia 1992, when 12...Qa4! is good for Black. In this line Gutman gives 10.c4! 0-0
11.Bb2 Re8 12.Re1 as an improvement; then if nothing else Black might play 12...d5 13.Bxf6 Qxf6
14.cxd5 Bd7 and is at least equal.
f) 6.Nb5?! is inaccurate due to 6...Ba5! (rather than 6...Qxe4+ 7.Be2, transposing to the main line)
and then:

f1) 7.Qd3 a6 (the immediate 7...Nf6 may be stronger, for which see 7.Nb5 in line A) 8.Nd4 was
A.Boncz-H.Nagy, Hungarian League 2003, when 8...Nf6! is line A with the extra moves ...Ba5 and
...a7-a6 which make little difference: 9.Nxc6 dxc6 10.Bd2 Bxc3 11.Bxc3 Nxe4 etc is much the same.
f2) 7.Qd5 Nf6 (or 7...h6!, preventing Bg5) 8.Bg5 Nxd5 9.Bxh4 Nxc3 10.Nxc3 d6 11.0-0-0 Bxc3
12.bxc3 Be6 and Black is no worse, L.Mezera-R.Pribyl, Czech League 1997.
f3) 7.Be3 a6! (for 7...Nf6 see 5.Be3 again) 8.Nd4 Qxe4 is good for Black, since White can’t play
Nb5 here; e.g. 9.Be2 Qxg2 10.Bf3 Qg6 11.Qe2 Nge7 12.0-0-0 Bxc3 13.bxc3, M.Jerkovic-
A.Zacinovic, Bosnian League 2010, and now 13...Nxd4! 14.cxd4 (or 14.Bxd4 Qg5+ 15.Kb1 Qb5+)
14...d5 15.Rhg1 Qd6 16.Rxg7 Be6, when Black is still one pawn up.
f4) 7.Be2 a6 8.Nd4 Qxe4 also wins the pawn safely: 9.Nxc6 (9.Be3 is ‘f3’ above) 9...Bxc3+
10.bxc3 Qxc6 11.0-0 Nf6 12.Bf3 (or 12.c4 0-0) 12...d5 13.Ba3 (or 13.Qd4 0-0 14.Re1 Bf5 15.c4 Be4,

281
M.Heika-H.Bredl, German League 2002) 13...Be6 14.Rb1 Ne4 15.c4, R.Ramesh-N.Babu, Madras
Zonal 1995, and now 15...0-0-0 with a clear advantage.
f5) 7.Bd2 a6 8.Na3 Bxc3 9.Bxc3 Qxe4+ is in Black’s favour too: 10.Be2 (or 10.Qe2 Qxe2+
11.Bxe2 Nf6 12.0-0 d6, M.Kevicky-K.Poulheim, corr. 2000) 10...Qxg2 11.Bf3 Qg6 (or 11...Qg5
12.h4 Qe7+ 13.Kf1 Nf6 14.Qd2 d5) 12.Qe2+ Nge7 13.0-0-0, M.Accattato Martin-Y.Gallardo de la
Torre, Mondariz 2003, and after 13...d6 White hardly has enough for two pawns; e.g. 14.Rhg1 Qh6+
15.Kb1 (or 15.Bd2 Qe6) 15...Rg8 16.Bxc6+ bxc6 17.Ba5 Ra7.
f6) 7.Bd3 secures the e-pawn but allows 7...a6 8.Na3 b5 9.Bd2 (on 9.Nab1 Stockfish suggests the
novelty 9...Ne5!?, followed by ...Bb7, as giving Black an edge; there is little wrong with 9...Bb6 or
9...Nf6 either) 9...Bb6 (not 9...b4? 10.Nd5) 10.g3 Qf6 11.0-0 Nge7 12.Nd5 (or 12.Qh5 Nd4)
12...Nxd5 13.exd5 Ne5 and Black is better, J.Hawes-H.Gretarsson, Elista Olympiad 1998.

6...Qxe4

Taking the pawn is the only consistent move and probably the only move full stop. Everything else
seems to fail. Summarizing briefly:
a) 6...Bxc3+ 7.bxc3 Qxe4 (or 7...Nf6 8.Nf5!) 8.0-0! Nxd4 9.cxd4 Ne7 10.Re1.
b) 6...Nf6 7.Nf5! Bxc3+ (or 7...Qxe4 8.Nxg7+ Kf8 9.Bh6) 8.bxc3 Qxe4 9.Nxg7+ Kd8 10.0-0.
c) 6...Nge7 7.0-0 Bxc3 (or 7...0-0 8.Ndb5 Ba5 9.Nd5!) 8.bxc3 Qxe4 9.Nb5 Kd8 and Black has been
move-ordered into the 9...Nge7 line below.

7.Ndb5

Now White sends the knight in. For 7.Nxc6?! and 7.Nf3?! see the respective moves at move 6
above.

282
7...Bxc3+

Not 7...Ba5?? 8.Nxc7+ Bxc7 9.Nxe4, or 7...Qxg2? 8.Bf3 Qh3 9.Nxc7+ Kd8 10.Nxa8 Ne5? 11.Be2
and White wins, J.Timman-S.Davidov, Baku 2008; while 7...Kd8?! 8.0-0 Bxc3 gives White the
choice between 9.bxc3 (returning to the main line) or playing 9.Nxc3! (which seems even better).

8.bxc3

The attack on c7 is more important than structure.

Bashing my moves out in a handicap blitz tournament once, I followed up with the usual 8...Kd8
and then noticed my opponent had actually recaptured on c3 with the knight. 8.Nxc3!? is more
correctly (!) answered by 8...Qxg2! 9.Bf3 with two options:

a) 9...Qh3 10.Nd5 (or 10.Rg1 Nge7 11.Rxg7 d6) 10...Kd8 (or 10...Ne5 – B&H; e.g. 11.Be2 Kd8
12.Bf4? Qe6 or 12.Qd4 Nf3+ 13.Bxf3 Qxf3 14.Qxg7 Qxh1+ 15.Ke2 Qe4+ 16.Kf1 Qh1+ with a draw)
11.Bf4 (or 11.Rg1 Ne5) 11...Qe6+ 12.Kf1 Ne5 13.Rg1, A.Dawson-J.Garratt, corr. 1891, and now
13...d6 14.Rxg7 Nxf3 15.Qxf3 Qh3+ 16.Qxh3 Bxh3+ 17.Kg1 Be6 leads to a level endgame after
18.Bg5+ Kd7 19.Nf4 Re8 20.Nxe6 Rxe6 21.Rxf7+ Ne7.
b) 9...Qg6 10.Qe2+ (not yet 10.Nd5?! Kd8 11.Bf4? d6 12.Qd2 Bg4 13.Bg2 Nge7 14.Ne3 Re8 and
Black won, H.Koronowski-Man.Keller, corr. 1980) 10...Nge7 11.Nd5 (not yet 11.Be3?!, B.Ksir-
L.Tkac, corr. 1996, in view of 11...d6 12.Nd5 Nxd5 13.Bxd5 Be6 or 12.0-0-0 Bd7 13.Rhg1 Qe6
14.Rxg7 0-0-0) 11...Kd8 12.Be3 and White has just enough play for the pawns; e.g. 12...Nxd5 (or
12...Nf5 13.0-0-0 Nxe3 14.Qxe3 Qh6 15.Qxh6 gxh6 16.Bh5) 13.Bxd5 d6 14.0-0-0 Qf6 (or 14...Bg4
15.f3 Be6 16.Rhg1 Qf5 17.Be4 Qa5 18.a3 g6 19.Rg5) 15.Rhg1 h6 16.Bxc6 bxc6 17.Bd4 Qf4+
18.Qe3 Qxe3+ 19.fxe3 Ke7 (or 19...Be6 20.Bxg7 Rh7 21.Bxh6) 20.Bxg7 Rh7 21.Rdf1 Be6 22.Bf6+
Kd7 23.h4 with a likely draw.

283
8...Kd8

The only way to defend c7, since 8...Qe5?? loses to 9.f4.

9.0-0

The obvious and best move. Not 9.Ba3?? Qxg2 10.Bf3 Qg5! and Black is winning; while 9.Be3!?
lets Black choose between 9...Nf6 10.0-0 (transposing to line C3 below) or 9...Qxg2 again, although
in this case 10.Bf3 Qg6 11.Qe2! (not 11.Bf4? d6 12.Qe2 Qe6, Ja.Martinec-N.Hanc, Mureck 2013)
11...a6 12.Nxc7! Kxc7 13.Bf4+ Kd8 14.0-0-0 seems to give White enough play for a draw; e.g.
14...Qe6 15.Qd2 Nge7 16.Rhe1 Qxa2 17.Qd6 Qa5 18.Bxc6 (or 18.Bg4 Qa1+ 19.Kd2 Qa5 20.Kc1
Qa1+, repeating) 18...Nf5 19.Qd5 Qxd5 20.Bxd5 d6 21.Bxf7 Rf8 22.Bd5.

9...Nf6!

The exclamation mark comes from GM Gutman, who identified 9...Nf6 as Black’s best defence.
Rather than worrying about structure either (White may yet capture on c6 and/or f6), Black hurries to
develop the kingside while leaving the e-file open for the rook.
The alternative development with 9...Nge7 is too passive; e.g. 10.Bd3 Qh4 11.Re1 f5 (or 11...a6
12.Re4 Qf6 13.Rf4 Qe6 14.Nd4 Nxd4, E.Schiffers-M.Chigorin, St. Petersburg 1897, and now
15.cxd4 – B&H) 12.Ba3 a6 (or 12...Qf6 13.Bc4 a6 14.Re6) 13.Nd4 d6 14.Bc4 Nxd4 15.cxd4 b5
16.Bf1 Bb7 17.Qd2 Bd5? 18.c4! Bxc4 19.Bxc4 bxc4 20.Rac1 with a very strong attack, P.Padilla
Cabero-J.Tait, Notts League 2009.
Note too that 9...a6 is not the best use of a tempo at this moment. While it is tempting to drive the
intruder away, the white knight is often better placed on d4 – and may indeed drop back there
voluntarily – so ...a7-a6 should be carefully timed if it is played at all. In this particular instance,
White doesn’t even need to retreat; e.g. 10.Bf3 Qc4 11.Nd6! cxd6 (or 11...Qe6 12.Nxc8 Rxc8 13.Re1

284
– Dembo & Palliser) 12.Qxd6 Nf6? (or 12...Nge7 13.Re1) 13.Be3 Ne7 14.Rfe1 Nfd5 15.Bxd5 Nxd5
16.Qxd5 1-0 T.Oral-M.Kantorik, Slovakian League 2000.

Most authors consider that White has the advantage here. What they don’t agree on is how that
advantage should be achieved, recommending variously 10.Rb1 (Lane, Emms), 10.Be3 (Wells, Flear,
Barsky), 10.Nd4 (Shaw), 10.Bd3 (Pavlovic), and 10.Re1 or 10.Bg5 (Khalifman & Soloviev). Of
course it could be the case that they’re all good for White. We’ll see.
C1: 10.Nd4
C2: 10.Bg5
C3: 10.Be3
C4: 10.Re1
Others:
a) 10.c4 prompts 10...a6 (now that the knight cannot go to d4) 11.Nc3 Qe5! (attacking the knight
again; this seems more precise than 11...Qd4 12.Qe1 Qe5, as in A.Grosar-D.Sermek, Maribor 1994)
12.Bb2 (now the knight is pinned) 12...d6 13.Rb1 Re8 (and now it can’t move because the other
bishop would drop) 14.Bf3 Qd4 15.Nd5 Qxd1 16.Bxf6+ gxf6 17.Rfxd1 Ne5 with equality.
b) 10.Ba3 looks premature: 10...Re8 11.Re1 (or if 11.Bf3 Qc4 12.Rb1, S.Reschun-G.Pfeiffer,
Austrian Junior Ch. 1993, then just 12...b6) 11...a6 (forcing the pace) 12.Nxc7!? (otherwise 12.Nd4
d6 13.Qd2 Qd5 14.Bf3 Qc4 is level, Ron.Fischer-R.Ohnmacht, corr. 2001) 12...Kxc7 13.Qd6+ Kd8
and now 14.Bf1! (not 14.Bc5? Re7 15.Qg3, S.Mareck-M.Heidenfeld, Schwäbisch Gmünd 1995,
when 15...Re6! 16.Qxg7 Kc7 17.Qxf7 Ne5 consolidates the piece) 14...Qxe1 (or 14...Re6 15.Qf8+
Re8) 15.Bc5 Re6 16.Qf8+ Kc7 (not 16...Re8?? 17.Bb6 mate) 17.Rxe1 Rxe1 18.Bd6+ (or 18.Qd6+
Kd8) 18...Kb6 19.Bc5+ Kc7 is a draw.
c) 10.Bd3 attempts to embarrass the black queen. It shouldn’t work: 10...Qh4 11.g3 (if 11.Be3 d6
12.Qd2, Wa.Schmidt-Frie.Meyer, Waldshut 1991, then 12...h6 – Gutman) 11...Qh5! (not 11...Qh3?

285
12.Bg5 Ne5 13.f4! Neg4 14.Qd2 and Black is in a fatal mess, as in O.Oussama-J.Brunet Martinez,
Mollerusa 1997, and analysis by GM Pavlovic) 12.Be2 (Black is fine if the queens come off, as with
12.Qxh5 Nxh5 or 12.Nd4 Qxd1 13.Rxd1 d6) 12...Qd5 reaches the same position as after 10.g3 Qd5,
which is really not the most testing; e.g. 13.Nd4 (or 13.Qxd5 Nxd5) 13...d6 14.Bf3 Qc4 15.Bg5 Qxc3
16.Bxc6 bxc6 17.Qh5 Be6 18.Nxe6+ fxe6 19.Qf7 Kc8 20.Qxg7 Ne4! 21.Qxc3 Nxc3 22.Bf6 Ne2+
23.Kg2 Rf8 and Black is still a pawn up.
d) 10.Bf3 leaves the knight undefended so will likely transpose to note ‘e’ after 10...Qc4 11.Rb1
Re8, given that other moves are inferior; e.g.11.a4 a6 12.Be2 Qd5 13.Nd4, J.Codina Tormo-
A.Chamocho Bellido, Madrid 2008, and now 13...d6; or 11.Be2 Qd5 (better than 11...Qe4, repeating)
12.Bf4 d6 13.Nd4, J.Munoz Chavez-R.Ugalde Garcia, Costa Rican Ch., Cenfotec 2003, and 13...Re8;
while 11.Nd4 d6 is much the same as 14.Bf3 in note ‘c’, and 11...Qxc3 is possible too.
e) 10.Rb1 should be answered by 10...Re8 11.Bf3

and 11...Qc4 12.Bg5 (if 12.Re1 Rxe1+ 13.Qxe1, P.Pflichthofer-T.Oparaugo, German League 1994,
then 13...h6; or 12.Nd4 d6) 12...a6 (not 12...Re5? 13.Qd2 a6 14.Nd4 Rc5? 15.Bxf6+ gxf6 16.Qh6 and
White won quickly, J.Hjartarson-P.Moulin, New York Open 1994) 13.Nd4 Nxd4 (nor 13...Qxc3? due
to 14.Bxc6 bxc6 15.Qh5! and wins, A.San Pedro Fraga-F.Osorio Curbelo, corr. 2002, in view of
15...Qxd4 16.Qxf7, followed by Qxg7 and Bxf6) 14.cxd4 and now, rather than defending b7 with
14...Rb8 (when Gutman’s 15.Qd2 gives White strong play for the pawn), Stockfish suggests 14...a5!?,
intending to develop the rook via a6 and letting the b-pawn go; e.g. 15.Bxf6+ (or 15.Bxb7 Bxb7
16.Rxb7 Kc8) 15...gxf6 16.a3 (or 16.Qd2 Qxa2 17.Qh6 Qe6 18.Qxh7 d5) 16...Ra6 17.Bxb7 Bxb7
18.Rxb7 Re2 and Black has sufficient counterplay.
In general, I think Black does best to meet Bf3 with ...Qc4 if possible, since the queen is more of a
nuisance there, whereas ...Qf5 allows White to develop an initiative undisturbed; e.g. 11...Qf5 12.c4
(or 12.Nd4 Nxd4 13.Qxd4 with an edge, A.Felsberger-K.Opl, Austrian League 1994; 13...Ke7!?
might have to be tried) 12...d6 13.Ba3 Qd7 14.h3 Ne5 15.c5 dxc5 16.Bxc5 Nxf3+ 17.Qxf3 Qc6

286
18.Qa3 b6 19.Be3 Bb7 20.f3 a6? (Stockfish prefers 20...Kc8) 21.Nd4 Qc4 22.Nf5 Nd5 23.Rfd1 g6
24.Ng7! Rg8 25.Rb4 Qe2 26.Qc1 Kc8 27.Re1 Nc3 28.Kh2 a5 29.Rxe2 Nxe2 30.Qf1 and White won
in Stoofvlees II-Igel 2, TCEC Cup 7 2020.

C1: 10.Nd4

The knight considers its job on b5 done and returns to the centre. GM John Shaw suggests this as a
“promising line” for White. “One idea is to threaten 11.Bf3 without allowing 11...Qc4 in response.”
No one else even mentions it.

10...Nxd4

Black has tried 10...Qg6?! with success; e.g. 11.c4 Nxd4 12.Qxd4 b6 13.Bd3 Qg4 14.Bf4 d6 15.h3
Qh5 16.Rfe1 Re8 (not 16...Bb7?? 17.Re5) 17.a4 Bb7 18.Be3?! (18.Bxd6 forces a draw: 18...cxd6
19.Qxd6+ Nd7 20.Rxe8+ Kxe8 21.Re1+ Kd8 22.Re7 Bc8 23.Rxd7+ Bxd7 24.Qf8+ Be8 25.Qd6+ and
so on) 18...Re7 19.a5 Ke8, followed by ...Kf8, and Black is a tad better, draco69-tsmenace,
ChessWorld.net 2017.
Unfortunately, 11.h4!, threatening both h4-h5 and Bg5, messes this up, when Black has nothing
better than to exchange on d4 anyway: 11...Nxd4 (11...h6? 12.h5 Qh7 13.Bd3 Qg8 14.Re1 is terrible)
12.cxd4, showing the queen has retreated too soon; e.g. 12...h6 13.a4 d6 14.Ra3! Bg4 15.Bxg4 Qxg4
16.Qxg4 Nxg4 17.Rg3 Nf6 18.Rxg7 with a clearly favourable endgame for White.

11.cxd4

287
 “For the sacrificed pawn, White has the bishop pair, much the safer king, a lead in development and
a flowing initiative; all that is more than enough” (Shaw) – more than enough to make this a decent
practical option for White. From the other side, and more concretely: is Black actually worse?

11...Re8!

Forcing the pace by attacking the e2-bishop. Instead:

a) 11...d5?! is positionally dubious since it weakens the dark squares and creates a target: 12.Bg5
(rather than 12.Rb1?! Re8 13.Bf3 Qf5 14.Ba3 b6 15.Re1 Rxe1+ 16.Qxe1 Be6 17.Qb4 Kc8 and Black
stands quite well, G.Braun-J.Krassowizkij, Illertissen 2012) 12...Re8 13.Re1 Qf5 14.Bh4 Bd7 15.Qd2
Kc8 16.Bd3 Ne4 17.f3! Nxd2 (or 17...Qh5 18.Qf4) 18.Rxe8+ Bxe8 19.Bxf5+ Bd7 20.Re1! and wins.
b) 11...d6?! allows White more freedom, not least to play 12.Bg5! Qf5 13.Bh4 (this is mentioned by
Gutman, via 10.Bg5) 13...Re8 (if 13...h6 14.Re1 g5 15.Bg3 Re8 16.c4 b6, juliangon-tsmenace,
ChessWorld.net 2017, then 17.a4 a5 18.c5 bxc5 19.dxc5 Qxc5 20.Rc1 is very dangerous) 14.Qd2
Bd7 (I tried 14...h6 again in Krambambuli-tsmenace, ChessWorld.net 2017, and won as my opponent
defaulted – just as well since 15.Bd3! Qg4 16.Bxf6+ gxf6 17.Qxh6 Qxd4 18.Rad1 or 18.h4 looks
good for White) 15.Rae1 (rather than 15.Rfe1?, when 15...Re4! 16.Bd3 Rxd4 17.Qe3 Qe5 18.Qf3
Bc6 19.Qh3 Qd5 equalizes) 15...Re4 (if 15...Re7, as in M.Dorer-P.Leisebein, corr. 2005, then 16.c4 is
very strong, planning to embarrass the black queen with h2-h3 and g2-g4 or Bd3) 16.f4 (threatening
17.Bd3 Rxd4 18.Qe3, which seems even stronger than 16.Bd3 Rxd4 17.Qe3 Qe5 18.f4) 16...c5 (to
give the king some air; if 16...c6 then 17.h3, threatening Bg4) 17.Qa5+ b6 18.Qc3 cxd4 (or 18...Re8
19.g4) 19.Qg3 Kc7 20.Bd3 Rae8 21.Bxf6 gxf6 22.Rxe4 Rxe4 23.Qf2 with Bxe4 to follow, and Black
hardly has enough for the exchange given the trashed pawn structure.

288
12.Bf3

Here 12.Re1 d6 13.Bg5 Qf5 14.Bh4 can be answered by 14...g5 15.Bg3, when Black has saved a
tempo on ...h7-h6 so can play something else, such as 15...b5!?.

12...Qg6!

This has been played just once – and Black lost – but it seems like an improvement on 12...Qf5
13.c4 d6 14.Re1, as in C.Koch-S.Sacerdotali, corr. 2002, given by Shaw, who also suggests 14.Qb3,
14.a4 and 14.c5 as attractive options for White.

13.c4 d6

289
14.Qb3

If we compare Shaw’s other moves here, the advantage of ...Qg6 is that ...Bg4 becomes possible;
e.g. 14.Re1 Rxe1+ 15.Qxe1 Bg4! 16.Bxb7 Rb8 17.Bc6 Rb1 equalizes; while 14.c5 Bg4 15.Bxg4
Qxg4 16.Qb3 Kc8 17.cxd6 Qe6! 18.Qg3 c6 19.Qxg7 Rg8 20.Qh6 Rxg2+ 21.Kxg2 Qg4+ is a draw; or
if 14.a4 then 14...Ke7 and ...Kf8 as below.

14...Ke7!

Running for cover.

15.Bf4 Kf8 16.c5

290
 Here 16...Qf5?! 17.Bg3 dxc5 18.dxc5 Qxc5 19.Rac1 Qb6 20.Qa3+ Kg8 21.Bxc7 Qe6 22.Qa5 b6?
23.Qd2 Ne4 24.Bxe4 Qxe4 25.Rfe1 Qa4 26.Rc4 Rxe1+ 27.Qxe1 Qa3 28.Qe4 led to a win for White
in Igel 2-Stoofvlees II, TCEC Cup 7 2020.
Stockfish rejects its lesser 3500-rated colleague’s move, preferring 16...Ne4, which it assesses as
equal; e.g. 17.Rae1 Qf6! 18.Bxd6+ cxd6 19.Bxe4 (or 19.Rxe4 Rxe4 20.Bxe4 dxc5 21.dxc5 Qe7
22.Qd3 Rb8 23.Bxh7 Be6) 19...Qxd4 20.Bd5 Be6 21.Bxe6 fxe6 22.cxd6 (or 22.Qxb7 Qxc5; or
22.Rxe6 Rxe6 23.Qxe6 Qxc5) 22...Qxd6 23.Qxb7 Kg8.
Practical tests await.

C2: 10.Bg5

Gutman wrote that “the pin of the knight seems to be the logical way of meeting the relatively new
variation.” Relatively new in 2001 that is. K&S mention it too, as an alternative route to line C4.

291
10...d6

Alternatives are not improvements:

a) 10...Re8? is wrong here, since 11.Bd3 Qd5 12.Qh5! is winning for White; e.g. 12...Qe6?
13.Qxh7 Ne7 14.Qxg7 Nh5 15.Qh7 f6 16.Qxh5 fxg5 17.Qxe8+ Kxe8 18.Nxc7+ Kf7 19.Nxe6 dxe6
20.f4, R.Torma-Z.Simonyi, Hungarian League 2012.
b) 10...Qe5 is mistimed: 11.Qd2 d6 12.Rad1 (not 12.Bd3?! h6 13.Bh4 Qh5) 12...Bd7 13.Bf4 Qe7
14.f3 Ne5 15.c4 Ng6 16.Bg3 Bxb5 17.cxb5 Re8 18.Bc4 Qe3+ 19.Qxe3 Rxe3 20.Bxf7 and White is
clearly better, Stockfish-LcZero, TCEC 18 Superfinal 2020.
c) 10...Qf5 also commits the queen, allowing Black to be move-ordered: 11.Bh4 d6 (not 11...Re8?
12.Bg3 d6? 13.Nxc7! and wins) 12.Nd4! Nxd4 (12...Qf4 13.g3 Qh6 14.f4 is good for White too)
13.cxd4 lands Black with 11...d6 in line C1; while 11.h4! d6 (neither 11...h6 12.Bd3 Qc5 13.Bxf6+
gxf6 14.Qg4, nor 11...Qc5 12.Qd2 improves for Black) 12.Re1 is 11...Qf5 in line C4.

292
11.Nd4

Instead:
a) 11.Re1! transposes directly to line C4 and is in fact the optimal move order.
b) 11.Rb1 Qe5 12.Bh4 (or 12.Qd2 h6 13.Bh4 a6 14.Nd4 g5 15.f4 Ne4! 16.Qe3 Qe7) 12...Re8
13.Bf3 (or 13.Re1 g5 14.Bg3 Qd5) 13...Qc5 14.Bxc6 Qxc6 (or 14...bxc6 15.Nd4 Qd5) 15.Qh5 Qe4
16.Bxf6+ gxf6 17.Qxf7 ½-½ Re.Möller-P.Leisebein, corr. 2001, since 17...Bd7 (or 17...Qe7)
18.Qxf6+ Kc8 is equal.
c) 11.Bxf6+ gxf6 12.Bf3 (or 12.Re1 Re8) 12...Qc4 13.Nxd6 (or 13.Rb1 Qc5 14.Bxc6 bxc6 15.Qf3
Be6! 16.Qxf6+ Kd7 17.Nd4 Bxa2 18.Ra1 Bc4 and Black is at least equal, U.Jansen-Mar.Richter,
corr. 2017) 13...cxd6 14.Qxd6+ Bd7 15.Rad1 (not 15.Qxf6+? Kc7 16.Rad1, H.P.Höfer-Er.Collins,
corr. 2001, because of 16...Rad8 17.Bd5 Qc5 with a clear advantage) 15...Qe6 16.Bxc6 bxc6 17.Qb4,
when 17...Re8 18.Qa5+ Kc8 19.Qa6+ Kd8 is a draw.
d) 11.Qd2 can be met by 11...Qf5 12.Rfe1 (alternatives, such as 12.h4 h6 or 12.Bh4 Bd7 or 12.Bd3
Qc5 13.Qf4 Be6, are no worse for Black) 12...h6 13.Bh4 (or 13.Bf4 Ne4 14.Qc1 Qf6) 13...Bd7
14.Nd4 Nxd4 15.Qxd4 (or 15.cxd4 g5 16.Bg3 Re8) 15...Qe5 16.Qb4 (or 16.Qxe5 dxe5 17.Bc4 Be6
18.Bxe6 fxe6 19.Rxe5 Re8) 16...Rb8 17.Bg4 Qb5 18.Bxf6+ gxf6 19.Qd4 Bxg4 20.Qxg4 Qg5 and
Black is defending.
Note too that the sacrifice on d6 doesn’t work here: 13.Nxd6? cxd6 14.Qxd6+ Bd7 15.Bf4 (or
15.Bh4 Qe6) 15...Rc8 16.Rad1 Re8 17.Bf3 b6 (intending ...Qc5) 18.Rxe8+ Nxe8 19.Qd2 Nf6 20.Re1
g5 21.Bg3 g4 22.Qxh6 gxf3 23.Qf8+ Be8 24.c4 Qd7 25.Qh8 Ne4 26.Qh4+ f6 27.Qxe4 Nd4 28.Bh4
Rc6 and Black consolidated, samurai-tsmenace, ChessWorld.net 2018.

11...Qe5

293
 A better moment for this move. Not 11...Nxd4?!, when 12.cxd4 transposes to 11...d6?! in C1 again,
and 12.Bxf6+ gxf6 13.cxd4 may be even stronger.

Another position requiring practical tests. Until then, here are a few variations:
a) 12.h4 h6 13.Be3 Bd7 14.c4 Re8 and Black is fine.
b) 12.Qd2 h6 13.Bf4 (or 13.Bh4 g5 14.Bg3 Qc5, intending 15.f4? Ne4) 13...Qd5 14.Bf3 Qc4
(14...Ne4!? is rather more risky but may be playable; e.g. 15.Bxe4 Qxe4 16.Nb5! Qg6 17.Rad1 Bd7
18.Nxc7!? Kxc7 19.Bxd6+ Kd8 20.Be7+ Kxe7 21.Qxd7+ Kf8 22.Rd6! Rd8 23.Qxd8+ Nxd8
24.Rxd8+ Ke7 25.Rxh8 Qxc2) 15.Be2 Qd5 16.Bf3 is a draw by repetition.
c) 12.Bh4 Re8 13.Bf3 (stronger than 13.Bh5 Re7 or 13.Re1 h6 14.Rb1 g5 15.Bg3 Qd5) 13...Nxd4
14.cxd4 Qf4 15.Bg3 Qg5 16.h4 (the immediate 16.c4 allows 16...Ne4; while 16.Qd3 is met by
16...Bg4! 17.Bxb7 Rb8, e.g. 18.Qa6 Be2 19.Qxa7 Rxb7 20.Qxb7 Bxf1 21.Rxf1 Qd5 with sufficient
counterplay) 16...Qg6 17.c4 (or 17.h5 Qg5) 17...Bg4 (not now 17...Ne4?! 18.Bh2) 18.c5 Ke7!
(running again) 19.cxd6+ cxd6 20.Re1+ (or 20.Rb1 b6) 20...Kf8 21.Bxd6+ Kg8 22.Be7 (or 22.Be5
Re7) 22...Nd5 23.Bxg4 Rxe7 and Black doesn’t seem to stand so badly; e.g. 24.Bf3 (or 24.Rxe7 Nxe7
25.Qf3 Rb8, or 24.Re5 Qd6 25.Bf3 Nb6) 24...Rxe1+ (or 24...Rd7 25.Re5 Nb6) 25.Qxe1 Qc6 26.Rc1
(or 26.Qa5 Qc3! 27.Qxc3 Nxc3 28.Bxb7 Rd8) 26...Re8! 27.Qd2 Qd7 28.Qa5 Nf4 29.Qxa7 Ne2+
30.Bxe2 Rxe2 with good chances to hold.

C3: 10.Be3

294
 This has been the putative main line ever since GM Zurab Azmaiparashvili won a nice game in the
early days of the variation. However, Black’s defences have been improved considerably.

10...d6

Generally, 10...d6 and 10...Re8 come to the same thing as Black plays both, but I now think the d-
pawn should be pushed first because it allows for ...Bg4 in one variation.
Gutman considered 10...Re8 to be “clearly the best continuation”.

Then:
a) 11.Re1 was Z.Azmaiparashvili-J.Hector, Donostia 1991, where Black got crushed after

295
11...Qd5?! (instead, 11...d6 is correct, transposing below) 12.Qc1! Qf5 13.Qa3 d6 14.Rad1 Re7?
(14...Rxe3 15.fxe3 Qe5 was relatively best) 15.c4 (or 15.Nxc7!? Rxc7 16.Qxd6+ Rd7 17.Qf8+ Kc7
18.Bd3! – Stockfish) 15...Be6 16.Bf3 Rd7 17.Bxc6 bxc6 18.Nd4 Qh5 19.Bg5! c5 (or 19...Qxg5
20.Rxe6) 20.Bxf6+ gxf6 21.Nc6+ Ke8 22.Qb2 Qh4 23.Qb7 Rad8 24.Rd3 1-0, as the rook swings
across to the kingside.
b) 11.c4! is more troublesome, since 11...d6?? now runs into 12.c5 d5 13.c4 and wins. Black has to
resort to 11...b6, which is less appealing; e.g. 12.Qd2 Bb7 13.Bd3 Qg4 14.f3 Qh5 15.a4! a6 16.Nd4
Qa5? (too soon) 17.Qf2 (and 17.Nf5 Qxd2 18.Bxd2 looks even stronger) 17...Nxd4 18.Bxd4 was
good for White in AndyAndyO-tsmenace, ChessWorld.net 2017. I should have played 16...Qe5
17.Nxc6+ Bxc6 18.Bf2 and then 18...Qa5. Even so, 19.Qxa5 bxa5 20.Bd4 gives White the better
chances, bearing in mind that the a5-pawn can be removed at any time.

11.Re1

Instead:
a) 11.c4 Bg4! is the reason for playing 10...d6 first; e.g. 12.c5 Bxe2 13.Qxe2 dxc5 14.Nc3 Qe5
15.Qc4 a6 16.Ne2?!, R.Pegg-A.Chigishev, corr. 2002, and after 16...b6! Black is clearly better.
b) 11.Bf3 Qc4 12.Rb1 is why Gutman preferred to develop the rook first, but 12...Re8 still seems
okay; e.g. 13.Nxd6 (or 13.Bg5 Qc5 14.Qd2 Ne5) 13...cxd6 14.Qxd6+ Nd7 (not now 14...Bd7?
15.Rxb7) 15.Bd5 Qxc3 16.Bxf7 Qf6 17.Qxf6+ gxf6 18.Bxe8 Kxe8 19.Rfd1 Nde5 20.Rd6 Kf7
21.Bd4 Nxd4 22.Rxd4 b6 and Black is not worse, S.Kronberg-K.Kögler, corr. 2000.
Or if 12.Nxd6!? cxd6 13.Qxd6+ Bd7 14.Rab1 Qe6 15.Qa3, AndyAndyO-draco69, ChessWorld.net
2017, then 15...Kc8! should hold; e.g. 16.Rfd1 Re8 17.Rd6 (or 17.h3 Qe5, or 17.Bf4 Re7) 17...Qc4
18.h3 Rxe3!? 19.fxe3 Rb8.
c) 11.Rb1 Re8 is transpositional: 12.Re1 is the main line, while 12.Bf3 Qc4 is note ‘b’ above. If
instead 12.Qd2 Re7 (Stockfish goes for 12...Qg6, and if 13.Bg5 then 13...Kd7!?, threatening ...Ne4)
13.Rfe1, S.Sepesi-R.Hros, Slovakian League 2018, then 13...Qd5 14.Bd3 Ne5 is level.
And 12.Qc1 can be met by 12...Qg6!, threatening 13...Bh3 14.Bf3 Ne5, or if 13.Rd1 Bh3 14.Bf1
then 14...Ne5; e.g. 15.Nd4 Bg4 16.Ne2 Bxe2 17.Bxe2 b6. In A.Borwell-W.Mai, corr. 2007, Black
reversed the move order with 12...Ne5 13.f3 Qg6 though should still be okay.

11...Re8

296
 Here GM Gutman wrote: “I feel that Black’s position is strong enough to confront White’s threats.”
Subsequent praxis and analysis appear to bear this out.

12.Rb1

This is Gutman’s main line and the move recommended by GM Vladimir Barsky. There are
numerous other possibilities, which we’ll deal with piece by piece:
a) 12.Nxd6? is premature: 12...cxd6 13.Qxd6+ Bd7 14.Rab1 (or 14.Bf3 Qe7! as in note ‘c’ below)
14...Qd5 15.Qg3 Kc8 0-1 V.Hergert-P.Leisebein, corr. 1997.
b) 12.Nd4?! meets with 12...Bg4! 13.Nxc6+ (or 13.Rb1 Bxe2 14.Qxe2 b6) 13...bxc6 14.Bxg4 Qxg4
15.Qd3, N.Weinberger-P.Leisebein, corr. 2000, when 15...Qe4 looks best, with a clear advantage.
c) 12.c4 again prompts 12...Bg4! 13.c5 (not 13.Nxd6?! cxd6 14.Qxd6+ Bd7 15.Bd3, V.Belikov-
R.Kholmov, Moscow 1996, where 15...Qe7! 16.Bb6+ axb6 17.Rxe7 Rxe7 is good for Black)
13...Bxe2 14.Qxe2, Eri.Rasmussen-H.Lykke, corr. 2000, and now 14...a6! 15.Nc3 Qb4 gives Black
an edge; e.g. 16.Qd3 (or 16.Qd2 Re6 17.Rab1 Qc4 18.Rxb7 Kc8 19.Rb3 dxc5) 16...Ne5 17.Qd2 Nc4
18.Qd4 Qxc5 19.Qxf6+ gxf6 20.Bxc5 dxc5 21.Nd5 Rxe1+ 22.Rxe1 f5 23.Nf6 Nd6.
d) 12.Qc1 is well met by 12...Qg6! (Leisebein); e.g. 13.c4 (not 13.Bd3? Bf5 14.Bxf5 Qxf5 15.Nd4
Nxd4 16.cxd4, B.Maassen-D.Rosner, corr. 1998, when 16...Ke7 and ...Kf8 consolidates; while if
13.Qa3, C.O’Bee-Mi.Lane, corr. 2000, Stockfish gives 13...Rxe3!? 14.fxe3 Ne5 as good for Black,
e.g. 15.Bd3 Qg5 16.Nd4 c5 and Black certainly has all the play) 13...Bd7 14.Qa3?! (or 14.c5 a6)
14...Ng4 15.Bd2 a6 16.Rad1?! Rb8 17.Nxd6 cxd6 18.Bf4 Nge5 19.Rxd6 Re6 20.Rdd1 Ke8 21.Bg3
Rd8 and Black won, chess1650-tsmenace, ChessWorld.net 2018.
e) 12.Qd2 asks for 12...Qd5! 13.Bd3 Ne5 14.Bg5 Be6!? (deviating from 14...Bd7 15.Nd4 Nxd3
16.cxd3 Rxe1+ 17.Rxe1 c5 18.Nf3 b5 19.c4 bxc4 20.Qa5+ with a draw, Reprimand-tsmenace,
ChessWorld.net 2018) 15.Nd4 Nxd3 16.cxd3 Kd7 17.Be3 Qa5 18.Qb2 Nd5 19.Bd2 Qb6 20.Nb3 Ne7

297
21.c4 Nf5 22.Rab1 Qa6, intending ...b7-b6, and Black was no worse, Pawnhunter1-tsmenace,
ChessWorld.net 2021.
f) 12.Bd3 Qh4 13.Qd2 (improving on 13.g3?! Qg4 14.Be2 Qg6 15.c4?! Ng4 16.Bf4 Qh5 17.Bxg4
Rxe1+ 18.Qxe1 Bxg4, when the light squares were a permanent problem, Rol.Markus-W.Baer, corr.
1999) 13...h6 14.Bf4 (or 14.c4 Bd7) 14...Be6 15.Be4?! (15.Bf1 is a better way to clear the file)
15...Nxe4 16.Rxe4 Qe7 17.Rae1, Rol.Markus-A.J.Reuter, corr. 1999, and now 17...Qd7 keeps Black
solid with an edge.
g) 12.Bf3 Qc4 13.Nxd6!? (or 13.Be2 Qd5) 13...cxd6 14.Qxd6+ Bd7 15.Rad1 shouldn’t gain more
than a draw after, for example, 15...Rc8 (or 15...Qe6 16.Bb6+ axb6 17.Rxe6 fxe6) 16.Bg5 Rxe1+
17.Rxe1 h6 18.Bxf6+ gxf6 19.Re3 b6 20.Bg4 Qxg4 21.Rg3 Qg5.
If White plays 13.Rb1 first, Black can pre-empt the sacrifice with 13...Bd7! (13...Re7 returns to the
main line); e.g. 14.Qd2 Ne4 15.Qc1, N.Janzen-J.Ryska, corr. 2003, and now 15...f6 16.Be2 (or
16.Qa3 b6) 16...Qf7 with an edge.
h) 12.Bg5, as in Jos.Jimenez-S.Aiguasenosa, corr. 1999, gives Black a whole tempo (...Re8) on line
C4, such that 12...Qe5!? is now a possibility; e.g. 13.Bh4 (or 13.Qd2 h6 14.Bf4 Qd5) 13...g5 14.f4 (or
14.Bg3 Qc5) 14...Qxf4 15.Bg3 Qe3+ 16.Bf2 Qf4 with at least a draw.

12...Re7!?

A useful move, defending the c7-pawn and enabling the king to run via e8. But it doesn’t prevent
the sacrifice on d6.
To that end, 12...Bd7 seems quite satisfactory; e.g. 13.Nd4 (or if 13.c4, G.Vescovi-A.Rodriguez
Vila, Sao Paulo 2005, then 13...Ng4 14.Rb3 Qf5 15.Nd4 Nxe3 16.Rxe3 Qg5) 13...Nxd4 14.cxd4,
C.Van Wieringen-B.Grabowski, corr. 2011, and now 14...Bg4 (despite the wasted tempo) 15.Bxg4
(or 15.h3 Bxe2 16.Qxe2 h6) 15...Qxg4 16.f3 Qf5 17.Rxb7 Nd5 18.Rb3 Ke7 and ...Kf8 with equality.

298
13.Bf3

Perhaps the best try in this variation, as it requires Black to defend precisely.
Other moves:
a) 13.c4 Bg4! 14.Nc3 (or 14.c5 Bxe2 15.Rxe2 d5) 14...Bxe2 15.Qxe2 Qg6 16.Rxb7?! Nd4
(Gutman) 17.Qd1 Nxc2 18.Re2 Nxe3 19.fxe3, Z.Stypka-K.Grabianowski, Poznan 2008, and now
19...Re5! is good for Black.
b) 13.Bd3 Qg4 (or 13...Qh4) 14.f3 Qh5 15.Qd2 (not 15.g4?! Nxg4 16.fxg4 Bxg4 17.Qd2 Ne5)
15...a6 16.Nd4 Ne5 17.c4 Ke8 18.Bg5 Nxd3 19.cxd3 h6 20.Bxf6 gxf6 21.Ne2 Kf8 22.Ng3 Qg6
23.Rxe7 Kxe7 24.c5 dxc5 25.Qe3+ Kf8 26.Qxc5+ Kg8 27.Qc3 f5 28.Qxc7 b5 led to a draw in
rooksac-tsmenace, ChessWorld.net 2018.
c) 13.Nd4 Nxd4 14.cxd4 Bg4 15.Bxg4 Qxg4 16.Qxg4 (not 16.Qd2?! b6 17.Bg5 Rxe1+ 18.Rxe1
Kd7 19.Bxf6 gxf6 20.Qb4 Qh5 21.Qa4+ b5 22.Qa5 Qd5 and Black is better, C.Callow-L.Henris, corr.
2002) 16...Nxg4 17.Rxb7 Kd7 is level, as shown by 18.h3 Nxe3 19.Rxe3 (or 19.fxe3 Kc6) 19...Rxe3
20.fxe3 Kc8 21.Rb3 Rb8 22.Kf2 Rb6, P.Leisebein-L.Gutman, corr. 1999.

13...Qc4

Not 13...Qh4??, as in M.Kreuzholz-A.Jochens, Kiel 2000, because of 14.Nxd6! cxd6 15.Qxd6+

Nd7 (or 15...Rd7 16.Qf8+ Kc7 17.g3) 16.Bxc6 bxc6 17.Bf4 and White wins (Barsky).

14.Nxd6 cxd6 15.Qxd6+ Bd7

299
 White had nothing better than to sacrifice anyway, but in this position the black queen can
participate fully in the defence. For example:
a) 16.Bg5 Rc8! (the key defensive move) 17.Bxf6 gxf6 18.Qxf6 Qc5! 19.Rxb7 Qd6 (19...Ne5 is
also okay) 20.Qh8+ (or 20.Rxe7 Qxe7 21.Qh8+ Qe8 22.Qf6+ Qe7 with a draw) 20...Re8 21.Rxe8+
Bxe8 22.Qxh7 Ne5 23.Be2 (or 23.Qh4+ f6) 23...Rxc3 24.Qh4+ f6 25.Rxa7 Rxc2 26.Ra8+ Rc8
27.Rxc8+ Kxc8 and the “0.00” assessment appears in all lines that don’t give something away.
b) 16.Rxb7 Qxc3 17.Rd1? (17.Kf1 is preferable) 17...Ke8?? (here 17...Rc8! is just good for Black)
18.Rxd7 Nxd7 19.Bxc6 Rd8 20.g3 Qe5 21.Qd3 Qc7 22.Ba4 Qa5 23.Bb5 f6? 24.Bc5 Re1+ 25.Kg2
Rxd1 26.Qe4+ 1-0 A.Niknaddaf-M.Asbahi, Urumia 2008.
c) 16.Bxc6 Qxc6 17.Bb6+ axb6 18.Qxe7+ Kc7 19.Qxf7 Qd5 20.Re7 (or 20.Qxd5 Nxd5 21.c4 Nf6)
20...Qxf7 21.Rxf7 Kc6!? (21...Rg8 should be a draw) 22.Rxg7 Rxa2 23.Rf7 Ne4 with an unclear
endgame (which Black won), B.Hjort-R.Zajontz, corr. 1999.

C4: 10.Re1

300
 K&S call this “White’s most logical move”, but I think their alternative move order 10.Bg5 d6
11.Re1 should be preferred if White wants to play the main line below, because 10.Re1 gives Black a
good way to avoid it in note ‘d’ below.

10...d6?!

The dubious mark is for missing the chance to deviate. I’ve actually played this seven times, so if
“?!” is a half-bad move, I deserve the annotation “????!”. Instead:
a) 10...Re8?! 11.Bg5! (rather than 11.Be3 d6, as in line C3) 11...a6 (for 11...d6 see 11...Re8 below)
12.Nd4 Nxd4 13.cxd4 Qf5 14.Bh4 is quite depressing for Black, who has forced no concessions; e.g.
14...h6 15.c4 a5 16.c5 a4 17.Rc1 g5 18.Bg3 Nd5 19.Be5 d6 20.cxd6 cxd6 21.Bxd6 and White won,
Rodent III-Bagatur 1, TCEC 16 Qualification 2019.
b) 10...h6?! prevents Bg5 but runs into 11.Bf3! (rather than 11.Rb1?! Re8 12.Be3 d6 13.c4 Bd7
14.Nc3 Qe5 15.Qd2 b6 16.Bf3 Qa5 17.Rb5 Qa3 18.Rb3 with a repetition, tsmenace-AndyAndyO,
ChessWorld.net 2017) 11...Qf5 (or 11...Qa4 12.Rb1 Re8 13.Rxe8+ Nxe8 14.Bd5 d6 15.Bxf7)
12.Bxc6 bxc6 13.Nd4 Qd5 14.Re5! Qc4 15.Qf3 Bb7 16.Rb1 Ba6 17.Bf4 and White is clearly better,
as shown by Confused_by_Theory at ChessPublishing.
c) 10...Qf5 asks for 11.Nd4! (11.Bd3 Qh5! isn’t so bad) 11...Nxd4 12.cxd4 with a promising
version of line C1 for White, who may continue with c2-c4 or perhaps a2-a4 and Ra3, bringing the
rook across to hassle the black queen and kingside.
d) 10...Qd5! is the strongest option here.

301
 Then:
d1) 11.Qxd5 Nxd5 12.Bc4 Nce7 13.Rd1 a6 14.Nd6 cxd6 15.Bxd5 Nxd5 16.Rxd5 b5 17.Be3 Bb7
18.Rxd6 Bc6 19.Re1 Rb8 20.Bf4 Rb7 was level in M.Mylnikov-S.Zvereva, St. Petersburg 2012.
d2) 11.Bd3 d6 (rather than 11...Re8?! 12.Be3 a6 13.c4 Qe5 14.Bd2 Qc5 15.Rxe8+ Nxe8 16.Nc3 h6
17.Ne4 Qf8 18.Bc3 with an edge for White – K&S, following C.Höher-Ru.Bauer, corr. 2014) 12.Bf4
(or 12.c4 Qh5) 12...Re8 13.Rxe8+ Nxe8 14.c4 Qc5 15.Be3 (or 15.Qd2 Be6 16.Be3 Qb4 17.c3 Qa5
18.Qc2 Nf6 19.Nd4 Ne5 20.c5 Nxd3 21.Qxd3 Ke7, B.Näter-C.Sausner, corr. 2014) 15...Qe5 16.c3
Bf5 17.c5 Qd5 18.Bxf5 Qxf5 19.c4 a6 20.Nc3 Kc8 21.Nd5 dxc5 22.Bxc5 Nf6, Mar.Richter-B.Näter,
corr. 2014, and Black drew without difficulty.
d3) 11.Bf4 d6 12.c4 (instead, 12.Bd3 is ‘d2’ above, while 12.Nd4 and 12.Qc1 can both be answered
by 12...Re8) 12...Qxd1 13.Raxd1 a6 (not 13...b6? 14.c5! bxc5 15.Nxc7! and wins) 14.Nxc7!? (14.Nc3
b6 15.Bf3 Bd7 16.Bg5 h6! 17.Bxf6+ gxf6 is equal) 14...Kxc7 15.Bxd6+ (not 15.Rxd6? g5!) 15...Kd8
(15...Kb6 16.Rb1+ Ka7 17.Bc5+ Kb8 18.Bd6+ is a draw) 16.Bf8+ Nd7 17.Bxg7 Rg8 18.Bb2 Re8 is
unclear but Black shouldn’t be worse.
d4) 11.Nd4 Re8 12.Be3

302
 12...Ne5! (not 12...d6?! 13.c4 Qa5 14.Rb1 with a “powerful initiative for the sacrificed pawn” –
K&S, who give the further 14...Qxa2?! 15.Nxc6+ bxc6 16.Ra1 Qb2 17.Bd4 Qb4 18.c3 Qb2 19.c5
with the clear advantage symbol, which is being kind to Black) 13.Nf5!? (perhaps the best try to show
something; 13.f4 Nc4 14.Bf2 Ne3 15.Bxe3 Rxe3 16.Bf3 Rxe1+ 17.Qxe1 Qc5 18.Qe3 Qe7 was level
in B.Näter-Mar.Richter, corr. 2014; and 13.Bg5 d6 14.Qd2 c5 15.f4 cxd4 16.fxe5 dxe5 17.Bf3 e4
18.Bxf6+ gxf6 19.c4 Qxc4 20.Bxe4 Qc3 didn’t get anywhere either, F.Fritsche-Mar.Richter, corr.
2014) 13...d6 14.Nxg7 Rg8 15.Qxd5 (or 15.Bg5 Nf3+! 16.Bxf3 Qxg5 17.Qe2 Ng4; not yet 15.Bh6??
Qxd1 0-1 Reg-UK-tsmenace, ChessWorld.net 2021, in view of 16.Raxd1 Nfg4 17.Bg5+ f6, winning a
piece) 15...Nxd5 16.Bh6 and now Stockfish proposes 16...b5!? as a way to activate Black’s forces;
e.g. 17.Bxb5 (or 17.Bh5 c6 18.f4 Nxf4 19.Bxf4 Rxg7 20.Rad1 Nc4 21.Bh6 Bg4 22.Bxg7 Bxh5)
17...Rb8 18.Ba4 (or 18.Be2 Nxc3 19.Bh5 f6) 18...Ng4 19.Bg5+ f6 20.Bh4 Ne5 21.Rad1 Ng6 22.Rd4
Nxh4 23.Rxh4 Nxc3 24.Nh5 Nxa4 25.Rxa4 Bd7 26.Rxa7 Bc6 27.g3 f5 with compensation, based on
...Re8 and the open b-file.

303
11.Bg5!

The strongest continuation.

White has also tried:
a) 11.Be3 Re8, which is line C3 once more.
b) 11.Ba3 allows 11...Qd5! again, intending 12.Nd4 Re8, or if 12.Qc1, K.Wolczyk-M.Holownia,
Polish League 2014, then 12...Re8 13.Nxc7 (or 13.c4 Qf5) 13...Kxc7 14.Rd1 Qe5 15.Bxd6+ Qxd6
16.Rxd6 Kxd6 with plenty for the queen.
c) 11.Nxd6 cxd6 12.Qxd6+ (for 12.Bf3 Qg6 see note ‘d’) 12...Bd7 13.Bf4 Rc8! enables Black to
defend; e.g. 14.Bf3 Qf5 (or 14...Qc4) 15.Bxc6 (or 15.g4 Qg6) 15...Rxc6 16.Qb8+ Bc8 17.Rad1+ Nd7
18.Re3 (not 18.Rd2? h6! 19.Bg3 Re6 20.Red1 Ke7 21.Rxd7+ Bxd7 22.Qxh8 Qxc2 23.Bh4+ g5
24.Qd4 Qa4 and Black comes out clearly better, tripoduk-jatait47, Chess.com 2020; this defence isn’t
possible after 18.Re3) 18...f6 19.Red3 Ke7 20.Rxd7+ Bxd7 21.Qxh8 Kf7 22.Qd8 Be6 23.Bd6 Qxc2
24.Re1 Qxc3 25.Qe7+ ½-½ remyrey-tsmenace, ChessWorld.net 2019.
d) 11.Bf3

304
 can be answered by 11...Qg6 (instead, 11...Qf5 12.Nd4! Nxd4 13.cxd4 Re8 14.c4 transposes to
12...Qf5 in line C1) 12.Nxd6 (now 12.Nd4 Nxd4 13.cxd4 Re8 14.c4 is 14.Re1 in line C1; 12.Rb1 Re8
13.Rxe8+ Nxe8 14.Nd4 Ne5 15.Bxb7 Bxb7 16.Rxb7, R.Do Amaral-K.Amann, corr. 2007, and
16...Qe4 17.Rb1 Ke7 is fine for Black; or if 12.Bf4 then 12...Re8!, e.g. 13.Qd2 Ne4 14.Bxe4 Rxe4
15.Rxe4 Qxe4 16.Re1 Qg6 17.c4 a6 18.Nc3 b6) 12...cxd6 13.Qxd6+ Nd7! (Gutman); e.g. 14.Qd5
Re8 15.Bf4 Re6! 16.Rad1 (or 16.Be4 Ne7!) 16...Ke8 17.Qd6 Qf6! 18.Bxc6 bxc6 19.Qxc6 Qxf4
20.Qxa8 Qc7 with messy equality.
e) 11.c4 is mentioned by Dembo & Palliser, with the comment “White has a strong attack”, and
might well be tricky to defend over the board.

Notice that after 11...Bd7 12.Be3 Re8 White has improved on 12.c4 in line E3, since Black has

305
played ...Bd7 rather than ...Bg4. Nonetheless, Stockfish isn’t overly concerned; e.g. 13.Qd2 Qg6 14.f3
a6 15.Nxd6!? Rxe3 16.Nxb7+ Kc8 17.Qxe3 Kxb7 18.Rab1+ Kc8 19.Qc5 Qxc2 20.Qf8+ Nd8
21.Qxg7 Qg6 22.Qh8 Ne8 looks quite unclear.

11...Bd7

I think this is best, aiming for some sort of solidity. Instead:

a) 11...Re8 12.Qd2 builds up straightforwardly; e.g. 12...Re5 (or if 12...Qa4 13.Nd4, as in J.Bucek-
P.Nekola, corr. 2002, and 13...Nxd4 14.cxd4 Be6, then 15.Rab1!, threatening Bb5 or Rxb7 – K&S)
13.Nxd6! cxd6 14.Bf3 Qf5 15.Qxd6+ Qd7 16.Qf8+ Qe8 17.Red1+ Bd7 18.Qxg7 Rxg5 19.Qxg5 and
White won, Ron.Fischer-P.Ster, corr. 2002.
b) 11...Qf5 was suggested by GM Gutman. I have both won with it and failed to win against it, but
after 12.h4! (rather than 12.Bh4 Re8; while for 12.Qd2 see 11.Qd2 in line C2), White is objectively
better:

b1) 12...Re8 13.Qd2! Re6 (or 13...Be6 14.Nd4 Nxd4 15.cxd4 – K&S) 14.Nd4 Nxd4 15.cxd4 h6
16.Bd3 Qg4 17.Bxf6+ Rxf6 18.Re4 Qh5 19.Rae1 Re6 (or 19...Bd7 20.Bc4 c6 21.d5) 20.Qf4 Rxe4
21.Bxe4 Rb8 22.c4 is terrible for Black, in view of the useless rook and vulnerable king.
b2) 12...h6 runs into 13.Nxd6! (simply 13.Bd3, as in R.Sherwood-S.Schere, corr. 2006, is not bad
either) 13...cxd6 14.Qxd6+ Bd7 15.Bd3 Qa5 16.Bxf6+ gxf6 17.Rab1 Qxc3 18.Bf5 Qd4 19.Qg3 Bxf5
(not 19...Kc8? 20.Qb3! Na5 21.Qb5! – K&S) 20.Red1 Qxd1+ 21.Rxd1+ Ke8 22.Rb1 and Black’s
uncoordinated pieces are not worth the queen, J.Van Gool-I.Kochan, corr. 2013.
b3) 12...Bd7 13.Nd4! Qc5 (or 13...Nxd4 14.Qxd4! Qe5 15.Qb4, threatening Qxb7 or a discovery on
the queen) 14.Rb1 Kc8 15.Bxf6 (not 15.Nxc6 Bxc6 16.Bf3 b6 17.Qd3? Bxf3 18.Qxf3 Qd5 19.Qxd5
Nxd5 20.c4 Nf6 21.Re7 Rf8 22.Bxf6 gxf6 with a level endgame, craig4000-tsmenace,
ChessWorld.net 2019) 15...gxf6 16.Bf3 (and not 16.Bg4?! Ne5 17.Bxd7+ Kxd7 18.f4 Rag8 19.fxe5

306
fxe5 20.Rb5 Qc4 21.Qe2 Qxe2 22.Nxe2 b6 23.Rf1 Ke6, when White cannot make progress,
tsmenace-juliangon, ChessWorld.net 2017) 16...Rb8 (or 16...Qxc3 17.Nxc6 bxc6 18.Rb3 Qa5
19.Ree3, intending Qd4) 17.Rb5 Qxc3 18.Nxc6 bxc6 (not 18...Bxc6? 19.Bg4+ Kd8 20.Rb3 Qc4
21.Rd3) 19.Re3 Qc4 20.Rxb8+ Kxb8 21.Qb1+ Ka8 (not 21...Kc8? 22.Re4 and Rb4) 22.Re7 Rb8 (or
22...Rd8 23.Qb2) 23.Qf1! Qxc2 24.Rxd7 Rb1 25.Bxc6+ Qxc6 26.Qxb1 Qxd7 27.Qxh7 Qg4 28.h5
and White should win with the passed h-pawn.

12.Nd4

As played by Leela. Instead:

a) 12.Bf3 Qf5 or 12.Bd3 Qg4 doesn’t get White anywhere.
b) 12.Rb1 Qf5 13.h4 h6 is an improvement for Black on note ‘b3’ above; e.g. 14.Bd3 Qc5
15.Bxf6+ gxf6 16.Nd4 (or 16.Qf3 f5) 16...Qxc3 17.Nxc6+ Bxc6 18.Qg4 b6 19.Bf1 (or 19.Qg7 Re8
20.Rxe8+ Kxe8 21.Qxh6 Ke7) 19...Qa5! 20.Qg7 Re8 21.Qxf6+ Kc8 22.Rxe8+ Bxe8 23.Qf3 (or
23.Bc4 Bc6) 23...c6 24.Qf6 Kc7 25.Qxh6 Bd7 26.Rd1 Qe5 is all rather unclear.
c) 12.Bxf6+ gxf6 13.Bf3 (if 13.Bh5 then 13...Qc4 14.Qf3 Be6 15.Qxf6+ Kd7) 13...Qg6 14.Rb1 (not
14.Qd2? Ne5 15.Bxb7 Rb8 or 15.Nd4 Nxf3+ 16.Nxf3 Bc6) 14...Rb8 (not 14...Re8? 15.Rxe8+ Bxe8
16.Nxd6! cxd6 17.Qxd6+ Kc8 18.Re1 Bd7 19.Rd1 Ne5 20.Qf8+ Kc7 21.Qc5+ Kd8 22.Bxb7 and
wins) 15.Nd4 Nxd4 16.cxd4 Re8 looks easier for Black than the main line below.
d) 12.Qd2 h6! 13.Bxf6+ gxf6 14.Bf3 (or 14.Rad1 Ne5) 14...Qf5 strangely improves on that for
Black; e.g. 15.c4 (or 15.Nd4 Qg5) 15...Rg8 16.Bd5 a6 17.Nd4 (or 17.Nc3 Ne5) 17...Nxd4 18.Qxd4
Bc6 19.Bxc6 bxc6 20.c5 (or 20.Qe3 Qxc2 21.Qe7+ Kc8 22.Qxf7 Qg6) 20...d5 21.Qe3 Kc8 22.g3 (or
22.Qxh6 Kd7) 22...Rb8 23.Qxh6 Rb2 with counterplay.

307
12...Nxd4

If 12...Qe5 then 13.h4! Qc5 returns to 11...Qf5 above; and 13...Re8 14.Qd2 is no better.

13.Bxf6+

Before 13.cxd4 Qf5 preserves Black’s pawn structure, intending 14.Bh4 (or 14.h4 h6) 14...Re8
15.c4 (or 15.Qd2 Re4) 15...b6.

13...gxf6 14.cxd4 Re8 15.Qd2

308
 This position may be critical and yet has hardly ever been tested – just twice to my knowledge. To
misquote K&S on a different position: “White has more than sufficient compensation for the pawn,
thanks to [a structural] advantage and the misplacement of Black’s king, which impedes [the] rook on
a8 to enter the actions.”
The limited praxis has seen:
a) 15...Rb8 16.Qc3 Qf4 17.Bf3 Rxe1+ (or 17...Re6 18.Bd5) 18.Rxe1 Qg5 19.Qb3 Qa5 20.Re3 Qb5
21.Qc3 Qg5 22.d5 h5 23.Qd4 c5 (not 23...b6? 24.h4! Qg6 25.Be2, intending Bd3 and Qe4; but
23...h4!? 24.Qxa7 Kc8 25.h3 b5 or 25.c4 f5 might be considered) 24.dxc6 bxc6 25.Re1 Kc7 26.h4 (or
26.Qxa7+ Rb7 27.Qd4) 26...Qc5 27.Qxf6 Qf5 28.Qxf5 Bxf5 29.Re7+ Kb6 30.Bxh5 Rh8 31.Bxf7
Rxh4 32.Bb3 and White won by (eventually) pushing the kingside pawns, LcZero-Stockfish, TCEC
18 Superfinal 2020.
b) 15...Qh4 16.Bf3 Bc6 17.g3 Qg5 18.Qd3 Qg6 19.Bxc6 bxc6 20.Qc4 (regaining the pawn, since
20...Kd7?? fails to 21.d5!) 20...Rb8 21.Qxc6 Re6 22.Qc4 Qf5 23.d5 Re5 24.f4 Rxe1+ 25.Rxe1 Rb6
26.h4 a6 (or 26...Qh3 27.Kf2 Qh2+ 28.Kf3) 27.Kh2 h6 28.Re2 Kd7 29.Re4 Kd8 30.Qd3! (forcing the
black queen to retreat) 30...Qd7 31.Re3 Rb5 32.a4 Rb4 33.a5 Rb5 34.Re2 (avoiding 34.c4? Rb2+)
34...Kc8 35.c4 Rb4 36.c5! dxc5 37.Qxa6+ Kb8 38.Qxf6 Qxd5 39.Qa6 (threatening Re8+ and mates)
39...Re4 40.Rc2 Rd4 41.Qb5+ Ka8 42.Rxc5 Rd2+ 43.Kh3 Qh1+ 44.Kg4 (the white king is perfectly
safe) 44...Qd1+ 45.Kf5 Rd6 46.Qe8+ Kb7 47.a6+! (a decoy/obstruction sacrifice; not yet 47.Rb5+?
Ka6) 47...Rxa6 48.Rb5+ Rb6 49.Qe4+ c6 50.Qe7+ Ka6 51.Rxb6+ Kxb6 52.Qb4+ Ka6 53.Kf6 Qg4
54.Ke7! (going across to remove the c6-pawn and any counterplay) 54...Qe6+ (54...Qxg3 55.Kd7
Qxh4 56.Kc7 forces mate) 55.Kd8 f5 56.Kc7 Qf7+ 57.Kxc6 Qe8+ 58.Kd5 Qg8+ 59.Ke5 Qxg3
60.Qf8 1-0 verdi-tsmenace, ChessWorld.net 2021, since 60...Qxh4 61.Qf6+ is a lost pawn ending.
c) 15...Qg6!? is another possibility, when 16.Bf3 (if 16.Bd3 then 16...Qg5, or 16.c4 Qg5 17.Qd3 c6
18.Qxh7?! Kc7 19.Qxf7 Qd2) 16...Bc6 17.Bxc6 bxc6 18.Qb4 Re6 19.Qb7 Rc8 20.Qxc6 Rb8 is note
‘b’ minus g2-g3, which means Black can try a more active defence; e.g. 21.Qc4 (or 21.c4 Qc2)
21...Rb2!? 22.Rxe6 fxe6 23.Qxe6 Qxc2 24.Qxf6+ Kc8 (or 24...Kd7!? 25.Qf7+ Kc6 26.h4 a5) 25.h4
Rb1+ (not 25...Rxa2?? 26.Qe6+ etc) 26.Rxb1 Qxb1+ 27.Kh2 Qxa2 28.Qh8+ Kb7 29.Qxh7 Qxf2.
Okay, White can be satisfied with the 2/2 there, but it’s clear that keeping control in these
endgames requires great precision, so I’d be disappointed to lose over the board. In any case, I note
that my own score after 9...Nf6 is P17, W11, D5, L1, which is close to 80% as Black. I like those
odds, even having just told my opponents what to do.

309
 Chapter Eight
Two Knights Defence

1.e4 e5 2.Nf3 Nc6 3.Bc4 Nf6

Knights before bishops! And that’s all I have to say about that.
White now has three main options. Of these, 4.Ng5 is the subject of the next chapter. Here we’ll
mostly look at:
A: 4.d4
B: 4.d3
Four lesser moves should be examined too:
a) 4.c3? looks as if White was expecting an Italian, replied automatically, and then noticed Black
developed the knight instead. After 4...Nxe4!

310
 White is already worse and will have to give up the c4-bishop just to regain the pawn; e.g. 5.0-0 (or
5.Bd5 Nf6 6.Bxc6 dxc6 7.Nxe5 Bd6 8.d4 0-0 with an edge; not 5.Qe2?! d5 6.Bb5 f6 7.d4 Bg4, or
5.Qb3? Nd6 6.Bd5 e4 7.Nd4 Nxd4 8.cxd4 c6 and White resigned quite soon, Da.Dubov-Ki.Georgiev,
Titled Arena blitz 2021) 5...d5 6.Bb5 Bd6 7.d3 Nf6 8.Re1 0-0 9.Bxc6 bxc6 10.Nxe5 Bxe5 11.Rxe5
Ng4! 12.Re2 Qd6 13.g3 Qg6 and White has permanently weak light squares, J.Arnous de Riviere-
G.Neumann, Berlin 1864.
b) 4.0-0?! also meets with 4...Nxe4.

b1) 5.Nc3 transposes to note ‘c3’ below.

b2) 5.d3 Nf6 6.Re1 regains the pawn but leaves White worse after 6...Be7 7.Nxe5 Nxe5 8.Rxe5 d5
9.Bb3 0-0.

311
 b3) 5.d4? d5 6.Bb5 exd4 7.Nxd4 Bd7 keeps the extra pawn, while 8.Re1 Bd6 9.Bxc6 bxc6 10.f3
fails to 10...Qh4! 11.fxe4 Bg4!; e.g. 12.exd5+ Kd7 13.dxc6+ Kc8 14.Qd2 Qxh2+ 15.Kf1 Bg3 16.Qe3
Qh1+ 17.Qg1 Qxg1+ 18.Kxg1 Bxe1 and wins.
b4) 5.Re1? d5 6.Bb5 Bc5 is pretty much winning too; e.g. 7.d4 (or 7.Re2 Nxf2 8.Rxf2 Bxf2+
9.Kxf2 e4 “with a crushing attack” – Lokander) 7...exd4 8.Nxd4 0-0 9.Be3 (or 9.Bxc6 Nxf2! 10.Kxf2
Qh4+ 11.Kg1 Qxd4+) 9...Nxd4 10.Bxd4 Nxf2! 11.Bxf2 Bxf2+ 12.Kxf2 Qf6+ 13.Qf3 Qb6+, followed
by ...Qxb5, with two extra pawns.
b5) 5.Qe2 (relatively best) 5...d5 6.Bb5 Bg4 (if 6...Bd6 then 7.c4! Be6 8.d3 Nf6 9.Nxe5) 7.Nc3!
(instead, 7.Bxc6+ bxc6 8.Qa6 f6! 9.Qxc6+ Kf7 10.d4 exd4 11.Nxd4 Bc5 12.c3 Qd7, H.Noordhoek-Li
Ruofan, Isle of Man 2004; and 7.d3 Bxf3 8.gxf3 Nd6 9.Qxe5+ Qe7 10.Qxe7+ Bxe7 11.Ba4 Nf5,
C.Mispireta Mercado-E.Landes, corr. 2017, or 10.Qxd5 Nxb5 11.Qxb5 0-0-0 are good for Black)
7...Bxf3 8.Qxf3 Nf6 9.Qe2 (rather than 9.Re1?! Qd6 10.Qf5 Nd7 11.d4, L.Dimapilis-T.Sambrook,
corr. 2000, when 11...Qe6! 12.Qxe6+ fxe6 13.dxe5 Nd4 gives Black an edge) 9...Be7 (or 9...Qd6
10.Bxc6+ bxc6 11.d4 e4 12.f3) 10.Qxe5 0-0 11.Bxc6 bxc6 12.d4 and White has more or less
equalized.
c) 4.Nc3 still allows 4...Nxe4!, thanks to the fork trick.

c1) 5.Bxf7+? is just bad, as after 5...Kxf7 6.Nxe4 d5 Black’s centre and bishops far outweigh the
slight inconvenience to the king; e.g. 7.Neg5+ (or 7.Ng3 e4 8.Ng1 h5!) 7...Kg8! 8.d4 h6! 9.Nh3 Bg4!
(threatening ...e5-e4) 10.dxe5 Nxe5 11.Nf4 (or 11.Nhg1 Qd6) 11...c6 12.h3 Bxf3 13.gxf3 Qf6 and
White didn’t last long, M.Novikov-V.Afromeev, Tula 2007.
c2) 5.Nxe4 d5 6.Bd3 (not 6.Bb5? dxe4 7.Nxe5 Qg5 and Black is winning – in comparison with line
A1 in Chapter Six, White lacks any tactical resources; e.g. 8.d4 Qxg2 9.Rf1 Bd6 10.Nxc6 Bd7
11.Nxa7 c6 – Pillsbury, or 11.d5 bxc6 12.Bxc6 Bxc6 13.dxc6 0-0 and Black won, G.Pollard-
W.Dixon, corr. 1903) 6...dxe4 7.Bxe4 is a reversed Three Knights position (see Chapter Six again).
The extra Nf3 is certainly useful, but not to the extent that it offers White the slightest advantage. For

312
example: 7...Bd6 8.0-0 (instead, 8.d4 exd4 9.Bxc6+ bxc6 10.Qxd4 0-0 is technically equal, though I
prefer Black with the two bishops; e.g. 11.0-0 c5 12.Qc3 Bb7 13.b3 Qd7 14.Ne5 Qf5 15.Nc4 Rfe8
16.Bb2? Qg6 17.g3 Re2 18.Rac1 Rae8 with a clear advantage, V.Hresc-M.Hebden, Cappelle la
Grande 1987; or 11.Be3 Be6 12.0-0 Bd5 13.Ng5?! f5 14.Nh3 Re8 15.Bg5?! Qb8! 16.Rfe1 Qb6
17.Qxb6 cxb6 and Black now has the bishops without the defects, J.R.Koch-Mi.Adams, French
League 2001) 8...0-0 9.Re1 Bg4 10.c3 (not 10.h3?! Bh5 11.g4 Bg6 12.d3 f5 13.gxf5? Bxf5 14.Bxf5
Rxf5 and White’s position is terrible, W.So-S.Mamedyarov, St. Louis rapid 2018) 10...Kh8 (the
immediate 10...f5? drops the e-pawn to 11.Bxc6 bxc6 12.Qb3+ etc) 11.h3 Bh5 and now 12.g4?, as in
Liu Guanchu-Peng Xiongjian, Tianjin 2020, is a mistake because of 12...f5! with a huge attack; e.g.
13.Bxc6 fxg4 14.hxg4 Bxg4 15.Be4 (or 15.Bxb7 Qf6) 15...Qd7 16.d3 Rf5.
c3) 5.0-0 Nxc3 6.dxc3 is a variation of the Boden-Kieseritzky Gambit. Searching the database finds
grandmasters opting for 6...Be7 7.Qd5 0-0 8.Nxe5 Nxe5 9.Qxe5 with complete equality and the
intention, presumably, to outplay White from there. Lacking such technique I’d prefer to keep the
material. To that end, I noticed a laconic comment in the eleventh world champion Robert Fischer’s
My 60 Memorable Games: “After 6...Q-K2! White has no compensation for the pawn.” Well, that’s
grand.
In algebraic notation: 6...Qe7 (which is also recommended by IM Lokander) reminds me of the
Nordwalde Defence to the King’s Gambit (where Black plays ...Qf6 and ...Qxf4 and then tries to shut
up shop with ...d7-d6 and ...f7-f6) but is much better because Black hasn’t wasted so much time with
the queen. For example:

c31) 7.Re1 d6 8.Ng5 (if 8.Nd4 Bd7 9.a4, aimed at ...0-0-0, Black can go the other way: 9...g6 10.a5
a6 11.Bd5 Bg7 12.Qf3 Nxd4 13.cxd4 c6 14.Bb3 0-0 with a clear advantage, G.Plume-W.Löffler,
corr. 2013) 8...Nd8 9.f4 f6 10.fxe5 (if 10.Nf3, M.Novikov-A.Khruschiov, Moscow 2006, then
10...Nf7! – Lokander, reinforcing e5 against any sacrifices) 10...dxe5 11.Qh5+ g6 12.Qh4 Bg7
13.Ne4 Be6 14.Bb3 Bxb3 15.axb3 Ne6 16.Be3 f5 17.Qxe7+ Kxe7 and Black won, Stuttgart-Munich,

313
match 1860.
c32) 7.Ng5 Nd8 8.Be3 (not 8.f4? Qc5+ and wins) 8...d6 (8...h6?! was given in an old edition of
ECO) 9.f4 f6 and now Stockfish comes up with 10.f5!? (rather than 10.Ne4 f5 11.Ng5 e4) 10...fxg5
11.Qh5+, while also supplying a defence: 11...Kd7 12.Bxg5 Qe8 13.Qh4 b5! (gaining time to secure
the king) 14.Be2 (if 14.Bxb5+ then 14...c6 and ...Kc7; or 14.Bb3 c5 15.Bd5 Bb7 or 15.a4 c4 16.Ba2
d5) 14...Be7 15.f6 gxf6 16.Bh5 fxg5 (16...Qf8 17.Bxf6 Rg8 18.Qh3+ Kc6 is a possible way to keep
everything) 17.Bxe8+ Rxe8 18.Qxh7 Bb7 with three good pieces for the queen.
d) 4.Qe2!? defends the e-pawn and has vague ideas of c2-c3, Rfd1 and d2-d4, if given time – which
White won’t be because Black can strike in the centre first. For instance, after 4...Bc5 5.c3 (not
5.Bxf7+? Kxf7 6.Qc4+ d5 7.Qxc5 Nxe4 and Black is clearly better, M.Albin-C.Schlechter, Vienna
1914; instead, 5.0-0 0-0 6.d3 d5 7.exd5 Nxd5 transposes to 7.Qe2 in line B, though 6...Na5!? 7.Nxe5
d5 is an interesting alternative, e.g. 8.exd5 Re8 9.Qd2 Nxc4 10.Nxc4 Nxd5 with activity and the two
bishops as compensation) 5...0-0 and then:

d1) 6.0-0 d5 7.exd5 Nxd5 8.d3 is a favourable version for Black of line B1 (where White would
never play 8.Qe2). Here 8...Bg4 looks like a good reply; e.g. 9.h3 Bh5 10.g4?! Bg6 11.b4 Bb6
12.b5?! Na5 13.Qxe5, O.Dolzhikova-M.Bottema, Gibraltar 2018, and now 13...Nxc4 14.dxc4 Nf6
15.Nbd2 Re8 16.Qf4 Qd3 leaves White’s position a complete shambles. Note that 8.Nxe5? is a
mistake because of 8...Nf4 9.Qe4, A.Sehm-D.Schiemann, Berlin 2010, and 9...Qh4! 10.d3 Nh3+!
11.gxh3 Qxe4 12.dxe4 Nxe5 13.Be2 Bxh3.
d2) 6.d3 d5 7.Bb3 reaches a semi-theoretical position, arising most often via 3...Bc5 4.c3 Nf6 5.d3
0-0 6.Bb3 d5 7.Qe2. Seeing as this is an obscure transposition, I’ll just note that Stockfish thinks
7...a5! is good for Black, playing as follows against White’s various tries: 8.a4 (or 8.0-0 a4 9.Bc2 Re8
10.Bg5 d4, or 8.Ba4 Rb8 9.0-0 b5 10.Bc2 d4, or 8.Bg5 dxe4 9.dxe4 h6 10.Bh4 a4) 8...h6 9.0-0 Be6
10.Bc2 Re8, and if 11.Nbd2 Qd7 12.exd5, S.Tiviakov-K.Godarz, Mashhad 2016, then 12...Nxd5
13.Nc4 f6; e.g. 14.Be3 (or 14.Rd1 Qf7) 14...Bf8 15.d4 exd4 16.Nxd4 Nxd4 17.Bxd4 (or 17.cxd4 Bf5)

314
17...Nf4 18.Qe4 g5! 19.Ne3 Qf7 and White’s pieces are in a dreadful mess.

A: 4.d4 exd4

The Scotch Gambit, which can also be reached via 3.d4 exd4 4.Bc4 (if Black plays 4...Nf6). White
now has two main options:
A1: 5.e5
A2: 5.0-0
Other moves are worse:
a) 5.Qe2? Bc5 6.e5 0-0! 7.0-0 d5 8.Bb5, W.Pietzsch-L.Masieev, corr. 1967, and now 8...Re8!
(Bologan) 9.Qd1 (or 9.Bg5 Bg4) 9...Ne4 is virtually winning already.
b) 5.Bg5? h6 6.Bxf6 (not 6.Bh4? g5 7.Bg3 Nxe4) 6...Qxf6 7.0-0 Bc5 8.e5 Qd8 9.Nbd2 0-0 and
...d7-d5; or if 9.c3!?, F.De Oliveira-A.Machado, Barueri 2019, then 9...dxc3 10.Nxc3 0-0, intending
...d7-d6, when White doesn’t have a whole lot for the pawn.
c) 5.c3? Nxe4 6.Qb3 (if 6.cxd4 d5 or 6.0-0 d5, White is just a pawn down; e.g. 7.Bb5 Bd6 8.Nxd4
0-0! 9.Nxc6 bxc6 10.Bxc6 Bxh2+ etc – Bologan) 6...Qf6 7.0-0 Nc5 8.Qc2 Ne6 9.cxd4 Ncxd4
10.Nxd4 Nxd4 11.Re1+ Be7 12.Qe4 c6 13.Bd3 d5 14.Qe3 Be6 and Black won, B.Wiesinger-
C.Roscher, Köstendorf 1998.
d) 5.Nxd4? Nxe4 6.Bxf7+!? (White gets nothing from 6.0-0 d5 7.Bb5 Bd7; or 6.Nxc6 bxc6 7.Qe2
Qe7 8.0-0 Nd6, E.Fleischer-G.Grassmehl, Bad Bevensen 2001; or 6.Qh5?! Qf6 7.Nf3 Qg6, or
especially 7.Nxc6 Qxf2+ 8.Kd1 bxc6 9.Qe5+ Be7 10.Qxg7 Rf8 11.Rf1 Bf6!, or 9.Rf1,
G.Vaicekauskas-A.Rimeikis, Plateliai 2014, and 9...Qxg2 10.Qxf7+ Kd8) 6...Kxf7 7.Qh5+ g6 8.Qd5+
Kg7 9.Nxc6 (not 9.Qxe4? Bb4+ 10.Kd1 d5) 9...Qe8! (better than 9...bxc6 10.Qxe4 Qe8 11.Qxe8
Bb4+, M.Chigorin-E.Von Schmidt, 6th matchgame, St. Petersburg 1879, as after 12.Nc3 Rxe8+
13.Be3 – Chigorin, Black has only a small edge) 10.Qe5+ (forced) 10...Qxe5 11.Nxe5 d5 12.0-0 Bd6

315
13.Nd3 Bf5 and Black has a serious advantage, based on the centre, two bishops, and lead in
development, Ja.Bem-L.Pospisil, corr. 1999.
e) 5.Ng5?! doesn’t enhance 4.Ng5 lines for White because of 5...d5 6.exd5 Qe7+! and then:

e1) 7.Kd2?? loses at once to 7...Qb4+.

e2) 7.Be2?! Nxd5 8.0-0 h6 9.Nf3 Bg4 is good for Black; e.g. 10.c4?! (or 10.Nxd4 Bxe2 11.Nxe2 0-
0-0) 10...Nf6 11.Nxd4 Rd8 12.Nxc6 bxc6 13.Bxg4 Rxd1 14.Bxd1 Qd7 15.Bf3 Be7 16.Nc3 0-0 and
Black won, F.Nicolo-D.Podetti, Milan 2015.
e3) 7.Qe2?! Qxe2+ 8.Kxe2 Nb4 is also quite simple; e.g. 9.Bb3 (not 9.d6? Nxc2 10.Na3 Nxa1
11.Nxf7 Bxd6 12.Nxh8 a6) 9...Bf5 10.Re1, Xue Minyinjie-M.Herrera, Oropesa del Mar 2000, and
now 10...0-0-0!? 11.Nxf7 Re8+ 12.Kf1 Rxe1+ 13.Kxe1 Rg8, intending ...Bc5 and ...Re8, gives Black
excellent play for a pawn that White won’t be keeping anyway.
e4) 7.Kf1 (in view of the above failures, this is pretty much forced) 7...Ne5 8.Qxd4 (not 8.Bb5+?!
c6 9.dxc6 bxc6 10.Ba4 h6 11.Bf4 Ba6+ 12.Kg1 hxg5 13.Bxe5 0-0-0! with a big advantage,
K.Dadello-D.Marholev, Avoine 2008) 8...Nxc4 9.Qxc4 h6 10.Nf3 (against the piece sacrifice
10.Nc3?! hxg5 11.Bxg5, I like Lokander’s 11...Qb4!, intending 12.Re1+ Kd8 and ...Bd7, or 12.Qxc7
Be7 13.Re1 Kf8 and ...Bf5, when Black’s piece is worth much more than White’s pawns) 10...Qc5
11.Qe2+ (after 11.Qxc5 Bxc5 12.c4 Bf5 “Black’s bishops cause White many problems” – Emms; e.g.
13.Ne5? 0-0-0! 14.Nxf7? Rhe8 15.Nxd8? Bd3+ and mates, or 13.Ke2 0-0-0 14.Be3 Rhe8 15.Nc3
Ng4 16.Nd1 c6 17.Rc1 Kb8 18.h3 Nxe3 19.Nxe3 cxd5 20.cxd5 Rxd5 with a clear advantage,
V.Bolzoni-G.Lane, Brussels 1990) 11...Be7 12.c4 (relatively best, even though it doesn’t defend the
pawn; whereas after 12.Be3 Qxd5 13.Nc3 Qa5 14.Re1 0-0 15.Bxh6 Re8 16.Bd2 b6 17.Kg1 Ba6
18.Qd1 Rad8 19.h3 Bb7, Black has a big advantage despite the missing pawn, F.Colombo Berra-
Y.Gadzinskiy, corr. 2017; and 14.Bf4 Qb4 15.Qb5+ c6 16.Qxb4 Bxb4 17.Re1+ Kf8 is good for Black
too with the two bishops, S.Kireev-M.Dard, corr. 2008) 12...Nxd5 (or 12...b5 13.Nc3 Qxc4 14.Qxc4
bxc4 15.Bf4 Bb7 16.Bxc7 Nxd5 17.Nxd5 Bxd5 18.Re1 Kd7 19.Be5 and White may hold,

316
Gu.Bjornsson-J.W.De Jong, Porto Mannu 2015) 13.Na3 Nb4 14.Nb5 (or 14.Be3 Qh5 15.Nb5 0-0)
14...Be6, and now 15.Be3! (not 15.b3?!, E.Titov-R.Butze, corr. 1974, since 15...Bf6 16.Be3 Qe7 is
good for Black) 15...Qxc4 16.Rc1 Qxe2+ 17.Kxe2 0-0 18.Nxc7 Nxa2 19.Nxa8 Nxc1+ 20.Rxc1 Rxa8
21.Rc7 Bf6 22.Rxb7 (or 22.Bd4 Bxd4 23.Nxd4 Re8 24.Kd2 Bd5) 22...Bc4+ 23.Ke1 a6 leaves Black
with just a slight edge with the bishops. Playing to draw this endgame may be the most White can
achieve in this whole line.

A1: 5.e5

This is known as the “Modern” variation. I’m not sure why, given that it dates back to 1837 in the
databases, which is earlier than the first game with 5.0-0. Well, anyway.

5...Ng4

I am indebted to IM Martin Lokander for this choice. Previously I mostly played 5...d5 6.Bb5 Ne4
7.Nxd4 Bc5 and did okay, but I was bored of it after some 30 years. Then 5...Ng4!? appeared in
Lokander’s 2015 book. A quick search sorted by Elo rating brought up games by Carlsen, Shirov and
Short on the Black side. Alright. It seems Keres was right when he wrote that “dieser Springerzug ist
besser als sein Ruf”.
White has two main replies:
A11: 6.Qe2
A12: 6.0-0
All other moves are bad:
a) 6.Nxd4? Ngxe5 is just an extra pawn, and if Black wants to make absolutely sure then 7.Nxc6
dxc6! 8.Qxd8+ Kxd8 9.Bb3 Bd6 is completely safe.

317
 b) 6.Bg5? Be7 7.Bxe7 (7.Bf4 d6 8.exd6 Bxd6 is note ‘c’ below) 7...Qxe7 only helps Black; e.g.
8.0-0 Ngxe5 9.Nxe5 Nxe5 10.Re1 0-0 11.Bb3 (11.f4 Qc5 or 11.Qe2 Qc5 safely escapes the pin)
11...Qf6 and Black is two pawns up, A.Goss-B.Connell, Bunratty Online (blitz) 2021.
c) 6.Bf4? is answered by 6...d6! 7.exd6 (after 7.Qe2 Ngxe5 8.Nxe5 Nxe5 9.Bxe5 dxe5 10.Qxe5+
Qe7 11.Qxe7+ Bxe7 Black keeps the pawn, M.Petursson-S.Lund, Reykjavik 2007; and 7.Bb5 dxe5
8.Nxe5 Ngxe5 9.Bxe5 Bd7 10.Bxc6 Bxc6 11.Qxd4 Qxd4 12.Bxd4 0-0-0 13.c3 Bxg2 regains it,
A.Viet-H.Silber, German League 2010) 7...Bxd6 8.Qe2+ Kf8! 9.Bxd6+ Qxd6 and “Black is simply
up a pawn” (Lokander); e.g. 10.0-0 Bd7 11.Nbd2 Re8 12.Qd3 Nge5 13.Nxe5 Nxe5 14.Qb3 Bc6
15.Rfe1 h5, followed by ...Rh6 and Black soon won, N.Azarova-A.Stefanova, Turin Olympiad 2006.
d) 6.Bxf7+? is a tempting trick which backfires after 6...Kxf7 7.Ng5+ Kg8!.

d1) 8.Qxg4 h6 9.Nf3 (not 9.Qf3? hxg5 10.Qd5+ Kh7 11.h4 g4 and wins, Vasiliev-Selivanovsky,
Moscow 1962; or 9.Nh3 d6 10.e6 Qf6 11.Nf4 Bxe6 12.Nxe6 Re8, regaining the pawn, T.Shadrina-
N.Pogonina, Russian Women’s Ch., Moscow 2011) 9...d6, when everything is bad for White: 10.Qe4
(after 10.e6 Qf6 11.0-0 Bxe6 12.Qg3 Bd5 White has less than nothing for the pawn, G.Nieto-
J.Banawa, St. Louis 2013; and 10.Qg3 dxe5 11.Nxe5 Qd5 12.Nxc6 Qe4+ 13.Kd1 Bf5 is no
improvement, K.Tkachuk-F.Erwich, German League 2011) 10...dxe5 11.Nxe5 (or if 11.0-0 Qf6
12.Nbd2 Bf5 13.Qd5+, E.Bogdanov-Mi.Bartel, Infoszach Masters blitz 2020, then 13...Qe6 should
win) 11...Qf6 (Pinski’s 11...Qe8 12.f4 Bd6 13.0-0 Nxe5 is good too) 12.Nxc6 bxc6 13.0-0 Bf5
14.Qe2 Bd6 15.Nd2 Kh7 and Black has a big advantage, F.Jacobsen-C.Menezes, Al Ain 2013.
d2) 8.Qf3 fails to 8...Bb4+! (not 8...Ngxe5?? 9.Qd5+ and mates) 9.c3 Ngxe5 (or 9...Qe7 first; not
9...Ncxe5?? 10.Qd5+ Kf8 11.Ne6+ 1-0 V.Muratov-Baranov, corr. 1964) 10.Qd5+ Kf8 11.0-0 (not
11.cxb4? Nxb4 or 11.f4? Nd3+ 12.Kd2 Qf6 13.Kxd3 d6 14.cxd4 Bf5+ 15.Kc4 h6 16.Ne4 Qg6 and
wins – A.Prik) 11...Qe7! 12.f4 (or 12.cxb4 Nxb4 – Pinski) 12...Ng4 13.cxb4 Nxb4 and Black is much
better, E.Hidegh-I.Kajari, Hungarian League 2008.

318
A11: 6.Qe2

The old theoretical recommendation, now generally (and rightly) discarded in favour of 6.0-0 (line
A12). I thought there might be an advantage here too, but the right side of equal is probably the best
Black can do.

6...Qe7

7.Bf4

The only alternative is 7.Bg5!?, encouraging a move Black was intending anyway: 7...f6 8.exf6
gxf6 (or 8...Qxe2+ 9.Bxe2 h6!? with ideas of ...g7-g5; not 9.Kxe2? gxf6 10.Bf4 Nge5 and Black is
clearly better, M.Dizdarevic-V.Harutyunyan, Melbourne 2003) 9.Bh4 (not 9.Qxe7+? Bxe7 10.Bf4
Nge5, again with a big advantage, J.Farinas Lucas-E.Recio Romaguera, corr. 2020) 9...Bg7!?
10.Nbd2 (or 10.Qxe7+ Nxe7 11.Nxd4 d5 12.Bb3 Ng6) 10...Qxe2+ 11.Bxe2 d6 12.0-0-0 (or 12.Nb3
d3! 13.Bxd3 Nb4) 12...Nge5 13.Nxe5 Nxe5 (pawn captures give White more play after Bh5+ and f2-
f4) 14.Rhe1 (or 14.Bh5+ Ng6 15.Rhe1+ Kd7!) 14...0-0 and I prefer Black, if not by much.

7...f6! 8.exf6 Qxe2+

By capturing on e2 straight away Black can decide how to recapture on f6 depending on how White
recaptures on e2, whereas capturing on f6 first gives White the privilege of prior knowledge:
a) 8...gxf6 9.Nbd2 Qxe2+ is answered by 10.Bxe2! Nge5 (or 10...d6 11.0-0) 11.0-0 d6 12.Nb3 Bd7
13.Nbxd4 Nxd4 14.Nxd4 0-0-0 15.Rad1 and White is fine, B.Siembab-V.Malaniuk, Mielno 2007.
b) 8...Nxf6 9.Nbd2 (not 9.Bxc7? d6 10.Bb5 Bd7) 9...Qxe2+ (Lokander goes for 9...d5 10.Bb5
Qxe2+ 11.Kxe2 which comes to the same thing; instead, 9...d6 10.0-0-0 Qxe2 11.Bxe2 transposes to

319
the main line and was mostly the route taken there, but here 10.Qxe7+ Bxe7 11.Nb3 is quite equal)
10.Kxe2! is now correct, intending Rhe1 and Kf1+. Although 10...d5 (if 10...Nh5 then 11.Bxc7! d6
12.Bb5 is sound) 11.Bb5 a6 has scored well for Black in practice, objectively White seems okay; e.g.
12.Bxc6+ bxc6 13.Rhe1 (or 13.Bxc7 c5 14.Rae1 Bf5 15.Kd1+ Kd7 16.Bg3 Bd6, L.H.Hansen-
Je.Andersen, corr. 2010, and 17.Ne5+ Kc7 18.b4! cxb4 19.Nb3) 13...Bd6 14.Kf1+ Kf7 (or 14...Kf8
15.Bxd6+ cxd6 16.b4!; and not 14...Kd7?, R.Wehner-K.Zeier, German League 2009, as 15.Ne5+
Bxe5 16.Rxe5 and Rhe1 is good for White) 15.Ne5+ Kg8 16.Nd3 c5 17.Bxd6 cxd6 18.b4! cxb4 (or
18...c4 19.Nf4) 19.Nxb4 with rough equality.

9.Bxe2

Not 9.Kxe2?! because of 9...gxf6!, gaining a knight outpost on e5 as well as open lines for the
rooks and bishops; e.g. 10.Bb5 (or 10.Nbd2 d6 11.Bb5 a6! 12.Bxc6+ bxc6 13.Nxd4 c5 14.N4f3,
D.Gjuran-L.Mazi, Murska Sobota 2006, and 14...Ne5; not 10.Bxc7? d6 11.Bb5 Bd7 12.Bxc6 bxc6
13.Nxd4 c5 14.Nf3 Bb5+, when the white king is caught in the crossfire) 10...Nge5 11.Nbd2 d6
12.Nxd4, F.Bentivegna-V.Tomescu, Cesenatico 2004, and now 12...a6! 13.Ba4 (or 13.Bxc6+ bxc6
14.Bxe5 fxe5 15.Nxc6 Bb7 16.Na5 Bxg2) 13...b5 14.Nxc6 bxa4 15.Nb4 Rb8 16.Nd5 Kf7 looks good
for Black, despite the wrecked pawn structure.

9...Nxf6!

Rather than 9...gxf6 10.Nbd2, which transposes to 8...gxf6 above.

10.Nbd2

The usual move, preparing long castling and heading for b3 to regain the pawn on d4. Instead:
a) 10.0-0 can be met by 10...Bc5 (10...d6 11.Re1! Be7 12.Nbd2 is pretty equal) 11.Nbd2,

320
J.Heppekausen-J.Snuverink, German League 2008, and 11...Bb6 12.Nb3 0-0, when 13.Rad1 Nd5
14.Bg3 (or 14.Bc4 Rxf4 15.Bxd5+ Kf8) 14...Nf4 15.Bxf4 Rxf4 might give Black a little something
with the two bishops.
b) 10.Bxc7!? is risky but not entirely bad: 10...d6 11.Bb5 Bd7 12.0-0 (if 12.Bxc6?!, R.Stranz-
K.Neumeier, Austrian League 2004, then 12...bxc6! 13.Ba5 Rb8, intending ...Rxb2 or ...c6-c5; while
13.Nxd4 c5 and 14...a5 lands the c7-bishop in trouble after all) 12...Rc8 13.Re1+ Ne7 14.Bxd7+
Kxd7 15.Ba5 Rc5 16.Bd2 Rxc2 with just the slightly better chances.

10...d6 11.0-0-0

If 11.Nb3 then 11...d3! 12.Bxd3 Nb4 13.0-0-0 Nxd3+ 14.Rxd3 h6! 15.Re1+ Kf7 and Black will
have an edge once the bishops come out.

11...Nh5

More promising than 11...Be7 12.Rhe1 Nd5 13.Bg3 Bd7 14.Nb3 0-0-0 15.Nbxd4 Nxd4 16.Rxd4
which was equal at this point, N.Karaklajic-R.Hess, Reggio Emilia 1979/80.

12.Bg3

Not 12.Bg5?! h6 13.Bh4 g5 14.Nxd4?! Nf4 15.Bb5 Bd7 and White is in trouble; e.g. 16.Rhe1+ (or
if 16.Bxc6 bxc6 17.Bg3 Nxg2 18.Rhg1, J.Slettebo-J.Hector, Helsingor 2019, then 18...c5 19.N4f3
Bh3) 16...Be7 17.Bxc6 bxc6 18.Bg3 Nxg2 (the knight is quite safe) 19.Re4 c5 20.Ne6 Bxe6 21.Rxe6,
K.Lie-Ma.Carlsen, Tromsø 2007, and now 21...Kd7 22.Re2 Rae8 23.Rg1 (or 23.Ne4 Kc6) 23...Nh4
24.Bxh4 gxh4 is good for Black.

12...Be7

Taking the bishop at once gives White more leeway; e.g. 12...Nxg3 13.hxg3 Bd7 14.Nh4,
threatening Ng6; while 12...Bd7 might allow White to keep it with 13.Bh4.

321
 After the text Black is, as indicated, on the right side of equal, the two bishops offering a slight pull
even after White regains the pawn.
a) 13.Bh4? is a mistake due to 13...Nf4 14.Bc4 Bxh4 15.Nxh4 Be6 and Black will keep the extra
pawn.
b) 13.Rhe1 Nxg3 (now that the white rook has left the h-file) 14.fxg3 (or 14.hxg3 Bd7 15.Bc4 Rf8
16.Nb3 Kd8!) 14...Bd7 15.Bb5 a6 16.Bxc6 bxc6 17.Nxd4 Kf7 18.N2f3?! (or 18.Rf1+ Kg6 19.Ne6
Rac8!) 18...Bf6 with a definite advantage, V.Gurevich-G.Ciolac, Werfen 1990 (½-½, 75).
c) 13.Nb3 Bd7 14.Nfxd4 Nxg3 15.hxg3 0-0-0 seemingly led to a win for Black in I.Zaitsev-
S.Smagin, Moscow 1983. Unfortunately, MegaBase doesn’t give any more moves, but the same
position arose in an engine game, which continued 16.c3 (or 16.Nxc6 Bxc6 17.Bg4+ Kb8 18.f3 Bf6)
16...Bf6 17.Bd3 Ne5!? 18.Bxh7 Ng4 19.Rdf1 c5 20.Nc2 g6 21.Bxg6 Rxh1 22.Rxh1 Nxf2 23.Rf1
Bg5+ 24.Kb1 Rg8, when Black had the two bishops and the initiative for a doubled pawn, Equinox-
Onno, TCEC 2 Division C 2011 (½-½, 61).

A12: 6.0-0

322
 The current theoretical recommendation for White.

6...Be7!?

IM Lokander’s proposed solution. The old books don’t mention it at all.

Formerly Black played 6...d6 7.exd6 Bxd6 8.Re1+ Kf8, but this has not stood up to engine scrutiny:
9.Nbd2! Nge5 10.Nxe5 Nxe5 11.Bb3 Qh4 (or 11...Bg4 12.f3 Bf5 13.Ne4) 12.h3 Bxh3 13.Rxe5! Bxe5
14.Qf3 Bf6 15.gxh3 and White is clearly better. I have won this way myself.
So Black has to resort to 7...Qxd6, when 8.Re1+ (not 8.Na3 a6 9.h3 Nf6 10.Ng5 Nd8 11.Re1+ Be7
12.Qe2 Ne6 13.f4? b5 14.Bd3, G.Sax-A.Gorchakov, Groningen 1972, as 14...Nxg5 15.fxg5 Nd5 is
very good for Black) 8...Be7 (or 8...Be6 9.Bg5!? – Ntirlis, though Stockfish isn’t worried) 9.b3 (here
9.h3 Nf6 10.b3 0-0 etc is much the same) 9...0-0 10.Ba3 Qd8 11.Bxe7 Nxe7 12.Qxd4 Qxd4 13.Nxd4
leads to a rather dreary endgame. That’s not to say it’s actually bad. GM Ernesto Inarkiev has
defended 13...Nf5 14.Nxf5 Bxf5 15.Re7 Rac8 four times and scored 3/4 as Black. I’m not sure how!
Note that 6...Ngxe5?? was not possible because 7.Nxe5 Nxe5 8.Re1 d6 9.f4 wins. With the text
Black forestalls the pin on the e-file and thus threatens to take the e-pawn. White has three ways to
defend it.

323
7.Re1!

The most natural and best response.

a) 7.Qe2?! is a poor choice as White has few concrete threats after 7...d6 8.exd6 Qxd6, making it
that much easier for Black to consolidate; e.g. 9.h3 (or 9.Re1 0-0) 9...Nge5 10.Bf4 Nxf3+ 11.Qxf3
Qf6 12.Nd2 Bf5 13.Rfe1 0-0-0 and Black is better.
b) 7.Bf4 is directed against 7...d6 (i.e. 8.exd6 cxd6 9.Nxd4 or 8...Bxd6 9.Re1+), so Black should
attack from the other side: 7...g5! (this is my kind of move; Lokander gives an alternative in 7...f6, “to
keep the game close to understandable”, should Black prefer that, e.g. 8.exf6 Nxf6 9.Nxd4 Nxd4
10.Qxd4 d5 11.Bb3 0-0 with equality, E.Selezneva-M.Lokander, St. Petersburg 2015)

324
 Now the play gets very complicated:
a) 8.Bxf7+? is no good: 8...Kxf7 9.Nxg5+ Bxg5 10.Qxg4 Bxf4 (or 10...h6 – M.Pap) 11.Qxf4+ Kg6
12.Qe4+ Kh6, followed by ...d7-d5, and the black king is quite safe.
b) 8.Nxg5? is a self-trap: 8...d5! 9.e6 (the only way to save the piece; if 9.exd6 Bxg5 10.Qe2+ Kf8
11.dxc7 Qf6 12.Bg3, R.Brauning-Ul.Neumann, corr. 2009, then 12...h5! 13.Nd2 Bxd2 14.Qxd2 h4
with ...h4-h3 to follow – Lokander) 9...Bxg5 10.Qxg4 Bxe6 11.Bxg5 (or 11.Qxg5 Qxg5 12.Bxg5
dxc4, D.Hersvik-R.Fyllingen, Gausdal 1999) 11...Bxg4 12.Bxd8 Rxd8, when Black is a pawn up.
Topdriver-tsmenace, ChessWorld.net 2020, continued 13.Bb5 Ke7 14.Nd2 Nb4 15.Rfc1 Bf5 16.Ba4
c5 17.Nf3 Na6 18.Bb5 Nc7 19.Bd3 Bxd3 20.cxd3 Kd6 and I won by pushing the queenside pawns.
c) 8.Bc1!? leaves the e5-pawn to its fate, hoping that ...g7-g5 will prove a serious weakness. I think
Black might as well take it: 8...Ngxe5 (the alternative is 8...d6 9.exd6 Qxd6 10.b3 0-0 11.Ba3 Qf6
12.Bxe7 Qxe7, followed by ...Nge5 – Lokander) 9.Nxe5 Nxe5 10.Qxd4 Bf6 11.Bxg5 (or 11.Qe4 c6
12.Bxg5 d5 13.Qf4 Nf3+ 14.Qxf3 Bxg5, followed by ...Be6, and Black has chances with the two
bishops and open g-file) 11...Nf3+! (not 11...d6? 12.Bxf6 Qxf6 and Black does just have weaknesses,
V.Sanal-L.Travadon, Podgorica 2021) 12.gxf3 Rg8 and Black regains the piece with advantage; e.g.
13.Qe3+ (or 13.Qe4+ Kf8 14.f4 Bxg5! 15.fxg5 d5 16.Bxd5 Rxg5+ 17.Kh1 Qxd5 with the superior
minor piece and greater activity) 13...Kf8 14.f4 h6 15.Nc3 hxg5 16.Kh1 b5! (opening up the long
light diagonal) 17.Be2 (not 17.Bxb5? Bb7+ 18.f3 g4; or 17.Qf3 Rb8 18.Bd5 g4; or 17.Bd5 c6 18.Bg2
gxf4) 17...Qe7 18.Qf3 (any of 18.Qxe7+ Bxe7, or 18.Nxb5 Rb8, or 18.Qd3 g4 enhances the ...Bb7
threat) 18...Rb8 19.Nd5 Qe6 20.f5! (or 20.Nxc7 Qd6) 20...Qxd5 21.Qxd5 Bb7 and Stockfish claims
an edge, though it looks more like a draw to the human eye.
d) 8.Bg3 prompts 8...h5!

325
 9.Bd5 (preventing ...d7-d5; everything else is worse: 9.h4? gxh4 10.Bf4 h3, intending 11.gxh3 d5
12.Bb5 Qd7!, heading for the kingside; or 9.Bb3? h4 10.Nxh4 gxh4 11.Qxg4 d5 12.Qg7 Rh5! 13.Bf4
Bf8 14.Qf6 Rf5 15.Qxd8+ Kxd8 16.Bc1 Rxe5 – M.Pap; or 9.h3?! h4 10.hxg4 hxg3 11.fxg3 d5,
R.W.Clark-M.Barkwell, corr. 2003, and 12.Bb5 Bxg4 13.Qxd4 Be6 with an edge; or 9.Nxd4?! h4
10.Qxg4 d5 11.e6 hxg3, R.Buckle-P.Varas, corr. 2011, and even after 12.Bb5 gxf2+ 13.Rxf2 Qd6
14.Bxc6+ bxc6 15.exf7+ Kf8 16.Qg3 Qxg3 17.hxg3 Black is for preference) 9...Nb4! (after 9...h4
10.Nxh4 gxh4 11.Qxg4 Nb4 12.Qg7 Rf8 13.Bxf7+! Rxf7 14.Qh8+ Rf8 15.Qh5+ White has at least a
draw; but 9...Kf8!? is worth considering) 10.Bb3 (here 10.Qxd4 Nxd5 11.Qxd5 h4 12.h3 hxg3
13.hxg4 d6 14.Nbd2 c6 15.Qd4 d5 16.e6 0-0 17.exf7+ Rxf7 was good for Black, docjan-tsmenace,
ChessWorld.net 2018) 10...h4 (improving on 10...d3!? 11.Nc3 h4 12.Nxh4 gxh4 13.Qxg4, Mis.Pap-
L.Pressman, Donostia 2011, when 13...d5 14.Qg7 Rf8 15.Bf4 is unclear, aka “0.00” – Stockfish)
11.Bxh4 (not 11.Nxh4? gxh4 12.Qxg4 d5 13.Qg7 Rh5! 14.Bf4 Bf8, followed by ...Bf5 with a clear
advantage) 11...gxh4 12.Qxd4 Nh6 13.Nc3 “is supposed to end in a draw according to the engines”
(Lokander). That was then (2015), this is now (2021) and Stockfish 13 gives Black the advantage; e.g.
13...Kf8 14.Rad1 (or 14.e6 Rh7 15.exf7 d5) 14...Nc6 15.Qf4 d6 16.Rfe1 (or 16.Nb5 Bg4 17.Bxf7
Kg7 18.exd6 Rf8 19.dxe7 Qxe7) 16...Kg7 17.Ne4 Nxe5 18.Nxe5 dxe5 19.Qxe5+ Bf6 20.Qa5 Qe7
21.Ng3 Qxe1+ 22.Qxe1 (or 22.Rxe1 b6) 22...hxg3 23.hxg3 a5 24.f3 a4 25.Bc4 Be6 26.Bxe6 Rhe8
27.Qe4 Rxe6 28.Qxb7 Rae8 and “Black is clearly better”, though it is hard to imagine a non-silicon
based lifeform getting so far, or converting it even if they did.

7...d6

White was threatening to embarrass the g4-knight with h2-h3, so the intrusive e-pawn must be
dislodged.

8.exd6

326
 8.e6? has been played a surprising number of times. After 8...fxe6 9.Bxe6 (if 9.Nxd4 Nxd4
10.Qxd4 Ne5 11.Bb3 0-0 12.f4 Nc6 13.Qd3 d5, White is just a pawn down, C.Kreuzer-W.Boulton,
4NCL 2012) 9...Bxe6 10.Rxe6 0-0 11.Re4 Qd7 Black was better in D.Velimirovic-L.Stein,
Yugoslavia-USSR match, Ohrid 1972; and 11...Nce5! looks virtually winning.

8...cxd6!?

IM Lokander’s almost-novelty. CCIM Nikolaos Ntirlis adds a word of caution: “I do not fully trust
the position, and in a practical game it would only take one or two small errors for Black to become
worse with the isolated d-pawn.”
Such concern is probably why most players have opted for 8...Qxd6, transposing to the 6...d6 line
above. Personally, I prefer Lokander’s move and have scored 3/5 with it (W1, D4, L0). As he says:
“[T]his is not a typical position with an isolated pawn. For now, the pawn is on d6 and not d5, which
increases Black’s control of the dark squares in the centre. What’s more, the knight on g4 might look
stupid, but it provides Black with excellent chances of counterplay.”

9.Nxd4

White has tried two other moves:

a) 9.Bb3 0-0 10.h3 Nge5 11.Nxd4 Nxd4 12.Qxd4 would be a good idea except for 12...Bxh3!
13.Nc3 (obviously not 13.gxh3?? Nf3+, or 13.Rxe5?? dxe5 14.Qxd8 Raxd8 15.gxh3 Rd1+ and Black
wins) 13...Bf6 14.Qd5 Bd7 (the simplest solution) 15.Qxd6 (or 15.Qxb7?! a5 with the initiative)
15...Bc6 16.Qxd8 Rfxd8 ½-½ RickF-tsmenace, ChessWorld.net 2018.
b) 9.Bg5!? (GM Alex Fishbein’s attempted improvement)

327
 9...Nge5 (after 9...f6 10.Bc1 Nge5 11.Bd5 Bg4 12.c3 “it is not clear if Black can complete
development effectively” – Fishbein; I’ve won from here as White) 10.Bxe7 Qxe7 11.Be2 (probably
best; 11.Nbd2?! 0-0 12.h3, Cl.Kuhn-N.Kanov, Bratislava 2019, and 12...Qf6 is worse for White, who
will have to work to regain the pawn; while 11.Na3 0-0 12.Nxd4, S.Avila Pavas-D.Arenas Vanegas,
Sabadell blitz 2019, and 12...Qh4 13.Nf3 Qf6 14.Be2 Nxf3+ 15.Bxf3 Qxb2 or 14.Nxe5 dxe5 15.c3
Bf5 looks equal to me) 11...Nxf3+ 12.Bxf3 Be6 13.Bxc6+ (or if 13.Nd2 0-0 14.Nb3 Qf6 15.Bxc6
bxc6 16.Qxd4 – Fishbein, hoping for a plus/equals position, then 16...Rfb8!? 17.Qxf6 gxf6, intending
...a7-a5-a4, and Black activity balances the inferior structure) 13...bxc6 14.Qxd4 0-0 15.Nc3 Rad8
16.b3 Qc7 17.Rad1 Bf5 18.Rd2 h6 and I drew by doing nothing, michaelbuss-tsmenace,
ChessWorld.net 2019.

9...0-0

328
10.h3

The engine’s choice. Instead:

a) 10.Bf4?? is a blunder: 10...Nxd4 11.Qxd4 Bf6, skewering the queen against b2, craig4000-
tsmenace, ChessWorld.net 2019.
b) 10.Nxc6 bxc6 gives the isolated pawn a colleague: 11.h3 Nf6 (or 11...d5, as in the main line)
12.Nc3 h6 13.Bf4 d5 14.Bd3 Be6 15.Ne2 c5 16.Ng3 Bd6 17.Bxd6 Qxd6 and Black was fine,
A.Jaumandreu Llopis-J.Atuán, corr. 2020.
c) 10.c3 can be met by 10...d5 (or 10...Bh4 11.g3 Bf6 – Lokander) 11.Be2 Nxd4, intending
12.Qxd4?! (or 12.cxd4 Bb4) 12...Bf6 13.Qd3 Qb6 14.Rf1 Be5 with the initiative.
d) 10.Nc3 can be met by 10...Bh4 (or 10...d5!? anyway; e.g. 11.Nxd5?! Bc5 12.Be3 Nxe3 13.fxe3
Ne5 14.Bb3 a5, intending ...Qg5 and ...Ra6-g6 – Lokander, or 11.Bxd5 Bc5 12.Be3 Qc7 13.Nf3,
when mass liquidations led to a draw: 13...Nxe3 14.fxe3 Bg4 15.a3 Qf4 16.Kh1 Bxe3 17.Qd3 Rae8
18.Ne4 Bxf3 19.Qxe3 Qxe3 20.Rxe3 Bxe4 21.Rxe4 Rxe4 22.Bxe4 Rd8 ½-½ rooksac-tsmenace,
ChessWorld.net 2020) 11.g3 Bf6 12.Ndb5 (other moves allow ...Bxc3; e.g. 12.Nb3 Bxc3 13.bxc3
Nge5 14.Be2 Re8, intending ...Bh3, ...d5, and ...Nc4 – Lokander; while 12.Nf3 Bxc3 13.bxc3 Nge5
14.Nxe5 dxe5 15.Ba3 Re8 is quite equal) 12...Nxf2 (or 12...a6 13.Nxd6 Bd4 14.Nce4?! Nxf2!
15.Nxf2 Qxd6 16.Be3 Qc5 17.Bxd4 Nxd4 and, having eliminated the isolated pawn while weakening
the white kingside, Black is slightly better, A.Ornstein-M.Lokander, Swedish League 2019; White
had to pre-empt that by 14.Nxf7! Rxf7 15.Bxf7+ Kxf7 16.Ne4 which gets the “0.00” from Stockfish)
13.Kxf2 a6 14.Qxd6 (not 14.Nxd6?! Bd4+, or 14.Na3?! b5 15.Bd5 b4 16.Bxc6 Qb6+ – Lokander)
14...axb5 15.Nxb5 (or 15.Qxd8 Rxd8 16.Nxb5 Nb4 17.Re2 Bg4 18.Rd2 Bf5 19.Bb3 Nxc2 20.Bxc2
Rxd2+ 21.Bxd2 Bxc2 22.Bc3 Be7, J.Simmelink-Bri.Perez, corr. 2017) 15...Bf5 16.c3 Na5 17.Qxd8
Rfxd8 18.Na3 Nxc4 19.Nxc4 Be6 with queenside pressure for the pawn, M.Auzins-M.J.Schröder,

329
corr. 2012.

10...d5!

It turns out there’s no need to retreat the knight.

11.Bf1

Or 11.Nxc6 bxc6 12.hxg4 dxc4 13.Qf3 Be6 14.Qxc6 Bxg4 15.Qxc4 Be6 16.Qe2 Re8 17.Nc3 Bf6,
again with strong pressure for the pawn, D.Grobler-A.Cannon, corr. 2012.

11...Nxf2!

330
 “A typical strike in this variation” (Lokander).

12.Kxf2

Or 12.Nxc6 bxc6 13.Kxf2 Bc5+ 14.Be3 d4! 15.Bd2 Qh4+ 16.g3 d3+ 17.Be3 Qf6+ 18.Kg2 dxc2
19.Qxc2 Bxe3 20.Rxe3 Bxh3+ 21.Kxh3 Qxf1+ 22.Qg2 Qc1 23.Qd2 and ½-½ Raffzahn-tsmenace,
ChessWorld.net 2019, in view of 23...Qf1+ 24.Qg2 Qc1, repeating.

12...Bc5 13.c3

Not 13.Be3? Nxd4 14.Bxd4? Qh4+ and wins.

13...Nxd4 14.cxd4 Qh4+ 15.Kg1 Bxd4+ 16.Kh1

After 16.Be3 Bxb2 17.Nd2 Bxa1 18.Qxa1 Bf5 “Black can’t complain [with] three pawns and a
rook for the two minor pieces” (Lokander).

331
16...Bg4!

I prefer this to 16...Qf2 17.Bb5 Bg4! 18.Qd2 Qg3 19.Qxd4 Qxe1+ 20.Qg1 Qxg1+ 21.Kxg1 Bf5
(Lokander), when 22.Nd2, followed by Nf3 and Bd2, might give White an edge.

17.Qd2 Bb6

Now Black keeps all attacking options open. Meanwhile White’s queenside is still at home and the
queen has to try and hold the defences together by herself.

For example:

332
 a) 18.b3? Rae8 19.Rxe8 Rxe8 20.Qf4 runs into 20...g5! 21.Qxg4 Qf2! 22.Qxg5+ Kh8, when even
23.Bb2+ Qxb2 24.Nc3 Qxa1 25.Qf6+ Kg8 26.Qg5+ Kf8 27.Qh6+ Ke7 28.Nxd5+ Kd7 doesn’t seem
to save White.
b) 18.Nc3 Rae8 19.Rxe8 Rxe8 20.Qf4 now meets with 20...Re1! 21.Qb8+ (or 21.Kh2 Qd8!)
21...Qd8 22.Qxd8+ Bxd8 23.hxg4 Rxf1+ 24.Kh2 Bg5 25.Bxg5 Rxa1 26.Be3 a6 and Black is not
worse.
c) 18.Na3 h6! (typical engine sangfroid; not now 18...Rae8? 19.Rxe8 Rxe8 20.Qf4 Re1? because of
21.Nc2!) 19.Nc2 (the alternative is to advance the b-pawn: then 19.b3 Bf5! threatens ...Rae8 or
20.Bb2 Be4 21.Qc3 f6 22.Rxe4 dxe4 23.Nc4 Rac8 24.Qe1 Qxe1 25.Rxe1 Bc5 26.Rxe4 Rfd8 with
sufficient counterplay; the same applies after 19.b4 Qg3!, threatening ...Bc7, and 20.Nb5 Bf3! or
20.Re3 Qxe3 21.Qxe3 Bxe3 22.Bxe3 Rfe8 23.Bd4 Be2 24.Nb5 Bxf1 25.Rxf1 Re4 26.Rd1 Re2)
19...Rac8! (threatening ...Rxc2) 20.Nb4 (if 20.Re5 then 20...Qf6, or 20.Qf4 f5!) 20...Rfe8 21.Nd3 (if
21.Rxe8+ Rxe8 22.Nd3 then 22...Bxh3! 23.gxh3 Re3 24.Nf4 Qxf4 25.Qg2 Rxh3+ 26.Qxh3 Qe4+, or
24.Qg2 Rg3 25.Qh2 Rg1+ 26.Qxg1 Bxg1 27.Kxg1 Qg3+ 28.Kh1 Qf3+ with a draw) 21...Bf5
22.Rxe8+ Rxe8 23.Qf4 Qf6 24.Bd2 g5 25.Qf3 Qd4 (threatening ...Qg1 mate) 26.Be2 Rxe2 27.Qxe2
Qxd3 28.Qxd3 Bxd3 and Black should be okay with two good pawns for the exchange.

A2: 5.0-0 Bc5!

The exclamation mark is for entertainment value. Most repertoire authors go for 5...Nxe4 because it
solves Black’s problems completely. For example: 6.Re1 d5 7.Bxd5 Qxd5 8.Nc3 Qh5 9.Nxe4 Be6
10.Bg5 Bd6 11.Nxd6+ cxd6 12.Bf4 Qd5 13.c3 Rc8 14.b3 0-0 15.Nxd4 Nxd4 16.Qxd4 Qxd4 17.cxd4
Rc2 18.Bxd6 Rd8 19.Be5 h5 20.h4 a5 21.Rec1 ½-½ tsmenace-michaelbuss, ChessWorld.net 2019. So
I think GM John Emms is correct to say that 5...Nxe4 “is the most reliable method of meeting 5.0-0.”
But it’s so very dull.

333
6.e5!

Initiating the Max Lange Attack. Note that White cannot delay taking action, otherwise Black might
start something:
a) 6.Nbd2?! 0-0 7.e5 (or 7.Nb3 Bb6 8.e5 d5) 7...Ng4! 8.Nb3 (or if 8.Re1, Ri.Rasmussen-J.Romsdal,
Alta 2005, then 8...Ne3! 9.fxe3 dxe3 10.Kh1 exd2 11.Qxd2 d6) 8...Ngxe5 9.Nxe5 Nxe5 10.Nxc5
Nxc4 11.Qxd4 Na5 and ...Nc6 consolidates the pawn.
b) 6.Bg5?! h6 7.Bh4? (for 7.Bxf6 Qxf6 see 5.Bg5 earlier) 7...g5! 8.Bg3 Nxe4 9.Re1 0-0! (not
9...d5? 10.Bxd5 Qxd5 11.Nc3) 10.Rxe4 d5 11.Bxc7 (or 11.Bxd5 Qxd5 12.Nc3 Qf5) 11...Qd7
12.Bxd5 Qxd5 13.Re1 Bf5 14.Nbd2 Rfe8 and Black is clearly better.
c) 6.Re1?! 0-0 7.e5 (or 7.Bg5 h6 8.Bxf6 Qxf6 9.e5 Qd8 and ...d7-d5; while 8.Bh4? g5 9.Bg3 Nxe4!
is note ‘b’) 7...d5 8.exf6 (or 8.Bd3 Ne4 – B&H; e.g. 9.Bxe4 dxe4 10.Rxe4 Bf5 11.Re1 Nb4 with a big
advantage, A.Masszi-L.Simai, Hungarian League 2006) 8...dxc4 9.Bg5? (or 9.fxg7 Re8) 9...gxf6
10.Bh6 Re8 and Black is two pawns up without the rook in the corner issues of line A221.
d) 6.c3 is also inferior, as it transposes to a sideline of the Italian which is more than okay for
Black. After 6...Nxe4 7.cxd4 (not 7.Re1? d5 8.cxd4 Bb4 or 7.Bd5? Nf6 8.cxd4 Bb6) 7...d5! 8.dxc5
(nor 8.Bb5? Bd6) 8...dxc4, White has tried various queen moves:

d1) 9.Qxd8+?! is both natural and bad after 9...Kxd8! (avoiding a pin on the e-file):
d11) 10.Be3 Be6 11.Na3 (or if 11.Nd4 Nxd4 12.Bxd4 Kd7 13.Re1 Nf6 14.Na3 Kc6 15.Rac1 Rad8
16.Be3 Ng4 17.Bf4, C.Engelbert-A.Neffe, German League 1994, then 17...Rhe8 – Bologan) 11...c3!
12.b3 (not 12.Nd4?! Nxd4 13.Bxd4 cxb2 14.Rad1 Kc8 15.Rfe1 Nf6 16.Bxb2 Rd8, Kar.Horvath-
Z.Hajnal, Paks 2000; or 12.bxc3?! Nxc3 13.Ng5 Re8 14.Nxe6+ Rxe6 15.Rfe1 Kd7 16.Rac1,
Thor.Müller-O.Schlesinger, Baunatal 1999, and 16...Nxa2 – Bologan) 12...a5 13.Nd4 a4 14.Nac2 (or
14.b4 Nxb4 15.Rfb1 Nd5 16.Rxb7 Nxc5) 14...axb3 15.axb3 Rxa1 16.Rxa1 Kd7 17.b4 Re8 18.Ra3
Nxd4 19.Bxd4 Bf5 20.f3 Ng5 21.Rxc3 Ne6 22.Be3 Bxc2 23.Rxc2 Kc6 and the active king gives

334
Black a clear advantage in the endgame, H.Herr-R.Wegelin, corr. 2014.
d12) 10.Rd1+ Bd7 11.Be3 (after 11.Na3 Nxc5 12.Nxc4 f6, or 11.Ng5 Nxg5 12.Bxg5+ Kc8 13.Na3
Be6, J.Blauert-M.Hebden, London Lloyds Bank 1991, Black is just a pawn up) 11...Kc8 12.Rc1 (if
12.Na3 c3 13.b4 Re8 14.Nc2, T.Cagasik-M.Zavadil, Brno 2009, then 14...a5 15.a3 Ne7 intending
...Ba4 or ...Nf5) 12...Be6 13.Na3 (not 13.Nbd2? Nxd2 14.Nxd2 Ne5 15.Bd4 Nd3 16.Rc3 Rd8
17.Bxg7 Nf4 18.Rc2 Bf5 19.Rxc4 Ne2+, followed by ...Rxd2, D.Lynch-M.Zielinski, Cappelle la
Grande 1989) 13...c3 14.bxc3, D.Hergott-Gild.Garcia, Linares 1994, and now 14...Re8! (Pinski),
“when White is desperately fighting for equality and will probably be unsuccessful.”
d13) 10.Ng5 Nxg5 11.Bxg5+ f6 12.Rd1+ (or if 12.Bf4 Nb4! 13.Na3 Nd3 14.Rfd1, D.Dumitrache-
Kr.Georgiev, Athens 1992, then 14...Be6 – Bologan) 12...Bd7 13.Bf4 (or 13.Be3 b5!) 13...Nb4!
14.Nc3 Nd3 15.Bg3, V.Okhotnik-A.Kovacs, Eger 1990, and now 15...Re8 – Bologan, who considers
that a pawn down endgame with opposite-coloured bishops is one of White’s best chances to survive.
Super.
d2) 9.Qe1? obstructs the f1-rook, so that 9...Qe7 (not 9...Qd3?? 10.Ng5 and wins) 10.Nc3,
A.Pashikian-G.Sargissian, Armenian Ch., Yerevan 2003, and 10...Bf5 is already very good for Black.
d3) 9.Qc2?! fails to pin the knight, allowing 9...Nxc5 10.Qxc4 Ne6 and Black keeps the pawn for
only vague compensation. In practice everyone has preferred 9...Qd3, when 10.Qxd3 cxd3 11.Re1 f5
12.Nbd2 0-0 13.Nxe4 fxe4 14.Rxe4 Bf5 is also at least equal.
d4) 9.Qe2 is relatively best, though Black still has the chances: 9...Qe7 (trying for more than
9...Qd3 – even though that’s a nice move to make – 10.Re1 f5 11.Nbd2 0-0 12.Nxe4 fxe4 13.Qxe4
Bf5 14.Qh4 Rae8 with equality) 10.Qxc4 (not 10.Re1? Nxc5 11.Qxc4 Be6! and ...0-0-0 and Black is
clearly better) 10...Bf5!? (and more than 10...Nxc5 11.b4 Ne6 12.b5! Na5 13.Qc3 b6 14.Re1 Bb7
15.Bg5 f6 16.Nd4 Bd5, T.Lagemann-J.Ball, corr. 2004, when 17.Bd2! 0-0-0 18.Qh3 or 17...0-0
18.Qg3 gives White enough for the pawn) 11.Be3 (or if 11.Re1 0-0-0 12.Nc3 Rhe8 13.Bf4,
M.Gluszko-M.Demidowicz, Koszalin 2005, then 13...Qe6!) 11...0-0-0 12.Nc3 Nxc3 13.Qxc3 Be4
14.Nd4?? (after 14.Nd2 Bd5 15.b4 Qe5 16.Qxe5 Nxe5 White is only slightly worse) 14...Nxd4
15.Bxd4 Bxg2! 16.Bxg7 Bc6 17.f3 Rhg8 18.Kh1 f6 0-1 P.Marik-R.Biolek, Klatovy 2003.

6...d5

335
 The central counter-thrust is the strongest option here.
A21: 7.Bb5
A22: 7.exf6
Other moves are weak; e.g. 7.Bb3?! Ne4, or 7.Re1?! 0-0 (as in 6.Re1 above), or if 7.exd6?! 0-0
8.dxc7 Qxc7 9.Bg5 Ne4 10.Bh4 Bg4 11.h3, R.Gagel-Ja.Neubauer, German League 1994, then
11...Bxf3 12.Qxf3 Rae8 13.Re1 (or 13.Bd5 Qe5) 13...Qe5 14.Bd3 Bd6 15.Bg3 Ng5! and Black is
clearly better.

A21: 7.Bb5

This system switch causes a slight move order issue since it transposes to a Modern (5.e5) line with
5...d5 which I’m no longer playing as Black. Fortunately, the early castling makes this an inferior
version for White.

7...Ne4 8.Nxd4 0-0!

Correctly ignoring the threat to c6.

9.Bxc6

The alternative capture 9.Nxc6?! bxc6 10.Bxc6 seems strong at first, winning a pawn while forking
a8 and d5, but Black has a stronger response in 10...Ba6!.

336
 Now White’s path to even a playable game is very narrow:
a) 11.b4? Bxb4 12.Qxd5 Bxf1 13.Qxe4 Qd1! (stronger than 13...Ba6 14.Bd2 Rb8, which led to a
win in J.Jires-J.Tait, Notts Championship 2017) 14.Nc3 Bxc3 15.Bb2 Rad8! and Black wins
(Stockfish).
b) 11.Nd2? Bxf1 12.Nxe4 dxe4 13.Qxf1 Rb8 14.Bxe4 Qd4 15.Bd3 Qxe5 only gave White a pawn
for the exchange, B.Belyakov-L.Suvorov, Titled Tuesday (blitz) 2021.
c) 11.Qxd5?! Bxf1 12.Qxe4 Bb5! (exploiting the weak back rank) 13.Nc3 Bxc6 14.Qxc6 Bd4
15.Bf4 Rb8! (Pálkövi) is also good for Black; e.g. 16.b3 (there is nothing better: 16.Rd1 Rxb2, or
16.Rb1 Qe8 17.Qxc7 Qe6 18.Rd1 Rb4!, or 16.Nd5 Rxb2 17.Rd1 Qh4 18.Bg3 Qh5 19.Rf1 Qe2
20.Ne7+ Kh8 21.Nf5 Bxe5 and so on, A.Meszaros-Z.Hajnal, Kecskemet 2014) 16...Qe8 17.Qf3 (or
17.Qc4 c5) 17...Rb6 18.Re1 Re6 19.Qg3 f6 20.Ne4, M.Raimann-L.Faber, corr. 2001, and now
20...Kh8! leaves White without a good move.
d) 11.Bxa8 Bxf1 12.Kxf1 (forced; not 12.Qxd5? Bc4, or 12.Bxd5? Bc4, or 12.Be3? Bxe3 13.fxe3
Bxg2 14.Qg4 Bh3 15.Qxh3 Qg5+ 16.Kf1 Rxa8 and Black won, O.Herrmann-P.Keres, corr. 1934)
12...Qh4 13.Qf3 Rxa8 14.Nc3 Nxc3 15.bxc3 Qxh2 16.g3 seems to be the best try, when Stockfish
considers White to be only slightly worse. Not 16.Qxd5? Re8! and Black won again in J.Maxfield-
M.Morss, corr. 1992, in view of 17.Qxc5 Rxe5 etc.

9...bxc6

337
10.Be3

The standard move in the Modern main lines, but the fact that Black has not spent a tempo on
...Bd7 means that White is now on the sharp end of the variations.
Instead:
a) 10.Nxc6 can be met by 10...Qd7 11.Nd4 Qe7 (Stockfish prefers 11...Re8!? and claims an edge for
Black, though this has yet to be tested) 12.Bf4 f6 13.e6 (if 13.Be3?!, I.Rogers-Wong Chee Chung,
Singapore 1998, then 13...Ba6! is strong, intending 14.Re1 Nxf2! 15.Bxf2? fxe5 16.c3 Rxf2! 17.Kxf2
Qh4+ 18.Kg1 exd4 and wins, or 15.Qf3 Ne4 with a clear advantage – Marin) 13...Bxe6 14.c3 (not
14.Nxe6 Qxe6 15.Bxc7? in view of 15...Nxf2! 16.Rxf2 Bxf2+ 17.Kxf2 Rac8 18.Ba5 Qf5+ 19.Kg1
Rxc2 – Pálkövi) 14...Bf7 15.Nd2 Rfe8 16.N2b3? (either 16.Re1 or 16.Nxe4 looks roughly equal)
16...Bb6 17.a4 c5 18.Nf3 a5, I.Rogers-M.Pavlovic, Sokobanja 1989, and Black is clearly better, with
attacking ideas of ...g7-g5 and ...f6-f5.
b) 10.b4?! Bb6 doesn’t enhance White’s chances; e.g. 11.Be3 (not 11.Nxc6?! Qe8! 12.Qxd5? Nxf2!
13.Rxf2 Bb7 14.b5 Bxc6 15.Qxc6 Qxe5 and wins, or 12.Nd4 Qxe5 13.Be3 Bg4 14.Qxg4 Bxd4
15.Bxd4 Qxd4 16.c3 Nxc3 17.Qxd4 Ne2+ 18.Kh1 Nxd4 with a clear extra pawn; and 11.a4 Ba6
12.b5 cxb5 13.axb5 Bb7 is good for Black too, G.Cesbron-C.Cruz Lopez Claret, Toulouse 1999)
11...Qe8 12.Nd2 c5!? (I prefer this to either 12...Qxe5 13.Nxc6 Qf6 14.Nxe4 dxe4 15.b5 Bb7 16.Nd4,
or 12...Nc3 13.Qf3 a5 14.bxa5 Bxa5, M.Santo-T.Baranowski, corr. 2011, when 15.N4b3 Qxe5
16.Rfe1 or 15...Bb4 16.Bc5 Bxc5 17.Qxc3 Ba7 18.Rfe1 gives White more control) 13.bxc5 Nxc5
14.Qf3 Qxe5 15.Nc6 Qe6 16.Qxd5 Bb7 17.Bxc5 Qxd5 18.Ne7+ Kh8 19.Nxd5 Bxc5 and Black is
better with the bishops.
c) 10.f3? runs into 10...f6!

338
 11.fxe4 (not 11.exf6? Qxf6 12.Be3 Ba6 13.Re1 Rae8 14.c3 Bd6, threatening ...Qh4, or 15.g3 Nxg3
16.hxg3 Bxg3 and wins, M.Chiburdanidze-Ma.Tseitlin, Moscow 1989) 11...fxe5 12.Rxf8+ Qxf8
13.Be3?! (or 13.c3 Rb8, intending 14.b4? Rxb4! etc, V.Konev-D.Hudak, corr. 2012) 13...exd4
14.Bxd4 Bg4 (14...Rb8 may be even better) 15.Qd2 dxe4 16.Qc3 Bxd4+ 17.Qxd4 Qd6 18.Qxd6 cxd6
with a winning endgame, Y.Estrin-V.Antoshin, Leningrad 1954.
d) 10.Nc3? asks for 10...Ba6! 11.Nxe4 (not 11.Re1? Nxf2! 12.Kxf2 Qh4+, A.Ergonen-A.Ataman,
Konya 2010; or 11.Nxc6? Nxc3 12.bxc3 Qd7 13.Nd4 Bxf1 14.Qxf1 f6) 11...dxe4 12.Nf5 (or 12.Nxc6
Qh4 13.g3 Qh3 14.Re1 Bxf2+! and so on, G.Letay-Be.Lengyel, Hungarian League 1994) 12...Bxf1
13.Qg4 g6 14.Bg5 gxf5 15.Qg3 f4 16.Qh4 (or 16.Qxf4 f6) 16...Qd7 and wins, C.Jahn-K.Dabrowska,
Dresden 1994.

10...Qe8! 11.Nd2

Other moves are worse:

a) 11.f3? is still a mistake because of the pin on the e-pawn; e.g. 11...Nd6! 12.f4? (but 12.Bf4 Nc4
and 12.Bf2 Nc4! are bad anyway) 12...f6! 13.Bf2 fxe5 14.fxe5 Ne4 15.c3 Nxf2 16.Rxf2 Rxf2
17.Kxf2 Qxe5 and Black soon won, J.Arnott-J.Tait, Sheffield League 2002.
b) 11.c3 Bb6 prepares what happens next: 12.Nd2?! c5 13.N4b3 Qxe5 14.Nxe4 dxe4 15.Qe2 a5
16.a4 Ba6 17.c4 Rad8 18.Rfd1 Rd3 19.Rxd3 Bxc4 and Black soon won, C.Caballo-R.Garbarino,
Buenos Aires 1995.
c) 11.Re1 doesn’t deter Black from taking the pawn: 11...Qxe5 12.Nxc6 Qd6 13.Nd4 Qg6 (or
13...f5) 14.Nf3 Bh3 15.Nh4 Qf6 16.Bxc5 Qxh4 17.gxh3!? (here 17.b4 Nxc5 18.bxc5 Qg5 19.Qf3
might be safer) 17...Nxc5 18.Qxd5 Ne6 19.Nc3 Rad8 20.Qg2 Rd2 21.Qg3 Qh6 with an ongoing
initiative for the pawn, and Black won... in 135 moves, T.Upton-M.Condie, Scottish Ch., St. Andrews
1989.

339
11...Bb6

Securing the bishop and freeing the c6-pawn. This looks better than 11...f5 12.f4 Bb6 13.Nxe4 fxe4
14.b4 a5 15.a3 and White is fine, A.Golizadeh-R.Khusnutdinov, Titled Tuesday (blitz) 2021; or
11...Nxd2 12.Qxd2 Bb6 13.a4 a5, F.Munoz-W.Espinoza Palomino, Medellin 2013, and then 14.Ra3!,
intending 14...c5? 15.Bh6!.

12.Nxe4 dxe4 13.Qh5

Defending the e-pawn for the moment.

13...f5 14.Qxe8 Rxe8

340
 Black can be more than happy here with the two bishops and a ready target on e5, so it’s a bit
annoying that Stockfish assesses the position as equal (assuming White is careful).
a) 15.Nxc6?! is wrong for a start, since 15...Bb7 16.Bxb6 (not 16.Nd4??, L.Schut-E.Ohanyan,
Titled Tuesday blitz 2021, when 16...f4 wins at once) 16...axb6 17.Nd4 Rxe5 is good for Black.
b) 15.Rfd1 c5 16.Nb3?! c4 17.Bxb6? axb6 18.Nd2 c3 19.bxc3 Rxe5 also gives Black a big
advantage, S.Zielinski-K.Pinkas, Polish League 2014.
Instead, Stockfish goes for 16.Nb5!, adding the further 16...Kf7 17.a4 a5 18.Rd2 h6 19.h4 Rxe5
20.Rad1 and, okay, it won’t be easy to make progress. All the same, I’d rather be Black.

A22: 7.exf6 dxc4

The main line. White now has two possibilities:

A221: 8.Re1+
A222: 8.fxg7

A221: 8.Re1+ Kf8!?

341
 A spoiler’s reply, avoiding the well-mapped complications of 8...Be6. GM Boris Alterman
considered 8...Kf8 to be “a tame move, which allows White to gain a terrific initiative”. We’ll see
about that.

9.Bg5

The traditional “refutation”. Three other moves need to be examined too:

a) 9.fxg7+ Kxg7 solves the problem of the h8-rook. White can count on reasonable play with the
black king now in the open, though no more than that: 10.Nbd2 (not 10.Ng5? Re8 11.Ne4 Bb4 12.c3,
M.Redondo-J.Ramis, San Fernando 1992, which fails to 12...dxc3! 13.Qh5 Re5 14.Qh6+ Kh8
15.Nbxc3 Bf8; while after 10.Ne5 Qf6 – Fishbein, Black is certainly not worse; e.g. 11.Nxc6 Qxc6
12.Nd2 b6 13.Nxc4 Bb7 14.Qg4+ Qg6 15.Qh3 f6) 10...Bg4 11.h3 (or 11.Ne4 Qd5 12.Bg5 Rhe8
13.Bf6+ Kg8 14.Nxc5 Qxc5 15.Qd2 Qf5 16.Qh6 Qg6) 11...Bh5 12.Ne4 Be7 (or 12...Qd5 again)
13.Ng3 Bg6 14.Bf4 (here 14.Ne5 Nxe5 15.Rxe5 Re8 16.Qf3 Bd6 17.Nh5+ Bxh5 18.Bh6+! Kxh6
19.Qxh5+ Kg7 20.Rxe8 Qxe8 21.Qg5+ draws) 14...Re8?! (14...Bf6 keeps it level) 15.Qd2?! (the
immediate 15.h4 h5 16.Ne5 was more promising) 15...Kg8 16.h4 h5 17.Ne5? (this no longer works)
17...Bb4! 18.Qe2? d3 19.cxd3 cxd3 20.Qf3 Bxe1 and Black won, J.Janecek-R.Dolezal, Tabor 2010.
b) 9.Ng5 seems never to have been tried, perhaps because no one has managed to improve on
Keres’ analysis:

342
 9...Bf5 10.fxg7+ Kxg7 11.Nxf7! (not 11.Ne4? Bxe4 12.Rxe4 Re8 13.Rg4+ Kh8 14.Bh6 Rg8)
11...Qf6 (not 11...Kxf7? 12.Qh5+ Bg6 13.Qxc5) 12.Nxh8 d3! and Black has sufficient compensation;
e.g. 13.Qf3 (or 13.Nd2 Kxh8 – Keres, and if 14.Nf3 then 14...Bg4 – Sovremenny Debyut) 13...Kxh8
14.Bd2 (or 14.cxd3 Ne5 15.Rxe5 Qxe5 16.Bd2 cxd3 17.Nc3 Rf8 18.Nd5 Bd4) 14...Nd4 15.Qd5 Qd6
(or 15...dxc2!? 16.Qxc5 Nf3+! 17.Kh1 Nxe1 18.Bc3 c1=Q 19.Bxf6+ Kg8) 16.Qxd6 (or 16.Qxb7 Rf8
17.Bc3 Kg8 18.Nd2 Ne2+ 19.Kh1 Bxf2) 16...cxd6 17.Bc3 (or 17.cxd3 Nc2 18.Nc3 Nxa1 19.Rxa1
Bxd3) 17...b5 18.cxd3 b4 19.Bxd4+ Bxd4 20.dxc4 Bxb2 21.Nd2 Bxa1 22.Rxa1 Re8, now with
enough play in the endgame.
c) 9.Nbd2 is recommended by IM Baker & FM Burgess, as well as GM Fishbein. B&B also
helpfully provide the antidote: 9...Bf5! (neutralizing the threat of 10.Ne4, the strength of which is
demonstrated by 9...Qd5? 10.Ne4 Bf5? 11.fxg7+ Kxg7 12.Bh6+! Kxh6 13.Qd2+ Kg7 14.Qg5+ Kf8
15.Qh6+ Ke7 16.Nxc5+ Be6 17.Rxe6+! 1-0 P.Ponkratov-V.Shinkevich, Khanty-Mansiysk 2013; or
9...Qxf6? 10.Ne4 Qf5 11.Nxc5 Qxc5 12.b4! Nxb4 13.Nxd4 with a big advantage – Fishbein)

343
 c1) 10.fxg7+ Kxg7 11.Nxc4 (now if 11.Ne4, Black has the added option of 11...Be7!?) 11...f6!
looks fine, taking control of key dark squares, when Black can develop easily with ...Qd7 and ...Rhe8;
e.g. 12.a3 (or 12.Nh4 Bg6 13.Qf3 Qd7 14.Bf4 Rhe8) 12...Re8 13.Rxe8 (or 13.b4 Rxe1+ 14.Nxe1 Bf8
15.Bb2 Qd7 16.b5 Ne5) 13...Qxe8 14.b4 Qe6 15.Na5 Nxa5 16.bxc5 Nc6 17.Bb2 Rd8 18.Nxd4 Nxd4
19.Bxd4 Bxc2 with a probable draw.
c2) 10.Ne4 Bxe4 11.fxg7+ Kxg7 12.Rxe4 h5! (forced, since the h6- and h5-squares need covering;
not 12...Re8? 13.Bh6+! Kxh6 14.Rg4 f5 15.Qd2+ Re3 16.Rh4+ Kg7 17.fxe3 and White “wins an
exchange” – Fishbein; or 12...Rg8? 13.Ng5! Qd6 14.Qh5 Qg6 15.Ne6+! fxe6 16.Rg4 and White just
wins, M.Volkovich-A.Martynenko, Dagomys 2009) 13.Nh4 (anything else and Black plays ...Qd5
with equality) 13...Qf6 (not now 13...Qd5?? 14.Qf3! and wins) 14.Bd2 (or 14.Rf4 Qe6 first) 14...Rae8
15.Rf4 Qe6 16.Qf3 Ne5 17.Qxb7 Qc6 “and Black engineers a queen exchange, but may still suffer
due to his many weak pawns” (B&B).
Or not, as the case may be; e.g. 18.Qxc6 (18.Nf5+ Kg6 19.Nh4+ Kg7 repeats) 18...Nxc6 19.c3 (or
19.Rd1 Bd6) 19...Re2 20.Rd1 Rb8 with quite enough activity to compensate for the structural
weaknesses. Stockfish declares “0.00” in this line – as well as after 17.Qg3+ Ng6 18.Nf5+ (or 18.Rf3
Rh7 19.Qxc7 Nxh4 20.Rg3+ Kh8 21.Qxc5 f6) 18...Kf8 19.Rf3 h4! 20.Qg5 (not 20.Qxc7?! b6,
threatening ...Rh5 or ...Ne5) 20...Rh7 (threatening ...Ne5) 21.Nxh4 Rxh4 22.Qxc5+ Qd6 23.Qxd6+
cxd6.

9...gxf6 10.Bh6+ Kg8

344
 White has sacrificed two pawns to curtail the black king and rook. In fact “Black’s king is one
check away from mate” (Alterman), so Black should be looking to play 11...Bf8 next move unless
there is a specific reason why not (as in notes ‘a’ and ‘c’ below).

11.Nc3?!

This predominant choice often receives an “!”, except for B&B who give it the more appropriate
text mark. It’s basically a trick, which only works with Black’s assistance.
White has tried numerous other moves, and most are better:
a) 11.Nxd4? (this isn’t one of them) 11...Nxd4! (11...Bxd4 12.c3 is less clear) 12.c3 Be6 13.Qh5 (or
13.cxd4 Qxd4 14.Qf3 Qh4) 13...Qd5 14.Qh4 Ne2+ 15.Kf1 Bf8 16.Rxe2 Bxh6 17.Qxh6 Qg5
18.Qxg5+ fxg5 is good for Black.
b) 11.Nh4 Bf8 was apparently mentioned by Lange himself. Then 12.Qh5! Ne5 13.Bxf8 Qxf8
14.Rxe5 fxe5 15.Qg5+ Qg7 16.Qd8+ Qf8 17.Qg5+ is a draw. Black can also play 11...Ne5 first,
followed by 12.Nd2 Bf8 13.Bxf8 Kxf8 14.Qh5 h6, or 12.h3 Bf8 13.Qh5 Bxh6 14.Qxh6 Ng6 15.Nd2
f5.
c) 11.Qd2 Bf5! (note that 11...Bf8?¸ as in Jo.Nelson-J.Tait, Sheffield League 2013, is a mistake
here because of 12.Bxf8! Qxf8 13.Nxd4 with a clear advantage, and 12...Kxf8? 13.Qh6+ Kg8
14.Nbd2 is even worse) 12.Na3 (or 12.Qf4 Bxc2 13.Nbd2 Bd6 14.Qh4 Bf8 15.Bxf8 Kxf8 16.Qh6+
Kg8 17.Rac1 Bg6 18.Rxc4 Qf8 19.Qh4 Kg7 20.Nxd4 Nxd4 21.Qxd4 Rd8 22.Qc3 Rd3 23.Qc1 Qd8
24.Nf1 c6 25.Ne3 Re8 and Black consolidated, H.Hirscheider-A.Rades, corr. 2006) 12...Bf8 (or
12...Bg6!? 13.Rad1 c3 14.Qc1 Bf8 15.bxc3 Bxh6 16.Qxh6 Qf8 17.Qxf8+ Kxf8 18.Nxd4 Kg7 ½-½
R.Higham-J.Tait, Notts League 1990, anticipating the mutual wreckage after 19.Nxc6 bxc6) 13.Bxf8
Qxf8 14.Nxd4 Rd8 15.Nab5 (or 15.c3 Bg6) 15...Nxd4 16.Nxd4 Bg6 17.Qf4 Qd6 is roughly equal;
e.g. 18.Qf3 Kg7 19.Nf5+ Bxf5 20.Qxf5 h5 21.Qf3 c6 22.Rad1 Rde8 23.Rxe8 Rxe8 24.Qxh5 Qd2!?

345
25.Kf1 Qxc2 26.Qg4+ Kf8 27.Qh5 Kg7 with a draw.
d) 11.Qc1 Bf8 is safe now because the d4-pawn is not en prise; e.g. 12.Bxf8 Qxf8 13.Qf4 Qd6
14.Qh6 Bf5. If instead 12.c3!?, as in W.Lattacher-M.Zödl, corr. 2010, then 12...Be6! 13.Bxf8 (or
13.Qf4 Qd6 14.Qh4 Bxh6 15.Qxh6 f5) 13...Qxf8 14.Qf4 (or 14.Nxd4 Nxd4 15.cxd4 Kg7 16.Re3
Rg8) 14...h5 (or 14...Qg7 15.Qxc7 Bd5) 15.Nxd4 (or 15.cxd4 Qg7 16.Nc3 Nb4; not 15.Qxf6?! Qg7
16.Qf4 Rh6, or 15.Qxc7?! dxc3 16.Nxc3 Qb8) 15...Nxd4 16.cxd4 c6 17.Qxf6 Rh6 and Stockfish
gives its favourite “0.00”.
e) 11.Nbd2 Bf8!

(otherwise 12.Ne4 might follow) 12.Bxf8 (not 12.Bf4?! Be6 13.Ne4 h6 14.Bxc7? Qxc7 15.Nxf6+
Kg7, K.Morton-J.Tait, Sheffield Cup 2011, as 16.Nh5+ Kh7 17.Nf6+ Kg6 is not a draw) 12...Kxf8
13.Nxc4 Be6 (or 13...Qd5 14.Ne3 Qc5 15.Nf1 Bf5 16.Qd2 Kg7 17.Ng3 Bg6 18.Ne4 Qf5, S.Cullen-
J.Tait, Scarborough 1989, when 19.Nxd4 Nxd4 20.Qxd4 Rhd8 21.Qc3 Qe5 is pretty equal) 14.Ncd2?!
(14.b3 is correct and roughly equal; e.g. 14...Rg8 15.Qd2 Rg4!? 16.Qh6+ Kg8 17.Rad1 Rg6 18.Qf4
Rg4 with a draw) 14...Rg8! 15.Ne4 (recovering a pawn has cost White two tempi on the main line)
15...f5 16.Ng3 (or 16.Nc5 Bd5, intending 17.Nxb7? Rxg2+! 18.Kxg2 Qg5+ 19.Kf1 Qg4 20.Re3 f4
and wins) 16...f4 17.Nh5 Qd6 18.Nf6 Rg6 19.Nxh7+?! (or 19.Ne4 Qd5) 19...Kg7 20.Nhg5 Bg4
21.Qd2 f6 22.h3 Rh8 23.hxg4 fxg5 and ...Rgh6 with a very strong attack, J.Szopa-P.Szymanski, corr.
2015.
f) 11.Na3 Bf8 12.Qd2 (otherwise 12.Bxf8 Kxf8 13.Nxc4 is note ‘e’) 12...Be6 13.Bxf8 Qxf8 (not
13...Kxf8?! 14.Qh6+ Kg8 15.Rad1 with an edge) 14.Nxd4 (or 14.Nb5 Qb4 15.Qxb4 Nxb4 16.Nfxd4
Rd8 17.Rad1 c6 18.Nc7 Bc8 with equality, G.Bischof-M.Hobert, corr. 2006) 14...Rd8 15.Nxe6 (or
15.Re4 h5) 15...Rxd2 16.Nxf8 Kxf8 17.Red1 Rxd1+ 18.Rxd1 c3 19.bxc3 Ke7 ½-½ Reprimand-
tsmenace, ChessWorld.net 2020.
g) 11.c3 is an untried suggestion by Stockfish, which gives the further 11...Bg4 (the usual 11...Bf8
is okay too) 12.Qc1 Bxf3 13.Qf4 Bh5 14.Qf5 Bf8 15.Qxh5 Bxh6 16.Qxh6 Qd5 17.Qxf6 Rd8 18.Nd2

346
dxc3 19.Nxc4 cxb2 20.Rad1 Qxd1 21.Qg5+ Kf8 22.Qh6+ Kg8 with another draw.

11...Bf8!

Other lines – such as 11...Bg4 12.Ne4 Bf8 – may be defensible, but Black can play for more than
that now.

12.Bxf8

Not 12.Qd2?? dxc3 13.Qxd8 Nxd8 14.Re8 Ne6 15.Nd4 Ng7 and Black wins.
The only alternative is 12.Nxd4 Bxh6 (not 12...Nxd4?? 13.Qxd4 Qxd4? 14.Re8 Qd6 15.Nd5! and
mates) 13.Nxc6 Qxd1 14.Ne7+ Kg7 15.Raxd1 Be6 16.Ned5, dating back to Bilguer’s Handbuch,
“with roughly equal chances” (Estrin). I don’t know about that. After 16...Rac8 Black still has an
extra pawn – and two bishops for the moment – so White will have to defend accurately even to
draw; e.g. 17.Ne4 f5 (another option is 17...Bxd5 18.Rxd5, as in P.Gayson-J.Tait, National Handicap
rapid 1994, and then 18...Rcd8 19.Rc5 Rhe8 20.g4 Rd4 21.f3 Bf4) 18.Nc5 Bxd5 19.Rxd5 Rhe8
20.Kf1 b6 21.Rxe8 Rxe8 22.Nd7 (or 22.Na6 Re7) 22...Bc1 23.b3 cxb3 24.axb3 Bb2 25.Rxf5 Bc3
(this plan allows White counterplay; Stockfish gives 25...f6! as good for Black) 26.g4 Re1+ 27.Kg2
Rc1 28.Rg5+ Kh6 29.Rh5+ Kg6 30.Nf8+ Kg7 31.Nxh7 Rxc2 32.Ng5 Bd4 33.Ne4 Re2 34.Rg5+ Kf8
35.Rd5 c5 36.Kf3 Rb2 37.Rd7 Rxb3+ 38.Kf4 a5 39.Nd6 Kg8 40.Rxf7 Rb2 41.Ra7 Rxf2+ 42.Kg5
Rf6 43.Nf5 Re6 44.Ne7+ Kf7 45.Ng6+ ½-½ G.Crook-J.Tait, corr. 2000. So my opponent did defend
accurately to draw, but I’d be more than happy to have another go.

12...Kxf8 13.Ne4

Instead, 13.Nb5 (or 13.Ne2) 13.Nb5 Bg4 14.Nbxd4 (or 14.Qd2 h5 15.Nbxd4 Nxd4 16.Nxd4 c5)
14...Nxd4 15.Qxd4 Qxd4 (not 15...Bxf3?!, as 16.Qf4! equalizes) 16.Nxd4, G.Beccaris-Ha.Karl,

347
Arvier 2005, is similar to the 12.Nxd4 line. After 16...Rd8 White would again have to play well to
draw.

13...f5!

B&H’s addition to Bilguer’s old analysis which ended one ply sooner. Estrin (1983) wrote that “it
is not easy for White to show that he has compensation for two pawns.” In 1998 IM Chris Baker
extended it by another ply:

14.Ng3

“With White’s lead in development his attacking chances should not be underestimated” (Baker),
perhaps influenced by beating me rather easily. But given the later “?!” to 11.Nc3, it seems he’s since
changed his mind.
Our game went 14...Rg8 15.Qd2 Rg6 16.Rad1 Rd6? (trying to hang on to everything is foolhardy;
Stockfish suggests several equalizing lines; e.g. 16...f4 17.Qxf4 Qd6 18.Qe4 Bg4 19.h3 Bxf3 20.Qxf3
Rd8) 17.Nh5 b5? 18.Qf4 a5? (rubbish) 19.Nxd4 Nxd4 20.Rxd4 Be6 21.Qh6+ Ke8 22.Ng7+ Kd7
23.Nxe6 fxe6 24.Rxd6+ cxd6 25.Qxe6+ Kc7 26.Qf7+ Kb6 27.Qxf5 and White won, C.W.Baker-
J.Tait, Nottingham Gambit (rapid) 1995.
Then I went home, studied it properly, discovered that 14...f4! is the correct move, and later
included it in an article for the BCCA magazine. For example: 15.Ne2 (or 15.Nh5 Bg4 16.Nxf4 Qd6)
15...Rg8! 16.Nxf4 (not 16.Nfxd4? Qd5 17.f3 Bh3, or 16.Nexd4?! Bh3 17.g3? Bg4 18.Nxc6 Qxd1
19.Raxd1 Bxf3 or 18.c3 Nxd4 19.cxd4 Qd5 and wins) 16...Bg4 17.h3 Qd6 and Black is clearly better.
This has received one test: 16.Qd2 Qf6 17.Nfxd4 (here 17.Qxf4 Qxf4 18.Nxf4 Bf5 was all I gave in
1998) 17...Bh3 18.g3 Rd8 19.c3 Nxd4 20.cxd4, J.Del Pozo Hernandez-G.R.McDonald, corr. 2000,
where Black made what I assume was a clerical error in 20...f3? and went on to lose. Instead,
20...fxg3 21.hxg3 Rg5 22.Rac1 b5 would likely have led to the opposite result, in view of Black’s
extra pawn and White’s weak light squares.

A222: 8.fxg7 Rg8

348
 Having put considerable effort into 8.Re1+ Kf8, the recent resurgence of 8.fxg7 is somewhat
annoying. Nevertheless, the black king can still enjoy a little stroll.

9.Bg5

Instead, 9.Re1+ Be7 10.Bg5 transposes to 10.Re1 below. Here 10.Na3 is also playable; e.g.
10...Qd5 (or 10...Rxg7 11.Nb5) 11.Bg5 Be6 12.Bxe7 Kxe7 13.b3! b5 (or if 13...c3, B.Greifer-
T.Willis, corr. 1980, then 14.Qd3 Rxg7 15.Nb5) 14.bxc4 bxc4 15.Rb1 a6 16.c3 Rxg7 17.cxd4 Kf8
with equality.
Other moves are very much worse; e.g. 10.Bh6? Be6 11.Nbd2 Qd5 12.Ne4 0-0-0, A.Rades-
M.Hobert, corr. 2006; or 10.Ng5? Rxg7 11.Qh5 Bf5 12.Qh6 Rg6 13.Qh5 Qd5 14.Bf4 0-0-0 and
Black is completely winning, L.M.De Silva-R.Leitao, Titled Tuesday (blitz) 2021.
The immediate 9.Bh6? is no better: 9...Qf6 10.Qd2 Be6 11.Qg5 Be7 12.Nbd2 0-0-0 13.Ne4? Qxf3!
14.gxf3 Bxg5 15.Nxg5 Bd5 and Black is winning again, S.Plukkel-H.Visser, Amsterdam 2011.

349
9...Be7

Should Black wish to avoid the main line then 9...f6!? 10.Re1+ (or 10.Bh6 Bg4) 10...Kf7 is
complicated and probably playable, though 10.Re1 etc in the next note is still relevant because White
could always play 9.Re1+ first.

10.Bxe7

White has a significant alternative in 10.Re1 Be6 11.Bxe7 Qxe7! (note the differing queen capture
here; 11...Kxe7 lands Black in the 11...Be6 main line) 12.Nxd4 Rd8 (not 12...0-0-0? 13.Nxc6 bxc6

350
14.Qf3) 13.c3 Rxg7 (“mit schwarzem Vorteil” – Lasker; not yet 13...Nxd4?! 14.cxd4 Rxg7 15.Nc3
Kf8 16.Qf3 c6 17.Re4, when White has a strategic edge based on the superior minor piece and pawn
structure). By delaying the capture, Black can organize stronger pressure on d4.

This position has been tested in three TCEC games, all of which were drawn without any real
problems for Black:
a) 14.Qa4?? runs into 14...Rxg2+! 15.Kxg2 Bh3+ 16.Kxh3 Qxe1 with a decisive attack; e.g.
17.Qxc4 Nxd4 18.cxd4 Rd6 19.Qb5+ Kf8 20.Qc5 Kg8 21.d5 Rg6 22.Qd4 Qg1 23.Qf4 f5 24.f3 Qf1+
25.Kh4 Rg4+ and wins, B.Piskaykin-E.Schmidek, Willingen 2015.
b) 14.g3 Nxd4 15.cxd4 Rg4 16.Nc3?! Rgxd4 17.Qh5 Kd7! 18.Ne2?! (or 18.Qxh7 Kc8) 18...Rd2
19.Qa5, P.Pisk-G.Halvax, Austrian League 2012, and now 19...Qd6! 20.Qxa7 Qb6 is very good for
Black.
c) 14.Nxe6 Rxd1 15.Nxg7+ Kf8 16.Rxd1 Kxg7 and the queen is the equal of the rooks, AllieStein-
Jonny 8, TCEC 15 Rapid Bonus 2019.
d) 14.Na3 Nxd4 15.cxd4 c5! 16.Qe2 cxd4 17.Nxc4 d3 18.Qd2 Rg4 and the alteration to the pawn
structure is not unfavourable for Black, KomodoMCTS-Komodo, TCEC Cup 3 Quarter-final 2019.
e) 14.Qf3 Nxd4 15.cxd4 c6 16.Nc3 Rxd4 17.Rad1 Rgg4 (Stockfish prefers 17...Rxd1 18.Qxd1 Rg6
19.Qd4 Kf8 20.Qxa7 b5 21.Qa8+ Qe8 with its “0.00”) 18.Ne4 Rxd1 19.Qxd1 Kf8 20.Qd2 Kg8 21.f4
h6 22.Qf2 f5 23.Nc5 Rg6 24.Qe3 Kf7 25.h4 h5 26.Qe5 Qf6 27.Nxe6 Qxe5 28.Nd8+ Ke7 29.Rxe5+
Kxd8 30.Rxf5 Rg4 31.Rxh5 b5 32.Kf2 Rxf4+ 33.Ke3 Rg4 and the rook endgame was drawn,
Komodo-KomodoMCTS, TCEC Cup 3 Quarter-final 2019.

10...Kxe7!

351
 Off he goes. Any other recapture would drop the d4-pawn.

11.Re1+

Nothing else is at all testing:

a) 11.Nbd2 (or 11.Na3) 11...Rxg7 12.Nxc4 Kf8 is fine for Black; e.g. 13.Re1 Bg4 14.Re4 f5
15.Re1 (or 15.Rf4?! b5!) 15...Qd5 16.Ncd2 Kg8.
b) 11.b4 cxb3 12.Nbd2 (or 12.Re1+ Kf6) 12...Rxg7 13.Nxb3 Bh3 14.Re1+ Kf8 15.g3 Qd5 and
Black is better, Ja.Balogh-Las.Szabo, Hungarian Ch., Budapest 1946.
c) 11.Qe2+ Be6 12.Nbd2 Qd5 13.Rfe1, I.Gorodetskij-F.Gibentif, Moscow 1996, and now 13...Kf6!,
intending ...Kxg7 with an edge.

11...Kf6

Continuing the king walk. 11...Be6 was considered good for Black in former times, but White’s
play has since been strengthened; e.g. 12.Re4 f5 (the best reply; not 12...Rxg7? 13.Nxd4 Nxd4
14.Rxd4 Qg8 15.g3 Rd8 16.Nc3 with a big advantage, Jo.Nelson-J.Tait, Sheffield Cup 2011) 13.Rh4
Kf7 14.Rxh7 Rxg7 15.Rxg7+ Kxg7 (H.Fahrni-S.Tartakower, Baden 1914) and now 16.Qc1 Qf6
17.Qf4 is good for White, Komodo-Chiron 2, TCEC 7 Stage 3 2014.
However, B&B give 13...Rxg7 14.Nxd4 Ke8! “with roughly level chances in a wild position”, and
15.Qh5+ Bf7 16.Qe2+ Kf8 17.Nxf5 Rg5 18.Rf4 Qf6 19.Qf3 Ne5 20.Qh3 Bd5 21.Qh6+ Qxh6
22.Nxh6+ Ke8 23.Nc3 Rxg2+ 24.Kf1 Rxh2 25.Nxd5 Rxh6 led to a draw, Jonny 8-AllieStein, TCEC
15 Rapid Bonus 2019. So perhaps 11...Be6 is okay after all.

352
12.Nbd2

The primary choice of Palliser & Williams in Grandmaster Gambits: 1.e4.

a) 12.Na3 gives Black the option of 12...c3!? (otherwise 12...Kxg7 13.Nxc4 transposes to the main
line) 13.bxc3 dxc3 14.Qe2 Kxg7, followed by ...Qf6 or ...Bg4 with equality.
b) 12.Re4 Kxg7 13.Nxd4

13...f5 (more forcing than 13...Kh8) 14.Rf4 Kh8 15.Nc3 (not 15.Qd2?! Qg5! 16.g3 Rd8 17.Qe3
Qf6 18.Nxc6? Rd1+ 19.Kg2 bxc6 20.Qe8+ Kg7 21.Qe2 Qd6 and White is in big trouble, Kl.Amann-
L.Korzh, corr. 2009) 15...Qg5! (the queen has to step nimbly and precisely, making constant threats,

353
to prevent White from consolidating) 16.Qf3 Ne5 17.Qg3 Qd8! (not 17...Qf6?, A.Riess-E.Schmidek,
Willingen 2015, due to 18.Nd5!) 18.Qe3 Ng6 19.Nce2 Qf6! (not 19...Nxf4? 20.Nxf4 Qf6? 21.Nh5
and wins) 20.Nf3 Qxb2 21.Re1 Qg7! 22.Rxc4 b6 23.Qc3 Qxc3 24.Rxc3 c5 ½-½ Honeybunch-
tsmenace, ChessWorld.net 2019.
c) 12.c3 “is interesting and untried” (B&B), who mention the line 12...Bg4 13.h3 Bxf3 14.Qxf3+
Kxg7 15.Re4. This has now been tried: 15...Re8 16.Rxe8 (if instead 16.Nd2, then 16...Rxe4 17.Nxe4
Qh4 18.Ng3 Kf8 19.Nf5 Qf6 20.Nxd4 Qxf3 21.Nxf3 Rd8, or 20.cxd4 Nxd4 21.Nxd4 Qxd4 22.Qxb7
Re8 with sufficient counterplay) 16...Qxe8 17.Na3 Rd8 18.Nxc4 Qe6 19.Nd2 (or 19.b3 Qf6)
19...dxc3 20.Qg3+ Qg6 21.Qxc3+ Qf6 22.Qxf6+ Kxf6 ½-½ drink1966-tsmenace, ChessWorld.net
2020.

12...Kxg7 13.Nxc4 Kh8 14.Nce5 Nxe5 15.Nxe5

If 15.Rxe5, as in A.Karpatchev-G.Flear, Saint Affrique 2011, then simply 15...f6! is fine; e.g.
16.Rh5 (or 16.Re4 c5) 16...Bg4 17.Rh4 c5 18.Ng5!? Rxg5 19.Rxg4 Rxg4 20.Qxg4 Qe7 with a
roughly level endgame.

15...Qd5

In this possibly critical position White has tried (or analysed) two moves:
a) 16.Qf3 Qxf3 17.Nxf3 c5 (17...Bh3 18.g3 transposes to note ‘b’) 18.Re7 Kg7! (coming back
again) 19.Rae1 Kf6 20.Rc7 Bh3 21.g3 and now 21...Rge8! (improving on 21...b6?! 22.Re4 Bf5
23.Rf4 Ke6 24.c3 dxc3 25.bxc3 with an enduring initiative for White, T.Lovholt-K.Kögler, corr.
2004) 22.Rxc5 Rxe1+ 23.Nxe1 Re8 24.Nd3 Re2, intending ...Rd2, when White can’t do anything.
b) 16.g3 Bh3 17.Qf3 (if 17.f3 then 17...Rad8, or 17.Nf3 f6, intending ...Bg4 or ...c7-c5) 17...Qxf3
18.Nxf3 c5 19.Re7 is suggested by Palliser & Williams, “seizing the initiative at the start of the

354
endgame”, which is fair comment. But as it happens, I’d already looked at this and concluded that
after 19...b6! Black is okay with precise play.

For example:
b1) 20.Rxf7 Rae8! 21.Rxa7 (not 21.Rd1 Re2 22.Rd2?? Rge8 and Black is winning) 21...Re2
22.Rc1 Bf5 23.Rc7 Be4 (or 23...Rf8) 24.Ne1 Rf8 25.f4 Re8 and Black’s control is enough for two
pawns, one of which can be regained at once by retreating the bishop.
b2) 20.Ne5 Rge8! 21.Nxf7+ Kg7! 22.Rb7 (not 22.Rae1? Rxe7 23.Rxe7 Kf6 24.Rc7 Re8) 22...Kg8!
23.Nh6+ (or 23.f4 Re2) 23...Kh8, intending ...Re2, or 24.Nf7+ Kg8 25.Nh6+ Kh8 with a draw.
arising most often via 3...Bc5 4.c3 Nf6 5.d3 0-0 6.Bb3 d5 7.Qe2

355
B: 4.d3

Defending the e-pawn thusly is White’s most popular choice nowadays, especially at grandmaster
level. I can see its attraction: spoiling Black’s fun and aiming instead for a strategic Closed Ruy type
of game after c2-c3, 0-0, Bb3 (in some order), possibly b2-b4 should the f8-bishop come to c5. I’ve
tried various moves in response: the standard 4...Bc5 or 4...Be7 (and done okay) and 4...h6 (and done
better), but allowing White what they want rankles with me. So in 2017 I went back to an old
favourite.

4...d5!?

“Risky” (B&H); “Premature” (Zagorovsky, Estrin, Lane, Pálkövi, Kravtsiv); “Dubious” (Müller &
Souleidis). IM Jan Pinski is the only author to have a favourable opinion of this move, writing:
“White usually plays [4.d3] in search for a quiet game. 4...d5 declines this suggestion [and] is
therefore interesting for that reason alone.”

5.exd5

The drawback to having played ...d7-d5 so soon is that it exposes the e5-pawn to attack via 0-0 and
Re1 before Black is ready to defend it. Other moves therefore make little sense:
a) 5.Bb3?! dxe4 6.Ng5 (not 6.dxe4? Qxd1+ 7.Kxd1 Nxe4) 6...Be6 7.Bxe6 (or 7.Nxe6 fxe6 8.Bxe6
exd3 9.cxd3 Qd6) 7...fxe6 8.Nxe4 (and not 8.Nxe6? Qd7 9.Nxf8 Rxf8, when White is seriously
behind in development) 8...Nxe4 9.dxe4 Qh4 and Black is at least equal, R.Salim-H.Gal, Mureck
2008.
b) 5.Bb5?! dxe4 6.Nxe5 Qd5! 7.Bxc6+ bxc6 8.d4 c5 9.Be3, M.Holly-J.Dano, corr. 2001, and now
9...Qe6! (avoiding 9...Bd6 10.Nc3!) 10.c3 (or 10.0-0 Bd6) 10...Bd6 11.Qb3 (or 11.Qa4+ Nd7) 11...0-

356
0 12.Qxe6 Bxe6 13.Nd2 cxd4 14.cxd4 Rfb8 15.b3 a5 with a strong initiative.

5...Nxd5

6.0-0

Little else makes much sense here either, though that hasn’t prevented White from playing all sorts
of different moves:
a) 6.Nxe5?? Nxe5 7.Qe2 fails to 7...Bb4+! 8.c3 0-0 9.0-0 Re8 10.Bxd5 Qxd5 11.cxb4 Bg4 12.Qc2
Nxd3 (or 12...Nf3+! 13.gxf3 Bh3 – Stockfish) 13.Nc3 Ne1 14.Nxd5 Nxc2 15.Rb1 Be2 and Black
won, D.N.Bowman-C.J.Purdy, Australian Ch., Melbourne 1949.
b) 6.Ng5?! is a Fried Liver Declined (i.e. 4.Ng5 d5 5.exd5 Nxd5 6.d3?!), which shows how silly it
is. After 6...f6! Black is already better; e.g. 7.Ne4 (or 7.Nc3 Bb4 8.Nge4 Be6; not 7.Qf3? Nd4 8.Qxd5
Qxd5 9.Bxd5 Nxc2+ 10.Kd2 Nxa1 11.Ne6 c6! and Black won, Cars.Andersen-M.Kotva, Pardubice
2004) 7...Be6 8.0-0 Qd7 9.Nbc3 0-0-0 10.Bb5 a6 11.Ba4 Nb6 12.Bb3 Bxb3 13.axb3 f5, J.Cubas-
M.V.Santos, San Bernardo 2005.
c) 6.Bg5?! also helps Black: 6...f6 7.Bd2 Be6 8.0-0 Qd7 9.Nc3 0-0-0 with much the better chances,
B.Brendel-M.J.Turner, Krumbach 1991.
d) 6.Bxd5?! surrenders the Italian bishop without even gaining a tempo on the queen: 6...Qxd5
7.Nc3 Bb4 8.0-0 (or 8.Bd2 Bxc3 9.Bxc3 f6) 8...Bxc3 9.bxc3 Bg4 10.c4 Qd6 11.Bb2 0-0-0 and Black
has the edge, w88lvo-tsmenace, ChessWorld.net 2020.
e) 6.Bb5?! wastes time to attack a pawn that is now easily defended: 6...Bd6 7.0-0 Bg4 8.Re1 0-0
9.Bxc6 bxc6 10.c4 Ne7 11.Bg5? f6 12.Bh4 Nf5 13.Bg3 Nd4 with a big advantage, Alf.Morris-
S.Crakanthorp, Australian Ch., Melbourne 1922.

357
 f) 6.h3 prevents a move (...Bg4) that was not yet intended, allowing Black to develop to taste.

Stockfish likes 6...Nb6 (alternatively, 6...Bc5 will likely transpose to the main line) 7.Bb3 Bf5 8.0-0
Qd7; e.g. 9.Re1 (if 9.Qe2 then 9...Bd6, or 9.Nxe5 Nxe5 10.Re1 f6 11.f4 0-0-0) 9...f6 10.Nh4?! (or
10.d4 0-0-0) 10...Be6 11.Qh5+ Qf7 12.Qxf7+ Bxf7 13.Bxf7+ Kxf7 with an edge for Black,
M.Sosovicka-Pe.Balogh, Lubovna 2013.
g) 6.c3 also allows a change of tack; e.g. 6...Nb6 7.Bb3 Bf5 8.d4 (what else? 8.Bc2 Qd7 and 8.Ng5
Bg6 are both good for Black) 8...exd4 9.Nxd4 Nxd4 10.cxd4 Qf6 and Black is on the right side of
equal.
h) 6.Nc3 is a bit simplistic: 6...Be6 (or 6...Nxc3 7.bxc3 Be7) 7.Ng5 (or 7.Bb5 f6) 7...Nxc3 8.Nxe6
Nxd1 9.Nxd8 Rxd8 10.Kxd1 Na5 11.Bb3 Nxb3 12.axb3, J.Bocska-D.Straubinger, Hungarian League
2006, and now 12...a6 is completely equal.
i) 6.Qe2 does at least threaten the e5-pawn. 6...f6 looks like the best response; e.g. 7.d4 (or 7.0-0
Nb6 8.Bb5 Bd6 9.d4 0-0) 7...Bg4 8.dxe5 (or 8.Bb5 Bxf3 9.Qxf3 exd4 10.0-0 Qd7, Z.Borosova-
F.Pasztor, Hungarian League 2011) 8...Nxe5 9.Qe4 Bxf3 10.gxf3 Nb6, and if 11.f4 Nbxc4 12.fxe5,
V.Valles-F.Lupi, Sabadell 1945, then 12...Nxe5!, since 13.f4? loses to 13...Qd4!.

358
6...Bc5

If the main line fails to hold up, there’s always 6...Be7!? 7.Re1 f6. Supposedly 8.d4 is good for
White, but 8...Nb6! may yet defend; e.g. 9.Bb5 0-0 10.Bxc6 bxc6 11.dxe5 Qxd1 12.Rxd1 Bg4
13.Nbd2 fxe5 14.Re1 Rae8 15.Nxe5 Bc5 16.Ndf3 Bxf3 17.gxf3 Bb4 18.Re4 Rxf3 19.Rxb4 Rxe5 and
Black drew this endgame in LcZeroCPU-Vajolet2, TCEC 16 Bonus 2019. Kravtsiv prefers 9.Bb3
Nxd4 10.Nxd4 Qxd4 11.Qxd4 exd4 12.Bf4 and “White’s advantage is clear” – maybe; 12...Bf5
13.Bxc7 Kd7 14.Bg3 Rac8 15.Rd1, J.Esquivel Leon-G.Vargas Cortes, corr. 2018, and 15...Bc5
doesn’t look that bad for Black.
Instead, 8.c3 Bg4 9.h3 Bh5 10.d4 exd4 11.cxd4 Qd7 12.Nc3 Bf7 13.a3 0-0-0 isn’t terrible either,
Y.Volodarsky-A.Sachuk, corr. 2019. Or if 8.h3 Nb6 9.Bb3 Bf5 10.Nh4 Qd7 11.Qf3, H.Odeev-
M.Buker, Adana 2006, then 11...Nd4! 12.Qxb7 Be6 equalizes; e.g. 13.Be3 Nxb3 14.axb3 Bd5 15.Qa6
g5 16.Nf3 Bxf3 17.gxf3 Qxh3.

359
 Once again White has played many moves here. The two main ones are:
B1: 7.c3
B2: 7.Re1
Others may transpose if White plays Re1 or c2-c3 soon, but there are plenty of independent lines:
a) 7.Nxe5?! is overhasty in view of 7...Nxe5 8.Re1 Be6! 9.Rxe5 Bxf2+! 10.Kh1 0-0 11.Qf3,
A.Lysakov-V.Burdinsky, Tomsk 2013, and then 11...Nf6!, intending 12.Qxf2? Ng4 or 12.Rxe6? fxe6
13.Bxe6+ Kh8 14.Qxf2 Ng4 15.Qe2 Nf2+ 16.Kg1 Qd4 and wins.
b) 7.Ng5 is better than on move six, but not so much as to give White an advantage; e.g. 7...0-0
8.Nc3 (not 8.Qh5? Bf5 9.Nc3 Bg6 10.Qf3? Ndb4 11.h4 Nd4 and Black should win, V.Inkiov-
R.Schuermans, Condom 2008) 8...Nf6 (not 8...Nxc3?? 9.Qh5) 9.Nce4 Nxe4 10.Nxe4 Be7 11.Qh5
Qd7, intending ...Qf5, or 12.f4 Na5 13.f5 Nxc4 14.f6 Bxf6 15.Nxf6+ gxf6 16.Bh6 Qd4+ 17.Kh1 Qg4,
which led to a draw in V.Inkiov-Be.Toth, Rome 1984.
c) 7.Bb5 can be met by 7...Bg4

360
 8.Re1 (Pinski didn’t like his own position after 8.h3 Bxf3 9.Qxf3 0-0 10.Bxc6 bxc6 11.Nd2 f5
12.Nb3, A.Deszczynski-J.Pinski, Warsaw 1997; it’d be nice to know how the game continued as
12...Qd6 13.Nxc5 Qxc5 14.c3 Rae8 doesn’t seem so bad to me) 8...0-0 9.Bxc6 bxc6 10.Rxe5 Qf6
11.Re4, A.Deszczynski-J.Sroka, Polish League 2003, and then 11...Qg6! with ideas of ...f7-f5,
...Rae8, ...Bd6 and good play for the pawn.
d) 7.Nc3 Nxc3 (not now 7...Be6? 8.Ne4 Bb6 9.Nfg5 and White is clearly better, P.Eljanov-
E.Ohanyan, Titled Tuesday blitz 2020) 8.bxc3 0-0 9.h3 (for 9.Re1 see 7.Re1) 9...Bf5 (or 9...Qd6)
10.Be3 Bb6 11.Re1 Qd6 12.Nh4 Be6 13.Bxe6 Qxe6 14.Nf3 Rad8 15.a4 Rfe8 and Black is at least
equal, S.Kosanski-M.Zelic, Croatian League 1997.
e) 7.h3 0-0 8.a4!? (for 8.Re1 see 7.Re1 again) 8...f6 9.Nbd2 Be6 10.Nb3 Be7 11.Re1 Kh8 12.Bd2
and now, rather than 12...Bg8 13.d4 exd4 14.Nbxd4 Nxd4 15.Nxd4 with an edge for White in
Ma.Carlsen-S.Karjakin, Paris GCT (rapid) 2017, Stockfish suggests 12...Bd6, intending 13.d4 Ne3!
with equality.
f) 7.d4!? might come as a shock, but there’s no need to panic:

361
 7...Nxd4! 8.Nxe5 Be6 9.c3 (or 9.Nd2 Qf6 10.c3 Qxe5 11.Re1 Qd6 12.cxd4 Bb6) 9...Nc6 10.Nxc6
bxc6 11.Nd2 0-0 12.Ne4 Bb6 13.Qf3 h6 and Black was fine, R.Shetty-R.Hegde, Indian League 2017.
g) 7.Qe2 0-0 was mentioned via 4.Qe2. It has only been seen once: 8.h3 was J.Infante-E.Cervetto,
San Fernando 1998, where 8...Bf5!? 9.Nxe5 Nd4 10.Qd1 Re8 11.Nf3 Ne2+ 12.Kh1 Bxh3! 13.gxh3
Qd7 14.Ng1 Nb6 is one solution.
The immediate 8.Nxe5 is more testing, when Black has to work harder to show compensation; e.g.
8...Nd4 (not 8...Re8?? 9.Nxf7!) 9.Qd1 (or 9.Qh5 g6 10.Qh6 Be6) 9...b5! 10.c3 (or 10.Bb3 Bb7)
10...bxc4 11.cxd4 (or 11.dxc4 Nb6 12.cxd4 Qxd4 13.Qxd4 Bxd4 14.Nf3 Bf6) 11...Bxd4 12.Nc6 Qf6
13.dxc4 Qxc6 14.Qxd4 Nb4 15.Na3 (or 15.Qc3 a5 16.Na3 Re8 17.Bf4 Bb7 18.f3 Qc5+ 19.Kh1
Rad8) 15...Re8 16.Be3 Bb7 17.f3 Rad8 18.Qc5 Nd3 19.Qxc6 Bxc6 20.Bxa7 Nxb2, still with a lot of
activity for the pawn.
h) 7.Qe1, avoiding ...Nd4 in note ‘g’, has its own drawback in that 7...f6! 8.c3?! (strong with the
queen on e2) 8...Nb6! 9.Bb3? drops the d3-pawn, and 9.Bb5 0-0 leaves the queen misplaced.
Otherwise 8.Be3 Nxe3 9.fxe3 Qe7 10.Nc3 Be6 11.Nd5 Bxd5 12.Bxd5 Qd7 is one way to equalize,
V.Markovic-Z.Lazhevskaya, St. Petersburg 2016.
i) 7.Nbd2 0-0 8.Ne4!? (attacking the bishop straight away is slightly unusual; for 8.c3 and 8.Re1
see lines B1 and B21 respectively) 8...Bb6 (or 8...Be7 as in line B1, leaving b6 for the d5-knight)

362
 9.c3 (instead, 9.Re1 f6 is B21 again; 9.h3 Kh8 prepares ...f7-f6 or perhaps ...f7-f5; while 9.Bxd5?!
Qxd5 10.c4, J.Medina Chirinos-G.Manrique, Dos Hermanas blitz 2004, is answered by 10...Qd7!,
covering f7 so that 11.c5 f5 12.Qb3+ Kh8 13.Neg5 doesn’t work) 9...f6 10.b4!? (threatening Bxd5
and c3-c4 etc; 10.a4 Be6 11.b4 a6 comes to the same thing; while anything like 10.h3 Kh8 11.Bb3
Bf5, Ma.Adams-M.Lyell, 4NCL 1996, is fine for Black) 10...Be6 11.a4 a6 12.a5 Ba7 13.Be3 aims for
d3-d4 with advantage, as after 13...Bf7? 14.Bxa7 Nxa7 15.d4, Li Ruifeng-M.Mulyar, Arlington 2016.
To prevent that Black had to play 13...Kh8!, intending 14.Bxa7 Rxa7 15.d4? Nxc3!. If instead 15.Re1
only then 15...Bf7, or 15.Qb3 (Müller & Souleidis) 15...Bg8 16.Rfe1 (or 16.d4 f5) 16...b6 and Black
seems okay; e.g. 17.Qc2 (or 17.axb6 cxb6 18.Rxa6 Rxa6 19.Bxa6 Ndxb4) 17...Qc8 18.Rad1 Rd8
19.d4 exd4 20.Nxd4 Nxd4 21.Rxd4 c5 22.bxc5 bxc5 23.Rd2 Rad7 with equality.

B1: 7.c3

An unusual move by this order (7.c3 occurs with roughly 6% frequency) but an annoying one, since
Black has little choice other than to enter an Italian main line.

7...0-0

363
 We’ve now transposed to 3...Bc5 4.c3 Nf6 5.d3 0-0 6.0-0 d5 7.exd5 Nxd5. Searching the databases
brings up hundreds of grandmaster games over the past decade, going right to the very top, and GM
Kravtsiv has 50+ pages on 8.a4 alone. I can’t compete with that; indeed, there seems little point in
doing so seeing as it’s a rare transposition. Instead, I’ll try and cram as much theory and practice into
as small a space as possible, for people (including me) to refer to (or not) as desired.

8.Re1

The main move, first played in 1893. White attacks e5 in the manner of line B2, though inserting
c2-c3 first can hardly be an improvement on that. As usual White has numerous other options, of
which the last four are most often seen:
a) 8.Ng5 (first played in 1851) 8...h6 9.Ne4 (if 9.Qf3 then 9...Nce7 is stronger than taking the piece;
after 10.Ne4 Bb6 Black is clearly better with ...f7-f5 coming) 9...Be7 (not 9...Bb6? 10.Bxh6!)
10.Nbd2, D.Iftime-Mari.Szabo, Buzias 2002, and now 10...Na5 (L&O) is good for Black.
b) 8.d4 exd4 9.cxd4 Bb6 10.Nc3 Be6 and Black already stands well, A.Heinrichsen-F.Gutmayer,
Berlin 1897.
c) 8.Bg5 f6 9.d4!? (either 9.Bh4 Kh8 or 9.Be3 Bxe3 10.fxe3 Na5 is fine for Black) 9...exd4 10.cxd4
Bb6 11.Nc3 Nce7 (or 11...Be6 12.Re1 Bf7 – L&O) 12.Re1 Kh8 13.Bh4 Bg4 14.h3 Bh5 15.Bg3 c6
and Black consolidated in a good position, M.Schula-Y.Gozzoli, Benasque 2017.
d) 8.Bb3 Bg4 9.h3 (if 9.Nbd2 or 9.Re1 then 9...Kh8 makes sense) 9...Bh5 10.Nbd2 (10.Bg5?! f6
11.Bh4 Kh8 is good for Black, J.Dourerassou-D.Wagner, Vandoeuvre 2014) 10...f5!? (or 10...Kh8
again) 11.Ne4, B.Jakab-L.Varga, Banska Bystrica 2019, and now 11...Bxf3 12.Qxf3 fxe4 13.Qxe4
Nce7 14.d4 exd4 15.cxd4 Bb6 16.Bg5 Rf6! keeps a material advantage.
e) 8.Qe2 Bg4 9.h3 Bh5 10.g4 Bg6 looks very risky for White; e.g. 11.b4 (or 11.Nxe5 Nxe5
12.Qxe5 c6) 11...Bb6 12.b5 Na5 13.Qxe5, O.Dolzhikova-M.Bottema, Gibraltar 2018, and now

364
13...Nxc4 14.dxc4 Nf6, when the extra doubled backward pawn is hardly compensation for White’s
wrecked position.
f) 8.h3 Nb6

9.Bb3 (this leaves the d3-pawn vulnerable; but neither 9.Bb5 Bd6 10.Re1 Re8 11.Nbd2 Bd7
12.Nc4 Nxc4 13.Bxc4 h6 14.Qb3 Qf6, D.Sermek-M.Mikac, Slovenian Ch., Radovljica 1993; nor
9.Bg5 Qd6 10.b4 Nxc4 11.bxc5 Qg6 12.dxc4 e4, I.Saric-S.Brunello, Szeged 2007, gives White
anything) 9...Bf5 10.Nxe5!? (a radical solution; after 10.Bc2 Re8 11.Qe2 Nd5 12.Re1 Qd7 13.Qf1
Rad8 Black is just better, E.Alekseev-M.Kazhgaleyev, Tashkent 2011) 10...Nxe5 11.d4 Nbd7!
12.Be3?! (12.Nd2 keeps White closer to equal), R.Muniz-K.Mekhitarian, Buenos Aires 2015, and
now 12...Bb6! 13.dxe5 Bxe3 14.fxe3 Qg5 is good for Black.
g) 8.Nbd2 Nb6 (the usual response to Nbd2 – see 10...Nb6 below) 9.Bb5 (as in the main line, 9.b4
Be7 10.b5 Na5 11.Nxe5 wins the e-pawn, but 11...Naxc4 12.Ndxc4 Nxc4 13.Nxc4 a6 provides quite
enough compensation; e.g. after 14.bxa6 Rxa6 15.Re1 Be6 16.Bf4, S.Melia-E.Atalik, Women’s
World Cup, Tehran 2017, and 16...b5 17.Ne5 c5; while the mutual exchange 9.Ne4 Nxc4 10.Nxc5
Nd6 is by no means bad for Black) 9...Bd6 10.a4 (instead, 10.Re1 Bg4 11.h3 Bh5 transposes to
11.Bb5 in the main line; as does 10.Ne4 Bg4 11.h3 Bh5 12.Ng3 Bg6 13.Re1 Re8) 10...a6 (otherwise
10...a5 transposes to note ‘i’ below) 11.Bxc6 bxc6 12.a5 Nd5 13.Re1 Re8 14.Nc4 f6 and Black stands
perfectly well, Wang Hao-E.Inarkiev, Danzhou 2019.
h) 8.b4 Be7 (intending ...Bf6 to secure Black’s position; not 8...Bb6??, J.Cochrane-M.Bonnerjee,
Calcutta 1856, because 9.Bxd5 Qxd5 10.c4 and c4-c5 wins)

365
 h1) 9.Re1 Bf6 10.b5 (or 10.Ba3 Re8) 10...Na5 11.Bxd5 (not 11.Nxe5?! Re8 12.Bxd5 Qxd5 13.d4
Bxe5 14.dxe5 Rxe5 and Black is better) 11...Qxd5 12.Bg5 Bg4 13.Bxf6 Bxf3 14.Qxf3 Qxf3 15.gxf3
gxf6 with equality, J.Siigur-P.Peuraniemi, corr. 2007.
h2) 9.a4 a6 doesn’t particularly improve this for White, e.g. 10.Re1 Bf6 11.Ng5!? Nb6 12.Qh5
Bxg5 13.Bxg5 Qd6 14.Bb3 Nd5 and Black is fine, Francesca-Invictus. TCEC 21 Qualification 2021.
h3) 9.b5 (consistent) 9...Na5 10.Bxd5 (not 10.Nxe5?! Nxc4 11.Nxc4? a6! 12.bxa6 Rxa6 with strong
play – L&O, or 11.dxc4 Nb6 12.Ba3 Bxa3 13.Qxd8 Rxd8 14.Nxa3 Na4 and Black is better, Ruan
Lufei-A.Kosteniuk, Women’s World KO Ch., Antakya 2010) 10...Qxd5 11.c4 Qd7! (an engine-
approved move; the queen supports potential attacks on d3/b5 with ...Rd8 or ...a7-a6, and also leaves
d8 free for the bishop in one variation) 12.Bd2 (not 12.Nxe5?? Qd4; while if 12.Nc3 f6 13.Nd5 then
13...Bd8! 14.Ba3 Re8 15.Rc1 a6 and Black is certainly not worse, M.Antipov-J.Van Foreest, World
Junior Ch., Khanty-Mansiysk 2015) 12...e4! (the tactical justification) 13.Bxa5 (not 13.dxe4?! Nxc4)
13...exf3 14.Qxf3 a6 15.a4 (L&O’s 2012 suggestion; all other moves are worse: 15.bxa6?! Rxa6
16.Bc3 c5! 17.Re1 b5! 18.cxb5 Rg6 19.Qe2 Bb7 20.f3 Bd6 with a strong attack, P.Schalkwijk-
F.Zabransky, corr. 2014; or 15.b6?! cxb6 16.Bxb6 Bf6 17.d4 Bxd4 18.Rd1?! Bxa1! 19.Rxd7 Bxd7
20.Qxb7 Rae8 and Black is clearly better, Z.Andriasian-A.Mastrovasilis, European Ch., Rijeka 2010;
or 15.Bc3?! axb5 16.cxb5 Qxb5 17.Qg3 Qg5 18.Qxc7, B.Macieja-L.Fressinet, German Bundesliga
2011, and now Stockfish trots out 18...Rd8! 19.Re1 Be6 20.Re5 Rac8! 21.Rxg5 Rxc7 22.Rxg7+ Kf8
23.Rg3 Bf6! 24.d4 Rxd4! 25.Na3 Rxc3 26.Rxc3 Rd5 27.Rcc1 Bb2 with a big advantage) 15...axb5
16.axb5 Rd8 17.Nd2 Bf6 18.Ra3 Qf5 19.d4 Qxf3 20.Nxf3 h6 21.Rfa1 Bxd4 22.Nxd4 Rxd4 23.Bb6
Rxa3 ½-½ F.Kubis-L.Romanov, corr. 2013, in view of 24.Bxd4 Rxa1+ 25.Bxa1.
i) 8.a4!? is the recommendation of GMs Delchev and Kravtsiv. In response I’d go for 8...a5, halting
the queenside expansion. This generally leads to an endgame where the GMs consider White to have
the chances, but those chances seem very thin against accurate defence: 9.Nbd2 (I’m going to ignore
nuances of move order and just head for the same position; for instance, 9.Re1 Bg4 10.Nbd2 Nb6

366
11.Bb5 Bd6 or 10.h3 Bh5 11.Nbd2 Nb6 12.Bb5 Bd6 transposes; other moves are weaker; e.g. 9.Qb3
Nb6 10.Bb5 Be6 11.Qd1, P.Senger-Elm.Zimmer, Koblenz 2002, and 11...Bd6 12.Nbd2 Na7; or
9.Bg5 f6 10.Qb3, T.Pioch-S.Sitta, Wiesbaden 2019, and 10...Nce7 11.Bh4 Kh8; or 9.h3 Nb6 10.Bb3
Bf5 11.Bc2 Qd7 12.Re1 Rfe8 13.Ng5 Bf8 14.Na3 Rad8 15.Ne4 f6 16.Qf3 Bg6 17.Nb5 Kh8 and
Black is better, F.Caruana-L.Aronian, Chess.com blitz 2018) 9...Nb6 10.Bb5 Bd6 (the best way to
defend the pawn since Ne4 will hit the bishop anyway)

11.Re1 (11.Ne4 is met the same way: 11...Na7 12.d4 Nxb5 13.axb5 exd4 14.Nxd6 Qxd6 15.Qxd4
Bd7 16.Bf4 Qxd4 17.Nxd4 Rfc8 18.Rfe1 a4 and Black has an even easier defence, Mi.Adams-
C.Aravindh, FIDE World Cup, Khanty-Mansiysk 2019; or if 11.Nc4 Nxc4 12.Bxc4 Qf6 13.Ng5,
M.Vachier Lagrave-F.Vallejo Pons, PRO League rapid 2019, then 13...Qg6 is equal) 11...Bg4 12.Ne4
(or 12.h3 Bh5 13.Ne4 Na7 transposing, since 13.Bxc6 bxc6 14.Ne4 f5 15.Ng3 Bxf3 16.Qxf3 Qd7
17.c4 Rae8 is fine for Black, I.Nepomniachtchi-L.Aronian, World Rapid Ch., Doha 2016) 12...Na7
(going after the white bishop) 13.h3 Bh5 14.d4 (not 14.Ng3 Bg6 15.Nxe5?? Bxe5 16.Rxe5 c6 and
Black wins a piece, Z.Andriasian-S.Vidit, Chess.com blitz 2018) 14...Nxb5 15.axb5 exd4 16.Nxd6
(inserting 16.g4 Bg6 is no improvement as Black then has ...Bg6 ‘for free’; e.g. 17.Nxd6 Qxd6
18.Qxd4 Qxd4 19.Nxd4 a4 20.Bf4 Rfc8 21.Rad1 Nc4 22.Re2 Re8 and White drifted into difficulties,
M.Vachier Lagrave-W.So, Paris GCT blitz 2018) 16...Qxd6 17.Qxd4 Qxd4 18.Nxd4 Bg6! (an
important move, preventing Nf5) 19.Bf4 (if instead 19.Nb3 a4 20.Nc5, D.Howell-E.Van den Doel,
European Team Ch., Batumi 2019, then 20...Rfc8 21.Nxb7 f6, intending ...Be8 to pick up the b5-
pawn; or 19.f4 f5 20.Re5 a4 21.Rc5 Rac8 22.Be3 Nd7 23.Rd5 Nb6 with a draw by repetition, Ding
Liren-W.So, Opera Euro rapid 2021) 19...Rfc8 20.Bg3 (if 20.b3 then 20...c5! 21.bxc6 bxc6 22.c4 a4
23.bxa4 c5 24.Nb5 Rxa4 equalizes, P.Harikrishna-S.Vidit, Prague 2019) 20...a4! 21.Re7 Nc4 22.b3
axb3 23.Rxa8 Rxa8 24.Nxb3 Rc8 25.Rxc7 (not 25.Bxc7?? Kf8 26.Rd7 Ke8 27.Nc5 b6 and wins)
25...Rxc7 26.Bxc7 b6 “with good chances for a draw despite being a pawn down” (Mikhalevski).
Indeed, the opposite-coloured bishops and weak b5-pawn make the draw odds on; e.g. 27.Kf1 Bd3+
28.Ke1 Kf8 29.h4 Ke7 30.f3 Kd7?? (oops; a passing move like 30...Ke8 will do, since 31.Nd2 Nxd2

367
32.Kxd2 Bxb5 33.Bxb6 Kd7 is a total draw) 31.Bg3?? (missing 31.Bxb6 Nxb6 32.Nc5+ Kc7
33.Nxd3 and wins) 31...Na3 32.Bf2 Kc7 with a draw in a few more moves, I.Nepomniachtchi-
P.Leko, Legends of Chess (rapid) 2020.

8...Bg4

The strongest reply. White can still win the e5-pawn but will have to compromise their position in
order to do so.

9.h3

It is worth White’s while to insert this move, so that g2-g4 can follow at any moment. 9.Nbd2 Nb6
will likely transpose if and when White plays h2-h3 and is much the same even without it; while 9.a4
a5 10.Nbd2 Nb6 11.Bb5 Bd6 returns to the 8.a4 variation above.
Not now 9.b4?! due to 9...Ndxb4!, when White has to work hard to stay in the game: 10.Ba3! (not
10.cxb4? Bxf3 11.Qxf3 Bd4) 10...Bxf3 11.Qxf3 Nc2 12.Bxc5 Nxe1 13.Qe2 b5! 14.Bb3 (not
14.Bxb5? Qd5) 14...Nxd3 15.Bxf8 Nc1 16.Qxb5 a6 17.Qc4 Nxb3 18.axb3 Qd1+ 19.Qf1 Qxf1+
20.Kxf1 Kxf8 with equality.

9...Bh5

368
10.Nbd2

The most common continuation, intending Ne4 and Ng3 to break the pin and take the e5-pawn
without weakening the kingside. Instead:
a) 10.a4 a5 transposes to the 8.a4 variation again; e.g. after 11.Nbd2 Nb6 12.Bb5 Bd6 13.Ne4 Na7
etc.
b) 10.g4 Bg6 11.Nxe5 Nxe5 12.Rxe5 c6 is an excellent Marshall for Black in view of the open
white kingside; e.g. 13.Bxd5 (or 13.Qf3 Bd6 14.Re1 f5!? 15.g5 f4; and not 13.d4? Bd6 14.Re1 Qh4
15.Qf3 Rae8, when White is lost) 13...cxd5 14.Qf3 Qc7 15.Rxd5 (or 15.Re2 Rae8 16.Rxe8 Rxe8
17.Nd2 Re6 18.Nf1 Rf6, Z.Borosova-N.Zhukova, European Women’s Ch., Gaziantep 2012)
15...Rae8 16.Bd2 f5 (or 16...Re6 again) 17.g5 Bf7 18.d4 Bd6 19.Rxf5? Be6 20.Rxf8+ Rxf8 21.Qe4
Qf7 22.Be3 Bd5 23.Qg4 Re8 24.Nd2 Rxe3!, G.Glidzhain-P.Ponkratov, Russian Team Ch. 2010, and
Black wins in view of 25.fxe3 Bh2+ etc.
c) 10.b4 Be7 (10...Bb6 is also possible) 11.b5 Na5 12.Rxe5 (instead, 12.g4 Bg6 13.Nxe5 Bf6
14.Nxg6 hxg6 15.Ba3 Nxc3 leads to mass exchanges and equality, R.Pruijssers-V.Babula, German
Bundesliga 2014) 12...Nf6 is an original Marshall type sacrifice (i.e. with ...Nf6 rather than...c7-c6)
where Black seems to have enough for the pawn; e.g. 13.Nbd2 Bd6 14.Re1 Re8 15.Bb3 (or 15.Rxe8+
Qxe8 16.Qa4 Nxc4 17.Qxc4 a6, J.Smolen-F.Bellini, Olbia 2008) 15...Qd7 16.c4 (or 16.Rxe8+ Rxe8
17.Bc2 Nd5 18.Ne4 Nxc3! 19.Nxc3 Bxf3, Ma.Carlsen-H.Nakamura, NIC Classic KO rapid 2021,
when White has nothing better than to allow 20.gxf3 Qxh3 21.f4 Re6 22.Ne4 Rg6+ 23.Ng3 Rxg3+
with a draw) 16...Rxe1+ 17.Qxe1 Bb4 18.Qf1 (or 18.Qe3 Re8 19.Ne5 Qd8 20.Bb2 Nd7 21.Qg3 Nxe5
22.Bxe5 f6 23.Bf4 Qd4 24.Rb1 Bxd2 25.Bxd2 Nxb3 26.axb3 c6 and Black drew with the opposite-
coloured bishops, L.Ljubicic-J.Zidu, corr. 2016) 18...c6 19.bxc6 Nxc6 20.Rb1 Bxd2 21.Nxd2 Re8
22.Ne4 Nxe4 23.dxe4 Rxe4 and the game was soon drawn, S.Ganguly-Lu Shanglei, Chinese Team
Ch. 2019.

369
10...Nb6

A purposeful reply, stepping back from tactical tricks, allowing the c5-bishop to defend at d6 (if
appropriate), uncovering the queen on the d-file and, not least, attacking White’s Italian bishop.

11.b4

The most direct option. White has three others:

a) 11.Ne4 Nxc4 12.Nxc5 Nd6 13.g4 Bg6 14.Nxe5 (after 14.d4 exd4 15.Bg5 Qc8 16.Nxd4 Nxd4
17.Qxd4 b6, A.Tzermiadianos-K.Markidis, Greek League 2009; or 14.Bg5 Qc8 15.Nh4 f6 16.Nxg6
hxg6 17.Be3 b6, M.Nurkiewicz-G.Gajewski, Polish League 2018, Black has the better chances)
14...Nxe5 15.Rxe5 b6 16.Bg5 (or 16.Nb3 f5!? with the initiative, which quickly turned into more:
17.f4? fxg4 18.hxg4 Nc4, G.Sarthou-A.Naiditsch, French League 2010) 16...Qc8 17.Be7 bxc5
18.Bxf8 Qxf8 19.Rxc5 Ne4 and Black is not worse, J.Kreutz-A.Jedinger, corr. 2014.
b) 11.Bb5 Bd6 (again the best response)

370
 12.Ne4 (neither 12.Nc4 Nxc4 13.Bxc4 Kh8, nor 12.a4 a6 is anything to worry about) 12...Re8 (for
consistency with 8.Nbd2 lines; via this move order 12...f5!? 13.Ng3 Bxf3 14.Qxf3 Qd7 is a strong
alternative with which I’ve won as Black) 13.Bg5 (here 13.Ng3 Bg6 14.Bg5 f6 15.Be3 a6 16.Bxc6
bxc6 is much the same, whereas 14.d4 exd4 15.Rxe8+ Qxe8 16.Nxd4 Rd8 is equal at best for White,
P.Klings-L.Schandorff, Helsingor 2012; similarly, 13.a4 a6! 14.Bxc6 bxc6 15.a5 Nd7 16.Ra4 Rb8
17.Rc4 c5 18.Ng3 Bg6, S.Kalugin-E.Alekseev, Russian Team Ch. 2011, when ...Rb5 will target the
a5-pawn) 13...f6 (I’m always pleased to get this in for free) 14.Be3 a6 15.Bxc6 bxc6 16.Ng3 (or
16.g4 Bf7 17.c4 Bf8 18.Kh2 a5 19.Rg1 Bg6 20.Qc2 Nd7 21.Rad1 c5, V.Sivuk-A.Khalifman,
Voronezh 2014) 16...Bf7 17.c4 (or if 17.d4, Ma.Carlsen-L.Aronian, Leuven GCT blitz 2016, then
17...exd4 18.Nxd4 Bxg3 19.fxg3 Nc4 20.Qc1 Qd5 – Roiz) 17...Bb4 18.Re2 Bf8 19.b3 c5 20.Qe1 a5
and “White is doomed to passive defence” (M.Roiz), H.Nakamura-A.Grischuk, Sinquefield Cup, St.
Louis 2018.
c) 11.Bb3!? is a recent try, presumably because nothing else works.

371
 It does require Black to be careful: 11...Qxd3 12.Nxe5 Qf5! 13.Nef3 (now threatening a g2-g4 fork)
13...Rad8! (allowing it) 14.Qe2 (after 14.g4 Bxg4 15.hxg4 Qxg4+ 16.Kh1 Qh3+ White has to defend
accurately: 17.Nh2 Bd6 18.f4 Bxf4 19.Re2 Ne5 20.Qf1 Qxf1+ 21.Ndxf1, emerging with a bishop
versus three good pawns, P.Michalik-A.Shirov, Prague 2019; no one has seen fit to repeat this line as
White, and from the other side I’d go for 21...Bxc1 22.Rxc1 Nd3 23.Rd1 Nc5, seeking to remove the
remaining bishop or render it passive) 14...Qc8 (and no one has tried 14...Rfe8!? 15.Qxe8+ Rxe8
16.Rxe8+ Bf8, probably because there are no weaknesses for the queen to attack, although Stockfish
says “0.00” anyway) 15.Qb5 (or 15.Ne4 Nd7 16.Bf4 Bb6 17.Rad1 Rde8 18.Qd3 Re7 19.Neg5 Bg6,
The.Schmidt-S.Caron, corr. 2020) 15...Nd7 16.Bc2 (or if 16.Nh4 a6 17.Qf1 Be7 18.Nf5, Ma.Carlsen-
W.So, Opera Euro Rapid KO 2021, then 18...Bf6 19.Ne4 Bg6) 16...Rfe8 17.Rxe8+ Rxe8 18.Bf5 a6
19.Qc4 Bd6 20.Qh4 Bg6 21.Bxg6 hxg6 22.Ne4 Nce5 and Black is fine, since 23.Nfg5 Nf8 gets White
nowhere, T.Radjabov-S.Mamedyarov, FIDE World Cup, Khanty-Mansiysk 2019.

11...Be7

372
 The bishop retreats all the way (as after 8.b4 above), heading for f6 and leaving the d-file open for
the queen.

12.b5

Or 12.g4 Bg6 first. Other lines are level as well; e.g. 12.Bb5 f6, or 12.Bb3 Qxd3 13.Nxe5 Bxd1
14.Nxd3 Bxb3 15.axb3 a6; while 12.Qe2 Nxc4 13.Nxc4 f6 14.b5 Nb8 15.d4 a6! has led to a lot of
draws in correspondence chess after 16.b6 (or 16.bxa6 Rxa6 17.dxe5 Nd7) 16...Nd7 17.dxe5 Nxb6
18.Nxb6 cxb6 19.Qe4 fxe5 20.Nxe5 Qc7.

12...Na5 13.g4

If 13.Rxe5 Bg6, Black has good play for the pawn with ...Bf6 coming.

13...Bg6 14.Nxe5 Bf6 15.Nxg6

Or 15.Ba3 Bxe5 16.Rxe5 Re8 17.Rxe8+ Qxe8 18.Bc5 Nbxc4 19.dxc4 Rd8 and “White’s extra
pawn is irrelevant” (L&O).

15...hxg6 16.Ne4

After 16.Bb2 a6! 17.bxa6 Naxc4 18.Nxc4 Nxc4 19.dxc4 Rxa6 Black has more than enough for the
pawn.

16...Naxc4 17.dxc4 Nxc4

373
 The position is equal, as evidenced by the 89.3% draw ratio in the databases. For instance, 18.a4,
18.Rb1, and 18.Qf3 can all be met by 18...a6, opening up for the a8-rook; 18.Bf4 Be5 19.Qe2 (or
19.Qf3 f5) 19...Bxf4 20.Qxc4 c6 21.Rad1 Qc7 22.bxc6 Rac8 is level (L&O); as was 18.Qb3 Ne5
19.Kg2 Re8 20.Rd1 Qc8 21.Nxf6+ gxf6 22.c4 Nxg4!? 23.Qf3 Ne5 24.Qxf6 Qe6 25.Qxe6 Rxe6,
R.Rutkus-R.Webster, corr. 2019.
My own experience in this line ended 18.Qxd8 Rfxd8 19.Nxf6+ ½-½ afms-tsmenace,
ChessWorld.net 2019. Play might have continued 19...gxf6 20.Bf4 Na3 21.Re7 (or 21.Bxc7 Rdc8
22.Re7 Nxb5) 21...Nxb5 22.Bxc7 Re8 23.Rxe8+ Rxe8 24.Ba5 Re2, as in F.Libiszewski-F.Vallejo
Pons, German Bundesliga 2019, which was one of the very few decisive games, and in Black’s
favour.

B2: 7.Re1 0-0

374
 The main line. White hurries to attack the e5-pawn and Black just ignores it. Here the two most
popular continuations are:
B21: 8.Nbd2
B22: 8.Nxe5
All others are “also rans” frequency-wise:
a) 8.c3 returns to line B1.
b) 8.Bxd5?! Qxd5 9.Nc3 Qd8 10.Ne4 Be7 (or 10...Bb6) 11.Ng3, M.Capellades Subirana-C.Vargas
Drechsler, Barcelona 2012, and here 11...f6 solidly defends e5, when White has given up the Italian
bishop for nothing.
c) 8.Nc3 Nxc3 9.bxc3 “with slightly the better chances” (Estrin). I don’t see it; e.g. 9...Qd6 10.Be3
(or 10.Qe2 Be6!?) 10...Bb6 11.a4 Bg4 12.h3 Bh5 13.Bxb6 axb6 14.Qe2 Rae8 led to a draw in docjan-
tsmenace, ChessWorld.net 2019.
d) 8.Bg5?! again gives Black 8...f6 for free. Then 9.Bh4 g5!? 10.Nc3 Be6 11.Bg3 Bf7 12.Nd2 a6
13.Nde4 Be7 14.h4 h6 15.hxg5 hxg5 was fine in J.Willow-J.Tait, Notts League 2019, at least before
16.Bh2?! Nd4??, after which Stockfish bashes out 17.Nxg5! fxg5 18.Nxd5 Bxd5 19.Qh5 Bxc4
20.Qg6+ Kh8 21.dxc4 and claims a win for White. In a blitz game I’d have played 9...Kh8! without
thinking. Having more time I was worried about 10.Nxe5? Nxe5 11.Rxe5, confused by the h4-bishop
covering the f2-square. Except that it doesn’t. Black can play 11...Bxf2+! anyway. Thinking can be
bad for you.
e) 8.a3 a6 9.h3 arose in another of my games, where I made a similar error in calculation.

375
 Black should reply 9...Be6! as in note ‘f’ below, since 10.Nxe5? again fails to 10...Nxe5 11.Rxe5
Bxf2+!, or if 10.Ng5 Ba7 11.Nxe6 fxe6 12.Be3 then 12...Nxe3 13.Bxe6+ Kh8 14.fxe3 Qf6, intending
...Bxe3+ etc. Instead, I blundered with 9...f6?? 10.Nc3 Be6 11.d4! Nxd4 12.Nxd4 Bxd4 13.Bxd5
Bxd5 14.Nxd5 Qxd5 15.c3 and White won a piece, S.Okhai-J.Tait, Notts League 2020 – my last
game before lockdown.
f) 8.h3 was suggested as an option for White by GM John Emms in Beating 1.e4 e5. The best reply
is 8...Be6!, which has since had considerable testing (generally arising from an Italian; e.g. 3...Bc5
4.0-0 Nf6 5.d3 0-0 6.h3 d5 7.exd5 Nxd5 8.Re1). Objectively, Black is fine, as long as the next few
moves are conducted with proper precision:

f1) 9.Nxe5? fails to 9...Nxe5 10.Rxe5 Bxf2+!, intending 11.Kxf2? Qf6+ and ...Qxe5.

376
 f2) 9.Ng5 Bb6 10.Nxe6?! fxe6 (the open f-file is more relevant than the weak e-pawns) 11.Be3
Nxe3 12.fxe3 (or 12.Bxe6+ Kh8 13.fxe3 Qf6 again; e.g. 14.Bd5 Bxe3+ 15.Kh1 Nd4 16.Rxe3? Qf1+
17.Qxf1 Rxf1+ 18.Kh2 Nxc2) 12...Qf6 13.Qe2 Ne7 14.Nd2 Nf5 15.Nf1 Kh8 16.c3 Rae8 17.Rad1
Qh6 18.d4 exd4 19.exd4 e5 and Black is clearly better, A.Zude-P.Zelbel, German Bundesliga 2019.
f3) 9.Nbd2 f6! (now that Nc3 is no longer possible and d3-d4 not yet a threat) 10.Nb3 (or 10.c3
Re8! again, defending one bishop against tactics and allowing the other to retreat to f8 if necessary)
10...Bb6 11.d4 (if 11.a4 then 11...a5! seems preferable to 11...a6 12.a5 Ba7, as in I.Nepomniachtchi-
H.Nakamura, Paris GCT blitz 2019) 11...Bf7! 12.dxe5 Nxe5 13.Nxe5 fxe5 14.Bxd5 (not 14.Rxe5?
Bxf2+) 14...Bxd5 15.Be3 c6 16.Qe2 Qh4 and Black’s activity balances the weak e-pawn, J.Van
Foreest-A.Bachmann, Teplice 2016.
f4) 9.c3 Bb6! (stepping back from any d3-d4 ideas; not yet 9...f6?! 10.d4! Bb6 11.Nbd2 Bf7
12.dxe5 Nxe5 13.Nxe5 fxe5 14.Nf3 with an edge – Mikhalevski) 10.Bb3 (or 10.Bb5 f6 11.d4 Nde7
12.Nbd2 Bf7, K.Gierulski-Ba.Dahl, Krakow 2017) 10...Re8 (and not 10...f6??, as in S.Karjakin-
L.Aronian, Wijk aan Zee 2017, because of 11.c4 Ndb4 12.c5 and wins – Mikhalevski) 11.Na3 (or
11.Ng5 Qf6) 11...Bc5! (an unexpected return to keep the bishop; 11...f6?! is still wrong, this time
because of 12.d4!, intending 12...exd4? 13.c4 Ndb4 14.c5 and wins, A.Volokitin-A.Shirov, Berlin
blitz 2018) 12.Nc4 f6 (finally) 13.d4 exd4 14.cxd4 Bf8 and Black is fine, Ding Liren-L.Dominguez
Perez, PRO League (rapid) 2020.

B21: 8.Nbd2

Another Italian by transposition, specifically 3.Bc4 Bc5 4.0-0 Nf6 5.d3 0-0 6.Nbd2 d5 7.exd5 Nxd5
8.Re1.

8...f6

This seems logical now that White has committed the knight to d2. Black strongpoints e5 and
intends to develop behind with, say, ...Kh8, ...Be6, ...Qd7, ...Rae8. The question is whether Black can
get away with it, since e5 isn’t really so very strong.
Nevertheless, I prefer the text to 8...Nf6, as played four times by GM Etienne Bacrot. This defends
e5 tactically in view of 9.Nxe5? Nxe5 10.Rxe5 Bxf2+! etc, but returning the knight from whence it
came feels wrong to me. GM Bacrot likely feels the same way about ...f7-f6.

377
9.d4

White attacks e5 at once, exploiting the vulnerable d5-knight. 9.Nb3 Bb6 10.d4 exd4 comes to the
same thing; and other lines with d3-d4 will probably transpose; e.g. 9.a4 a5 10.c3 Kh8 11.d4 exd4
12.Nb3 Bb6 13.Nbxd4 Nxd4 14.Nxd4 c6 (see 14.c3 below); or if 9.c3 Kh8 10.d4, A.Chudinovskih-
K.Novosadova, Brno 2015, then 10...exd4 11.Nb3 Bb6 12.Nbxd4 Nxd4 13.Nxd4 (13.cxd4 makes no
sense) 13...c6. And 9.Ne4 Bb6 10.Nc3 Be6 11.d4, V.Jacko-V.Zemlicka, Ostrava 2010, can be
answered by 11...Bf7!, intending 12.dxe5 (or 12.Nxd5 Bxd5 13.Bxd5+ Qxd5 14.dxe5 Qxd1 15.Rxd1
Nxe5) 12...Nxc3 13.Qxd8 Nxd8 14.e6 (or 14.Bxf7+ Nxf7 15.e6 Nd5) 14...Nxe6 15.Bxe6 Bxe6
16.Rxe6 Na4 with equality.
White cannot hope for much, if anything, without pushing the d-pawn; e.g. 9.Ne4 Bb6 10.c3 (the
attempted Noah’s Ark with 10.Bxd5+ Qxd5 11.c4? fails to 11...Qf7 12.c5 Ba5 13.Re3 Be6 14.a3
Bb3) 10...Kh8 11.b4!? (threatening Bxd5, followed by c3-c4-c5; if instead 11.a4, S.Savitskiy-
A.Shirov, Russian Team Blitz Ch. 2017, then 11...a5 looks best; similarly 11.h3 Be6 12.a4,
R.Pruijssers-P.Zelbel, Belgian League 2019, and 12...a5; while 11.Bb3 Bg4 12.h3 Bh5 13.Ng3 Bf7
14.Nh4 Qd7 15.Nhf5 Rad8 gave Black the edge in E.Cormack-J.Tait, Sheffield League 2020) 11...a6
12.Qb3, An.Sokolov-S.Feller, Nimes 2009, when 12...Be6 and ...Qd7 seems fine for Black; ...Bf7 and
...Bg8 also come into consideration.

9...exd4

Black cannot maintain the central bulwark because 9...Nxd4 10.Nxd4 Bxd4? 11.Nb3 c6 12.Nxd4
exd4 13.Qxd4 breaks it down anyway and in a worse position.

10.Nb3

378
 After 10.Ne4 Bb6 11.Nxd4 White can at best regain the material; e.g. 11...Nxd4 12.Be3 (or 12.c3
Kh8) 12...Nxc2 13.Qxc2 (or 13.Bxd5+ Kh8 14.Qxc2 Qxd5) 13...Bxe3 14.Rxe3 c6 15.Nc3 Kh8
16.Nxd5 cxd5 17.Rd1 d4 18.Rg3 Re8 and Stockfish says “0.00”.

10...Bb6 11.a4

Inserting two a-pawn moves gives White an extra option. The alternative is 11.Nbxd4 Nxd4
12.Nxd4 c6 and then:

a) 13.c3 Kh8 14.Qh5 (or 14.Bd2 Bxd4 15.cxd4 Bf5) 14...Bxd4 15.cxd4 Bd7 16.Bxd5 (or 16.Bd3
f5) 16...Be8! 17.Qf3 Qxd5 with equality.
b) 13.Bf4 Kh8 14.Bg3 Re8 15.Rxe8+ (or 15.Qh5 Bd7 16.Bd3 Rxe1+ 17.Rxe1 Qg8! 18.Nf3 Re8
19.Rxe8 Bxe8 20.Qh4 Ne7 21.Qe4 ½-½ Capalaskine-tsmenace, ChessWorld.net 2018) 15...Qxe8
16.c3 Bd7 17.a4 a5 18.h3 Qc8! (making way for the bishop manoeuvre that follows) 19.Qd2 Be8
20.Re1 Bg6 21.Nb5!? Qd7 22.Nd6 Bc5 23.Bxd5 cxd5 24.Qxd5 Bxd6 25.Qxd6 Qxa4 26.Re2 ½-½
afms-tsmenace, ChessWorld.net 2018.
c) 13.Ne6 Bxe6 14.Rxe6 Qd7 15.Re2 Kh8 16.Qd3 Rad8 17.Bd2 Qg4! 18.h3 Qh4 19.Bxd5 Rxd5
20.Qf3 h6 21.Rae1 Rf7 and Black had everything defended, pgcbies-tsmenace, ChessWorld.net 2018.

11...a5 12.Nbxd4 Nxd4 13.Nxd4 c6

379
14.Nb5!?

And this is it. Other moves are much the same as before:
a) 14.c3 Kh8 15.Qh5 Bxd4 16.cxd4 Bd7 17.Bd3 (or 17.Bxd5 Be8 again) 17...f5 18.b3 Qb6 19.Ba3
Rf6 20.Bc4!? Qxd4 21.h3 (preventing 21.Rad1 Qg4) 21...Rg6 22.Qf3 Re6 23.Rad1 Rxe1+ 24.Rxe1
Re8 25.Rd1 Qf6 26.Bc1 (or 26.Bxd5 cxd5 27.Qxd5 Bc6 with equality, since 28.Qxa5?! Qg6 29.g3 f4
30.Qc3 h5 is only dangerous for White) 26...Be6 27.Re1 Nc7 28.Qf4 Bxc4 29.Rxe8+ Nxe8 30.Qxc4
h6 31.Qc5 Qd6 32.Qxf5 Qd1+ 33.Kh2 Qxc1 34.Qf8+ Kh7 35.Qxe8 Qf4+ ½-½ docjan-tsmenace,
ChessWorld.net 2018.
b) 14.Bf4 Kh8 15.Bg3 Re8 16.Qd2 (otherwise, 16.Rxe8+ Qxe8 or 16.Qh5 Bd7) 16...Bg4 17.Rxe8+
Qxe8 18.h3 Bh5 19.Nf5 Bc5 20.Nd6 Qd7 21.Bxd5 cxd5 22.Qxd5 Bxd6 23.c3 f5 24.Qxd6 Qxd6
25.Bxd6 Rd8 26.Bc7 Rd1+ 27.Rxd1 Bxd1 28.Bxa5 Bxa4 with an opposite-coloured bishops draw,
sinjoor1-tsmenace, ChessWorld.net 2020.

380
14...Bf5

Seeing as the white knight has left d4. Obviously Black can’t take it: 14...cxb5? 15.Qxd5+ and
a4xb5 wins.

15.Be3 Bxe3 16.Rxe3 Bg6

Better than 16...Kh8?! 17.Bxd5 Qxd5 18.Qxd5 cxd5 19.Nd4 Bd7 20.Rb3 Rfb8 21.Rb6, when Black
has misplaced the king for the endgame.

17.Nc3

Increasing the pressure on d5 before Black has escaped the pin. Now after 17.Bxd5+ Qxd5
18.Qxd5+ cxd5 19.Nd4 Rac8 Black has far more activity, and while the white knight is a good piece,
it’s not up against a bad bishop.

17...Kh8

381
 Now is the right moment. (Not 17...Bf7? 18.Rd3 and d5 falls for nothing.) Black maintains material
equality due to the possibility of ...Bxc2. For example:
a) 18.Bxd5 cxd5 19.Nb5 (or 19.Nxd5 Bxc2 20.Qxc2 Qxd5) 19...Qd7 20.Rb3 b6 21.Rc3 Rac8
22.Rxc8 Qxc8 23.c3 Qc5 24.h4 Re8 25.Qd2 h6 26.g3 Re5 27.b3 Qe7 28.Nd4 Qe8 and again Black
had everything defended, pgcbies-tsmenace, ChessWorld.net 2019.
b) 18.Nxd5 cxd5 19.Qxd5 (or 19.Bxd5 Bxc2 again) 19...Bxc2 20.Qxb7 Bxa4! 21.b3 (not 21.Rxa4??
Qd1+) 21...Rb8 22.Qa7 (slightly more testing than 22.Qf3 ½-½ pgcbies-tsmenace, ChessWorld.net
2019; presumably my opponent had concluded that it didn’t lead to anything either) 22...Bb5 23.Rg3
(or 23.Be6 Ra8 24.Qc5 Rb8) 23...g6 24.Rh3 (or 24.h4 Bxc4 25.bxc4 Ra8) 24...h5 25.Rg3 (or 25.g4
Bxc4 26.bxc4 Qc8) 25...Bxc4 26.bxc4 Qb6 27.Qxb6 Rxb6 28.Rxg6 Kh7 29.Rg3 Rc8 and Black will
hold the rook endgame.

B22: 8.Nxe5

382
 Taking the pawn is surely critical. GM Alexander Delchev calls this whole variation “a speculative
pawn sacrifice, which [White] should accept if [they] want to punish Black’s bizarre move order.”

8...Nxe5

Or 8...Bxf2+ 9.Kxf2 Nxe5 10.Rxe5, transposing.

The old idea was 8...Qh4?!. I once surprised IM (later GM) James Howell with this (Warrington
rapid 1990). He thought for a couple of minutes and then played 9.Nf3! (9.Qf3 is good too) 9...Qxf2+
10.Kh1, which basically refutes the whole thing; e.g. 10...Nf6 11.Re2 Ng4 (the only worthwhile try; I
didn’t see it) 12.c3 b5!? 13.Bf4! (not 13.Bxb5? Nce5! 14.Nxe5 Bb7 15.Rxf2 Nxf2+ 16.Kg1 Nxd1+
17.d4 Bd6; or 13.Bd5?! Bf5! 14.Bxc6?! Bxd3 15.Rd2 Rad8, while 14.d4 Rae8 15.Rxe8 Rxe8
16.Bxc6 Bc2! 17.Qg1 Qe2 18.Bxe8 Nf2+ 19.Qxf2 Qxf2 20.Bxb5 Bd6 is “0.00” according to
Stockfish) 13...bxc4 14.Bg3 cxd3 15.Rxf2 Bxf2 16.Qa4 Bd7 17.Bxf2 Nxf2+ 18.Kg1 Ng4 19.Nbd2
and Black doesn’t have enough for the queen.

9.Rxe5 Bxf2+!

I’ve started so I’ll finish. 9...c6 10.Qf3 Be6 11.Nd2 Bb6 12.a3 Bc7 13.Re1 (Müller & Souleidis) is
good for White.

10.Kxf2 Qh4+

This is an important zwischenzug. Not 10...Qf6+? 11.Qf3 Qxe5 12.Bxd5 Qxh2 because, with no
...Qh1+ threats to worry about, White can develop and consolidate straightforwardly; e.g. 13.Nc3 c6
14.Bb3 Qh4+ 15.Qg3 Qf6+ 16.Kg1 Bf5 17.Bg5 Qg6 18.Rf1 and White won, V.Rodchenkov-
E.Doluhanova, Titled Tuesday (blitz) 2020.

383
11.Kf1

The cheapo 11.Kg1?? Qd4+ can catch people out in blitz games.

11...Qf6+ 12.Qf3 Qxe5 13.Bxd5

Not 13.Qxd5? Qxh2, when White needs to think about a draw; e.g. 14.Qd4 Be6 15.Bf4 Qh1+
16.Qg1 Qh4 17.Qf2 Qh1+ etc.

13...c6

Not 13...Be6? 14.Nc3 (or just 14.Bxb7!) 14...Bxd5 15.Nxd5 Rad8 16.Nc3 Qxh2 17.Rb1 Rd4
18.Be3 Rh4 19.Ke2 f5 20.Rf1 Rg4 21.Rf2 c6 22.Kd2 and White consolidated, G.Jones-T.Kriebel,
German Bundesliga 2019.

14.Bb3

The most natural retreat. I’ve faced 14.Bc4!? too, allowing ...b7-b5 with tempo, which changes the
play in subtle ways. If Black continues as in the main line with 14...Qxh2 15.Be3 a5 16.a4 Be6
17.Nd2 Qh1+ 18.Bg1 Rae8 19.Qg3 Qh6, then 20.Bd4! is good for White; or if 15...b5 16.Bb3 a5
17.a4 Be6 18.Nd2 Qh1+ 19.Bg1 Rae8 20.Qg3 Qh6, then 21.Nf3! is possible because 21...Bxb3
22.cxb3 Qh5?! runs into 23.b4! axb4 (or 23...bxa4 24.Bc5) 24.a5 with a strong passed pawn.
So Black should focus on development and leave regaining material for the moment: 14...b5
15.Bb3 a5 16.a4 Be6! 17.Qf4 (or 17.Nc3 Bxb3 18.cxb3 Qxh2) 17...Qh5 (threatening ...Qd1+) 18.c4
(or 18.Nc3 Bxb3 19.cxb3 b4) 18...Rad8 with an unclear position; e.g. 19.axb5 cxb5 20.Nc3 Qg6!
21.Kg1 Qxd3 22.Qe3 Qxe3+ 23.Bxe3 Rd3 24.Kf2 b4 25.Bc2 Rxe3 26.Kxe3 bxc3 27.bxc3 Bxc4
28.Rxa5 Rd8 29.Be4 Kf8 30.Ra8 ½-½ remyrey-tsmenace, ChessWorld.net 2020.

384
14...Qxh2

15.Be3

“White’s pieces could prove to be stronger than the rook, but White should not overestimate his
chances,” says GM Delchev, who adds a short continuation beginning 15...a5. Black will play the
same way against other moves; e.g. 15.Kf2 a5, or 15.Nc3 a5, or 15.Bf4 Qh1+ 16.Kf2 a5.

15...a5!

Seeking to undermine the b3-bishop so that ...Be6 comes with a threat to White’s structure. Not yet
15...Be6?! 16.Bg1 (or 16.Nd2 Qh1+ 17.Bg1 Rfe8 18.Qg3 Bxb3 19.axb3 h6 20.Qh2 Qxh2 21.Bxh2,
A.Moshovakis-D.Tsekmes, Greek League 2018) 16...Qh4 17.Nc3 a5 (too late) 18.Re1 a4 19.Bf2
Qh1+ 20.Ke2 Qh6 21.Bxe6 (or just 21.Nxa4) 21...Rae8 22.Kd1 Rxe6 23.Ne4 “and White had
successfully negotiated the complications to reach an advantage in R.Tischbierek-I.Donev,
Liechtenstein 1995.” (Emms)

385
16.a4

Halting the black a-pawn. Note that 16.Nc3?! doesn’t do that because of 16...a4! 17.Bxa4 Qh1+
18.Bg1 Ra5! and the rook swings across with strong counterplay; e.g. 19.Qg3 Rf5+ 20.Ke2 Qh5+
21.Kd2 Rg5 22.Qf3 Qh1 23.Rf1 Qxg2+ 24.Qf2 Be6 25.Ne4 Rg4 26.Bb3 Bxb3 27.axb3 f5 and Black
is not worse, V.Oliynyk-D.Tylecek, corr. 2015.

16...Be6 17.Nd2 Qh1+

To undevelop White’s dark-squared bishop and gain the h6-square for the queen.

18.Bg1 Rae8 19.Qg3

Intending Qh2 to exchange queens and neutralize Black’s initiative.

19...Qh6

386
 Delchev ends here, writing that White “is only slightly better in a sharp position.” (I’m assuming
this is what he meant anyway, since his 17...Rae8 18.Qg3 Qh6 has serious drawbacks for both sides.)
I can live with “slightly better”, especially as White has to get here first, and even then I’ve drawn
four games out of four.

20.Rd1

Simply defending the knight. Instead:

a) 20.Be3 forces a draw if White wants one: 20...Qh1+ 21.Bg1 Qh6 22.Be3 Qh1+ 23.Bg1 Qh6 ½-½
ChessPleasure-tsmenace, ChessWorld.net 2020.
b) 20.Nf3 Bxb3 21.cxb3 Qh5 prepares to attack the new targets at b2, b3 and d3.
c) 20.Nc4 Bd5! 21.Nxa5? concedes a stronger initiative; e.g. 21...Re6 22.Bxd5 cxd5 23.Nb3 (or
23.Bd4 Rg6 24.Qe3 Qh2 25.Qg1 Qc7 26.Bc3 Re8) 23...Rg6 24.Qh3 Rf6+ 25.Bf2 Qf4 26.Qf3 Qc7
27.Qe3 Qxc2.
d) 20.Bd4 can be met by 20...c5! 21.Be3 (or if 21.Bc3 then 21...Bxb3, followed by ...Re3; not
21.Bxc5? Qxd2 22.Bxf8 Rxf8) 21...Qh1+ 22.Bg1 Qh6 23.Nc4 (or 23.Qf2 b6) 23...Bd5 24.Bxc5 Qh1+
25.Bg1 Re6 26.Qh2 Bxg2+ 27.Kf2 Rfe8 28.Qxh1 Bxh1 29.Kf1 Rf6+ 30.Bf2 g5 31.Re1 Rxe1+
32.Kxe1 h5 and Black has strong counterplay with the rook and kingside pawns.
e) 20.Bxe6 allows 20...Qxd2 21.Bg4 Qb4 (21...Qxc2?! 22.Bd4 is good for White) 22.Rd1 (or 22.b3
Qc3) 22...Qxb2 with rook and two pawns for two bishops, unless White goes for the exchange:
23.Bc5 Qxc2 24.Bxf8 Rxf8, in which case Black has three pawns for one bishop; e.g. 25.Qf4 b5
26.axb5 cxb5 27.Qc1 Qxc1 28.Rxc1 Rb8 and Stockfish gives White only a slight edge.

387
20...b5!?

Not necessarily best. Stockfish prefers 20...Qh5 21.Re1 Bxb3 22.Nxb3 f5 23.Nd4 f4 24.Qf3 Rxe1+
25.Kxe1 Re8+ 26.Kf1 Qxf3+ 27.Nxf3 h6 with ideas of ...g7-g5, ...Kg7-g6, ...h6-h5 and so on.

21.Bxe6

Otherwise 21.Nf3 Bxb3 22.cxb3 wrecks White’s queenside again.

21...Rxe6 22.Nf3 Rfe8 23.Qh4

Finally forcing queens off. 23.Qg5 Qh1 24.Qh4 comes to the same thing.

23...Qxh4 24.Nxh4

388
 White seems close to consolidating since the doubled rooks are not doing much on the e-file. But
Black isn’t finished yet.

24...g5! 25.Nf5

Not 25.Nf3?? g4 26.Nd4 Rf6+ 27.Bf2 g3 and Black wins.

25...h5 26.Ng3

After 26.Bd4?! c5 27.Bc3 b4 28.Ng7 bxc3 29.Nxe8 Rxe8 30.bxc3 Black is not worse in the rook
endgame.

26...h4 27.Ne4

Centralizing the knight, preventing ...Rf6+, as well as hitting the g5-pawn.

27...g4 28.Bd4

Threatening Nf6+.

28...Rd8 29.c3 Rd5!

Finding another route to the f-file.

30.Ra1 Rf5+ 31.Kg1 bxa4 32.Rxa4 g3

The cramping g-pawn now gives Black sufficient counterplay against the white king to hold the
draw. For example:

389
 a) 33.b3 Kh7 (or ½-½ right here in another game) 34.Bf6 Rexf6 35.Nxf6+ Rxf6 36.Rxh4+ Kg7
37.Rh3 ½-½ Capalaskine-tsmenace, ChessWorld.net 2018, in view of something like 37...Rg6
38.Kf1 c5 39.Ke2 f5 40.Kf3 Rb6 41.Rh4 Rxb3 42.Rc4 a4 43.Rxc5 a3 44.Kxg3 a2 45.Ra5 Rxc3
46.Rxa2 Rxd3+ and draws.
b) 33.Ra1 Kh7 34.Bb6 Kg6 35.c4 Rh5 36.Bxa5 Rxa5! 37.Rxa5 f5 38.Nd2 Re1+ 39.Nf1 f4 40.Ra8
f3! 41.gxf3 Kg7 (threatening ...h4-h3 and wins; not 41...h3?? 42.Rg8+) 42.Ra7+ (or 42.Kg2 Re2+
43.Kg1 Re1) 42...Kg6 ½-½ lysol-tsmenace, ChessWorld.net 2019, since 43.Kg2 (or 43.Ra8 Kg7)
43...Re2+ 44.Kg1 Re1 repeats.
I was actually White in both those games, which reached the exact same position with reversed
colours; e.g. via 1.e4 Nf6 2.Nc3 e5 3.a3 Bc5 4.Nf3 d6 5.d4!? with a later 16.a4.
We’ve now transposed to 3...Bc5 4.c3 Nf6 5.d3 0-0 6.0-0 d5 7.exd5 Nxd5.
generally arising from an Italian; e.g. 3...Bc5 4.0-0 Nf6 5.d3 0-0 6.h3 d5 7.exd5 Nxd5 8.Re1
Another Italian by transposition, specifically 3.Bc4 Bc5 4.0-0 Nf6 5.d3 0-0 6.Nbd2 d5 7.exd5 Nxd5
8.Re1.

390
 Chapter Nine
Ulvestad Variation

1.e4 e5 2.Nf3 Nc6 3.Bc4 Nf6 4.Ng5

Having examined other fourth moves in the previous chapter, all that remains is this richtiger
Stümperzug. I was intending initially to write about the Traxler, my main weapon for over thirty
years. Unfortunately, issues have arisen with 4...Bc5, not least 5.d4! for which IM Panayotis Frendzas
has been proselytizing online. It seems the best Black can do is to try to defend a dreary endgame a
pawn down after 5...d5 6.Bxd5 Nxd5 7.dxc5 Ndb4 8.a3 Qxd1+ 9.Kxd1 Na6 10.b4. Although I have
managed that each time as Black, the prospect of having to do so repetitively is profoundly
unappealing.
So I’ve switched to my second string: the Ulvestad Variation. An online thematic tournament
prompted me to go through published theory with Stockfish which turned up quite a lot of interesting
ideas. Ultimately, Black may still be obliged to defend pawn-down endgames, but I think there are
excellent prospects for success.

4...d5 5.exd5 b5!?

This move was discovered by the US master Olaf Ulvestad and analysed in his openings booklet
Chess Charts. Ulvestad took the view that 5...Na5 was suspect (!) because it sidelines the knight and
thus sought to find an improvement. He came up with 5...b5, which he justified as follows:

“It meets the Pawn attack on the Knight by a Pawn attack on White’s most effectively placed piece.
White’s original attack is completely broken up and Black definitely obtains a powerful initiative.
Instead of committing the same mistake as White, by moving the same piece twice in the opening,

391
Black attacks with a move which furthers his general development. 5...P-Kt4 provides useful freedom
to his QB and QR.”
This led to a discussion with another American master, Albert Pinkus, in the 1941 issues of Chess
Review (incidentally, the Traxler line above comes from Pinkus’ articles), where Ulvestad doubled
down on behalf of 5...b5: “It stands out head and shoulders above all the other moves as the strongest,
sharpest and best. The last word on the play may not be given for a long time but I have complete
confidence in the move itself. Analysis will support it—now or later.”
Last words are hard to come by in chess, so the following should only be regarded as the latest
word. Until it no longer is.
A: 6.Bxb5
B: 6.dxc6
C: 6.Bf1
The most natural response for White is to take the b-pawn, so we’ll look at 6.Bxb5 first, followed
next by 6.dxc6. Neither of these direct captures should trouble Black. The critical move is 6.Bf1!, as
recommended by Pinkus. Note that the more obvious retreats, 6.Be2 and 6.Bb3 are inferior, as
6...Nd4 then hits the bishop, neutralizing 7.c3 as a threat:
a) 6.Be2 Nd4 7.Nf3 (the best try; 7.c3?! Nxe2 8.Qxe2 Qxd5 is already good for Black, e.g. 9.d4
Qxg2 10.Qxe5+ Be7 11.Qxb5+ Bd7 12.Qf1 Qd5 13.h4 h6 14.Nh3 Bb5 15.Qg1 Bd6 16.Be3 Bc6 and
Black won, V.Malvinski-Mir.Lazic, Bela Crkva 1989; as is 7.0-0 Bd6 8.c3?! Nxe2+ 9.Qxe2 0-0,
since 10.Qxb5? loses to 10...h6 11.Nf3 e4 12.Nd4 Bxh2+! etc, B.Kiviaho-J.Van Leusden, Toronto
1978) 7...Qxd5 (Ulvestad went for 7...Nxf3+ 8.Bxf3 e4 9.Be2 a6) 8.Nxd4 (or 8.Nc3 Qc5 9.Nxd4 exd4
10.Nxb5 Bb7 with compensation; e.g. 11.c3 d3! 12.Bxd3 0-0-0 13.Be2 Qg5 14.Rg1 Qh4 15.Qb3 Re8
16.Nxa7+ Kb8 17.Nc6+ Kc8 18.Na7+ Kb8 is a draw) 8...exd4 9.0-0 Rb8 10.Bf3 Qc5 11.Qe2+ Be6
and Black is fine, S.Petz-S.Titgemeyer, Dortmund 2006.
b) 6.Bb3 Nd4 and now:

392
 b1) 7.c4? Nxd5 8.Nxf7? fails spectacularly to 8...Qh4! 9.0-0 Bg4 10.Qe1 (or 10.f3 Nf4 11.Kh1 Nf5
12.Qe1 Bh3!) 10...Ne2+ 11.Kh1 Ndf4 12.Nxe5 (or 12.f3 Bxf3! 13.gxf3 Qh3) 12...Bd6 13.f3 Bh3!
14.gxh3 Qxh3 15.Rf2 Ng3+ 16.Kg1 Qg2+! 17.Rxg2 Nh3 mate (L.Radchenko).
b2) 7.c3?! Nxb3 8.Qxb3 Qxd5 is good for Black; e.g. 9.0-0 Bb7 10.Qxd5 Nxd5 11.Re1 f6 12.Ne6
Kf7 13.Nxf8 Rhxf8 14.d4 exd4 15.cxd4 Rfe8, T.Warakomski-M.Kanarek, Warsaw (rapid) 2006.
b3) 7.d6 Nxb3 8.dxc7 is met by 8...Qd5! 9.axb3 Qxg2 10.Qf3 Qxf3 11.Nxf3 Bb7 12.Ke2 a6, which
leaves White in a mess.
b4) 7.d3 Bc5 (or 7...Nxb3 8.axb3 Qxd5 – Ulvestad) 8.0-0 Bg4 9.Qe1 0-0 10.Ne4 Nxe4 11.dxe4
Qh4!? (11...f5! is stronger) 12.Kh1?? (12.Nd2 was forced) 12...Nf3! 0-1 A.Touret-V.Petkov, Metz
2007.
b5) 7.Nc3 h6 8.Nge4 Nxe4 (8...a5! looks even better) 9.Nxe4 Qh4!? 10.Ng3 Bg4 11.f3 Bd7 12.c3
Nxb3 13.axb3 Bc5 14.Qe2 0-0 15.Qe4 Qd8! with excellent play for the pawn, F.Raab-A.Karpatchev,
Selestat 2008.

A: 6.Bxb5 Qxd5

393
 This was Ulvestad’s main idea. Deflecting the bishop has allowed Black to recapture the d5-pawn
with the queen, which now attacks both b5 and g2, thus forcing White to make a concession – either
in wasting time or relinquishing the pawn or bishop pair.

7.Bxc6+

Relatively best, in that White has nothing better:

a) 7.Qe2? Qxg2 8.Qxe5+? (too greedy; despite the wasted tempo on Qe2 White has to concede
8.Qf3 Qxf3 9.Nxf3 Bd7 with a slight edge for Black) 8...Be7 9.Rf1 0-0 10.Bxc6? (or 10.Qxc7? Nd4)
10...Qxc6 11.Qc3 Qb5 wins for Black (Ulvestad), and even after 10.Qg3 Qxg3 11.hxg3 Nd4 12.Bd3
(or 12.Ba4 Bd7) 12...Re8 White is unlikely to survive.
b) 7.Bf1? (too passive) 7...Bc5! 8.d3 (or similarly 8.Nc3 Bxf2+! 9.Kxf2 Qd4+ 10.Ke1 Qh4+ 11.g3
Qxg5, I.Barchuk-M.Bezkorovaina, Lvov 2017) 8...h6 9.Nc3 Bxf2+! 10.Kxf2 Qd4+ 11.Ke1 hxg5
12.Bxg5 Rb8 13.Bxf6 gxf6 14.Qf3 Qh4+ 15.g3 Nd4 16.Qf2 Qh6 and Black has more than enough for
the pawn, J.Tanti-K.H.Hansen, corr. 2008.
c) 7.Nc3?! (too simplistic) 7...Qxg2 8.Qf3 Qxf3 9.Nxf3 Bd7 sees Black regain the pawn straight
away.

394
 Ulvestad stopped here, assessing the position as equal. Practice has shown Black to be on the right
side of that. For example: 10.0-0 (or 10.d3 Nd4 11.Bxd7+ Kxd7 12.Nxd4 exd4 13.Ne2 Re8 14.Kf1 c5
15.Bd2 Bd6 16.Ng3 Nd5, M.Nezar-L.Fressinet, French League 2010; or 10.Rg1 Nd4 11.Bxd7+ Kxd7
12.Nxd4 exd4 13.Ne2 d3!? 14.cxd3 Nd5, J.Bourges-A.Karpatchev, Guingamp 1999) 10...Bd6
11.Bxc6 Bxc6 12.Nxe5 Bxe5 13.Re1 0-0-0 14.Rxe5 Rhe8 15.d4 (or 15.Re2 Rxe2 16.Nxe2 Re8
17.Kf1 Ng4 18.Nd4 Nxh2+ 19.Kg1 Nf3+ 20.Nxf3 Bxf3, D.Norris-Tsang Hon Ki, Genting 1995)
15...Rxd4 16.Rxe8+ Nxe8 17.Be3 Rg4+ 18.Kf1 Rh4 (or 18...Rg2!?) 19.Ke2 a6 20.Rd1 Rxh2 21.Nd5
Bxd5 22.Rxd5 Rh4, L.M.Kovacs-A.Hennings, Sarajevo 1969. White was worse in all these games.
d) 7.Be2 at least has the idea 7...Qxg2?? 8.Bf3, so Black must rely on activity: 7...Bb7 (more
accurate than 7...Nd4, if only because 8.Nf3 Bb7 9.Nc3 Qc6?? then loses to 10.Nxd4 exd4 11.Bb5)

395
 d1) 8.0-0 Nd4 9.Nf3 (both 9.f3? h6 10.Nh3 g5 and 9.Bf3?! e4! 10.Nxe4 Nxf3+ 11.gxf3 0-0-0 are
bad for White, despite the extra material) 9...Bd6 (or 9...0-0-0 10.Nc3 Qc6 – Ulvestad) 10.d3 0-0
11.c4 Qc6 12.Nc3 Rfe8 13.Be3 Rad8 14.Rb1 a5 and Black’s control is worth more than White’s
pawn, J.Splíchal-Wo.Bauer, corr. 2010.
d2) 8.d3 Nd4 9.Nf3 (if 9.Bf3 then 9...Nxf3+ 10.Qxf3 Qa5+ 11.Bd2 Qb6 – Estrin; or 9.f3?! Bc5
10.Nc3 Qc6 11.Nge4 Bb6 12.Bg5 0-0-0 13.Qd2 h6 14.Bh4 g5 15.Bf2 Nh5 and White’s position is
horrible, H.Wademark-M.Keller, corr. 1986) 9...0-0-0 10.Nc3 Qc6 11.0-0, F.Renault-L.Lejarre,
French League 2008, and now Black could play 11...g5! at once, since 12.Bxg5? Rg8 13.Bxf6 Qxf6
14.Kh1 Qg7 15.Rg1 f5 is a crushing attack.

7...Qxc6

Having neutralized one threat (at b5) White can now look to the other (g2) and has two ways of
doing so:
A1: 8.Qf3
A2: 8.0-0
Not 8.Nf3? Ba6! 9.d3 e4 and White is already in serious difficulties.

A1: 8.Qf3

A counter-attacking try analysed by Pinkus.

8...e4

The usual reply. Black seems never to have grabbed the c2-pawn with 8...Qxc2!?, perhaps because
9.Qxa8 Qxc1+ 10.Ke2 Qxh1?? 11.Qxc8+ Ke7 12.Qxc7+ Nd7 13.Nc3 gives White a decisive attack

396
(Pinkus). Nonetheless, 10...Qc4+! 11.d3 (or 11.Kd1 Qa6) 11...Qg4+ 12.Nf3 Bd6 looks okay, as does
9.0-0 Rb8.

9.Qb3

Targeting the weak f7-square which Black cannot defend without making unfavourable
concessions.
Opposing queens again with 9.Qc3 doesn’t bring White anything: 9...Bb7 (or 9...Qd5 10.d3 Bb7, if
Black prefers to keep the queens on) 10.Qxc6+ Bxc6 11.0-0?! h6 12.Nh3 0-0-0 13.Nc3?! g5 and
White’s casual play landed him in trouble, L.Fassio-A.Gallucci, Turin 2010.

9...Bc5!

Fortunately, Black has no need to worry about f7. Even 9...h6!? 10.Nxf7 (not 10.Qxf7+? Kd8
11.Nh3 e3! 12.f3 Qxc2) 10...Be6 11.Ne5 Bxb3 12.Nxc6 Bxc2, first seen in P.Tumurbator-K.Harandi,
Tehran 1972, should be okay for Black, who has control of the light squares.

10.Qxf7+

If White throws in 10.d4 (“!” – Pinkus) first, then 10...exd3 11.Qxf7+ Kd8 12.0-0 Rf8 13.Qb3 dxc2
14.Qxc2 (or 14.Na3 Ng4 15.Nf7+ Kd7 16.Nxc2 Bb7) 14...Bb7 15.Nf3 Ng4 gives Black excellent
play for the pawn.

10...Kd8 11.0-0

Not yet 11.Qxg7?? Rg8 12.Qh6 Bxf2+ and wins; or similarly 11.Nc3? h6 12.Qxg7? Rg8 13.Qxh6
Bxf2+; while 11.Qb3?! allows 11...Re8! 12.0-0 Ba6.

397
11...Rf8

Instead, 11...h6!? 12.Qxg7 Rg8 13.Qxh6 Bb7 14.Kh1 e3 15.f3 e2 16.Re1 Ne4, G.Bardi-O.Kenez,
corr. 1976, leads to a draw after 17.Qxc6 Nf2+ 18.Kg1 Nh3+ 19.Kh1 etc. But White might try
12.d4!? exd3 13.Nf3 dxc2 14.Nc3 with a favourable version of Pinkus’ idea above.

12.Qxg7

Retreating the queen is objectively safer: 12.Qb3 Ng4! 13.d4! (not 13.Nh3?, Y.Berghaell-
G.Lagland, corr. 1964, because of 13...Rb8! 14.Qc3 Ba6 15.Qg3 Qd7 or 14.Qg3 Bd6 15.Qh4+ Be7
16.Qg3 h5 with a huge attack) 13...Bxd4 14.Nc3 (White can play 14.h3 as well; but not 14.Rd1? Qd6
15.Nxe4 Qxh2+ 16.Kf1 Ba6+ 17.c4 c5 and Black is winning, K.Essegern-C.Brauer, corr. 1984)
14...Nxf2 (14...Bxf2+?! 15.Rxf2 Nxf2 16.Be3 Ng4 17.Rd1+ Ke8 18.Nd5 is only dangerous for Black)
15.Bf4! (the key resource) 15...Rxf4 16.Ne2 Nh3+ 17.Kh1 Rxf1+ 18.Rxf1 Nf2+ 19.Rxf2 Bxf2
20.Qg8+ with a draw.

12...Rg8 13.Qh6 Bb7

For the three-pawn investment Black has a big lead in development, nearly all of which is lined up
against the enemy king. White needs to be very precise to survive.

14.Kh1!

The only move, as shown by 14.Rd1? Bxf2+ 15.Kf1 Ke7 16.Nxh7 Bd4 17.b3 Qxc2 0-1 S.Sandin-
W.Weiss, corr. 1962; or 14.Nf7+? Ke7 15.Ne5 Qd5 (Stockfish bangs out 15...Rxg2+! at once) 16.Kh1
e3 17.f3 e2 18.Re1 Rxg2! and Black won, N.Kirchner-Joh.Holzer, Postbauer 2007.

398
 And now:
a) 14...Rg6 15.Qh4 Qd5? is given in all my books, intending 16.Nh3? e3 17.Nf4 Qxg2+! “with a
quick win” (Estrin); but White can disrupt this with 16.d4! exd3 (or 16...Bxd4 17.Rd1) 17.f3 dxc2
18.Nc3 and seems to be consolidating. In this line 15...Ke7! 16.Nc3 Rag8 is a better try, when White
will again look to defend after 17.d4!. Stockfish says “0.00”.
b) 14...e3 forces a draw: 15.f3 e2 16.Re1 Ng4 (or 16...Ne4) 17.Qxc6 Nf2+ ½-½ J.Grau Ribas-
W.Weissleder, corr. 1997, in view of the continuation 18.Kg1 Nh3+ 19.Kh1 Nf2+ etc.
c) 14...Bf8!? is one way to keep the game going. After 15.Qh4 h6 16.d3 (forced) 16...hxg5 17.Bxg5
Bg7 (or 17...Be7) 18.Nc3 Kc8, say, White has four extra pawns, Black an extra bishop. I’m not sure
which side I’d prefer to play. Perhaps either side.

A2: 8.0-0

Castling provides only a very temporary lull. Once Black builds the battery on the long diagonal,
White will be forced to fight once again.

8...Bb7

The most logical, if not the only move. Another option is 8...h6!? 9.Nf3 (or 9.Qf3 Qb7!? 10.Qxb7
Bxb7 11.Nf3 e4) 9...e4 10.Ne5 (or 10.Re1 Be7) 10...Qe6 11.d4 Bd6, when Black has the usual
activity for the pawn, V.Hanjs-Z.Csapo, Budapest 1998.

399
9.Qf3

White is virtually forced to enter the complications.

a) 9.f3? h6! (even stronger than 9...Bc5+ 10.Kh1 h6, which is also terrible for White) 10.Nh3 g5
and ...g5-g4 already looks crushing; e.g. 11.d3 g4 12.Nf2 gxf3 13.Re1 Rg8 14.Rxe5+ Kd7 15.g3 Bd6
and so on.
b) 9.Nf3?! is no good either as “Black’s attack develops unhindered” (Estrin); e.g. 9...0-0-0 (Black
has plenty of choice here: 9...Bc5, 9...Bd6, and 9...g5!? are all options too) 10.Nc3 (or if 10.d3 e4
11.Ne5 Qd5 12.d4 e3 13.Nf3 exf2+ 14.Rxf2 Ng4 15.Re2?, Myosotis1984-Qaurial, playchess.de
2001, then 15...Qh5 16.h3 Bxf3 17.gxf3 Rxd4! wins) 10...Bc5 11.d3 e4 12.Ne1 exd3 13.cxd3 Rhe8
14.Nf3, J.Del Arco Gimeno-F.Destruels Moreno, Pref 1992, and now 14...Ng4! again is very strong.

9...e4

Ulvestad’s 9...Qa6!? has hardly been tested, perhaps due to 10.Qf5 Bd6 11.d3.

10.Qb3

Swinging across to attack f7 again, and the inclusion of 0-0 and ...Bb7 on line A1 means that
10...Bc5?? now loses to 11.Qxf7+ Kd8 12.Qe6. No worries; Black can just castle. Note that 10.Re1
Be7 is also more than acceptable for Black.

10...0-0-0! 11.Qh3+

A necessary intermediary move to defend the kingside. The immediate 11.Nxf7? runs into 11...e3!
12.f3 e2 13.Re1 Bc5+ 14.Kh1 Rde8 and Black is winning; e.g. 15.d4 Bxd4 16.Qd3 Bf2 17.Rxe2 Ba6

400
18.Qf5+ Kb8 19.Ne5 Qd6, A.Salazar-J.Davila Liceaga, Villa Giardino 2002.

11...Kb8

12.Nxf7

It’s too late to back out now.

a) 12.d3? exd3 13.Nxf7 fails to the rook lift 13...Rd4! 14.Be3 (or 14.Nxh8 Rh4 15.Qf3 Qxc2
16.Qg3 Rg4) 14...Rg4 15.f3 Bc5 16.Bxc5 (equivalently, 16.fxg4 Re8! 17.Bxc5 Qxc5+ etc; or if
16.Ne5 Bxe3+ 17.Kh1 Qc5 18.Nxg4 then 18...dxc2 19.Nc3 Bc8) 16...Qxc5+ 17.Kh1 Re8 18.cxd3 (or
18.fxg4 Re3) 18...Rg6 19.Nc3 Qe7 and Black wins.
b) 12.Nc3?! can be answered by the computer move 12...Rg8! (rather than 12...Rd7, I.Kan-
A.Konstantinopolsky, USSR Ch., Moscow 1945, when 13.d3 exd3 14.Nf3 is more or less equal); the
point is simply to nullify the threat to f7 by removing the h8-rook from danger. For instance, after
13.d4 (or 13.Nxf7? Re8) 13...exd3 14.cxd3 Rxd3 15.Be3 Bc5 Black has regained the pawn with a
definite initiative.

12...Rd5!

401
 This time the rook is heading for h5. Black has won every game from this position:
a) 13.Nxh8? Rh5 “with a very strong attack” (Estrin), which Stockfish upgrades to “winning”; e.g.
14.Qc3 Bc5 15.Kh1 e3 16.f3 Ng4 (or 16...e2 17.Re1 Qd6 18.h3 Qg3 19.d4 Ng4 – Cramer) 17.dxe3
Nxh2 (17...Rxh2+ 18.Kg1 Rh1+ etc was quicker) 18.Kg1 Nxf1 19.Kxf1 Rh1+ 20.Kf2 Rxc1 21.b3
Rxc2+ 0-1 A.Kotlar-A.Veksler, Essen 1999.
b) 13.Nc3? Rh5 14.Qg3 (or 14.Qe3 Bc5 15.Qg3 Rf8 16.Ne5 Qe6 and the knight goes,
B.Ramanamurthy-J.Van der Kooij, corr. 1995) 14...Qe8! (trapping the knight again) 15.Nxh8 Bd6
16.Qxg7 Bxh2+ 17.Kh1 Bg3+ 18.Kg1 Rh1+ 0-1 V.Marincas-L.Csilcser, Felix Spa 2007.
c) 13.d4! exd3 14.Nxh8 is the only way to bail out, when 14...Rh5 (or if 14...Bd6 15.Nf7 Rh5 then
16.Nd8! and White survives) 15.Qf3 Bd6 (or 15...Qxc2 16.Qg3 Ne4 17.Qg4 Nf6 18.Qg3) 16.Qxc6
Bxh2+ 17.Kh1 Bg3+ 18.Kg1 Bh2+ is a draw.

B: 6.dxc6 bxc4

402
 Taking on c6 is a rather impractical variation for White. For a meagre investment Black has gained
the two bishops (including an unopposed light-squared bishop), easy development, and a cramping
pawn on c4, and can hope to pick up the c6-pawn anyway. Nevertheless, the engines say “equal”, so
it may not be objectively so bad.
The two main continuations are:
B1: 7.Qe2
B2: 7.Nc3
Note that 6.Qe2 bxc4 7.dxc6 and 6.Nc3 bxc4 7.dxc6 transpose to each of these lines. Instead:
a) 7.0-0? allows Black to attack almost automatically; e.g. 7...h6 8.Nf3 (or 8.Nh3 Qd5 9.Re1 Bd6
10.d4 cxd3 11.Qxd3 Qxc6 with a clear advantage, H.Bänsch-H.Krongraf, corr. 1993) 8...e4 9.Ne5 (if
9.Re1 then 9...Be6, or 9.Qe2 Bd6) 9...Bd6! (9...Qd4 10.Nd7 is less clear) 10.Nxc4?, Ch.Roberts-
C.McCarty Snead, London 2012, and now the classic sacrifice 10...Bxh2+! 11.Kxh2 Ng4+ 12.Kg3
(or 12.Kg1 Qh4) 12...h5 would have won.
b) 7.d3 (or 7.d4) 7...cxd3 removes the cramping c4-pawn at the cost of making d3 a target: 8.cxd3
(or 8.Qxd3 Qxd3 9.cxd3 Ba6) 8...Ba6 9.0-0 (or 9.Qb3 Qd5 10.Nc3 Qxb3 11.axb3 Bxd3 with an edge,
S.Lode-A.Rainfray, Romans 1999) 9...Qxd3 (9...Bc5 10.Re1 0-0 is also fine) 10.Qxd3 Bxd3 11.Re1
Bd6 12.Bf4, S.Reklaitis-V.Kazakovskiy, Palanga 2011, and now 12...Ng4 13.h3 0-0 is equal, since
14.Bg3 Nh6 15.Bxe5?? loses to 15...Bxe5 16.Rxe5 Bxb1 17.Rxb1 f6.

B1: 7.Qe2

Attempting to justify White’s play by immediately attacking c4 and e5. Black has two main replies:
B11: 7...Qd5
B12: 7...h6!?

403
 Even 7...Bd6 8.Qxc4 0-0 is possible, though the onus is then on Black to prove sufficient
compensation; e.g. after 9.Nc3 Rb8 10.a3 (or 10.0-0 Rb4 11.Qe2) 10...h6 11.Nge4 Be6 12.Qe2 Nxe4
13.Nxe4 f5, M.Bocangel Chavez-J.Perez Rodriguez, Lima 2016, and simply 14.Nxd6 cxd6 15.0-0.

B11: 7...Qd5

The simplest solution. The centralized queen defends both pawns while attacking c6 and g2 in
return.

8.0-0

Other moves are worse:

a) 8.Nc3?? Qxg2 9.Qxe5+ Be6, K.McEwan-B.Keogh, Scottish League 1996, and White is lost in
view of 10.Rf1 Bd6 11.Qe3 0-0 and ...Rfe8.
b) 8.Nf3 Bd6 9.Nc3 Qxc6 sees Black regain the pawn with advantage. L.Knutsson-J.Gunnarsson,
Reykjavik 2002, continued 10.h3 0-0 11.0-0 Re8 12.Re1 Ba6 13.Qe3? Bc5 14.Nxe5? Rxe5 15.Qxe5
Re8 16.Qxe8+ Nxe8 and White resigned.
c) 8.f3 Be7 (obviously 8...Bd6 9.Nc3 Qxc6 is fine too) 9.Nc3 Qxc6 10.Qxe5 0-0 11.0-0 (or if
11.Qb5, P.Bartron-D.Arganian, Redmond 2016, then 11...Qxb5 12.Nxb5 h6 13.Ne4 Nxe4 14.fxe4
Bb7 15.Nxc7 Bxe4 is good for Black; e.g. 16.Nxa8? Bxg2 17.Rg1? Bh4+ 18.Ke2 Re8+ 19.Kd1 Bf3
mate) 11...Bc5+ 12.Kh1 Re8 13.Qf4 Bd6 14.Qh4 Bf5 15.Nce4 Bxe4 16.fxe4 h6 and Black is
winning, M.Caravan-P.Melmuk, Brno 1987.
d) 8.f4 Bd6 (not 8...h6?, as in J.Mieses-H.Steiner, Hastings 1945/46, because of 9.Nf3 e4 10.Ne5
Be6 11.b3 and White is better, V.Sivuk-N.Getz, Fermo 2009) 9.0-0 (both 9.fxe5 Bg4 10.Nf3 Bxf3
11.gxf3 0-0-0 and 9.Nc3 Qxc6 10.fxe5 Bg4 11.Nf3 0-0-0 12.0-0 Rhe8 are also good for Black) 9...0-0
10.fxe5? Bxe5 11.Qf3? Bd4+ 12.Kh1 Qxg5 and Black won, P.Jazdzyk-Mo.Herman, Polanica Zdroj

404
2009.

8...Bd6

9.b3

Challenging the cramping c4-pawn straight away is probably best.

a) 9.Qf3 is well met by 9...e4! 10.Nxe4 Nxe4 11.Re1 0-0 12.Qxe4 (or 12.Rxe4 Bf5) 12...Qh5 13.g3
Bf5 14.Qh4, J.Garwell-N.Hoiberg, Haifa Olympiad 1976, and now 14...Qxh4! 15.gxh4 Rfe8 with a
clear advantage, since swapping rooks is terrible for White; e.g. 16.Rxe8+ Rxe8 17.Kg2 Re4 18.d3
cxd3 19.cxd3 Rg4+ 20.Kf3 Rxh4 and so on.
b) 9.Re1 Bg4 10.f3 Bf5 11.Nc3?! (but if 11.Ne4 Nxe4 12.fxe4 then 12...Qc5+ 13.Kh1 Be6 with
...Qxc6 to follow) 11...Qc5+ 12.Kh1 Bxc2 and Black was clearly better, P.Grott-P.Leisebein, corr.
1998.
c) 9.Nc3 Qxc6 is also good for Black, who again isn’t even a pawn down; e.g. 10.b3 (or 10.d4 cxd3
11.cxd3 Bb7 12.Nge4 Nxe4 13.Qxe4 f5 14.Qxc6+ Bxc6 with an edge, A.Borstnik-M.Skrinjar,
Slovenian League 2009) 10...Bb7 11.f3 Ba6 12.b4? (but if 12.bxc4 then 12...Bc5+ 13.Kh1 Bd4
14.Bb2 h6 15.Nge4 Nxe4 16.Qxe4 Qxe4 17.fxe4 Bxc4) 12...0-0 (or 12...h6) 13.Rb1 Bxb4! is another
terrible position for White, G.Mahia-A.Conde, Buenos Aires 2013, where the Argentine IM had to
rely on a swindle.

9...e4!?

An enterprising, if unnecessary response. Any of 9...Qxc6, 9...Ba6, or simply 9...cxb3 is fine for
Black; e.g. 9...cxb3 10.Nc3 Qxc6 11.axb3 0-0 12.Qb5 Bb7 13.Qxc6 Bxc6 14.Ba3 Rfd8, A.Sousa-
A.Antoniou, Moscow Olympiad 1994.

405
10.Nxe4 Qxe4 11.Qxe4+ Nxe4 12.Re1 f5 13.f3 Bc5+ 14.Kh1 Bf2 15.Re2

This was M.Manik-T.Olsarova, Brno 2011, where Black needed to find 15...Ba6! 16.fxe4 Bd4
17.exf5+ (or 17.c3 cxb3) 17...Kf7 18.c3 Rae8, exploiting the weak back rank. Then after 19.Re6
(forced) 19...Bc8 20.cxd4 Bxe6, the game would have been roughly equal, since 21.fxe6+?? Rxe6 is
fatal for White.

B12: 7...h6!?

More ambitious than 7...Qd5, in that Black offers up the e5-pawn as well.

8.Qxe5+

406
 White should probably decline in favour of note ‘b’.
a) 8.Ne4 Qd5 9.Nxf6+ (or if 9.Nbc3 Qxc6 10.0-0 Nxe4 11.Qxe4 Qxe4 12.Nxe4 f5 13.Nc3,
H.Sleeman-R.Walmisley, corr. 1998, then 13...Ba6 14.Re1 e4 and ...0-0-0) 9...gxf6 10.Qf3 Qxf3
11.gxf3 Bf5 12.d3 cxd3 13.cxd3 Bxd3 and Black is clearly better as the c6-pawn is a liability, G.Coy-
W.Korn, corr. 1942.
b) 8.Nf3 e4 (or immediately 8...Bd6!? 9.Nxe5 0-0 10.Nxc4 Re8 11.Ne3 Ng4 with a useful initiative
for the pawns) 9.Nc3 (here 9.Ne5 Qd4 10.Nxc4 Ba6 11.d3 0-0-0 12.Be3 exd3 leads to equality)
9...Bd6 (better than 9...Qe7 10.Nd4 Rb8 11.Ndb5 Qe5, as in G.Beresovsky-A.Scuderi, corr. 1992,
when 12.f4 Qc5 13.b4! Qxc6 14.Nxa7 Qa6 15.Nxc8 Bxb4 16.Rb1 Qxc8 17.Qxc4 Bxc3 18.Rxb8
Qxb8 19.Qc6+ Ke7 20.Qxc3 is somewhat in White’s favour) 10.Nxe4 (or 10.d4 cxd3 11.cxd3 Ba6
12.Nxe4 0-0 13.Nxd6 Qxd6 14.0-0 Bxd3) 10...0-0 11.Nxd6 (or 11.0-0 Ba6) 11...Qxd6 12.0-0 (or
12.d4 Qxc6 13.0-0 Bg4) 12...Ba6, when Black’s activity and light square clamp should be enough for
the pawn(s), as it’s difficult to see how White will develop successfully.

8...Be7

9.Nf3

Here 9.Ne4 0-0 10.0-0 is well met by 10...Ng4! 11.Qb5 (similarly 11.Qa5 f5 12.Nec3 Bd6 and so
on; or if 11.Qc3 Qd5 12.Qf3 f5 13.Ng3, M.Böhnisch-J.Reich, Ruhla 1957, then 13...Qxf3 14.gxf3
Ne5 is very good for Black) 11...f5! (stronger than Ulvestad’s 11...Qd4 12.Nbc3 Be6, when 13.d3
gives White the edge; or 11...Be6, P.Oldrati-R.Kurylo, corr. 2002, where 12.h3 would be unclear)
12.h3 (not 12.Nec3 Be6 or 12.Qxc4+ Kh8 13.Nec3 Bd6 14.h3 Qh4 with a very strong attack)
12...fxe4 13.hxg4 Qd4 14.Nc3 Bh4 15.Qd5+ Qxd5 16.Nxd5 Bxg4 and Black’s development is worth
more than White’s pawn.

407
9...0-0 10.0-0

An exchange of queens with 10.Qd4 Qxd4 11.Nxd4 Bc5 doesn’t help White at all.

10...Bg4

No one has yet tried 10...Bd6!, which forces the white queen to scurry home: 11.Qe2 (or 11.Qe1 –
anywhere else and 11...Bg4 is very strong, as we’ll see by transposition in the next note) 11...Re8
12.Qd1, when Black has definite compensation; e.g. 12...Bf5 (discouraging a d-pawn break) 13.b3
(not 13.Re1? Rxe1+ 14.Qxe1 Bxc2; while 13.Na3 Bxa3 14.bxa3 Qd6 and 13.Nc3 Ng4 14.h3 Ne5
15.Nxe5 Bxe5 both give Black good play) 13...Ng4 (13...Nh5!? looks promising too) 14.Bb2 Be4
15.g3 Bc5 16.bxc4 Rb8 17.Bc3 Nxf2 18.Rxf2 Bxf3 19.Qxf3 Re1+ 20.Kg2 Bxf2 21.Kxf2 Qe7 and
Black is certainly no worse.

11.Qf4?

Going for the c4-pawn as well is far too greedy. Enabling development with d2-d4 is an urgent
requirement.
a) 11.Qd4? Bd6 12.Ne5 (for 12.Qxc4 see the main line) 12...Bxe5 13.Qxe5 Re8 14.Qa5 Be2 15.Re1
Bd3!? 16.Re3 Ng4 is good for Black.
b) 11.Qc3? Bd6 12.d4 (or 12.Qxc4 again) 12...Nd5! 13.Qxc4 Bxf3 was H.Röder-G.Dietz, corr.
1971; in view of 14.gxf3 Qh4, Black is already winning.
c) 11.d4 cxd3 12.cxd3 was Holzschuh-G.Cramer, Bad Neumheim 1984, where 12...Bxf3 13.gxf3
Bd6 14.Qf5 Re8 (threatening ...Re5) 15.d4 (or 15.f4 Re6) 15...Nd5 appears to give Black full
compensation, if no more than that.
d) 11.Qe2 Re8 12.Qxc4? (12.d4 is critical, when 12...cxd3 13.Qxd3 Qxd3 14.cxd3 Bxf3 15.gxf3

408
Red8 is probably okay for Black; all the same, 10...Bd6 seems preferable) 12...Bd6 13.Nc3 Bxf3
14.gxf3 Re5 15.d4? (the immediate 15.f4 was necessary) 15...Rh5 16.f4 Ng4 17.Qd3 Nxh2 and Black
is winning, J.Holwell-H.Tiemann, corr. 1989; e.g. 18.Re1 f5 19.Kg2 Ng4 or 19.Qg3 Rg5!.

11...Bd6 12.Qxc4 Bxf3 13.gxf3 Rb8

Threatening ...Rb4, which White must prevent.

14.a3

Now:
a) 14...Nh5 (Ulvestad) doesn’t offer an advantage, as after 15.d3 the dark squares f4, g5 and h4 are
all covered. Black must resort to the deflection sacrifice 15...Rxb2!, when 16.Bxb2 Bf4 17.Kg2 Qg5+
18.Kh3 Qf5+ is only a draw.
b) 14...Rb5!? was very imaginative in K.Burger-Zweiburg, corr. 1963: 15.Qxb5 Nd5 16.Qc4? (after
16.d4! Qh4 17.Re1 Qxh2+ 18.Kf1 Bg3 19.fxg3 Qh1+ 20.Kf2 Black can only draw again) 16...Nf4
17.Kh1 Qh4! (more accurate than the game’s 17...Re8? 18.Rg1? Qh4, in which Pinski’s 18.d3 Qg5
19.Bxf4 Bxf4 20.Rg1 Qh4 21.Rg3 might have spoiled Black’s attack) 18.Rg1 (18.Qe4 f5! 19.Qc4+
Kh7 doesn’t help) 18...Re8 19.Qf1 (or 19.Qd4 g6 20.d3 Qh3 – Pinski) 19...Nd3 20.f4 Nxf2+ 21.Kg2
Bc5 and White resigned.
c) 14...Qc8! (Estrin) is the strongest continuation: 15.Kg2 (if 15.f4 Qg4+ 16.Kh1 Qf3+ 17.Kg1
Bxf4, Black is winning – Cramer; e.g. 18.Qd3 Qg4+ 19.Kh1 Qh5 20.h3 Rfd8 21.Qc3 Ne4) 15...Qf5
(Cramer) 16.d3? (missing the threat; the only defence was 16.Rg1, though after 16...Qg6+ 17.Kf1
Qh5 Black has all the chances) 16...Qg6+ 17.Kh1 Qh5 18.f4 Ng4 0-1 D.Shershakov-Rom.Zakharov,
Dombay 2014.

409
B2: 7.Nc3

A more sober approach than 7.Qe2.

7...h6

Seeking clarification before committing to any specific piece placement. Other moves allow White
the benefit of information; in particular, Qe2 may prove more effective. For example:
a) 7...Bc5 8.Qe2 0-0 9.0-0 Qd6 (or 9...Ba6 10.Qxe5 Bd6 11.Qa5) 10.Qxc4 Qxc6 11.d3 Bb7
12.Nce4 Bxf2+ (or 12...Nxe4 13.Qxe4) 13.Rxf2 Qxc4 14.Nxf6+ gxf6 15.dxc4 fxg5 16.Rf6 Be4
17.Bxg5 Rab8 18.b3 Bxc2 19.Re1 and White has a definite advantage despite the opposite-coloured
bishops, A.Streltsov-O.Kinberg, Netanya 2019.
b) 7...Bf5 8.Qe2 (not 8.0-0 h6 9.Nf3 e4 10.Ne5 Bd6 11.Nxc4? Bxh2+! 12.Kxh2 Ng4+ 13.Kg3 h5
and Black is probably winning) 8...Bxc2 9.Qxc4 (rather than 9.Qxe5+? Be7 10.0-0 0-0 11.Re1 Bd6,
C.Groenendal-J.Wijker, corr. 1986; or 9.Nb5? Be7 10.Qxc4 Bg6 11.Nf3? a6 12.Nc3 Bd3, when
White has huge problems, D.Scheglmann-W.Weiz, corr. 1989) 9...Bg6 10.0-0 and White keeps the
extra pawn, though the d3-square will be a permanent weakness.
c) 7...Bg4!? 8.f3 Bf5 9.Qe2 Bxc2 10.Qxc4 Bg6 improves on note ‘b’, as 11.0-0 is now well met by
11...Qd3, intending 12.Qxd3 Bxd3 13.Re1 Bc5+ 14.Kh1 Bd4; while 11.Nge4 Nxe4 12.Nxe4 Qh4+
13.g3 Qh3 14.Qe2?! (14.b3, Bb2 and 0-0-0 was a better plan) 14...Be7 15.d3 0-0 16.Bd2 Rad8 17.0-
0-0 Qe6 18.Kb1 Qxc6 gave Black a very promising position in J.Havumaki-E.Orak, Iasi 2013.

8.Nf3

The half retreat with 8.Nge4 leaves White struggling to equalize: 8...Nxe4 9.Nxe4 Qd5 10.Qf3
(defending with a threat; other moves are worse, e.g. 10.d3 cxd3 11.cxd3 Bb4+ or 10.Qe2 f5 11.Ng3

410
Qxg2 12.Qxe5+ Kd8) 10...Be6 (not 10...Qxc6?? 11.Nf6+ etc) 11.0-0 (here 11.b3 0-0-0 12.Nc3,
M.Belica-R.Müller, German League 2008, 12...Qxf3 13.gxf3 Bb4; and 11.Nc3 Qxf3 12.gxf3 Rb8
13.b3 Rb6 14.bxc4 Bxc4, H.Niese-J.Carlstedt, Bad Wiessee 2011; and 11.d3 cxd3 12.cxd3,
S.Dyrhaug-C.McCarty Snead, London 2012, 12...0-0-0 13.0-0 f5 14.Nc3 Qxf3 15.gxf3 Rxd3 all give
Black the advantage) 11...0-0-0! and now:

a) 12.Qe3?! Qxc6 13.Nc3 Bc5 14.Qxe5 Rd4 15.h3 Re8 16.Qh5? (avoiding the discovery lands
White in greater trouble) 16...g5! 17.Qxh6 g4 18.hxg4? (or 18.h4 g3) 18...Rxg4 19.Qh3 Reg8 0-1
K.Hanusek-M.Chmiel, Murzasichle 2011.
b) 12.Re1?! Qxc6 13.b3 f5 14.Nc3?! (but 14.Ng5 e4 15.Qh3 Bd7! 16.Nf7 Qf6 17.Nxd8 Qxa1
18.Nf7 Rg8 is still good for Black) 14...e4 15.Qh5 Bc5 16.bxc4?! (16.Na4 is preferable) 16...g6
17.Qe2?! (and here 17.Qh4 – Pinski) 17...Bxc4 18.d3 exd3 19.cxd3 Rxd3 20.Qc2 Ba6 (20...Bd4!
21.Bb2 Ba6 is even stronger) 21.Nd1? Bb7 22.Ne3 Rxe3 0-1 N.Bahram-J.Hector, Stockholm 1998, in
view of 23.fxe3 Bxe3+ etc.
c) 12.b3 is suggested by Pinski as best; e.g. 12...Qxc6 13.bxc4 f5 14.Ng5 e4 (or if 14...Qxf3
15.Nxf3 Bxc4 16.d3 e4 then 17.Ne5) 15.Qh3 Bxc4 16.Qxf5+ Qd7 (or 16...Rd7 17.Qxe4) 17.Qxd7+
Rxd7 18.Nxe4 Bxf1 19.Kxf1 and White should probably survive with two pawns for the exchange.

8...Bd6

411
9.0-0

Pinski’s one-move, natural improvement. The only reasonable course of action is to accept that
White has nothing, break with d2-d4, and get the pieces out. Instead:
a) 9.d4 cxd3 10.cxd3 0-0 11.0-0 transposes below.
b) 9.h3?! e4 (or just 9...0-0) 10.Qe2 0-0 11.Nd4, M.Kajan-S.Sigfusson, European Cup, Panormo
2001, and 11...Be5! 12.Ndb5 Re8 gives Black a very strong initiative; e.g. 13.Qxc4 (or 13.0-0 Nh7,
intending ...f7-f5) 13...a6 14.Na3 Be6 15.Qe2 Nd5 16.Nxd5 Qxd5 – and for just a single pawn (once
the c6-pawn drops).
c) 9.Qe2?! 0-0 (compare this with 8.Nf3 e4 9.Nc3 Bd6 in line B12; clearly there’s no need for 9...e4
here) 10.Qxc4 (similarly 10.0-0 Re8 11.Qxc4? e4 12.Ne1 Bxh2+! 13.Kxh2 Ng4+ 14.Kg3, Malu-
Kibitzer, playchess.com 2003, and now 14...h5! 15.Qb4 h4+ 16.Kh3 a5 17.Qa3 Ne3+ 18.Kh2 Nxf1+
wins – Pinski; or if 15.Rh1 then 15...Qd6+ 16.f4 g5 17.d4 h4+ etc) 10...Rb8 11.0-0? (11.a3 was
required) 11...Rb4 12.Qe2 e4 13.Ne1 Bxh2+! 14.Kh1 (or 14.Kxh2 Ng4+ 15.Kg3 Qd6+ 16.f4 exf3+
etc) 14...Bg4 15.f3 Bd6 16.fxg4 Nh5 17.Qe3 Qh4+ 18.Qh3 Ng3+ 0-1 M.Brokko Olde-J.Hector,
Jyvaskyla 1994, in view of 19.Kg1 Bc5+.

9...0-0 10.d4

10.Re1 Rb8 11.d4 would be okay too, rather than the greedy 11.Nxe5?! Bxe5 12.Rxe5 Ng4
13.Re4?! (13.Rh5 prevents the next move) 13...Qh4 14.h3 Qxf2+ 15.Kh1 Bf5 16.hxg4? (16.Re2 Rfe8
17.d4 was necessary) 16...Bxe4 17.Nxe4 Qh4+ 18.Kg1 f5 19.g3 Qh3 20.gxf5 Rxf5 21.d3 Rbf8 with a
decisive attack, J.Kramlová-L.Lisetskaya, corr. 2014.

10...cxd3 11.Qxd3?!

412
 Thinking about structure is more unwarranted avarice. 11.cxd3 would allow White to mobilize with
moves like Re1, Qc2 and Ne4. Stockfish says “0.00”. All the same, humans would probably prefer
Black.

11...a5 12.Re1 Ba6 13.Qd1 Re8 14.Be3 Re6 15.Qd2 Qe8 16.Rad1 Qxc6

Black has: regained the pawn, more space, more active pieces, the two bishops, therefore a clear
advantage, which was converted in L.Powell-J.Picchiottino, corr. 1998.

C: 6.Bf1!

The only good move for White. There are several possible replies. The question is which offers
Black the most compensation for the pawn.

6...Nxd5!?

This was Ulvestad’s second idea; it is supposed to be bad because of Pinkus’ counter-analysis and a
later suggestion by GM Reuben Fine (see 12.Qg7?! below). Interestingly, the fourth edition of ECO
Vol.C (2000) gives 6...Nxd5 once more as the main line, this time on the basis of a 1997 game by
GM Alexander Obukhov.
I think other moves are worse:
a) 6...h6 (Ulvestad’s first idea) is met by 7.Nxf7! (stronger than 7.Nf3 Qxd5 8.Nc3 Qe6) 7...Kxf7
8.dxc6 Bc5 9.Be2!, when Black has struggled to show compensation.
b) 6...Nd4 7.c3 transposes to the Fritz Variation (5...Nd4)

413
 where 7...Nxd5 8.cxd4! Qxg5 9.Bxb5+ Kd8 10.0-0 Bb7 (or 10...exd4 11.Bc6! Rb8 12.d3; while
10...Nf4, as in Gi.Goldsztejn-Ma.Carlsen, Chess24.com blitz 2020, is also met by 11.Bc6!, as the
world champion bantered at the time) 11.Qf3 Rb8 12.dxe5! Ne3 13.Qh3 Qxg2+ 14.Qxg2 Nxg2
15.d4! has been shown to be good for White. Sverre’s Johnsen’s untested 10...Qg6!? 11.dxe5 Rb8
may be the best try, though 12.Nc3 Nf4 13.Qf3 Bb7 14.Qg3 Nxg2 15.d4 Nh4 16.Be2 still seems in
White’s favour.
c) 6...Bg4? fails to 7.f3 (not 7.Be2? Bxe2 8.Qxe2 Qxd5 9.0-0 Nd4 10.Qd1 h6 11.Nh3 Bd6 and
Black had a very nice position, which he later lost in E.Paoli-O.Ulvestad, Reggio Emilia 1960/61 –
the inventor’s only game with his variation in the databases) 7...Nxd5 8.Nxf7 Kxf7 9.fxg4 Nf4
10.Nc3 Bc5, K.Honfi-P.Pitters, corr. 1962, and now 11.Ne4 (rather than 11.g3 e4!?) 11...Qd5 12.d3,
followed by g2-g3, is probably winning for White.
d) 6...Qxd5!? is worth considering. Then 7.Nc3 Qc5 8.Bxb5 is worse for Black than 6.Bxb5 Qxd5,
because the queen has first been deflected from its target on g2. All the same, with activity and rapid
development Black has decent practical chances, and it’s not such a simple matter for White to
consolidate. Ulvestad thought that it was “probably a good answer” as well.

414
7.Bxb5

Both 7.Qf3 Qd7 (not 7...f6? 8.Bxb5 Bb7 9.Ne6) 8.Bxb5 Bb7 and 7.Qh5 g6 8.Qf3 Qd7 9.Bxb5 Bb7
transpose to notes at move eight.
White punted 7.d4? and won after 7...h6? 8.Bxb5 Bb7 9.Nf3 e4 10.Ne5 Qf6 11.Qg4 in D.Ledesma
Claros-A.Campos Parejo, Gran Canaria (online blitz) 2021; but 7...Nxd4 8.c3 a6! 9.cxd4 Bb4+ is
good for Black.

7...Bb7

The first recorded game with the Ulvestad saw 7...Bd7? 8.d3 Bb4+ 9.c3 Be7 10.Nf3 Bd6 11.0-0 0-0
12.Re1 Re8 13.Nbd2, when Black had nothing for the pawn and duly lost in A.Pinkus-S.Bernstein,
Ventnor City 1941. Very satisfying for Pinkus no doubt, even if theoretically irrelevant. The Fried
Liverish sacrifice 8.Nxf7!? is tempting too: 8...Kxf7 9.Qf3+ Ke6 10.Bc4 (better than 10.Nc3?! Ncb4
11.Bc4 c6 – Pálkövi) 10...Ncb4 11.0-0, intending d2-d4.

415
 In this fundamental position, White has two main continuations:
C1: 8.d3
C2: 8.d4
Other moves have been tried or analysed:
a) 8.Bc4 Qxg5 9.Bxd5 0-0-0 10.d3 Qg6 11.Bf3 f5 (Ulvestad) is good for Black, and throwing in
10...Bb4+ may be even better.
b) 8.Bxc6+ Bxc6 9.Nf3? (an instructive mistake; 9.d3 is preferable, when 9...Qd7 10.0-0 0-0-0
gives Black quite enough play for the pawn, with plans of pushing on the kingside) 9...Nf4! (stronger
than Ulvestad’s 9...Bd6, which is fine for Black) 10.0-0 Nxg2! 11.d4 (forced, since 11.Kxg2? Qg5+
12.Kh1 Qg4 just wins) 11...Nh4 12.Nxe5 Bb7 leaves the white king wide open.
c) 8.Nf3 Bd6 and now:

416
 c1) 9.0-0 0-0 10.Nc3 (or 10.Bxc6 Bxc6 11.d3 f5) 10...Nd4 11.Nxd4 exd4 12.Nxd5 (not 12.Ne2?,
A.Chiari-M.Numanoglu, World Junior Ch., Kemer 2007, because of 12...Qh4 13.g3 Qh3 14.f3 Nf6
with powerful threats) 12...Bxh2+ 13.Kxh2 Qxd5 14.Qg4 f5 15.Qg3 Qxb5 is messily equal.
c2) 9.d4 exd4 (Ulvestad’s line 9...e4 10.Ne5 Bxe5 11.dxe5 0-0 12.Bxc6 Bxc6 13.Qd4 Qe7 14.0-0
looks better for White) 10.Nxd4 0-0 11.Nxc6 Qe8+ 12.Qe2 Bxc6 13.Bxc6 Qxc6 14.0-0 leaves White
a pawn up for seemingly not very much, but Stockfish considers Black to be perfectly okay; e.g.
14...Rad8 15.c3 (or 15.Qf3 Qxc2) 15...Rde8 (now Nc3 has been ruled out) 16.Qf3 Re6 17.g3 f5.
c3) 9.Nc3 a6! (here Ulvestad’s 9...Nxc3 seems too helpful, and 9...Nf4?!, M.Cerrato Torrijos-
J.Nunez Alonso, Padron 2008, can be met by 10.d4 exd4 11.Bxf4 dxc3 12.Bxd6 cxd6 13.Qe2+ Qe7
14.bxc3 with an edge) 10.Ba4 (now if 10.Bc4 then 10...Nf4, or 10.Bxc6 Bxc6 11.0-0 Nxc3 and 12...0-
0) 10...Nb6 (as we’ll see, targeting the light-squared bishop is a standard plan in this whole line)
11.Bxc6+ (or 11.Bb3 Nd4 12.Nxd4 exd4 13.Qe2+ Qe7 14.Qxe7+ Kxe7 15.Ne2 a5 16.a4 d3)
11...Bxc6 12.0-0 0-0 13.Re1 Re8 14.d3 f5 with full compensation due to the bishop pair and general
activity.
d) 8.Qf3

417
 8...Qd7 (not 8...f6?? 9.Ne6 Qd6 10.Qxd5 Qxd5 11.Nxc7+ Kd8 12.Nxd5 Nd4 13.Bc4 and wins,
P.Ofstad-B.Heggheim, Oslo 1978; nor 8...Qxg5?! 9.Qxd5 Qf6 10.Nc3 Be7 11.0-0 0-0 12.Re1 Rad8
13.Bxc6 Bxc6 14.Qxe5 Rfe8 15.Qxf6 Bxf6 16.Rf1, M.Novikov-U.Sevdimaliev, Bogoroditsk 2011, as
I don’t think Black has enough for two pawns) 9.Nc3 (or 9.0-0 h6 10.Ne4 0-0-0 11.d3 f5 12.Ng3 g6 –
Ulvestad) 9...Ndb4! (not 9...Nxc3?! 10.bxc3 f6 11.Ne4 0-0-0 12.Rb1 f5 13.Ng5 and White is clearly
better, A.Rodriguez Vila-D.Izquierdo, Montevideo 2015; note that 13...Bc5 can be met by 14.Qxc6!
Qxc6 15.Bxc6 Bxc6 16.Nf7, winning the exchange) 10.Ba4 (or 10.Qd1 h6; not 10.0-0?! Nxc2 11.Ba4
N2d4!) 10...h6 (rather than 10...f6?! 11.a3 Nxc2+ 12.Bxc2 fxg5, which messes up Black’s kingside,
and 11...fxg5? 12.axb4 is just bad, S.Sorbe-S.Peray, Amiens 2001) 11.Nge4 (not now 11.a3?! Nxc2+
12.Bxc2 hxg5) 11...0-0-0 12.a3 Nxc2+ (or 12...Nd5!?) 13.Bxc2 Nd4 14.Qd1 f5 with a lot of play; e.g.
15.Bb1 (or 15.d3 fxe4 16.dxe4 Qe6 17.0-0 Ba6; and not 15.Ng3? Bxg2) 15...Be7! 16.0-0 h5 17.d3 g5
and so on.
e) 8.Qh5!? g6 (I prefer this to 8...Qd7 9.d4 g6 10.Qe2 Be7 11.dxe5, even if Black did go on to win
in rooksac-remyrey, ChessWorld.net 2019) 9.Qf3 Qd7 (all as in note ‘d’ above; the question is how
the extra ...g7-g6 changes things) 10.Nc3 Ndb4! (not 10...Rd8? 11.Nxd5 Qxd5 12.Qf6 Rg8 13.c4 Qd7
14.0-0 Bg7 15.Qf3 h6 16.Nxf7 and White won, rooksac-mizuzul, ChessWorld.net 2019) 11.0-0!?
(this is the difference: 11...Nxc2? 12.Ba4 N2d4 now fails to 13.Qf6; otherwise 11.Ba4 h6 is much the
same as before) 11...h6 leads to complications, but I think Black is okay.

418
 Here are some supporting variations from analysis:
e1) 12.Nge4?! 0-0-0 13.Qd1 f5 14.a3 Nd5 15.Nxd5 Qxd5 16.Nc3 Qe6 with superb play for the
pawn.
e2) 12.a3 hxg5 13.axb4 0-0-0 14.Rxa7 f5! 15.Ba6 Bxa6 16.Rxa6 e4 17.Qe2 Qh7 18.Ra8+ (18.h3
Nd4 19.Qc4 Nf3+! is much the same) 18...Kd7 19.Rxd8+ Kxd8 20.h3 Nd4 21.Qe3 Nf3+ 22.gxf3 (not
22.Kh1?? Qxh3+! 23.gxh3 Rxh3+ 24.Kg2 Rh2+ 25.Kg3 Bd6+ and mates) 22...f4! 23.Qd4+ Kc8,
followed by ...Qxh3, so White must hurry to take the draw.
e3) 12.d4 hxg5 13.d5 Nxd5 14.Rd1 0-0-0 15.Rxd5 (not 15.Nxd5?! Qe6) 15...e4 (or 15...g4!?)
16.Qxe4 Bd6 17.Bxg5 Rde8 18.Rxd6 (not 18.Qd3?! Bxh2+ 19.Kf1 Bd6 20.Kg1 Qg4) 18...Qxd6
19.Qf4 Qxf4 20.Bxf4 f6 is roughly equal.
e4) 12.Qf6 Rg8 (or even 12...hxg5!? 13.Qxh8 0-0-0) 13.Nge4 Be7 14.Ba4 Rf8 (obviously the
queen can’t be taken yet) 15.a3 0-0-0 16.Nc5!? (if 16.Qf3 then 16...Nd5 17.d3 f5 or 17.Qd3 Qe6)
16...Qg4 17.h3 Qh5 18.Qg7 Bxc5 19.axb4 Nd4 20.bxc5 Bxg2! 21.Kxg2 Qf3+ 22.Kh2 Qf4+ with a
draw.
e5) 12.Nce4

419
 12...Qf5! (the only move in fact) 13.a3 hxg5 14.axb4 Qxf3 (14...0-0-0!? 15.c3 Nxb4 is also
possible) 15.gxf3 a6 16.Bxc6 (or 16.Ba4 Bxb4 17.d4 exd4 18.Bxg5 Be7) 16...Bxc6 17.d4 exd4
18.Bxg5 (or 18.Nf6+ Kd8 19.Bxg5 Kc8) 18...f5 19.Bf6 Rh5 20.Ng3 Rh7 is unclear. Stockfish says
equal, but “0.13” equal, so Black perhaps shouldn’t be too complacent.

C1: 8.d3

Rather than force the pace by opening the position with d2-d4 at once, White plays in a more
restrained fashion, hoping to consolidate the extra pawn.

8...Be7

420
 Developing the bishop while renewing the attack on the white knight. Black cannot afford to play
casually:
a) 8...Bc5 9.0-0 0-0 10.Nc3 Nd4 11.Bc4 Nb6?? 12.Qh5 and White won, E.Paoli-K.Ojanen,
Trencianske Teplice 1949.
b) 8...Qd7 9.Nc3 0-0-0 10.0-0 h6 11.Nf3 Qf5, P.Ptacnik-M.Muron, Czech League 2011, and now
12.Nxd5 Rxd5 13.Bc4 leaves White a safe pawn up.
c) 8...h6 9.Nf3 Bd6 (Ulvestad) can be compared with 8.Nf3 Bd6 above. Here Black’s ...h7-h6 looks
like a wasted tempo. After 10.Nc3 Nxc3 11.bxc3 0-0 12.0-0 Qf6 13.Nd2, as in R.Denda-W.Ries,
Koblenz 1998, White is at least a little better.

9.Qh5!?

Critical, if not entirely consistent. Having maintained the knight on g5, White uses it to prompt a
weakening in Black’s kingside dark squares. Others:
a) 9.Qg4 supports the knight again and has ideas of Qc4. Black can interfere with both by playing
9...Nf4! 10.Bxf4 (or 10.Qf3 Qd6 11.Ne4 Qh6) 10...exf4 11.Nf3 (or 11.h4 h6) 11...0-0 12.0-0 (not
12.Qxf4?! Nb4) 12...Nd4 13.Nxd4 Qxd4 14.Nc3 f5 15.Qh3 Rf6 with compensation.
b) 9.Nf3 f6!? (slow-playing the position: Black secures the centre and will develop behind it;
instead, 9...Qd6 10.0-0 0-0-0 11.Nbd2 Nf4? 12.Nc4 Qf6 13.Bxc6 Bxc6 14.Ncxe5 won quickly in
L.Dek-N.Cheng, Vlissingen 2007; while 9...0-0 10.0-0 Nd4 11.Bc4 Nxf3+ 12.Qxf3 Rb8 13.Qd1 Qd6
14.Kh1 Qg6 15.f3 Bc5 16.Re1 Bd6 17.Qd2 h6 18.Qf2 gave White a slight edge, though activity
enabled Black to hold in tsmenace-remyrey, ChessWorld.net 2019)

I’ve had the opportunity to test this three times:

b1) 10.c4?! Ndb4 11.d4? (11.c5 was necessary) 11...e4 12.d5 exf3 13.0-0 fxg2 14.Re1 0-0 15.dxc6

421
Qxd1 16.Rxd1 Nxc6 and Black was clearly better, tripoduk-jatait47, Chess.com 2019.
b2) 10.Nc3 a6 11.Ba4 (if 11.Nxd5 axb5 12.Nxe7 Nxe7 or 11.Bc4 Nb6 12.Be6 Qd6 13.Bh3 g6 14.0-
0 f5, Black has reasonable compensation) 11...Nb6 12.Bb3 Qd7 13.Be3 0-0-0 14.Qe2 Rhe8 15.0-0-0
Bb4 16.Ne4 Nd4 17.Nxd4 exd4 18.Bd2 a5! was Reprimand-tsmenace, ChessWorld.net 2019. Here
White overstepped the time limit; play might have continued 19.a3 a4 20.Ba2 Bxd2+ 21.Qxd2 Bxe4
22.dxe4 Rxe4 and Black should be okay, having regained the pawn.
b3) 10.0-0 Qd7 11.d4! (possibly critical; 11.Be3 a6 12.Ba4 0-0-0 13.Nbd2 g5 14.Nc4 Qe6 15.c3 h5
16.Qb3 h4 17.Rad1 g4 18.Nfd2 Rdg8 19.g3 f5 led to massive complications in tripoduk-jatait47,
Chess.com 2020, which went on 20.Bxc6 Bxc6 21.Ba7 Bb7 22.Na5 Ba8 23.Nac4 Qh6 24.Qb8+ Kd7
25.Nxe5+ Ke6 26.Rde1 Qh7 and Black eventually triumphed) 11...a6 12.Be2 (more testing than
12.Ba4 Nb6 or 12.Bxc6 Qxc6 13.dxe5 0-0-0) 12...exd4 13.Nxd4 Nf4! 14.Bf3 (or 14.Bxf4 Nxd4
15.Bd3 0-0-0 16.Nd2 f5 with ideas of ...g7-g5 and ...f5-f4) 14...Qxd4 15.Qxd4 Nxd4 16.Bxb7 Nfe2+
17.Kh1 Rb8 18.Bxa6 Nxc2 19.Bxe2 Nxa1 20.Bd3 (or 20.Bd1 Bd6) 20...Rd8 21.Be4 0-0 (intending
...f6-f5) 22.g4 Rd4 23.Nc3 Bb4 and again I think Black should be okay.

9...g6 10.Qh6 Qd7

Improving on the only previous game to get so far, which saw 10...Bf8? 11.Qh4 Be7 12.Qc4 0-0
13.Bxc6 Bxg5 14.Bxb7 and White was clearly better, J.Roos-C.Caminade, French Ch., Vitrolles
1981.

11.0-0

11.Nc3 a6 is similar and could indeed transpose after 12.Ba4 f6 13.Nge4 0-0-0 14.0-0.

11...a6!

422
 This little move proves very useful in these positions. Generally, an exchange on c6 is desirable for
Black, whose light-squared bishop would be left unopposed, so White is likely to retreat to a4, but the
bishop isn’t entirely safe there either.

12.Ba4

Black should be okay after:

a) 12.Bxc6 Qxc6 13.Nf3 (or 13.Qh3 f6) 13...f6 14.Nc3 (or 14.Qg7 Rf8) 14...0-0-0.
b) 12.Bc4 Nd4 13.Nc3 0-0-0 14.Bxd5 (not 14.Nxf7?? Nxc3 15.bxc3? Ne2+ 16.Kh1 Bxg2+! and
mates) 14...Bxd5 15.Nxd5 Qxd5 16.c3 Nc2 17.Rb1 Qxa2.

12...f6 13.Ne4 0-0-0 14.Nbc3 Nb6

White’s c3-knight is “superfluous” (Dvoretsky) in that only one can occupy e4, so the black knight
moves away while also hitting the a4-bishop (thanks to the earlier ...a7-a6).

15.Bxc6

White opts to give up the bishop after all since retreating would see Black’s initiative increase; e.g.
15.Bb3 Nd4 16.Be3 Qc6 17.f3 Nf5 18.Qh3 Kb8 19.Bf2 h5.

15...Qxc6

Creating a dangerous-looking queen and bishop battery, so White logically erects a bulwark on the
long diagonal.

16.f3 Rdg8

423
 Planning to harass the queen with ...g6-g5 and ...Rg6, while preparing a kingside pawn storm, for
which White’s f-pawn now provides a hook.
This was rooksac-tsmenace, ChessWorld.net 2019. Stockfish says “0.00” persistently; White has an
extra pawn, Black an obvious initiative. I think most human players would be happy as Black.
Here 17.Be3 might be White’s best, intending 17...g5 18.Rab1 Rg6 19.Qh3+ Kb8 20.b4 Bc8
21.Qg3 and b4-b5. Instead, 17...Kb8! is a useful insertion, vacating the c8-h3 diagonal, when Black
looks to have good play; e.g. 18.Rab1 (or 18.Kh1 g5 19.g4!? Rg6 20.Qh5 Nd7 21.Rab1 Bf8, and not
20.Qh3? h5) 18...Bf8 19.Qh3 (or 19.Qh4 f5) 19...Bc8 20.Qh4 g5 21.Qf2 f5 22.Nd2 h5 and so on.
My game saw 17.Kh1 g5 18.Be3 Rg6 19.Qh3+ (not 19.Qh5?! f5! 20.Nxg5 Bxg5 21.Bxg5 Rhg8)
19...Kb8 20.a4!? (planning to mobilize the rook via the fourth rank) 20...h5 21.a5 Bc8! (to gain time
on the queen, since after 21...Nd7?! 22.b4 or 21...Nd5 22.Qf5 White should be at least okay) 22.Qg3
Nd7 23.Ra4 h4 24.Qf2 f5 (the pawns roll forward, and either 25.Nc5? g4 or 25.Nd2? f4 wins for
Black) 25.Rc4 Qe6 26.Nc5 (or 26.Ba7+ Ka8 27.Rxc7 fxe4 28.dxe4 g4 29.Qe2 Bb7 30.Bg1 Kb8 and
White’s counter-attack is over) 26...Bxc5 27.Bxc5 Bb7 28.Ba7+ Ka8 29.Rxc7 g4

30.Rxb7 (removing the bishop is the only way to secure the long diagonal) 30...Kxb7 31.fxg4 h3
32.Be3 (not 32.gxf5? hxg2+ 33.Qxg2 Rxg2 34.fxe6 Rhxh2 mate) 32...Qc6 33.Qf3 hxg2+ 34.Qxg2
Qxg2+ 35.Kxg2 Rxg4+ 36.Kh1 f4 37.Bg1 (White has survived for the moment but lacks any
counterplay, so Black can organize a breakthrough at leisure) 37...Nf6 38.Rf3 Kc6 39.Bf2 Kd7
40.Nb1 Ke6 41.Nd2 Kf5 42.Nc4 Rg6 43.Bg1 Ng4 (provoking a weakness which the knight will run
back round to attack) 44.h3 Nf6 45.Bc5 Nh7 46.Nd6+ Ke6 47.Ne4 Ng5 48.Nxg5+ Rxg5 49.Bf2 Kf5
50.b4 Rhg8 51.Bc5 Rg2 and White resigned.

C2: 8.d4

424
 The supposed refutation of Ulvestad’s 6...Nxd5.

8...exd4!

Other moves are worse for Black. In particular:

a) 8...Qf6 (Obukhov) 9.0-0 0-0-0?! (for 9...exd4 see 9...Qf6 below) 10.Qh5! g6 11.Qh3+ Kb8
12.Bxc6 Bxc6 13.dxe5 wins another pawn.
b) 8...Be7 9.Nf3 exd4 (or 9...e4 10.Ne5 Qd6 11.0-0 0-0 12.Bxc6 Bxc6 13.c4, A.L.Souza-P.Palozi,
Curitiba 2013) 10.Nxd4 shows Black’s move order to have been inaccurate. After 10...Qd6 11.0-0 0-0
12.Nf5 Qe5 13.Nxe7+ Ndxe7 14.Nc3 White is just a pawn ahead, B.Ahlander-O.Kristinsson,
Icelandic League 2014.

9.0-0

“etc” wrote Tartakower (in 1955). Not 9.Qxd4? because of 9...Qe7+! 10.Qe4 Qxe4+ 11.Nxe4 0-0-0
12.Bxc6 Bxc6 (Ulvestad) and Black is clearly better; e.g. 13.Ng5? f6 14.Nf7 Bb4+, followed by
...Rhe8+ and wins.

9...Be7

As in line C1, I think this simple developing move is best. “Now White undoubtedly has the better
chances,” wrote Estrin (in 1983).
GM Obukhov played 9...Qf6? and achieved a good position after 10.Qf3?! Nde7? (here 10...Qxf3
11.Nxf3 0-0-0 may be okay) 11.c3?? (Stockfish bangs out 11.Qb3! and wins) 11...h6! 12.Qxf6 gxf6
13.Ne4 0-0-0 14.Nxf6 d3 15.Rd1 Ne5 in S.Nadyrhanov-A.Obukhov, Krasnodar 1997 (the line given
in ECO).

425
 A much earlier game saw 10.Re1+ Be7 11.c4 (Obukhov himself suggests 11.Ne4 Qg6 12.Nc5 0-0-
0 13.Nxb7 Kxb7 14.Qf3, which also looks good for White) 11...Ndb4? (or 11...Nb6 12.Nf3) 12.a3
Na6 13.b4 Nab8 14.Ne4 Qg6 15.Nc5 Bc8 16.Nd3, which led to a win for White in Na.Kellner-
J.Stagpoole, corr. 1963. Pálkövi’s simple 10.Nf3 seems promising too.

10.Qh5

Pinkus gave 10.Nf3 0-0 11.Bxc6 Bxc6 12.Nxd4 as good for White, to which Estrin added 12...Bb7
13.c3. White does seem to be a clear pawn up with a solid position, so it’s surprising to discover that
the engines all prefer Black. After 13...Qd7, for instance, it’s quite difficult for White to mobilize
satisfactorily: 14.Nd2? Nf4 15.N2f3 Nxg2!; or 14.Na3? Bxa3 15.bxa3 Nxc3; or 14.Be3? Nxe3
15.fxe3 Bg5; or 14.Qd3 Rfe8 15.Bd2 g6 16.Na3 Nb6, threatening ...c7-c5. One version of Stockfish
suggested 14.Nf3 Qc6 15.Ne5 Qe6 16.Re1 Rfe8 17.Na3 Bf6 18.Nf3 Nxc3 19.bxc3 Bxf3 20.Rxe6
Bxd1 21.Rxe8+ Rxe8 22.Be3 Ba4, but even here the two bishops give Black an edge in the ending.
Other moves, too, fail to consolidate; e.g. 13.c4 (or 13.Nf5 Bf6) 13.c4 Nb6 14.Nf5 (on 14.Qg4,
F.Kreideweiss-B.Knorr, corr. 1989, Stockfish retorts with 14...h5! 15.Qe2 Re8) 14...Bf6 15.Qc2?!
Re8 16.Ne3 Be4 17.Qb3 Rb8 18.Nd2? Bd3 19.Re1 Nxc4 and Black won, J.Roscher-B.Knorr, corr.
1989.

10...g6

The only move. Not 10...Bxg5? 11.Bxg5 Nf6 12.Re1+ Kf8 13.Bxf6 Qxf6 14.Nd2 Rd8 15.Ne4 Qg6
16.Qc5+ Kg8 17.Rad1 h6 18.Rd3 and White won quickly, M.Wills-S.Isaksson, corr. 1965.

11.Qh6 Qd7

And not 11...Qd6?! 12.Qg7 Qf6 (since 12...0-0-0? 13.Nxf7 wins) 13.Qxf6 Bxf6 14.Re1+ Kf8

426
15.Ne4 (Pinski), when Black’s position has notable flaws.

12.Re1

Here 12.Qg7 (?) 12...0-0-0 13.Qxf7 (Fine) is given as clearly better for White in all my old books.
Stockfish is less than impressed: 13...Rhf8 (13...Ne3!? and 13...a6!? are possible too) 14.Qe6 (or
14.Qxh7? Nf4! 15.Bxf4 Rxf4 16.Nf3 Rxf3! 17.gxf3 Qf5 18.Bd3 Qxf3 19.Nd2 Ne5! with a big
advantage) 14...Ne3! 15.Bxc6 (or 15.Qxd7+?! Rxd7 16.Nxh7 Rf5!) 15...Bxc6 16.Bxe3 dxe3 17.Qxe3
Qf5 (or 17...Rf5!? 18.Ne4 Re5) 18.Qe6+ Kb7 19.Qxf5 Rxf5 20.Nf3 (not 20.Nh3? g5) 20...Bxf3
21.gxf3 Rxf3 and Black has more than enough for the pawn.

12...a6!

Once again this little pawn move, asking the question of the b5-bishop, is very useful for Black.
The point is seen after the bishop retreats. Note here that 12...Qf5? would be met by 13.Bd3!; while
after 12...0-0-0? 13.Nxf7 Bb4 14.Nxd8 Bxe1, as in Dar.Pena-C.A.Fonseca, Pamplona 2012, simply
15.Nxb7 Kxb7 16.Bd2 is very good for White.

13.Ba4

White has numerous other options with the bishop, each of which can lead to a flurry of tactics. It’s
hard to know where to begin when analysing such positions from scratch, even with engine
assistance. What follows is an outline of my own investigations, which may be useful as a starting
framework for your own:
a) 13.Bc4 0-0-0 14.Nxf7 (or 14.Bd2 Kb8, intending 15.Nxf7? Bf8!) 14...Rhe8! 15.Nxd8 Bg5! (a
key resource) 16.Rxe8 Bxh6 17.Re1 (or 17.Rh8 Bg7) 17...Bxc1 (or 17...Nxd8 18.Bxh6 g5!) 18.Nxb7
Bxb2 is good for Black.

427
 b) 13.Bd3 0-0-0 14.Nxf7 (or 14.Nd2 Nf4) 14...Rhe8! 15.Nxd8 Bg5 16.Rxe8 Bxh6 17.Re1 Bxc1 etc.
c) 13.Bf1!? f6 (note that 13...0-0-0? 14.Nxf7 Rhe8 no longer works because ...Rxe1 isn’t check)
14.Ne6 (or 14.Nf3 0-0-0) 14...Ne5 15.Nxd4 0-0-0 with compensation; e.g. 16.Nc3 Ng4 17.Qd2 (or
17.Qh4 Bd6) 17...Qd6 18.Nf3 Nxc3 19.bxc3 Qxd2 20.Bxd2 Bc5 21.Nd4 Rxd4! 22.cxd4 Bxd4
23.Reb1 Nxf2 24.c3 Nh3+ with a draw.
d) 13.Bxc6 Bxc6 leaves Black with an unopposed light-squared bishop and hence what should be
decent play for the pawn. For example, 14.Bd2 (instead, 14.Nf3 0-0-0 15.Ne5 doesn’t remove the
bishop in view of 15...Qe6 16.Nxc6?? Qxe1 mate; while if 14.Na3 0-0-0 15.Nxf7 then 15...Bxa3!
16.Nxd8 Bf8 17.Qh4 Qxd8 is fine) 14...Rf8! and now:

d1) 15.Nf3 Bb7 16.Nxd4 (or 16.Ne5 Qf5) 16...0-0-0 17.Nb3 Nb6 18.Kh1 (or 18.f3 g5) 18...Qg4
(18...Bf6 plays for more) 19.f3 Bxf3 20.gxf3 Qxf3+ is a draw.
d2) 15.Nxh7 Rg8 16.c4!? (now if 16.Ng5 0-0-0 17.Nxf7 then 17...Rdf8 18.Ne5 Qf5 19.f3 Bf6 with
a strong initiative; or 16.Bg5 0-0-0 17.Bxe7 Rge8! 18.Nd2 Rxe7 19.Nf3 Nb4 is equal) 16...dxc3
17.Nxc3 0-0-0 18.Bg5 (or 18.Nxd5 Qxd5 19.Qh3+ Qf5 20.Rxe7 Qxh3 21.gxh3 Rxd2 22.Rxf7 Rgd8)
18...Rge8 19.Nxd5 Bxg5 20.Qxg5 Rxe1+ 21.Rxe1 Qxd5 22.Qxd5 Rxd5 with sufficient play even in
the endgame; e.g. 23.a3 Rd2 24.h4 Rxb2 25.Re7 Bd5 26.Nf6 Ba2 27.g4 Rb6 28.Ne4 Kd8 29.Re5
Bb1.
d3) 15.Na3 0-0-0 16.Nc4 Bb7 with various ideas of ...Bb4, ...Nb4, ...Qf5, ...Rfe8, and ...Bf8.

428
 For example: 17.Rac1 (the engines’ preference; other lines include 17.Na5 Ba8, or 17.Re2 Bb4
18.Nf3 Qb5, or 17.Nf3 Nb4 18.Nce5 Qf5, or 17.a3 Qf5 18.Nf3 Bf6 19.Nce5 Bxe5 20.Nxe5 f6 21.Nd3
g5) 17...Ne3!? (or 17...Qf5) 18.Bxe3 (the point is 18.fxe3 Qd5, and if 19.Qh3+ Kb8 20.Nf3 Qxc4
21.exd4 then 21...Bxf3 22.Qxf3 Qxd4+ 23.Be3 Qc4 is roughly equal; or 18.Nxe3 dxe3 19.Bxe3 f6
20.Nf3 Rfe8) 18...dxe3 19.Nxe3 Kb8 and the two bishops are quite enough for the pawn; e.g. 20.Nf3
(or 20.Rcd1 Qb5) 20...Bc5 21.Rcd1 Bxe3!? (a concrete way to equalize) 22.Qxf8 (or 22.fxe3 Qf5; not
22.Qxe3?? Qxd1 23.Rxd1 Rxd1+ 24.Ne1 Rfd8 and ...R8d6-e6) 22...Qxd1 23.Rxd1 Bxf2+ 24.Kxf2
Rxf8 25.Ne5 Kc8 with a probable draw.
As they say, practical tests are required.

13...Qf5

Now this move is possible. Instead, 13...0-0-0!? is not a blunder, but 14.Nxf7 Bb4 15.Nxd8 Bxe1
16.Nxb7 Kxb7 17.Nd2 Re8 18.Qh3 Bxd2 19.Bxd2 (not 19.Qxd7?? Re1 mate) 19...Qxh3 20.gxh3
should be somewhat better for White, who now has two bishops vs. two knights, even if the extra h-
pawn is fairly irrelevant.

429
14.Nf3

14.Bxc6+ Bxc6 15.Nf3 0-0-0 transposes.

White can bring the queen’s knight to f3 too: 14.Nd2!? 0-0-0 15.Ndf3 Bb4 (Black should also be
fine after 15...Bf8 16.Qh4 Bg7 or 16.Qh3 Qxh3 17.Nxh3 Bb4, if not 17...Bg7?! 18.Nf4!) 16.Bxc6 (or
16.Rd1 Bf8 17.Qh3 Qxh3 18.Nxh3 Bg7 now that 19.Nf4 can be met by 19...Ne5) 16...Bxc6 17.Nxd4
Qg4 18.Nxc6 Bxe1 19.Nxd8 Bxf2+ 20.Kxf2 Qd4+ 21.Ke2 (if 21.Kg3 then 21...Qe5+ 22.Kh3 Qf5+,
or 21.Kf3 Rxd8) 21...Re8+ 22.Nde6 (or 22.Nge6 fxe6 23.Bg5 Qxb2) 22...f6 23.Qh3 fxg5 24.Bxg5
Qe4+ 25.Kf1 Rxe6 26.Kg1 Kb7 (breaking the pin) 27.c4 Nb6 28.Rc1 Nxc4 29.Qc3 Rc6 30.Qb4+
Ka7 31.h3 a5 32.Qc3 Kb7 33.Kh2 Qe5+ 34.Qxe5 Nxe5 35.Rxc6 Kxc6 36.b3 Nd3 and the endgame
was drawn, Hypotenuse-tsmenace, ChessWorld.net 2020.
The main alternative is to send the queen in with 14.Qg7, which has at least had one test. Black
needs to defend the f-pawn, so 14...Rf8 and then:

430
 a) 15.Qxh7 0-0-0 16.Qh3 escapes with bounty at the cost of a serious deficit in development; after
16...Bxg5 17.Qxf5+ gxf5 18.Bxg5 f6, I think White will do well to equalize.
b) 15.Nxh7 0-0-0 16.Nxf8 picks up the exchange too, though the white royalty is less happy about
that: 16...Bxf8 17.Qh8 (or 17.Qh7 Ne5 18.Qh4 f6, intending ...Bc5) 17...Bd6 18.Re8 (or 18.Qh4 Nf4
19.f3 Nxg2! 20.Kxg2 Ne5 21.Nd2 g5 22.Qh5 d3) 18...Rxe8 19.Qxe8+ Nd8 20.Nd2 Nf6 21.Qe1 Ng4
22.Nf1 (not 22.h3? Bh2+ 23.Kh1 Ne3! 24.fxe3? Qxh3 and wins) 22...d3! 23.Bb3 (not 23.cxd3? Qd5
24.f3 Qd4+) 23...Nxh2 24.Nxh2 Bxh2+ 25.Kxh2 Qh5+ 26.Kg1 Bxg2! 27.Kxg2 Qg4+ with perpetual
check.
c) 15.Qxd4 0-0-0 16.Qe4 Nd4!? (16...Rg8!? isn’t so silly either) 17.Qxd4 (if 17.Qxf5+ gxf5 18.c3
Bxg5 19.Bxg5 then 19...Nf3+! 20.gxf3 Rg8 equalizes) 17...Nb6 18.Qf4 (or 18.Qe3 Bc5) 18...Nxa4
19.Qxa4 Bxg5 20.Nc3 (not 20.Bxg5? Qxg5) 20...Bf6 21.Be3 Rfe8 and Black should be okay with
control of the light squares.
d) 15.Re5!?

431
 was tried in remyrey-tsmenace, ChessWorld.net 2019, which continued 15...Qd7 16.Nxf7 (if
16.Nxh7 0-0-0 17.Nxf8 Bxf8 18.Qh7 then 18...Ne3! 19.Bxe3 Qg4 20.f3 dxe3! 21.Nc3 Qb4 22.Bxc6
Bxc6 23.Qxf7 Qxb2 24.Qe6+ Rd7 25.Rb1 Qxc3 and White should take the draw) 16...Ne3! (not yet
16...Rxf7? 17.Qg8+ Rf8 18.Qxd5) 17.fxe3 Rxf7 18.Qg8+ Rf8 19.Qxh7 (note that 19.Qd5? 0-0-0 now
lands White in trouble due to the displaced f2-pawn; instead, 19.Qe6 Qxe6 20.Rxe6 Rf6 21.Rxf6
Bxf6 22.e4 0-0-0 23.Bh6 Ne5 24.Nd2 manages to keep the pawn, but Black is very active and
24...Ng4 25.Bf4 Ne3 is one way forward) 19...0-0-0 20.Rxe7 (or 20.Qxe7 Qxe7 21.Rxe7 dxe3)
20...Qxe7 21.Qxe7 Nxe7 22.e4! (despite the exchange for two pawns Black has to be careful since the
white bishops are quite strong) 22...Nc6 (not 22...Bxe4? 23.Bg5 Rf7 24.Nd2 Bd5 25.Re1 etc) 23.Bg5
Rde8 24.Nd2 Re5 (24...Re6 may be more accurate) 25.Bh6 Rf7 26.Bb3 Rh7 27.Bf4 Re8 28.Bc4 Nb4
29.Rc1 Bxe4 30.Nxe4 Rxe4 31.Bg3 Rd7 32.a3 d3! 33.cxd3 Nxd3 34.Bxa6+ Kd8 35.Bxd3 Rxd3
36.Bxc7+ Kd7 37.Ba5 g5 38.Bc3 g4 (setting a lock on the position such that, despite the material
advantage, White is unable to make progress) 39.Rc2 Kc6 40.Kf2 ½-½.

432
14...0-0-0 15.Bxc6

The only serious try; and this time it doesn’t (necessarily) surrender the two bishops, since Nxd4
and Nxc6 will get a bishop back if desired. Whereas after something like 15.Nbd2 Nf4 16.Bxc6 Bxc6
17.Rxe7 Qg4, White can easily end up in difficulties.

15...Bxc6 16.Nxd4 Qf6 17.Nxc6

The alternative is 17.c3 Rhe8 18.Na3 and then:

a) 18...Bxa3? leads to a very forcing line: 19.Bg5 Qd6 20.Nxc6 Qxc6 21.Bxd8 Bxb2 22.Qh3+ f5

433
23.Rxe8 Qxe8 24.Rd1 Nxc3 25.Qxh7 Qc6 26.Re1 Ne2+ 27.Kf1 Nf4 28.Qxc7+ Qxc7 29.Bxc7 Kxc7
30.g3 Nd5 31.Re6, where White’s rook and pawn(s) are far superior to Black’s bishop and knight.
b) 18...Bb7!? 19.Nc4 Bf8 20.Qh3+ Kb8 21.Be3 Nxe3 22.Rxe3 Bc5 was Honeybunch-tsmenace,
ChessWorld.net 2019. White has consolidated the extra pawn at the cost of conceding the two
bishops. Had White not defaulted here, the game might have concluded 23.Rd1 (if 23.Rae1 then
23...Rxe3 24.fxe3 Bd5, or 23.Rxe8 Rxe8 24.Qd3 Qg5) 23...Re6 (threatening ...Bxd4) 24.Rxe6 fxe6
25.Na5 (if 25.Qe3 then 25...Rd5, or 25.Qd3 Ka8) 25...Bxd4 26.cxd4 Rxd4 27.Rxd4 Qxd4 28.Qb3
Qb6 29.Qxb6 cxb6 30.Nxb7 Kxb7 with a draw.
c) 18...Bf8 produces an endgame where Black’s activity and bishops should be enough to hold:

19.Bg5 (not 19.Qh3+?? Bd7 or 19.Qd2? Nf4) 19...Bxh6 20.Bxf6 Rxe1+ 21.Rxe1 Re8, when
AndyAndyO-tsmenace, ChessWorld.net 2019, went on 22.Nac2 Rxe1+ 23.Nxe1 Bb7 24.Bh8 Nb6
25.Ne2 Be4 26.Bd4 Kb7 27.b3 Bb1 28.a4 (or 28.a3 Bf8 29.a4 c5) 28...Nd7 29.f4 c5 30.Be3 Ba2
31.Nc1 c4! 32.b4 (not 32.Nxa2?? cxb3) 32...Bb3 33.a5 (and not 33.Nxb3?! cxb3 34.Nd3 Kc6)
33...Bd1 34.Nf3 Bxf3 35.gxf3 Nf6 36.Kf2 Nd5 37.Ne2 f5 38.Bd2 ½-½.

17...Qxc6

434
 This time White has the extra pawn and has eliminated the two bishops as well, but there is still the
small issue of development. I’ve defended this position twice without much trouble:
a) 18.Na3 Rhe8 (threatening ...Bxa3; not yet 18...Bxa3? 19.Qh3+) 19.Bg5 (obviously not
19.Qxh7?? Bxa3; and 19.Qh3+ f5 20.Bd2 Nf6 21.Bg5 Bc5 doesn’t get White anywhere) 19...f6
20.Qh3+ (or 20.Bd2 f5) 20...Qd7 21.Qxd7+ Kxd7 (21...Rxd7? 22.Bd2 leaves the e7-bishop pinned)
22.Bd2 Bxa3 23.bxa3 Rxe1+ 24.Bxe1 (or 24.Rxe1 Rb8) 24...Re8 25.Rd1 Kc6 26.Kf1 Re4 27.Rb1
Kd6 28.Bd2 c5 29.Rb7 Rc4 0-1 Hullplayer50-tsmenace, ChessWorld.net 2019, as White again
defaulted. Clearly Black is fine in this endgame since White’s extra a-pawn is meaningless. It’s even
a “wrong pawn for the bishop” should things get that far.
b) 18.Bd2 Qxc2 19.Nc3 Nxc3 (19...Bc5?! 20.Nxd5 Rxd5 21.Be3 Qxb2 22.Rab1 Qf6 23.Bxc5 Rxc5
24.Qh3+ Rf5 25.Rb4, as in remyrey-rooksac, ChessWorld.net 2019, is rather more difficult for Black)
20.Bxc3 (or 20.Qh3+ Kb8 21.bxc3 Qxd2 22.Rxe7 Qd1+! 23.Re1 Qd3 and the black king should be
safe enough on the a-file) 20...Bc5 (the counter-attack at f2 saves both bishop and rook) 21.Qf4 Rhe8
22.Qf6 Bb6 23.Rxe8 Rxe8 24.Rf1 Re2 25.h3 Qf5!? 26.Qxf5+ gxf5 (Black’s central control means
that the structural wreckage cannot be exploited) 27.g3 a5 28.a4 Kd7 29.Kg2 Ke6 30.Kf3 Re4 31.Ra1
c5 32.b3 Kd5 ½-½ rooksac-tsmenace, ChessWorld.net 2019.
c) 18.Nc3 Nxc3 19.bxc3 Qxc3 20.Be3 may be the best practical try, since the open files towards the
black king could have some significance. On the other hand, the a- and c-files are both closed, and
White has no disruptive knight charge or pawn break available, so conducting a defence with the
heavy pieces is a bit easier. Black might play ...f7-f5 and ...Bf6, followed by ...Rd6, or ...Rd5-b5, or
...Bg7. For example: 20...f5 (Stockfish goes for 20...Rd5 and says “0.00”) 21.Rab1 (or 21.Qf4 Qc6)
21...Bf6 22.Ba7 Rd5 23.a4 Kd7. The engines offer no constructive advice on how White might make
progress, so I think Black should be okay.

435
 Chapter Ten
Jaenisch Gambit

1.e4 e5 2.Nf3 Nc6 3.Bb5 f5

Black sees 3.Bb5 – the Ruy Lopez, the Opening of Champions – and pushes the f-pawn at it
contemptuously. And this is only borderline disreputable. Searching the databases by Elo rating
brings up an array of elite grandmasters on the Black side: Aronian, Carlsen, Ivanchuk, Nisipeanu,
Radjabov, Sokolov, Zvjaginsev, et al. However, proper players mostly play proper lines such as
4.Nc3 fxe4 5.Nxe4 Nf6 and 4.d3 fxe4 5.dxe4 Nf6 6.0-0 Bc5. I won’t be bothering with that sort of
thing. For anyone who wants to, there is Junior Tay’s book The Schliemann Defence: Move By Move.
Yes, it’s the same opening.
The main line with 4.Nc3 (played in over 55% of games) is examined in the next chapter. In this
one we’ll look at all the rest – of which 4.d3 is the most important and most frequently seen.
A: 4.exf5
B: 4.Qe2
C: 4.Bxc6
D: 4.d4
E: 4.d3
Castling is definitely inferior: 4.0-0? fxe4 5.Bxc6 (or 5.Ne1 Nf6 6.d3 Nd4) 5...dxc6! (much
stronger here than 5...bxc6 as in line C) 6.Nxe5 Qd4 7.Qh5+ (or 7.Ng4 h5 8.Ne3 Be6 and ...0-0-0)
7...g6 8.c3 (not 8.Nxg6? hxg6 as the black queen defends the rook) 8...Qd5 9.Qg5 Bh6 (or 9...Be7
and 10...Bf5) 10.Qg3 Ne7 and White is worse; e.g. 11.Re1 Nf5 12.Qg4 Ne3!, or 11.Ng4 Bg7 12.d4
exd3 13.Bh6 Bxg4, while after 11.d4? Nf5 12.Bxh6 Nxg3 13.fxg3 Be6 14.Bg7 Rg8 15.Bf6 c5 Black

436
won easily in Kr.Andreev-G.Filev, Teteven 2014.

A: 4.exf5

The Gambit Accepted, which can lead to very interesting positions. Black’s initial response is to
attack the knight, gaining time and prompting White to give up the light-squared bishop.

4...e4

5.Qe2

Pinning the pawn to the king temporarily; Black’s equivalent will obstruct the f8-bishop
temporarily. Instead:
a) 5.Bxc6 bxc6 6.Qe2 Qe7 transposes below. Not 6.Nd4? Qf6! and White is already worse; but
6.Ne5!? is a reasonable option. Black should reply 6...Qg5! 7.0-0 (or 7.g4 Qe7!) 7...Qxf5 8.d4,
T.Warakomski-M.Olszewski, Brody Ilzeckie 2003, and 8...exd3, intending ...Bd6, ...Ne7, ...0-0 with
equality.
b) 5.Ng1 is met well too by 5...Qg5! (5...Nf6 6.d3! is okay for White – I&K) 6.g4 (or if 6.Ne2 Qxg2
7.Ng3?, N.Talov-V.Grechihin, Kazan 2008, then 7...Bc5! 8.Qe2 Nf6) 6...h5 7.d3 Qxg4 8.Qxg4 hxg4
9.dxe4, as in K.Zavilin-Al.Ivanov, St. Petersburg 1992, when 9...Nb4! 10.Na3 (not 10.Ba4?! b5!
11.Bb3 Bb7 12.Nc3 Nf6; or if 10.Kd1 Nf6 11.Nc3 then 11...a6 12.Be2 d5 13.a3 d4 14.axb4 dxc3
15.Bd3 Bd7 16.Ne2 0-0-0 – Tay) 10...a6 11.Be2 b5 12.Bxg4 Bb7 13.Bf3 0-0-0 gives Black excellent
play for the pawns.

5...Qe7

437
6.Bxc6

The only serious way to save the knight; 6.Ng1? Nd4 is obviously terrible.

6...bxc6!

As well as being consistent with line C (and note ‘a’ above), I think this capture is best. Black gets
an extra pawn in the centre, plus the chance to play ...Ba6. The pawn deficit can be ignored for the
time being.

7.Nd4 Nf6

One of my earliest games with 3...f5 finished 7...Qe5 8.Nf3 Qe7 9.Nd4 Qe5 10.Nf3 Qe7 ½-½
R.Forey-J.Tait, Derbys League 1989. Then I went and studied it.
Note that 7...c5 is premature in view of 8.Nb5 d5 9.d3 c6 (or 9...a6 10.Bg5!) 10.dxe4! cxb5 11.Nc3
Nf6 (or 11...Bb7 12.Bg5) 12.e5, as after 10...c6? in the next note.

8.0-0

Safeguarding the king before attacking the central pawn with d2-d3 or f2-f3 or both. It is too soon
yet for either: 8.d3?! exd3 9.cxd3 c5 10.Nf3 Qxe2+ 11.Kxe2 d6 and 8.f3? c5! 9.Nb5 exf3 10.Nxc7+
Kd8 11.Qxe7+ Bxe7 12.Nxa8 fxg2 13.Rg1 Bb7 are good for Black.
Similarly, 8.Nc3?! allows Black to take the initiative with 8...c5 and then:

438
 a) 9.Nb3 a5! (stronger than 9...d5, as in J.Szily-D.Bronstein, Budapest-Moscow 1949, where 10.0-0
is “0.00” according to Stockfish) 10.d4? (but 10.a4! d5 11.0-0 Ba6 12.d3 c4 is still good for Black,
S.Doyle-T.Mecke, corr. 2003) 10...exd3 11.Qxe7+ Bxe7 12.cxd3 a4 13.Nd2 d5 14.0-0 Bxf5 15.Nf3
0-0 16.Re1 Rfe8 17.d4 cxd4 18.Nxd4 Bd7, followed by ...c7-c5 with a big advantage, cnile-tsmenace,
ChessWorld.net 2020 (in this game 9...a5 was actually 10...a5, having arisen via 3...a6 4.Ba4 f5!?
5.exf5 e4 and so on).
b) 9.Ndb5 d5 10.d3 (here 10.g4 c6 11.g5 Ng8 12.d3 a6 13.Nd6+!? Qxd6 14.dxe4 d4 15.e5? Qb8
16.Qh5+ Kd8 led to a win for Black in M.Lambert-J.Tait, Notts Championship 1995; Stockfish
prefers 10...h6!, intending 11.d3 a6! again) 10...a6! (the natural 10...c6? is much weaker since
11.dxe4! cxb5 12.e5 is fine for White, especially after 12...Nd7?? as in M.Lambert-J.Tait, Notts
League 1991) 11.dxe4 axb5 12.e5 Bxf5! 13.exf6 Qxe2+ 14.Kxe2 c6 (see how useful that pawn is)
15.fxg7 Bxg7 16.Kd1 0-0 with massive, probably decisive, compensation for the pawn.

439
8...Qc5!

A suggestion from S&S, after which “Black is ready to work on the weakness in the white camp.”
It’s surprising it’s not been played more – just a handful of games since 1983.
The standard 8...Qe5 (not 8...c5?! 9.Nb5 d5 10.f3 c6 11.N5c3 Bxf5 12.fxe4 Bxe4, A.Neumann-
R.Spielmann, Barmen 1905, when 13.Re1 c4 14.Nxe4 dxe4 15.Na3 should give White an edge)
8...Qe5 9.Nf3 Qxf5 (not now 9...Qe7? 10.Re1) 10.d3 Ba6 is merely okay; e.g. 11.Re1 (or 11.c4 d5)
11...0-0-0 12.dxe4 Bxe2 13.exf5 Bxf3 14.gxf3 Re8 15.Be3 Kb7 and White’s extra tripled f-pawn is
meaningless, D.Lopes-W.Bravo, corr. 2012.

9.c3

Everything else is worse: 9.Nf3? Be7, or 9.Qe3?! Bd6! (threatening ...Bxh2+), or 9.Nb3?! Qxc2 –
all S&S; e.g. 10.Nc3 (or 10.Re1 a5) 10...Qd3! 11.Re1 Qxe2 12.Rxe2 d5 13.d3 Ba6 with a big
advantage, J. Raymond-tsmenace, ChessWorld.net 2020.

440
 In this critical position Black has tried three moves. I’m not sure which is best:
a) 9...Bd6 10.f3 (or 10.d3 Ba6 – S&S, though 11.b4 Qb6, T.Støre-R.Bergquist, corr. 1998, and
12.b5! keeps things balanced) 10...Qe5 11.g3 Bc5 12.Qe3? (12.fxe4 d5 is just unclear) 12...d5 13.d3
0-0 14.Nd2 Bxf5 15.N2b3, V.Malischew-J.Tait, corr. 1992, when 15...Bb6 16.fxe4 Bh3 17.Rf4 dxe4
18.dxe4 Rae8 would have been very strong.
b) 9...Qb6 (threatening ...Ba6 and making way for ...c6-c5) 10.Nc2?! (10.Nb3 Ba6 11.c4 d5 12.d3
0-0-0 13.Nc3 is preferable and unclear) 10...Ba6 11.c4 d5 12.b3 0-0-0 13.Nc3 d4?! (here 13...c5
improves) 14.Na4 Qa5 15.d3 Qxf5, V.Malischew-H.D.Vötter, corr. 1995, and at this point 16.dxe4
Qxe4 (or 16...Nxe4 17.Qd3) 17.Qxe4 Nxe4 18.Ba3 looks better for White with the a6-bishop shut out
of the game.
c) 9...a5 (threatening ...Ba6 again) 10.Re1 Ba6 11.d3 Be7 12.Qd1! Bxd3 13.f3 Bxb1 14.Rxb1 exf3
15.Qxf3 0-0, when Black has safely regained the pawn and castled. Ares777-tsmenace,
ChessWorld.net 2018, continued 16.Bg5 Rab8 17.Kh1 a4 18.b4 axb3 19.axb3 Rbe8 20.Qd3 Ng4
21.Bxe7 Rxe7 22.Qg3 Nf6 23.Rxe7 Qxe7 24.Re1 Qd6 25.Qxd6 cxd6 26.h3 d5 27.Re3 c5 28.Nf3 Rb8
29.Nd2 Ra8 and Black was more active in the endgame, but White defended stoutly and made the
draw.

B: 4.Qe2

441
 This Worrallish move originates from the early days of the opening. White plays to win a pawn and
usually can, but Black gets more than ample compensation.

4...fxe4 5.Qxe4

The alternative is 5.Bxc6 bxc6!? (for consistency with line C; here 5...dxc6 is more usual, when
6.Qxe4 Bd6 7.d3 Nf6 8.Qh4 is roughly equal, while 7.Nxe5?! Nf6 8.Qe2 0-0 transposes below)
6.Qxe4 (instead, 6.Nxe5 Nf6 7.Nc3 Qe7 or 7.d4 exd3 8.Nxd3+ Be7 9.0-0 0-0 is equal, or perhaps
7...a5!?) 6...Bd6 7.0-0 (after 7.Nxe5 Qe7 8.d4 Ba6 Black has excellent compensation, Joh.Hall-
Ph.Ferguson, Canada 1992; if 7.Qc4 a5 8.d3 Ba6 9.Qg4?!, as in B.Kharchenko-N.Yaremko, Kiev
2004, then 9...Nf6! 10.Qxg7 Rg8 11.Qh6 e4 12.dxe4 Rg6 gives Black strong play; not 7.d4?! Nf6
8.Qe2 e4 9.Bg5?! 0-0 10.Ne5 Qe8 and Black is clearly better, A.Hussein-J.Tait, Derbys League 1992)
7...Nf6 8.Qc4 (or 8.Qe2 0-0 9.Nxe5 a5) 8...a5 9.Re1 (or 9.d3 Ba6 10.Qa4 0-0 – Pruijssers) 9...Ba6
10.Qh4 0-0 11.d3 was M.Chlost-K.Nemcova, German League 2011. Here Stockfish suggests
11...Qe8!?, intending 12.d4 (or 12.Nc3 Nd5) 12...e4 13.Nbd2 (or 13.Nc3 Bb4) 13...e3! 14.fxe3 (or
14.Rxe3 Qg6 15.c4 Rae8) 14...Nd5 15.a3 Qg6, again with strong play for the pawn.

5...Nf6 6.Qe2 Bd6!

442
 The key move, preventing Qxe5+ and planning to castle quickly. Winning the pawn will cost White
the two bishops and cede the initiative.

7.Bxc6?!

White should probably think again.

a) 7.d4 e4 8.c3 (not yet 8.Ng5?! Qe7 9.c3 h6 10.Nh3 g5!, A.Konstantinopolsky-V.Zak, Leningrad
1959) 8...0-0 9.Ng5, R.Yankovsky-M.Khachiyan, Los Angeles 2012, when 9...Kh8! 10.0-0 Qe8
11.Re1 Qg6 is fine for Black, since 12.Nxe4?! is answered by 12...Nxe4 13.Bd3 Bxh2+! 14.Kxh2 d5
15.Bxe4 dxe4 16.Nd2 Bg4 17.Qe3 Rae8 “and Black can already think about a kingside attack with
moves like ...Ne7-f5 or ...Re6/f6-h6. Maybe even a move like ...Bf3 is possible” (Pruijssers).
b) 7.0-0 0-0 8.Nc3 (not 8.b3?? e4! 9.Ng5 Nd4 and the bishop is soon trapped: 10.Qc4+ Ne6
11.Nxe6 dxe6 12.Qe2 c6 13.Bc4 b5 and wins) 8...a6! 9.Bxc6 (or 9.Ba4 b5 10.Bb3+ Kh8 11.d3 Nd4
12.Nxd4 exd4 13.Ne4 Qe8 “with the idea to take on e4 at the right moment and play ...Bb7 and ...c6-
c5” – Pruijssers) 9...dxc6 10.Nxe5 (not 10.d4?! exd4 11.Qc4+ Kh8 12.Qxd4 Bg4 or 12.Nxd4? Qe8
13.f4 c5 14.Nf3 b5 and White was soon in big trouble, J.Mikoska-S.Firla, corr. 2002) 10...Re8 11.f4
Bxe5 (simpler than 11...Ng4 12.d4 c5 as in G.Isea Fernández-H.Buczinski, corr. 2020) 12.fxe5 Qd4+
(Pruijssers) and Black regains the pawn with equality.

7...dxc6!

This capture makes most sense with the bishop already on d6, though 7...bxc6!? looks acceptable
too; e.g. 8.Nxe5 0-0 9.0-0 a5 and Stockfish says “0.00”.

443
8.Nxe5 0-0 9.d4 Re8

There is no need for 9...Kh8?! 10.0-0 c5 11.c3 cxd4 12.cxd4, J.Löwenthal-E.Falkbeer, Birmingham
1858, where White is better since 12...c5 13.dxc5! Bxc5 doesn’t achieve very much.

After 9...Re8 GM Sabino Brunello writes: “One does not have to be Mikhail Tal to see that Black
has excellent compensation”. Black will follow with ...c6-c5 and likely ...c7-c5 as well. For example:
a) 10.f4?! c5 11.0-0 (not 11.dxc5?? Bxe5 12.fxe5 Qd4) 11...cxd4 12.Qc4+ Be6 13.Qxd4 b6!
(improving on J.Neill-G.Bellingham, corr. 1892, where Black took on e5 at once), and if 14.Kh1
Bxe5! 15.Qxd8 Raxd8 16.fxe5 Bc4 17.Re1? then 17...Rxe5 just wins due to the back rank mate.

444
 b) 10.Be3 c5! (not 10...Bxe5 11.dxe5 Rxe5 with equality, V.Kozlov-Mi.Tseitlin, Moscow 1976)
11.c3 (for 11.0-0 see note ‘d’) 11...cxd4 12.cxd4 c5! and White is struggling.
c) 10.Nd2 c5 11.0-0 cxd4 12.Ndf3, M.Garcia Carbo-M.Munoz Muriel, Ponferrada 1992, can be
answered by 12...Bxe5! 13.Nxe5 Qd5 14.Bf4 Bg4 15.f3 Bf5 with advantage; the immediate threat is
...d4-d3 and ...Qd4+.
d) 10.0-0 c5! (again 10...Bxe5 11.dxe5 Qd4 is only equal) 11.c3 (if 11.Be3 then 11...cxd4 12.Bxd4
c5 13.Bc3 b5 with a strong initiative – Brunello; or similarly 11.Bf4 cxd4 12.Qc4+ Be6 13.Qxd4
Nh5! 14.Be3 c5 15.Qc3 b5 – Pruijssers) 11...cxd4 12.cxd4 c5 13.Bf4 Bb8! 14.Qc4+ Qd5 15.Qxd5+
(not 15.Qc1? cxd4, which is why the bishop retreated all the way) 15...Nxd5 16.Bg3 cxd4, when “the
bishop pair and active pieces give Black very nice play.” (Pruijssers); e.g. 17.Re1 (or 17.Nf3 d3)
17...Nb4 18.Nf3 (or 18.Na3 Be6) 18...Rxe1+ 19.Nxe1 Bxg3 20.hxg3 Bf5 with a slight pull in the
endgame.

C: 4.Bxc6 bxc6!?

Other authors who mention 4...bxc6 at all do so disparagingly. S&S said it “would be pointless
here”. Flear said that “4...bxc6?! is probably not good.” Sokolov wrote: “Taking with 4...bxc6?! is not
good for Black.” Personally I think 4...bxc6 is neither not good nor pointless. It leads to both
acceptable and interesting play for Black.
Far more interesting than 4...dxc6 5.Nc3, when White can grind at no risk of losing; e.g. 5...Nf6
6.Qe2 fxe4 7.Nxe4 Bg4 8.h3 Bxf3 9.Qxf3 Nxe4 10.Qxe4 Qd5 11.d3, G.Milos-Ma.Carlsen, Caxias do
Sul (rapid) 2014 and ½-½, 30. I’m sure Magnus enjoyed that game very much.
The only thing Black has to hope for is 5.Nxe5 Qd4! 6.Qh5+?? g6 7.Nxg6 hxg6 and wins, as the
queen defends the rook in the corner – hoping, in other words, that White doesn’t play 6.Nf3 Qxe4+
7.Qe2, as in T.B.Turner-J.Tait, Derbys League 1990, which I almost lost by trying too hard to win.

445
That was the very last time I took with the d-pawn.

5.Nxe5

The most common reply. White has tried six other moves too, most of which transpose elsewhere:
a) 5.exf5 e4 is 5.Bxc6 bxc6 in line A.
b) 5.Qe2 fxe4 is 5.Bxc6 bxc6 in line B.
c) 5.d4 is supposedly best, when 5...fxe4 is line D1. Alternatively, 5...exd4!? is also playable; e.g.
6.Nxd4 Qh4 (not yet 6...Qe7?! 7.0-0 fxe4 8.Nc3 Bb7, A.Colovic-K.Shirazi, Sautron 2006, when
9.Qh5+! g6 10.Qa5 is annoying) 7.0-0 (now if 7.Nc3 fxe4 8.g3 then 8...Qe7 9.0-0 Nf6, or 7.Nxf5
Qxe4+) 7...fxe4 8.Nf5 Qf6 9.Ng3 d5 (or 9...Qg6) 10.c4 Bb7 and ...0-0-0 looks fine for Black.
Probably 6.e5! is critical, when White has a decent reversed Falkbeer.
d) 5.d3 fxe4 (5...Nf6 is 5.Bxc6 bxc6 in line E1) 6.dxe4 Nf6 is a poor 4.d3 line for White, who has
given: up the bishop pair and Black an extra central pawn; e.g. 7.0-0 (for 7.Nxe5?! see line E1 again)
7...d6 8.Nc3 Be7 9.h3 (if 9.Ne1 0-0 10.f3, A.Romero Holmes-M.Gimeno Oteo, Saragossa 1994, then
even 10...d5 becomes possible) 9...0-0 10.Bg5 Qe8 11.Bxf6 Bxf6, followed by ...Qg6, ...Be6, ...Rf7,
...Raf8 and so on, as in Ant88888888-tsmenace, ChessWorld.net 2020.
e) 5.Nc3 fxe4 is 5.Bxc6 bxc6 at the beginning of Chapter Eleven.
f) 5.0-0 can lead to independent play: 5...fxe4 (the Philidoresque 5...d6!? is possible too: 6.d4 fxe4
7.Nxe5!? dxe5? 8.Qh5+ Ke7 9.Bg5+ Nf6 10.dxe5 Qd5 11.f4 exf3 12.exf6+ gxf6 13.Bxf6+! Kxf6
14.Rxf3+ led to a crushing loss for me in K.Pinkas-J.Tait, corr. 1998, but 7...Ne7 8.Ng4, D.Gomez
Anadon-C.Ferron Garcia, Mataro 2004, and 8...Ng6! should be okay; from the other side, 6.exf5 and
7.d4 might be White’s best try) 6.Nxe5 and then:

f1) 6...Nf6 7.d3 (here 7.d4! is line D1; not 7.Re1 d5 8.Nxc6?? Qd6 9.Nd4 Ng4 10.g3 Qf6 and wins,

446
or 8.d3?!, as in J.Tait-W.Van Vugt, corr. 2003, because of 8...Bd6!, when my intended 9.dxe4? loses
to 9...Bxe5 10.exd5 0-0 11.Rxe5 Ng4 – apparently) 7...Bd6! 8.Nc4 (or 8.Ng4 0-0) 8...Bc5 9.dxe4 0-0
10.Be3 Nxe4 11.Nbd2 Bxe3 12.Nxe3 Nc5 13.Nb3 Ne6 14.Nd4 Nf4 15.Ne2 d6 16.Nxf4 Rxf4 and
after that mutual knight tour Black is fine, 66Kolya99-tsmenace, ChessWorld.net 2018.
f2) 6...Qh4!? is another option; e.g. 7.d3 (for 7.d4! see line D1 again; or if 7.Qe2 Nf6 8.Nc3 then
8...d5! 9.d4 a5, and not 9.Nxc6? Bd6 10.f4 Ba6! 11.Qxa6? Ng4 with mate in eight) 7...Bd6 8.f4 Nf6,
when 9.Qe1 (or 9.Nc3 Bc5+! 10.d4 Bb6) 9...Qxe1 10.Rxe1 0-0 11.dxe4 Bxe5 (or 11...Bc5+ 12.Be3
Nxe4) 12.fxe5 Ng4 13.Nd2 Nxe5 led to a draw in J.Lacasa Diaz-L.Comas Fabrego, Sanxenxo 2004.

5...Qe7

Hitting the knight and the pawn behind it. There’s no need to fear the queen check on h5, which is
just as well because every other fifth move is bad for Black.

6.d4

The natural response, considering that ...d7-d6 will leave the c6-pawn loose.
a) 6.Nf3?! is no easy draw now: 6...Qxe4+ (or 6...fxe4 7.0-0 d5 8.Re1 Bg4) 7.Qe2 Qxe2+ 8.Kxe2
d6 and Black is better with the two bishops.
b) 6.f4 is less useful than d2-d4 but may be okay; e.g. 6...d6 (or 6...Nf6!? 7.Nc3 fxe4 8.0-0, as in
D.Skocibusic-D.Udovicic, Valpovo rapid 2011) 7.Qh5+ (not 7.Nxc6?! Qxe4+ 8.Qe2 Qxe2+ 9.Kxe2
Bb7 10.Nd4 Bxg2 11.Rg1 Bb7 and Black is better, especially after 12.Nxf5 0-0-0 13.Nxg7 Nh6)
7...g6 8.Nxg6 Qf7!? (or 8...Qxe4+ 9.Kd1 hxg6 10.Qxh8 Qxg2 11.Re1+ Kf7 12.c4) 9.exf5 Ne7
10.Qh3 (or 10.g4 Rg8) 10...Qxg6 11.fxg6 Bxh3 12.gxh3 hxg6 seems roughly equal. Black will get
one pawn back on h3 and has quite enough for the other.
c) 6.Qh5+ g6 7.Nxg6 should be a draw at best for White: 7...hxg6 (if Black wants to draw then

447
7...Qxe4+ 8.Kd1 Nf6 9.Qh4 Qxg2 10.Re1+ Ne4 11.Nxh8 Qf3+ is the safest way to go about it, as in
T.Kokkila-T.Vaatainen, Helsinki 1996) 8.Qxh8 Qxe4+ 9.Kd1 (not 9.Kf1?? Qxc2 10.Nc3,
E.Schröder-A.Nimzowitsch, Berlin 1903, as 10...Ba6+! 11.Kg1 Qd3 wins; or similarly 10.Qe5+ Kf7
11.Qe1 Ba6+ 12.Kg1 Re8 0-1 jamtaylo-tsmenace, ChessWorld.net 2004) 9...Qxg2 (or if 9...Qg4+
10.f3 Qxg2, M.Bajkovic-A.Dimovski, Kragujevac 2000, then 11.Re1+ Kf7 12.Qh7+ Bg7 13.c4
brings up the usual “0.00”) 10.Re1+ Kf7 11.Qe5!? (otherwise 11.Qh4 Qf3+ 12.Re2 Qh1+ is a draw)
11...Qxf2?! (Stockfish suggests that 11...Bb7 12.Qg3 c5 might give Black an edge – or, rather, might
give itself an edge) 12.Qe8+ Kg7 13.b3 Qf3+ 14.Qe2!? Qxe2+ 15.Rxe2 Ba6?! 16.Bb2+ Nf6 17.d3
Bd6? (still 17...Bb7 and ...c6-c5) 18.Nd2 Rh8 19.Nf3, when the moderate risk-taking paid off and
White went on to win, M.J.Turner-J.Tait, Notts League 1994.

6...d6

7.Nxc6

Not now 7.Qh5+?? g6 8.Nxg6 Qxe4+, as the white king has nowhere to hide, and 9.Be3 Nf6 just
wins.

7...Qxe4+ 8.Qe2 Qxe2+ 9.Kxe2 Bb7 10.d5

The only line to be wary of is 10.Na5!? Bxg2 11.Rg1, when 11...Bd5?! 12.c4 Bf7 13.d5 gave White
an edge in J.Ibarra Jerez-V.Lillo Castay, La Massana 2008. Black should play 11...Bh3, not fearing
12.f3 as 12...f4! 13.Bxf4 g6, followed by ...Bg7 and ...Kd7, offers good counterplay with the two
bishops.

448
 After 10.d5 Black has a superb score in my database: P16, W9, D7, L0.
a) 10...Bxc6 11.dxc6 Ne7 12.Nc3 Nxc6 13.Nd5 Kd7 was quite equal and later drawn in Jo.Shaw-
B.Lalic, Edinburgh 1999.
b) 10...Nf6 11.c4 (or 11.Nc3 Kd7, intending ...Nxd5; e.g. 12.Rd1 Nxd5 13.Nxd5 Bxc6 14.c4 g6
15.b3 Bg7 16.Rb1 a5 17.Bb2 Bxb2 18.Rxb2 a4 19.b4 a3 20.Rbd2 Ra4 and Black is no worse,
M.Hatipoglu-Mis.Pap, Anogia 2014; while if 12.Na5 then 12...Re8+ 13.Kf3 Bxd5+ or even 13...Ba8)
11...Ba6 (regaining the pawn due to the pin) 12.b3 (or 12.Na3 Nxd5 13.Rd1 Kd7 14.Rxd5 Kxc6
15.Rxf5 d5!) 12...Nxd5 13.Rd1 Nf6 14.Kf1 (or 14.f3 d5! 15.Kf2 Bc5+ 16.Be3 Bxe3+ 17.Kxe3 dxc4,
I.Teran Alvarez-S.Del Rio Angelis, Seville 2006) 14...d5! 15.Re1+ Kd7 16.Ne5+ Kc8 17.Ba3 dxc4
18.Nxc4 Bxc4+ 19.bxc4 Rb8 20.c5 Ne4 21.f3 Bxc5!? (21...Rxb1 is equal) 22.fxe4 Bd4 23.Nd2 Bxa1
24.Rxa1 Rd8 created practical chances which led to a win for Black in M.R.Savic-Mis.Pap, Serbian
League 2010, though Stockfish favours White slightly at this point (with 25.Nc4).
c) 10...Kd7 has been my choice: 11.b3?! (instead, 11.c4 Bxc6 12.dxc6+ Kxc6 13.Nc3 Re8+ 14.Kd3
Nf6 is again equal; e.g. 15.h3 Nd7 16.b3 Nc5+ 17.Kc2 g6 18.Be3 Bg7 19.Rad1 Rxe3!? 20.fxe3 Bxc3
21.Kxc3 Ne4+ 22.Kd4 Nf2 23.Rdf1 Nxh1 24.Rxh1 Re8 with a draw in due course, A.Rodriguez
Cespedes-F.Gomez Fontal, Cuban Ch., Matanzas 1998) 11...Ne7! 12.Nxe7 Bxe7 13.Bb2 (or 13.Nc3
Bf6) 13...Bxd5 14.f3? Rhe8 15.Kf2 Bh4+! (forcing White to weaken) 16.g3 Bg5 17.Nc3 Be3+
18.Kg2 Bc6 19.Rae1 g5 0-1 W.Wittmann-J.Tait, corr. 1991. IM Wittmann never took postal games
very seriously, playing hundreds at a time, I think just to see what happened.

D: 4.d4

449
 This move would be very strong, were it not for a key trick in line D2. Naturally Black starts off by
taking the e-pawn.

4...fxe4

Now White has two options:

D1: 5.Bxc6
D2: 5.Nxe5

D1: 5.Bxc6

Enabling White to take on e5 safely. As usual, we’ll recapture with the b-pawn for consistency with
line C.

5...bxc6!? 6.Nxe5 Nf6

The most natural continuation, preventing Qh5+ and protecting the e-pawn. Given time Black can
kick the white knight away with ...0-0, ...Qe8 and ...d7-d6, or perhaps ...Bb7 and ...d7-d6.
GM Brunello suggests that 6...Qh4!? “could be considered. The idea is quite enterprising, although
it may not stand up to the highest scrutiny.”

450
 If Stockfish is the highest scrutiny, then it stands up pretty well. One idea is to follow up with ...Ne7
and ...d7-d6. For example:
a) 7.Be3?!, as in A.Franco Valencia-C.Munoz Monroy, Palmira 2017, when 7...Ne7! 8.Nc3 d6
9.Nc4 d5 10.Ne5 Nf5 is just good for Black.
b) 7.Qe2 can also be answered by 7...Ne7! 8.Qc4 (or 8.g3 Qh3 9.Qxe4 d6 10.Nxc6 Bf5 11.Qf3 Bg4
12.Qe4 Bf5 with an unexpected repetition) 8...Nd5 9.0-0 a5 10.Qe2 Bb4!? (10...Ba6 11.c4 Bd6
12.Nd2 Nf4 13.Qxe4 0-0 seems more straightforward) 11.c4 Nf6 12.a3 Bd6 13.f4 c5 14.Nc3 0-0
15.Nxe4 cxd4 16.Nxd6 cxd6 17.Nf3 Qh5 with a mess which ended in a draw, C.Tanis-M.Jarecki,
corr. 2016.
c) 7.Nc3 Bb4 (not now 7...Ne7? 8.d5!) 8.0-0 (or if 8.Qe2?! then 8...Bxc3+! 9.bxc3 Ne7, intending
10.g3 Qf6 11.Qxe4 0-0 with excellent play for the pawn) 8...Bxc3 9.bxc3 Nf6 and 10...0-0 is fine for
Black.
d) 7.0-0 looks critical,

451
 when Black should play 7...Nf6 (not 7...Ne7? 8.f3 d6 since 9.Nf7! Kxf7 10.fxe4+ Kg6 11.Rf4 is
strong; e.g. 11...Qh5 12.g4 Qh3 13.a4! and Ra3 wins) 8.f3 (or 8.Nc3 Bb4 again) 8...Ba6! 9.Re1 (or
9.c4 Bd6 10.fxe4 0-0 11.Rf4 Ng4! – yes, really) 9...d5 10.fxe4 Nxe4 11.Nd2 Bd6 12.Nxe4 dxe4
13.g3 Qh3 14.Qg4 (or 14.Rxe4 0-0) 14...Qxg4 15.Nxg4 0-0 16.Nf2 (not 16.Rxe4?? Rf1+ 17.Kg2
Raf8) 16...Rae8 17.Nxe4 c5! 18.dxc5 (or 18.Bd2 cxd4 19.Nxd6 cxd6) 18...Bxc5+ 19.Be3 Rf1+!
20.Rxf1 Bxe3+ 21.Nf2 Bxf1 22.Rxf1 Bb6, followed by ...Re2 with a drawn endgame.

7.0-0

Castling to safety and planning to pile up on the e4-pawn before Black consolidates. Nothing else
should cause any trouble:

452
 a) 7.f3?!, J.Szen-E.Falkbeer, Vienna 1852, is well met by 7...Ba6! 8.fxe4 (or 8.c4 exf3 9.Qxf3 Bd6)
8...Bd6 and ...0-0.
b) 7.Qe2, A.Czebe-J.Micic, Balatonbereny 1993, asks for 7...a5! 8.0-0 (8.Qc4 Nd5 9.Qe2 Nf6
repeats) 8...Ba6, when even 9.c4 d5!? 10.Nxc6 (or 10.f3 Bd6 11.fxe4 0-0) 10...Qc8 11.Qc2 (or 11.b3
Qe6 12.Ne5 c5) 11...Bd6 12.Re1 Bxc4 13.Ne5 Bb5 14.Nc3 Qb7 looks fine.
c) 7.Nc3 allows 7...Bb4 8.0-0 0-0, followed by ...Qe8; whereas 8.Qe2?! a5 9.Qc4? Nd5 10.a3 Ba6
11.Qa2 Bxc3+ 12.bxc3 0-0 13.c4 Qh4 led to a win for Black in M.Nepeina Leconte-V.Inkiov, Clichy
1999.
d) 7.Bg5 Be7 8.Qe2 0-0 9.Nc3 (or 9.Qc4+ d5! 10.Qxc6 Rb8 11.0-0 Rb6 12.Qa4 Ng4 with a strong
initiative, R.Burridge-G.Hughes, corr. 2018) 9...Qe8 10.0-0-0, R.Pereira-L.Reis, Lisbon 2008, and
now 10...d5 is correct, intending 11.f3 c5! 12.dxc5 Bd8 13.Ng4 Bxg4 14.fxg4 c6, when the strong
centre provides excellent compensation.

7...Bd6!

Only S&S mention this move, first played by IM Artur Hennings in 1972.
Alternatively, 7...Bb7 (preparing ...d7-d6) has scored well for Black. The critical response is 8.Nc3
(if 8.c3 then 8...Rb8) 8...d6 9.Qe2!? (if 9.Ng4 d5 10.Nxf6+ Qxf6, V.Krasivskaya-M.Petrov,
Ruzomberok 2017, and 11.f3, then 11...Ba6 12.Rf2 Bd6 13.Be3 Qh4 14.g3 Qh3 15.fxe4 Bxg3
16.hxg3 Qxg3+ draws) 9...dxe5 10.dxe5 Nd5 11.Nxe4, when White has a dangerous attack for the
piece. Stockfish appears to defend it, but the path is very narrow; e.g. 11...Qe7 12.c4 Nb6 13.Bg5 Qe6
14.f4 h6! 15.Bh4 Be7 16.f5 Qxc4 17.Qh5+ Qf7 18.Qg4 Bxh4! 19.Qxh4 c5! 20.e6 (or 20.Ng5 Qc4!
21.Qh5+ Kd7) 20...Qe7 21.f6 gxf6 22.Nxf6+ Kd8 23.Qg4 h5! 24.Nxh5 Bd5! 25.Rad1 Kc8 and so
forth.

453
8.Re1

8.Nc3 0-0 9.Re1 comes to the same thing. Again, nothing else is troublesome; e.g. 8.Bg5 (or 8.Nc4
Be7 – it’s usually worth a tempo to keep the bishop) 8...0-0 9.Nc3 (if 9.Nd2 Qe8 10.Nec4 then
10...Be7 again, or 10...Qg6 11.Nxd6 cxd6 12.Bh4? d5! and Black stood well with the snake of pawns,
L.Varnam-J.Wittmann, 4NCL 2010) 9...Qe8 10.Nc4 (or if 10.Bxf6 gxf6 11.Nc4, W.Rocha-
K.Ferreira, Sete Lagoas 2009, then 11...Be7!) 10...Bb4! 11.Re1 d5 12.Ne5 Bd6 and Black is clearly
better, Mu.Smith-J.Tait, Notts Championship 2017.

8...0-0 9.Nc3

If 9.Nd2, T.Warakomski-M.Tazbir, Laczna 2002, Black can play 9...c5! 10.Nxe4 Nxe4 11.Rxe4
Bb7 12.Re2 Qf6 13.f4 Qe6! with the two bishops and activity for the pawn.

9...Bb4 10.Re3!

After 10.Bg5 Bxc3 11.bxc3 Qe8 12.Qe2 (not 12.Bxf6 Rxf6 13.Rxe4?? d6 and White loses a piece)
12...d6 13.Bxf6, as in P.Roydov-I.Nikolov, Plovdiv 2009, simply 13...Rxf6 is fine.

10...Bxc3 11.Rxc3

A nice idea – apart from preserving White’s own structure the increased latent pressure on c6
makes it harder to achieve ...d7-d6. Black also has potential weaknesses at e4 and a7, as well as on
the c-file.
In AussieTal-tsmenace, ChessWorld.net 2018, I tried 11...Rb8!? 12.b3 Qe8 13.Bg5 Rb6, which
failed to solve the problem: after 14.Bxf6 Rxf6 15.Ng4 Rf8 16.Qd2 Black still has inferior pawns,
though I managed to draw.

454
 I also looked at 11...Qe8 12.Bg5 Qe6 (intending ...c6-c5) 13.Bxf6 Qxf6 14.Qe2 c5!? 15.Ng4 (White
can’t take the pawn) 15...Qxd4 16.Rd1 Ba6 17.Qe1 Qb4 18.Ne5, when the structural issue remains,
even after 18...e3!? 19.Rxe3 Qxe1+ 20.Rdxe1 d6 21.Nc6.
Instead, Stockfish goes for 11...a5, which it assesses (after a couple of hours) as basically equal
(“0.09”), giving the variation 12.a4 Qe8 13.Bg5 d6 14.Bxf6 Rxf6 15.Nxc6 Bd7 16.d5 Bxc6 17.Rxc6
Qe5 18.Rxc7 e3 19.fxe3 Qxe3+ 20.Kh1 Re8 21.Rc3 Qf2 22.Qg1 h6 23.Qxf2 Rxf2 24.Rc6 Ree2. So
perhaps I’ve been worried about nothing.
In any event, with 6...Qh4!? also looking viable (or even 7...Bb7!?), there seems little justification
for writing off 5...bxc6 – or indeed 4...bxc6 – as inevitably bad for Black.

D2: 5.Nxe5!?

The full-on aggressive approach. It’s important to know how to handle this as Black. Yet if the
Jaenisch could be blown away so easily, it wouldn’t have lasted a week.

5...Nxe5 6.dxe5 c6!

455
 This is the trick alluded to earlier. If the bishop retreats, then ...Qa5+ follows, picking up the e5-
pawn. it’s surprising how often White seems surprised by this move.

7.Nc3

White goes all-in and sacrifices a piece. Okay, if they play this, they probably weren’t surprised.
Instead:
a) 7.0-0? cxb5 8.Nc3 d5 9.exd6 Qxd6 10.Qh5+ (or 10.Nxb5 Qe5) 10...g6 (or 10...Qg6) 11.Qxb5+
Bd7 12.Qxb7 Qc6 and White had nothing much for the piece, G.Marco-F.Marshall, Monte Carlo
1902, yet somehow won.
b) 7.e6? d5 (or 7...cxb5 8.Qh5+ g6 since 9.exd7+ fails to 9...Qxd7! 10.Qe5+ Kf7 11.Qxh8 Nf6 and
...Bg7, W.Buchanan-J.Stevenson, Scottish Ch., Largs 1998) 8.Qh5+ g6 9.Qe5 Nf6 (or 9...Qf6 since
10.Qxd5 fails to 10...Ne7) 10.Bg5 Be7 (or 10...Bg7) 11.c4 0-0 12.Bh6 Re8 (or just 12...cxb5) 13.Nc3
a6, Tereshchuk-V.L.Ivanov, USSR 1991, when 14.Ba4 b5 and ...Qd6 still offers Black a massive
advantage.
c) 7.Bc4?! Qa5+ 8.Bd2 (after 8.Nd2? Qxe5 9.Bxg8 Rxg8 10.Qe2 d5 White had nothing in five
games, including D.Twine-J.Tait, Bristol rapid 1991; similarly 8.Nc3? Qxe5 9.Bxg8 Rxg8 10.0-0 d5
– this time in 54 games, including S.Milson-J.Tait, Sheffield League 2019) 8...Qxe5 9.Bxg8 Rxg8
10.c4! (throwing up a breakwater; otherwise the black centre is too strong; e.g. 10.Nc3? d5 11.Qe2
Bd6 12.0-0-0 Bf5, J.Thorvaldsson-J.Boey, Varna Olympiad 1962, when 13.f3 achieves nothing in
view of 13...0-0-0 14.fxe4 d4!) 10...Bc5 (now after 10...d5 11.cxd5 cxd5 12.Nc3 Black is nowhere
near as secure; e.g. 12...d4 13.f4! exf3+ 14.Kf2 or 12...Be6 13.0-0 0-0-0? 14.Qc1!) 11.0-0 d5?!
12.cxd5 cxd5 13.Bf4! and White came out okay, I.Zakurdjaeva-E.Jussupow, Erfurt 2010, since
13...Qxf4?! 14.Qxd5 Qf8 15.Nc3 would be very risky. So Black should restrict themself to Pruijssers’
11...d6 and a far lesser edge – which is still an edge, mind.

456
 d) 7.Be2!? (another way to give up the pawn) 7...Qa5+ and then:

b1) 8.Bd2 Qxe5 9.Bh5+ (or 9.Bc3 Qg5 10.Bh5+ Kd8) 9...Kd8 (not 9...g6? 10.Bc3) 10.Bc3 (if 10.c4
Nf6 11.Be2 Bc5 12.Bc3, M.Mikadze-M.Arabidze, Ureki 2015, then 12...Qd6 13.Nd2 Re8 14.0-0 Kc7
– Tay) 10...Qg5 11.0-0 (after 11.g3 d5 12.h4 Qf5 13.g4 Qe6 14.g5 Ne7, R.A.James-J.Tait,
Eastbourne Major Open 1991; or 12.Bd4 Nf6 13.Be2 Bd6 14.c4, J.Polgar-S.Milliet, Cap d’Agde
rapid 2010, and 14...Kc7 – Tay, White is struggling to show anything for the pawn) 11...Nf6 12.f4 (or
12.Be2 d5 13.f4 exf3 14.Bxf3 Bd6, Sergeev-Mi.Tseitlin, Moscow 1975) 12...Bc5+ 13.Kh1 Qh6
14.Bxf6+ (or 14.Be2 Nd5! 15.Bd4 Ne3, S.C.Pereira-M.Macedo, Fortaleza 2011) 14...gxf6 15.f5 Kc7
16.c4 was M.Tantsiura-I.Feduk, Ukraine 2010, where 16...e3 17.Nc3 Bd6 18.Qe2 b6 19.Rad1 Be5
20.Rd3 Ba6 21.b3 Rad8 22.Rxe3 d5 is good for Black (Tay).
b2) 8.Nc3 Qxe5 9.Be3 Nf6 (or 9...d5 10.Bd4) 10.Bd4 (or 10.Qd2 Bc5! – Brunello) 10...Qg5 (if
10...Qe6, H.Dutschak-M.Breutigam, German Bundesliga 1996. then 11.0-0 d5 12.f3 e3 13.Qd3 or
11...Be7 12.f3 0-0 13.fxe4 Nxe4 14.Bd3 d5 15.Rxf8+ Bxf8 16.Qh5 g6 17.Qh4 Nxc3 18.Bxc3 with
some play for the pawn) 11.f3!? (not 11.0-0?! d5 12.Qc1 Qxc1 13.Raxc1 Bd6 14.f3 exf3 15.Rxf3 0-0,
C.Bünjer-P.Leisebein, corr. 2001; while after 11.Bxf6 Qxf6 12.Nxe4 Qe5! – Pruijssers, Black is
better with the two bishops and ...d7-d5 coming) 11...Qxg2 (if 11...d5 12.fxe4 dxe4 then 13.h4!?
Qg3+ 14.Bf2 Qc7 15.Qd4 or 13...Qf4 14.Bh5+ g6 15.Ne2 Qd6 16.0-0) 12.Rg1 Qxh2 13.fxe4 with
some compensation; e.g. 13...Bd6 (or 13...Bb4 14.Qd3 Qh4+ 15.Bf2 Qxe4 16.Rxg7) 14.Qd3 Bf4
15.Rxg7 Qh6 (or 15...Qh4+ 16.Bf2 Qh6 17.Bg3!) 16.e5! Qxg7 17.exf6 Qg5 18.Qe4+ Kd8 19.Qe7+
Kc7 20.Ne4 Qh4+ 21.Bf2 Qh1+ 22.Bf1 b6 23.Rd1! Re8! 24.Qxe8 Qxf1+ 25.Kxf1 Ba6+ 26.Kg2
Rxe8 27.Kf3 Be5 28.Rh1 and so on. But Stockfish is having to work hard, find all sorts of resources,
just for White to remain only slightly worse.

7...cxb5

Don’t even think about anything else. Just take the piece.

457
8.Nxe4

Not 8.Nxb5? Qa5+ 9.Nc3 Qxe5.

8...d5!

The key move in Black’s defence, returning a pawn to facilitate development. Not 8...Nh6? 9.0-0
Nf7 because of 10.Re1! Bb4 (or 10...Be7 11.Nd6+ Kf8 12.Qd5 Bxd6 13.exd6 Qf6 14.Re7 with a
crushing position, J.Bertorello-D.Beaumont, corr. 2001) 11.c3 Ba5 12.Nd6+ Ke7 13.Qd5 Qg8
14.Nf5+ Ke8 15.e6 and mates, M.J.Turner-J.Tait, Notts League 1993.

9.exd6 Nf6

10.Bg5

A critical continuation.
White has also tried:
a) 10.Nxf6+? Qxf6 11.Qh5+ Qg6 12.Qxb5+ Kf7 13.Bf4 (or 13.Qd5+ Qe6+) 13...Bxd6 and Black is
already winning, A.Gunsberger-W.Weinwurm, Lignano Sabbiadoro 2005.
b) 10.Qd4? (or 10.Qe2?) 10...Nxe4! 11.Qxe4+ Kf7 12.Bf4 (similarly 12.Qd5+ Kg6 13.Bf4 Qe8+
14.Be5 Qc6!, L.Milde-B.Hanison, corr. 2003; or if 13.g4 h6! 14.Qe4+ Kf7 15.Qd5+, H.Pecorelli
Garcia-S.Tatai, Havana 1985, then 15...Be6 16.Qxb7+ Kg8 17.Bf4 Qc8 – Tay) 12...Qe8 13.Be5 Qc6!
(not 13...Bxd6? 14.Qd5+ Qe6 15.Qxd6 Qxd6 16.Bxd6) 14.Qf4+ Kg8 15.0-0-0 Bd7 and White has
nothing for the piece. The game T.Ishkanov-J.Geller, Novaya Ladoga 2001, showed how Black can
consolidate: 16.Kb1 Rc8 17.Bc3 h5 18.Rd3 Rh6 19.Rhd1 Rg6 20.g3 Rg4 21.Qd2 b4 22.Be5 Rc4
23.Rc1 Bf5 24.d7 Rd8 25.Qg5 Rxc2 0-1.

458
 c) 10.0-0 is unappealing after 10...Nxe4! 11.Qh5+ (not 11.Re1? Bf5) 11...g6 12.Qe5+ (not
12.Qxb5+? Qd7 13.Qe5+ Kf7 14.Qxh8 Nf6 15.Re1 Bg7 16.Qxg7+ Kxg7 17.Re7+ Qxe7 18.dxe7 Be6
– Tay; or 14.Qxe4 Bxd6 15.Bg5 Qf5 16.Qd4 Qe5 17.Qh4 h6 18.Be3 Qh5 19.Qxh5 gxh5 and Black
won, M.Jablonicky-L.Novosadova, Slovakian League 2018) 12...Kf7 13.Qxh8 Nf6 threatens to trap
the queen with ...Qd7 and ...Bg7, so White has to enter an endgame with rook and pawn vs. bishop
and knight where Black all the chances; e.g. 14.Bh6 (or 14.Bg5 Be6 15.Rad1 Bg7 16.Qxd8 Rxd8
17.Rfe1 Bf8 18.Bf4 Nd5 19.Rd4 Bxd6 20.Bxd6 Rxd6, A.Skawinski-T.Kriebel, Polish League 2016)
14...Be6 15.Qxf8+ Qxf8 16.Bxf8 Rxf8 17.Rfd1 Rc8 18.c3 a6 19.f3 Ke8 20.Kf2 Kd7 and the d6-pawn
will soon fall, N.Bijlsma-M.Lyell, Groningen 2010.

10...Qa5+!

Black needs to know the first few moves below.

11.Nc3

Retreating the bishop is no good at all: 11.Bd2? b4! 12.Nxf6+ (if 12.Qf3 then 12...Qf5; or 12.Qe2
Qe5, e.g. 13.Nxf6+ gxf6 14.Qxe5+ fxe5 15.Bxb4 Rg8 16.0-0-0 a5 17.d7+ Bxd7 18.Bxf8 Rxf8
19.Rhe1 Rxf2 20.Rxe5+ Kd8 and Black won in Ma.Carlsen-J.S.Christiansen, Chess24.com blitz
2020, while 16...Bh6+ 17.Kb1 Rxg2 was even better, Mar.Thomas-J.Tait, Notts League 1990)
12...gxf6 13.0-0 Bxd6 14.Bh6 Bc7! 15.Re1+ (or 15.Bg7 Rg8 16.Bxf6 Qf5 17.Re1+ Be6, as in
E.Sparenberg-T.Beerdsen, Vlissingen 2018) 15...Kf7 16.Qd3 Qc5 (Brunello) 17.Qb3+ Kg6 18.Re4
Kxh6 19.Qf7 Bxh2+! 20.Kxh2 Qh5+ and wins (Tay).

11...b4! 12.Bxf6

There is nothing else for White with both bishop and knight en prise.

459
12...gxf6

And I’d thought there was nothing else for Black here either. Certainly no one has ever played
anything else.
However, several months after writing this section, I discovered a Chessable course on the Jaenisch
by GM S.P.Sethuraman, who gives 12...Qf5! 13.Bh4 (or 13.Nd5 gxf6 14.Nc7+ Kf7 15.Nxa8 Qe5+,
followed by ...Bxd6) 13...bxc3 14.0-0 Kf7 and “White clearly has not enough compensation for the
piece” – a quite brilliant idea, which may refute 10.Bg5 entirely.

13.Nd5

Not 13.Qe2+? Kf7 or 13.0-0? Be6! 14.Re1 0-0-0 and Black wins.

13...b3+!

A timely zwischenschach. The point is seen in comparison with 13...Be6 14.Qh5+ Kd8 15.0-0-0,
when 15...b3 16.axb3 (or 16.cxb3 Rc8+ 17.Kb1 Rc5) 16...Qa1+ 17.Kd2 Qa5+ 18.Kc1 Qa1+ is an
immediate draw.

14.c3 Be6 15.Nc7+

Not 15.Nxf6+? Kf7, nor 15.Qh5+? Kd8 16.0-0-0 since 16...bxa2! now wins for Black.

15...Kd7 16.0-0

Not 16.Nxa8? Qe5+ 17.Qe2 bxa2 18.Qxe5 fxe5 19.Nc7 Bc4 20.b3 Bxb3 21.Nb5 a6 22.Na3 Bxd6
and the knight is lost, Zlochevsky-V.L.Ivanov, Moscow 1987.

460
16...Bxd6 17.Nxe6

Still not 17.Nxa8? bxa2! (Pruijssers); e.g. 18.b4!? Qe5 19.Qa4+ Kc8 20.g3 Kb8 21.Rad1 Bg4!
22.Rd4 Bh3 23.Ra1 Kxa8 24.Qxa2 Bb8 and Black consolidates (Tay).

17...Qe5! 18.Re1!

Giving the knight up is unlikely to succeed:

a) 18.g3?! Qxe6 19.Re1 Qf7 20.Qg4+ (or similarly 20.axb3 Kc7 21.b4 h5! 22.Qa4 a6 23.Qa5+ b6
24.Qf5 h4 – Tay) 20...Kc7 21.axb3 h5 22.Qd4, J.Klisurica-G.Hamiti, Kosovan Ch., Pristina 2020, and
now 22...h4! is consistent; e.g. 23.Ra4 (or 23.Rxa7? hxg3 24.hxg3 Rxa7 25.Qxa7 Qd5 and mates)
23...Rad8! 24.Qxa7 Ra8! 25.Qd4 Rxa4 26.bxa4 hxg3 27.hxg3 Qh5.
b) 18.Nf4?! Qxf4 19.g3 Qf5 20.Qxb3 Kc6 21.Rad1 Rad8 22.Qc4+ Qc5 23.Qe4+ (or if 23.Qf7,
A.Rechinsky-L.Yagupova, St. Petersburg 1996, then 23...Rd7 24.Qe6 Re7 25.Qxf6 Rhe8 26.b4 Qe5
27.Qh6 Kc7 28.Rd4 Qe6 – Tay) 23...Kc7 24.Rd4 Qe5 25.Qf3 Rhe8 26.Rfd1, M.Justin-B.Markun,
Kranj 1999, when 26...Kb8 and ...Bc7 again consolidates.

18...Qxh2+ 19.Kf1 Qh1+ 20.Ke2 Qxg2 21.Qxb3 Rae8 22.Kd3 Kc8

Stronger than 22...Qf3+ 23.Re3 Qf5+ 24.Ke2, as in D.Pavasovic-V.Srebrnic, Ljubljana 1992, when
24...Rxe6 25.Qxb7+ Ke8 26.Qc8+ Kf7 27.Qxh8 Qb5+ 28.Ke1 Rxe3+ 29.fxe3 Bg3+ 30.Kd2 Qxb2+
31.Kd3 Qxa1 32.Qxh7+ is another draw (Brunello).

23.Qc4+ Kb8

This critical position has been reached in three games, with Black winning all three.
a) 24.Qd4? (the queen was better where it was) 24...Qh3+ 25.Re3 Qf5+ 26.Qe4 Qxf2 27.Qd5 Bc7

461
28.Rae1? (28.Re2 was necessary – Tay) 28...Rd8 29.Nxd8 Rxd8 0-1 S.Grodzensky-V.Filippov, corr.
1987.
b) 24.Re2! is suggested by Tay as offering good drawing chances; e.g. 24...Be5 25.Kc2 Qg6+ (or
25...Qg8 26.f4) 26.Kb3 Qf7 27.Nd4. Alternatively, if 24...h5, Stockfish proposes 25.Rd1 Qg6+ (or
25...h4 26.Kc2 Qg6+ 27.Kb3 Rc8 28.Qd3 Qxd3 29.Rxd3 h3 30.Rxd6 h2 31.Re1 h1=Q 32.Rxh1 Rxh1
33.Nd4 Rf8 34.f3 – Tay, via 25.Kc2 h4 26.Rd1 etc) 26.Kd2 Be5 27.Kc1 Rxe6 28.Qxe6 Bf4+ 29.Re3
h4 30.Qe7 Qe8 31.Qxf6 Bxe3+ 32.fxe3 Qxe3+ 33.Kc2 Qe8 34.Rh1 and White can be congratulated
on their sterling defence. It makes a change anyway – far more often it’s the other way round.
in this game 9...a5 was actually 10...a5, having arisen via 3...a6 4.Ba4 f5!? 5.exf5 e4 and so on

462
E: 4.d3

Statistics can be deceptive. When referring earlier to 4.Nc3 featuring in 55% of games, I neglected
to mention that that applied to all games, including those by correspondence where White has access
to theoretical tomes and analytical engines. Over the board things are rather different, especially
nowadays. A scan of the past ten years in the TWIC database puts 4.d3 on 53%, with 4.Nc3 lagging
behind on 34%. This also reflects recent repertoire books where 4.d3 has often been recommended.
The attraction is that 4.d3 is a great deal simpler than 4.Nc3 and yet still offers White chances to play
for a plus.

4...Nf6!?

463
 The standard (and strongest) sequence is 4...fxe4 5.dxe4 Nf6, when 6.0-0 transposes to line D2
below. White has options on move six, but none is as good as castling. So what’s the justification for
this alternative route which offers White a more significant option on move five? Very little. It’s
mostly psychological: holding off the exchange on e4 can disturb White, who then responds with a
secondary move.
As for the rest, GM Fabiano Caruana said: “Basically the position is just bad for Black. There’s not
much... We just win a pawn and he will have a hard time getting this back. And we’ll castle and e5
might fall at some point too. So...” We’ll see.
E1: 5.exf5
E2: 5.0-0
Secondary moves:
a) 5.Bxc6 bxc6! (of course) 6.Nxe5 (instead, 6.0-0 d6 7.Nbd2 fxe4 8.Nxe4 Be7 9.Ng3 0-0 10.Re1
c5 gave Black a nice position in J.Barker-J.Tait, North Midlands League 1993; 6.exf5 d6, psly-
tsmenace, ChessWorld.net 2004, and 7.Nh4 Be7 looks quite acceptable too) 6...fxe4 7.0-0 (or 7.Bg5
Qe7 8.d4 Bb7, Follofollo-tsmenace, ChessWorld.net.2004; note that 7.dxe4? would give White a bad
main line, where 7...Qe7 8.Bf4 g5 9.Bg3 d6 10.Nd3 Qxe4+ 11.Kd2 Qf5 is very good for Black,
E.Thiele-J.Boey, corr. 1972) 7...Bd6 8.Nc4, V.Skara-T.Kriebel, Banska Stiavnica 2012, and now
8...Bc5 looks fine.
b) 5.Nbd2 d6!? (keeping it independent; for 5...fxe4 6.Nxe4 see note ‘e’ below) 6.0-0 Be7 7.Re1 0-
0 8.c3 Kh8 9.d4 fxe4 10.Nxe4 Nxe4 11.Rxe4 d5 12.Re1 e4, J.Dementieva-I.Prihodko, Dmitrov 2011,
when 13.Bxc6 bxc6 14.Ne5 Qe8 is equal.
c) 5.Bg5 prompts a switch with 5...fxe4 6.dxe4 and then 6...Bb4+!. I like this idea very much,
messing up White’s intended Nc3 in the style of line F2 in Chapter Two. It should be even better here
because the white bishop is somewhat misplaced on b5. For example: 7.c3 (or 7.Nbd2 Bxd2+ 8.Nxd2

464
0-0 9.c3 Qe8 10.Bh4 Qg6 11.Bg3 d6 12.Qe2 Be6 13.Bc4 Nd8 with typical play for Black, A.Bouzidi-
D.Camacho Campusano, Batumi Olympiad 2018) 7...Bc5 (Black can now follow with ...d7-d6 and
...0-0; or if White plays Bc4, then ...Qe7 and ...Nd8) 8.Nbd2 (note that 8.Bxc6 bxc6 9.Nxe5?? loses at
once to 9...Bxf2+; and neither 8.0-0 d6 9.b4 Bb6 10.Qb3 Qe7, nor 8.Bxf6 Qxf6 9.Qd5 Bd6! gets
White anywhere) 8...d6 9.0-0 (or 9.Qe2 a6! – Pruijssers, pre-empting an attack on the bishop by Nc4
and a2-a4; not yet 9.b4 Bb6 10.Nc4? due to 10...Bxf2+ again, R.Hamid-A.Erigaisi, Bangladeshi
League 2021) 9...0-0 10.b4 Bb6 11.a4 a6 12.Bc4+ Kh8 13.Qe2 Ne7 and I much prefer Black,
O.Letreguilly-H.J.Hecht, St. Denis 2012.
d) 5.Qe2 also asks for 5...fxe4 6.dxe4 Bb4+! (6...a6!? is reasonable too) 7.c3 (or 7.Nbd2 Bxd2+
8.Bxd2 d6 9.0-0 Qe7 and ...0-0) 7...Bc5 8.Nbd2 (similarly, 8.0-0 d6 9.a4 a6, N.Short-H.J.Hecht,
London 1981; or 8.b4 Bb6 9.Nbd2 a6! again; and 8.Bxc6?! bxc6 9.Nxe5 Qe7 is no good for White,
J.Godois-A.Souza, Balneario Camboriu 2002) 8...a5! (Tay’s improvement on 8...Qe7 9.Nc4 d6 10.b4
Bb6 11.a4 from E.Ermenkov-Mi.Tseitlin, Pamporovo 1977) 9.0-0 (or 9.Bxc6 bxc6 10.Nxe5 Bb6
11.0-0 0-0 12.Qc4+ d5 13.Qxc6 Qd6! 14.Qxa8 Qxe5 “and Black’s initiative is at least good enough
for a draw” – Tay) 9...d6 10.a4?! 0-0 11.Bc4+ Kh8 12.Ng5?! Qe7 and Black now stood very well,
M.Souza dos Santos-M.Wahlund, Titled Tuesday (blitz) 2020.
e) 5.Nc3 is the second most common choice (quite a way ahead of 5.exf5).

This time 5...fxe4 6.dxe4 Bb4 transposes to a main line with 6.Nc3, known to be okay for Black;
while 6.Nxe4 d5 is the same as 4.Nc3 fxe4 5.Nxe4 Nf6 6.d3 d5 and also fine; e.g. 7.Nxf6+ gxf6 8.d4
e4 9.Ne5!? fxe5, A.Werle-R.End, Stockholm 1977, when 10.Bxc6+ bxc6 11.Qh5+ Kd7 12.Qf5+ Ke8
is a draw.
Alternatively, there is 5...Bb4!?, still holding back the exchange on e4, and then:

465
 e1) 6.exf5 0-0 7.0-0 (if 7.Bxc6 bxc6! 8.0-0 d6 9.Nh4!?, Black can regain the pawn at once via
9...Bxc3 10.bxc3 Nd5 or play it more slowly with 9...d5, intending ...Ne8) 7...d6 8.Bg5 Bxc3 9.bxc3
Bxf5 10.Nd2 (if 10.d4, as in K.Brameld-J.Tait, Grantham rapid 1997, then 10...h6 11.Bxf6 Qxf6
12.Bxc6 bxc6 13.dxe5 dxe5 creates mutual wreckage) 10...Qe8 11.Qf3 Bd7 12.Rae1 Qg6 13.Bxf6
Rxf6 14.Qd5+ Kh8 15.Re3?! Raf8 was typically good for Black in H.Weenink-B.Kostic, London
Olympiad 1927.
e2) 6.Bd2 is unimpressive: 6...d6 7.Qe2 (after 7.exf5 Bxf5 8.Ne4 Bxd2+ 9.Nfxd2 0-0 10.Bxc6 bxc6
11.Nxf6+ Qxf6 12.0-0 Qg6 Black is at least equal, R.Calapso-R.Spielmann, Merano 1926) 7...0-0
8.0-0 Kh8 (8...f4!? is tempting) 9.a3 Bxc3 10.Bxc3 fxe4 11.dxe4 Bg4 12.Qe3 d5 13.Bxe5 Bxf3
14.Bxc6 Ng4 15.Qd4 Nxe5 16.Qxe5 bxc6 17.gxf3 Rxf3 headed for a draw, Y.Melnikova-K.Shirazi,
Sautron 2006.
e3) 6.0-0 Bxc3 7.bxc3 fxe4 8.Bxc6 (unbalancing the pawn structure, whereas 8.dxe4 d6 is a good
version for Black of line B1 in Chapter Three; e.g. 9.Qd3 Qe7 10.Rb1 0-0 11.Bxc6 bxc6 12.Bg5 Kh8
13.Nd2 a5 14.c4 Qe8 15.f3 Nd7 16.Be3 Ba6 17.Rb3 c5 with advantage, S.Khanin-E.Solozhenkin,
Chelyabinsk 2021; or if 9.Bc4 then 9...Na5) 8...bxc6 9.Nxe5 0-0 with ...Qe8 to follow is fine for
Black; e.g. 10.c4 (or similarly 10.Bg5 Qe8 11.Bxf6 Rxf6 12.Ng4 Rf8 13.Rb1 d5, pickle47-tsmenace,
ChessWorld.net 2009; not 10.dxe4 Qe8 11.Qd4?? due to 11...c5! 12.Qc4+ d5 and wins) 10...Qe8
11.Ng4 Nxg4 12.Qxg4, N.Przymusinski-K.Shirazi, Proville 2007, when 12...exd3 is safe and equal.

E1: 5.exf5

This is the “more significant option”, usually given as the refutation of 4...Nf6.

5...Ne7!?

466
 The viability of 4...Nf6 relies on this imaginative idea from CCGM Sandor Brilla Banfalvi. Black
leaves the b5-bishop gazing into the void before being poked at with ...c7-c6.
Nothing else is any good. For 5...Bc5 6.0-0 see 5...Bc5 in line E2; 5...Be7 is too passive; and
5...Nd4 6.Ba4 c6 7.0-0 d5 8.Nxe5 Bd6 9.c3 Bxe5 10.Re1 0-0 11.Rxe5 Nxf5 is only good enough for a
blitz game, which Black lost anyway in E.Najer-I.Sokolov, World Blitz Ch., Dubai 2014.

6.d4

Seeing Black engaged in obscure manoeuvres White opens the game up at once. In The Modernized
Ruy Lopez GM Dariusz Swiercz gives 6.d4 an exclamation mark.
Other moves:
a) 6.g4?! c6 (not yet 6...Nxg4?? 7.Ng5 c6 8.Qxg4 h5, F.Libiszewski-I.Nataf, French Ch., Besancon
2006, due to 9.Qf3 cxb5 10.f6 and White is winning) 7.g5 (if 7.Ba4, Black would take the g-pawn)
7...Nxf5 8.gxf6 cxb5 9.f7+ Kxf7 10.Nxe5+ Kg8 11.Qf3 Qf6 12.Qd5+ Qe6 13.Qxe6+ dxe6 and ...Bd6
is fine for Black, N.Audet Bouchard-N.Kraiouchkine, Montreal 2008; and 11...d5 12.Nc3 Be6 may be
even better.
b) 6.Nxe5 c6 has shades of line D2 – and if the bishop retreats here, then ...Qa5+ wins a whole
knight. This looks decisive, but White has a resource: 7.Bg5! h5! (the only move; not 7...cxb5??
8.Bxf6 gxf6 9.Qh5+ and wins) 8.0-0 cxb5 9.Re1 d6 10.Ng6 Rh7, reaching a random position where
White has sufficient play for the sacrificed piece.

467
 For example: 11.Nc3 (similarly 11.Qf3 Kf7! 12.Bxf6 gxf6 13.Nxe7 Bxe7 14.Nc3 Bd7; rather than
11...Qd7?!, R.Polaczek-J.Boey, corr. 1991, when Stockfish throws out 12.Re6!) 11...Kf7 12.Bxf6
gxf6 13.Nxe7 Bxe7 14.Nd5 Bf8 15.Qf3 Bd7 has led to a draw by repetition in two of my own games:
16.Re4 (or 16.Re3 Bc6 17.c4 Kg8 18.Rae1 Kh8 19.cxb5 Bxb5 20.Nf4 d5 21.Qg3 Bd6 22.Ng6+ Kg8
23.Ne7+ ½-½ K.Pettersson-J.Tait, corr. 2003) 16...Kg8 17.Rh4 Bc6 18.Re1 Rc8 19.c3 Rcc7 20.Re6
Rcf7 21.h3 Rh6 22.a3 a6 23.Rd4 Bd7 24.Ree4 Bc6 25.Re6 Bd7 26.Rh4 Bc6 27.Rd4 Bd7 ½-½ lysol-
tsmenace, ChessWorld.net 2019. Essentially Black gets tied down to defence and is unable to
capitalize on the extra piece.
c) 6.Bg5 c6 7.Nxe5 transposes to ‘b’ above; while 7.Ba4 d6 8.Nc3 Bxf5 9.d4 e4 10.Nh4 is similar
to note ‘e’, except that Black has to play 10...Bg4 (since 10...d5?? 11.Bxf6 gxf6 12.Nxf5 Nxf5
13.Qh5+ wins a piece) 11.Qd2 d5. Capalaskine-tsmenace, ChessWorld.net 2018, continued 12.h3
Be6 13.f3 exf3 14.Nxf3 h6 15.Qe3 Bf5 16.Bf4 g5 17.Bh2 Bg7 18.0-0-0 0-0 19.Rhe1 Ng6 and Black
was okay.
d) 6.Nc3 c6 7.Ba4 d6

468
 8.d4 (for 8.0-0! see note ‘e’) 8...e4 (unusually 8...exd4! is stronger here; e.g. 9.Nxd4 Nxf5 10.0-0
Be7) 9.Ng5?! (9.Nh4 is preferable) 9...d5 10.f3 was V.Hadraba-G.Veltkamp, Decin 1996, where
10...exf3 and ...Nxf5 gives Black the edge.
e) 6.0-0 c6 7.Ba4 (instead, 7.Bc4 d6 8.Re1 Bxf5 9.Bg5 Qd7 10.Nc3 h6 11.Bxf6 gxf6 was level in
the much-quoted game S.Kindermann-M.Geenen, Thessaloniki Olympiad 1988; as was 8.Be6 Nxf5
9.Bxc8 Qxc8 10.Re1 Be7 11.Nc3 0-0, M.Kaminski-S.Zhurov, Warsaw 1993; or if 8.Bg5 h6 9.Bxf6
gxf6 10.Nh4 then 10...h5 11.Qf3 d5 12.Bb3 Qd7) 7...d6 8.Nc3 (not yet 8.d4 e4 9.Nh4 Nxf5! or 9.Ng5
Bxf5; nor 8.Nh4 Nxf5 9.Nxf5 Bxf5 10.f4?, when 10...e4 11.dxe4 Qb6+ 12.Kh1 Nxe4 is good for
Black, P.Zerdo-R.Dabo Peranic, Croatian League 1992) 8...Bxf5 (not 8...Nxf5? 9.Re1 Be7 10.Bb3,
when Black is in a mess with nowhere to put the pieces, including the king) 9.d4 is a “more sensible
approach” (Tay), which “is much better for White” (Jos.Doknjas). Certainly the position after 9...e4
10.Nh4 looks promising.

469
 White is about to open the kingside with f2-f3, while Black is a long way from completing
development, never mind securing the king. But it’s not all bad news: Black does have a strong centre
to hide behind, and White’s pieces are not particularly well placed. My record is played four, drawn
four.
After 10...d5 (not 10...Bg4?! 11.Qe1!, when 11...d5 12.f3 exf3 13.Nxf3 leaves Black in another
mess; or if 11...Qb6, AussieTal-tsmenace, ChessWorld.net 2009, then 12.Bb3 d5 13.h3! Bd7 14.f3 is
much the same) 11.f3 (if 11.Bg5 then 11...Bg4 12.f3 exf3 13.Nxf3 h6 14.Bxf6 gxf6 and ...Qd6; or
11.Nxf5 Nxf5 12.f3 Qd7 13.fxe4 dxe4 and so on) 11...Qd7 12.fxe4 dxe4 13.Bg5 (in my other games,
13.Bb3 0-0-0 14.Qe1 Bg6 15.Bf4 Nf5 16.Nxf5 Bxf5 17.Be5 Bg6 18.Rd1 Be7 19.h3 Rhe8 20.Qe3
Bb4 21.a3 Ba5 22.Na4 Bc7 23.Bxc7 Qxc7 24.d5 cxd5 25.Qxa7 Re6 turned out okay in Capalaskine-
tsmenace, ChessWorld.net 2018; while 13.Qe1 0-0-0 14.Bg5 h6 15.Bxf6 gxf6 16.Rd1 Bg4 17.Rd2 f5
18.h3 h5! 19.hxg4 hxg4 20.g3 Bh6 21.Rd1 Bg5 gave Black a strong kingside initiative for the piece,
schwabu-tsmenace, ChessWorld.net 2019) 13...0-0-0 14.d5 (or 14.Nxf5 Nxf5 15.d5 b5!) 14...Bg4
15.dxc6 Qxd1 16.cxb7+ Kxb7 17.Nxd1 Ned5,

470
 Black has surprising counterplay for the pawn, based on White’s vulnerable king and king’s knight;
e.g. 18.h3 (or 18.c4 h6 19.Bxf6 Bc5+ 20.Kh1 Nxf6) 18...Bc5+ 19.Kh1 Be6 20.Re1 (instead and
briefly: 20.c4 Nh5!; or 20.Bb3 Nh5 21.Re1 Ng3+ 22.Kh2 Bd6 23.c4 Nf4; or 20.Nf5 Nh5 21.g4 Bxf5
22.gxh5 g6! 23.Bxd8 Rxd8, planning ...e4-e3 and ...Be4+) 20...Nh5 21.g4 Nhf6 (preparing to attack
with ...h7-h5) 22.Bb3 (to answer 22...h5?! with 23.Nc3) 22...Rhe8 23.Ne3 (or 23.Nc3 Nxc3 24.bxc3
h6) 23...h6 24.Bxf6 (or 24.Bxd5+ Nxd5 25.Bxd8 Nxe3 26.Ba5 g5 27.Nf5 Nxf5 28.gxf5 Bxf5 29.Kh2
e3 with compensation) 24...Nxf6 25.Nhg2 Bxb3 26.axb3 Nd5 27.Rad1 Nxe3 28.Nxe3 Bxe3 29.Rxd8
Rxd8 30.Rxe3 Rd1+ 31.Kg2 Rd2+ 32.Kg3 Rxc2 finally ended up in a drawn rook endgame,
PonderStibbons-tsmenace, ChessWorld.net 2018.

6...e4 7.Nh4 c6 8.Be2 Nxf5

471
 The familiar trick of ...Qa5+ enables Black to regain the pawn.

9.Nxf5 Qa5+ 10.Nc3

If instead 10.Nd2 Qxf5 11.0-0 d5 12.f3, Black can play 12...e3! (not 12...Qg6? 13.fxe4 dxe4
14.Nc4 and White wins) 13.Bd3 (or if 13.Nb3 Bd6 14.Bxe3 0-0 15.Qd2, F.Wantiez-G.Veltkamp,
Belgian League 2015, then 15...Nh5!? intending 16.g4 Qe6 17.gxh5 Bxh2+ 18.Kxh2 Qh3+ and
draws) 13...exd2 14.Bxf5 dxc1=Q 15.Qxc1 Bxf5 with three pieces for the queen, D.Korze-
R.Trubetskoi, corr. 2004.

10...Qxf5 11.0-0

GM Swiercz writes here that “White plays f3 next, with a big advantage.” Well, maybe. Note that
11.f3 d5 12.0-0 comes to the same thing; and 11.g4 Qe6 12.g5 Nd5 is also fine for Black, E.Kance-
J.Tait, corr. 1998.

11...d5

Not 11...Bb4?! 12.f3 Bxc3 13.bxc3 0-0 14.fxe4 Qxe4 15.Bd3 Qe7 16.Bg5 d5 17.Qf3 and White is
better, especially after 17...Be6?! 18.Qg3 Kh8? 19.Rae1 (I resigned here in J.Horner-J.Tait, Blackpool
1998) 19...Qd7 20.Bxf6 gxf6 21.Bxh7!, followed by Rxe6 and wins, M.Mikac-V.Srebrnic, Slovenian
Ch., Postojna 1992.

12.f3 Qg6 13.fxe4 dxe4

Black is a move or two away from safety, as with 14.Bf4 Bb4 15.a3 Ba5 16.b4 Bb6 17.Na4 Bd8
18.Nc5 0-0, H.Gnirk-J.Tait, corr. 1998. So White should probably start something.

472
14.Rxf6!?

This exchange sacrifice makes it tricky. Obviously 14...gxf6?? loses the queen.

14...Qxf6 15.Nxe4 Qh4

Improving on 15...Qe7? from the original game E.Hintikka-S.Brilla Banfalvi, corr. 1981, where
16.Bd3! is very good for White; e.g. 16...Bf5 17.Bg5 Qe6 18.Qf1! Bb4 (or 18...h6 19.Re1) 19.c3 Rf8
(or 19...0-0 20.Bc4) 20.Nf6+ Rxf6 21.Re1 and wins.

473
16.Bh5+!

Another disruptive move. After 16.Bf3 Be7 and ...0-0 Black is fine, whereas now the weakened
dark squares will be an issue.

16...g6 17.Bf3

The tempting 17.Qe2? fails to 17...Bf5! (not 17...gxh5?? 18.Bg5 or 17...Qxh5?? 18.Nf6+) 18.Bf3
(the discovered check is harmless; e.g. 18.Nd6+ Kd7) 18...Bxe4 19.Bxe4 0-0-0 20.g3 Qf6 and Black
consolidates the exchange, LeHorla-tsmenace, ChessWorld.net 2020.

17...Be7

Black can’t fianchetto since 17...Bg7?? 18.Bg5 traps the queen.

I’ve had this position twice:

a) 18.c3 0-0 19.Qb3+ Kg7 20.Be3 (planning d4-d5 and Bd4 to embarrass the black king on the long
diagonal) 20...Bd8 21.Bf2 (now if 21.d5 then 21...Bb6 neutralizes the bishop; e.g. 22.Bxb6 axb6
23.dxc6 bxc6 24.Qxb6 Be6 25.Qxc6 Rxa2) 21...Qf4! (a Stockfish solution; 21...Qe7?! 22.Re1 Bc7
23.c4 Qd8? 24.d5 Be5 25.d6 gave White a powerful position which was converted in tarby-tsmenace,
ChessWorld.net 2009) 22.Re1 (instead, 22.Bg3 Qe3+ 23.Bf2 Qf4 repeats; while after 22.c4 Bh4
23.Bg3 Bxg3 24.hxg3 Qf7 the dangerous bishop has gone, or if 23.Be3 Qc7 24.d5 then 24...Bf5 25.d6
Qd7 26.c5 Bxe4 27.Bxe4 Rae8 defends) 22...Bh4! 23.Ng3 (obviously not 23.g3?? Qxf3; and either
23.Bxh4? Qxh4 or 23.Be3?! Qc7 helps Black) 23...h5 24.d5 (24.Re4 Qc1+ 25.Re1 Qf4 repeats again)
24...Bd8 25.Bd4+ Kh6! and Black is safe with ...Bb6 coming.
b) 18.Be3 0-0 19.d5 Bd8 (blockading d5-d6-d7 in advance) 20.d6 Be6 21.c4! (a clever decoy)
21...Bxc4 (Black can’t allow c4-c5) 22.g3 Qh3 23.Bg4 (the point: with the queen trapped Black has

474
to return the exchange) 23...Rf1+ 24.Qxf1 Bxf1 25.Bxh3 Bxh3 and the endgame was drawn,
drink1966-tsmenace, ChessWorld.net 2021.
So that’s the sum of my experience with 5.exf5 Ne7!?. It’s an exercise in brinkmanship for sure and
I don’t always feel up to it. No worries – on such occasions I just play 4...fxe4!.

E2: 5.0-0

Castling is the most common move and is IM Andrew Greet’s choice in Play the Ruy Lopez. If
White is intending the main line with 6.0-0, it makes things simpler to play 5.0-0 here as well since
Black has nothing better than exchanging on e4.

5...fxe4

Over the board I’ve mostly played 5...Bc5?! which is very much inferior: 6.exf5! 0-0 7.Nbd2! d6
(Black does no better with 7...Nd4 8.Nxd4 exd4 9.Ne4 Be7 10.g4 or 8...Bxd4 9.Nf3; or 7...d5 8.Bxc6
bxc6 9.Nxe5 Qe8 10.Nb3 Bd6 11.Re1 Bxf5 12.Nd4, O.Graham-B.Wood, corr. 2019) 8.Ne4 Bb6
9.Ng3 Ne7 10.Nh4 and Black has a dreadful kind of reversed Schallopp; e.g. 10...c6 11.Ba4 Bd7 (or
11...Nfd5?! 12.Bb3! Nxf5? 13.Nhxf5 Bxf5 14.c4, winning a piece, G.Quillan-J.Tait, 4NCL 2001)
12.Bb3+ Kh8 13.Qe2 Qc8 14.Ne4 d5 15.Nxf6 gxf6 16.g4 Rg8 17.Kh1 c5 18.f3 and White is a solid
pawn ahead, afms-tsmenace, ChessWorld.net 2018. The nice-looking black centre isn’t actually going
anywhere, while White can build up to attack on any part of the board.

6.dxe4 d6

I like this old-fashioned move, defending the e-pawn and planning ...Be7 and ...0-0. Once – or
rather if – castling is achieved safely, Black will generally have a good game. So White’s task is to
interfere with this plan. I was going to boast that I’ve never lost with 6...d6. Then I noticed a bad loss

475
to (future IM) James Cobb in 1998. Nuts.
The more modern and active 6...Bc5 is twice as popular and a lot of analytical work has gone into
proving its viability, even against 7.Bxc6 bxc6 8.Nxe5 0-0 9.Nc3, but I’d rather play White there.
After 6...d6 White has tried a great many moves, including:
E21: 7.Nc3
E22: 7.Bc4
Significant others:
a) 7.Bxc6+?! bxc6 returns to 5.d3 etc in line C.
b) 7.h3 is the type of nothing move that people who know nothing about the Jaenisch tend to make.
Black is given a free tempo to get castled; e.g. 7...Be7 8.Nc3 (or 8.Bc4 Na5) 8...0-0 9.Ng5?! Kh8
10.Bc4 Qe8 11.f4 (at least playing as if they meant it) 11...h6 12.Ne6 Bxe6 13.Bxe6 Nd4 (or 13...exf4
14.Bxf4 Nh5 15.Bh2 Rxf1+ 16.Qxf1 Bf6, K.Alekseenko-G.Oparin, Titled Tuesday blitz 2020)
14.Bc4 b5 15.Bd3, M.Schurade-G.Möhring, Leipzig 1981, and now 15...c5 looks good for Black.
c) 7.Re1 defends the e-pawn and might prepare some Ruy Nbd2-f1 business, but the rook looks
misplaced here in the Jaenisch; e.g. 7...Be7 8.Nbd2 (or if 8.Bc4 Na5 9.Qd3 Nxc4 10.Qxc4,
A.Wahedi-L.Do Valle Cardoso, Titled Tuesday blitz 2021, then 10...Qd7, followed by ...Bd8 and
...Qe6, is one way to sort out the slight problem on the a2-g8 diagonal) 8...0-0 9.c3 Kh8 10.Nf1 Qe8
11.Ng3 Nh5 12.Nxh5 Qxh5 13.Be3 Nd8 14.Be2 Qg6 15.Nd2, A.Gurevich-S.Chapygin, Moscow
2008, when 15...Bg5 leaves Black very much for preference.
d) 7.Bg5 aims to gain control of d5 with a quick exchange on f6, but Black has little to fear:

7...Be7 8.Bxf6 (for 8.Nc3 see 8.Bg5 in line E21; or if 8.Qd3 then 8...a6! as in note ‘g’) 8...Bxf6
9.Qd5 Bd7 10.Nc3 Qc8! 11.Rad1 a6 (not yet 11...Nd8?! 12.Bxd7+ Qxd7, S.Wolf-R.Spielmann,
Vienna 1928, as 13.Qb3 and 14.Nd5 is annoying) 12.Ba4 (or 12.Bc4 Nd8) 12...Rb8 covers the

476
queenside and White can’t really increase the pressure; e.g. 13.a3 (not 13.Nd2? b5 14.Bb3 Nd4)
13...Be6!? 14.Bxc6+ bxc6 15.Qxc6+ Bd7 16.Qc4 Rxb2 17.Nd2 Be6 18.Nd5 0-0 with equality since
the rook can’t be trapped: 19.Nb3 Qb8 20.Qc3 Ra2.
e) 7.c4!? plans a quick c4-c5 to weaken Black’s solid structure.

It doesn’t look like anything much to worry about; e.g. 7...Be7 8.Nc3 0-0 (or 8...a6 9.Ba4 0-0) 9.c5
Kh8 10.cxd6 (or 10.h3 Bd7 11.cxd6 Bxd6 12.Re1 Qe8 13.Nd5 Qh5, J.Alonso Bouza-M.Outerelo
Ucha, Mondariz 2016) 10...Bxd6 11.Bg5 Qe8 12.Qd3 Bd7 13.Rfe1 Qg6 14.Bxf6 Rxf6 and Black is
fine, especially after 15.Nd5?? Rxf3! 16.Qxf3 Nd4 and wins, R.Jober-B.Osolin, Slovenian League
1996.
f) 7.Qe2 intends a timely Qc4 to cause annoyances on the light squares, but putting the queen on e2
encourages a pin: 7...Bg4 (if 7...a6 8.Bc4 Na5, White can play 9.Bd3 Be7 and 10.b4 Nc6 11.Bc4
before Black can castle) 8.Nc3 (here 8.h3 Bh5 9.Nc3 a6! comes to the same thing; if instead 8.Qc4
then 8...Qd7; or 8.Rd1 Be7 since 9.h3 Bh5 10.g4 fails to 10...Nxg4!) 8...a6! (asking the question) 9.h3
(asking one back) 9...Bh5 10.Bxc6+ (or 10.Ba4 Be7 11.Qc4 Qd7 and ...Bf7; not 10.Bc4?! Nd4)
10...bxc6 11.Qc4 Qd7 12.Rd1 and now 12...Bf7 (12...Bxf3!? 13.gxf3 Nh5 14.Kg2 Be7 15.Rd3! was
very unclear in AlanJM-tsmenace, ChessWorld.net 2018) 13.Qa4 (or 13.Qe2 Qc8 14.Ng5 Bh5)
13...c5 14.Qa5 Qc8 was safest, since 15.Ng5 Bh5 16.g4?! is well met by 16...h6!.
g) 7.Qd3 has the same idea as in ‘f’ and without allowing the pin. The drawback this time is that the
queen blocks the b5-bishop’s retreat and so is well met by 7...a6! (the exclamation mark comes from
Sokolov) and then:

477
 g1) 8.Bc4 Na5, and if 9.a4!? (after 9.Nc3 Nxc4 10.Qxc4 Qd7, Black “will finish development
having a solid position and the potential advantage of the bishop pair” – Sokolov) 9...Nxc4 10.Qxc4
(“intending Nc3 and Bg5 with a nice position – Black still needs to solve the problem of how to
castle” – Greet), then 10...Qe7 11.Nc3 Qf7 looks fine; e.g. 12.Qxf7+ Kxf7 13.Ng5+ Kg8 14.f4 h6
15.Nf3 exf4 16.Bxf4 Be6 17.Rad1 Kh7 and Black had no problems, T.Ruiter-K.Faber, corr. 2020.
g2) 8.Bxc6+ bxc6 (“Black has yet to finish his development, but it is rather difficult for White to
open up the position” – Sokolov)

9.Qc4 (if 9.Nc3 h6 10.Rd1 Bd7 11.Nd2 g5 12.Nc4 Bg7 13.Na5 Qe7 14.Qc4 Qf7 15.Qxf7+ Kxf7,
Black is better with the bishops, R.Kasimdzhanov-T.Nyback, Khanty-Mansiysk Olympiad 2010; or
9.Rd1 Bd7 10.Nc3 Qe7 11.Bg5 h6 12.Bxf6 Qxf6 13.Qc4 g5 14.Ne1 Qf7 15.Qb4 Bg7 ½-½

478
P.Smirnov-V.Zvjaginsev, Moscow 2009, when “Black is definitely not worse and could have played
on here” – Sokolov) 9...Bd7 10.Nc3 h6 (or 10...Qe7 – Sokolov) 11.b3 g5 12.Ba3 Nh5 13.Rad1 g4!?
(either 13...Qf6 or 13...Nf4 looks more natural) 14.Nxe5! dxe5 15.Bxf8 Kxf8 16.f4 should have led to
a draw until White went wrong in E.Sutovsky-C.Hanley, Gibraltar 2012.
g3) 8.Ba4 can also arise via a Schliemann Deferred (3...a6 4.Ba4 f5). I’ve had this via both routes;
e.g. 8...Be7 9.Nc3 (or 9.Bb3 Na5) 9...Bd7 (or 9...0-0!? 10.Qc4+ Kh8 11.Bxc6 bxc6 12.Qxc6 Rb8
13.b3 Rb4 with compensation, D.Wells-J.Tait, Notts League 2008) 10.a3 0-0 11.Bb3+ Kh8 12.Ng5
Qe8 13.Ne6 Bxe6 14.Bxe6 Nd4 15.Bh3 Qg6 16.Be3, when 16...Nf3+!? 17.Kh1 Qh5 18.Nd5 Nxd5
19.Qxd5 Bg5 20.Qxb7 Bxe3 21.fxe3 Ng5 22.Qxa6 Nxh3 23.gxh3 Qxh3 led to a draw in Glawurtz89-
tsmenace, ChessWorld.net 2018
h) 7.a3!? plans to prevent castling with Bc4 and first ensures the bishop’s safety on the a2-g8
diagonal while waiting to see how Black responds. This is more tricky than it appears. After normal
moves like 7...Bg4 8.h3 Bh5 9.Nc3 or 7...Be7 8.Bc4 Bg4 9.h3 Bh5 10.Nc3, it’s hard to find a
convincing plan of development for Black; e.g. 10...Nd4 (10...Qd7 11.Qd3 Nd8 is too slow in view of
12.Nh4 Bf7 13.Nf5) 11.g4 Nxf3+ 12.Qxf3 Bg6 13.g5 Nd7 14.Qg4 Nc5 15.f4 exf4 16.b4 Nxe4
17.Nxe4 d5 18.Nc5 and White was winning in K.Kulaots-M.Timmermans, Barcelona 2013.
Instead, 7...Be6 is probably correct, occupying the key diagonal. It feels wrong to create a target for
Ng5 and f2-f4, but Black seems to be okay. For example:

h1) 8.Ng5 Bg8 9.f4 Be7 10.Nc3 a6! 11.Bxc6+ bxc6 transposes to ‘h3’ below and none of 11.Ba4
h6 12.Nf3 Bc4, or 11.Be2 Nd4 12.Bd3 Ne6, or 11.Bd3 h6 12.Nf3 Bf7 seems to improve for White.
h2) 8.Qd3 a6! (such a useful little move) 9.Bc4 (or 9.Ba4 Be7 10.Ng5 Bg8 11.f4 Nd7) 9...Bxc4
10.Qxc4 and now even 10...d5 11.exd5 Qxd5 12.Qxd5 Nxd5 is possible, since 13.Re1 0-0-0
14.Nxe5?! fails to 14...Nd4.
h3) 8.Nc3 a6! 9.Ng5 (after 9.Ba4 Be7 10.Qd3 h6 11.Ne2 0-0 12.Bb3 Qd7 13.Ng3 Nd8 Black was

479
fine in W.Bachmann-D.Nightingale, corr. 2019) 9...Bg8 10.Bxc6+ bxc6 11.f4 Be7! (not 11...h6
12.Nf3 Bh7? 13.Qe2 Be7 14.Qc4 and White was winning, T.Kosintseva-V.Cmilyte, Tbilisi 2012)
12.Qe2 (if 12.Nf3 then 12...Bc4, or 12.fxe5 dxe5 13.Qe2 Bd6) 12...Nd7 (this is the key idea:
defending e5 and uncovering against the white knight) 13.Nf3 (or 13.Qh5+ g6 14.Qh6 Bxg5 15.fxg5
Bc4 16.Rf2 Qe7) 13...Bf7 14.fxe5 Nxe5 15.Nxe5 dxe5 16.Be3 0-0 17.Na4 Qb8 and Black looks okay,
despite the shattered queenside.
i) 7.a4!? has the same idea as 7.a3, and the pawn might advance further too. Black should reply in
the same way; i.e. 7...Be6, intending 8.Nc3 a6, or 8.Ng5 Bg8 9.f4 Be7 10.Nc3 a6 again (Stockfish
also suggests 9...Qd7!? here, followed by ...a7-a6 and ...0-0-0).
The only games in my database with this both saw 8.Qe2 and then 8...a6 9.Ng5 Bg4! (since the
queen is there to hit) 10.Bxc6+ (if 10.f3 then 10...axb5 11.fxg4 Nd4) 10...bxc6 11.f3 Bd7 (better than
11...Bh5?! 12.Ne6 Qd7 13.Nxf8 Rxf8 14.Nd2 Kf7 15.Qc4+ d5 16.Qc3 Rfe8 17.b3 Kg8 18.Re1, when
Black had to defend for 138 moves before getting a draw, Laser-Houdini 6, CCC 2.Blitz Battle 2018
– the programmers will have cared even if the engines didn’t) 12.f4 exf4 (not 12...Be7? here because
the other bishop is missing from the key diagonal, allowing 13.fxe5 dxe5 14.Qc4) 13.Bxf4 Be7
14.e5!? dxe5 15.Bxe5 Bg4 (back again!) 16.Nf3 (or 16.Qc4 Qd5) 16...0-0 17.Nc3 Qd7 18.Rae1 Bh5
19.Kh1 Bd6 and Black was fine in Z.Rusznák-H.Buczinski, corr. 2020.

E21: 7.Nc3

The traditional main line. Developing the knight is the most natural move and has been seen more
often than all the others put together.

7...Be7

Getting ready to castle, after which Black should be quite okay. White now has two main ideas:

480
 1. To discourage castling by placing the queen on e2 or d3, so that 8...0-0 would then lose a pawn
to Qc4+, Bxc6 etc
2. To allow castling and answer it with 9.Bc4+ Kh8 10.Ng5 – in this case White will have to decide
which eighth move might be most useful.
From the other side, there is little reason to fear either plan. Black gets good compensation for the
pawn in the first, and there is no eighth move which makes the second particularly strong.

8.Qd3

This, too, is the traditional move. White has tried numerous others:
a) 8.Qe2 0-0 9.Qc4+ or 9.Bxc6 bxc6 10.Qc4+ transposes to the main line.
b) 8.Bg5 Bg4 (not yet 8...0-0?!, in view of 9.Bxf6 Bxf6 10.Qd5+ Kh8 11.Bxc6 bxc6 12.Qxc6, when
Black may not have quite enough for the pawn) 9.Bxf6 (or 9.h3 Bh5 10.Bxf6 Bxf6 11.Qd5 Qd7
12.Nd4!? Bf7 13.Nxc6 Bxd5 14.Nxe5 Qxb5 15.Nxb5 Bxe4 16.Ng4 Kd7) 9...Bxf6 10.Qd5 (or 10.Nd5
0-0 11.Ne3 Bh5 12.c3 Kh8 13.Nf5 Ne7 14.Ng3 Bg4, followed by ...Ng6 with a strong kingside
initiative, mizuzul-tsmenace, ChessWorld.net 2020) 10...Qd7 11.Rad1 Be6 12.Qd2 0-0 and Black was
fine here, G.Schwietzer-E.Kemeny, St. Louis 1904.
c) 8.Nd5 offers a pawn which it seems more thematic to decline:

8...0-0 (8...Nxe4!? is okay if White rushes to win the pawn back, i.e. 9.Nxe7 Qxe7 10.Qd5 Nf6
11.Bxc6+ Kf8! 12.Qb5 a6, K.Honfi-J.Boey, Den Haag Zonal 1966; but something less forcing, such
as 9.Bc4 or 9.Re1 might cause more problems) 9.Ng5 (or 9.Bc4 Kh8, while if 9.Nxf6+ then 9...Rxf6!)
9...Kh8! 10.Bc4?! (this doesn’t work out; nor does 10.Nxf6 gxf6 11.Nf3 f5 with an edge, G.Blaich-
S.Hasecic, German League 1994; instead, the logical 10.f4 Bg4 11.Qe1 Nxd5 12.exd5 Nd4 13.Bd3
looks about equal) 10...Nxd5! 11.Nxh7 Rf4! 12.Bxf4? (but if 12.Qxd5 then 12...Nb4! 13.Qa5 Rxe4
14.Qxb4 d5 is good for Black) 12...Nxf4 13.g3 Nh3+ 14.Kh1 Qe8 15.f4 exf4 16.gxf4 Kxh7 and Black

481
won, Ma.Tseitlin-V.Arbakov, Gomel 1983.
d) 8.a3 0-0

9.Bc4+ (or 9.Ng5 Kh8 10.Bc4, transposing; for 9.Qd3 see 9.a3 in the main line) 9...Kh8 10.Ng5
Qe8 11.f4 (note that 11.Ne6 doesn’t gain the two bishops because of 11...Qf7! 12.Nxc7 Qxc4
13.Nxa8 Nd4 and the a8-knight will not escape; we’ll see more of that idea in other variations; and
11.Nb5?! is pointless due to 11...Bd8 and ...a7-a6) 11...exf4 12.Bxf4 Ne5 (or 12...Qg6) 13.Bxe5 (or
13.Bb3 Bg4 14.Qe1 Nh5 15.Bxe5 Rxf1+ 16.Qxf1 Bxg5 17.Bd4 Qe7 18.Qc4 c6 19.Rf1, G.A.Thomas-
F.Lazard, Paris 1929, and now 19...Nf4 is fine) 13...dxe5 14.Kh1 h6 (or 14...c6) 15.Nf3 Qh5 16.Nd5
Bd6 17.Nxf6 Rxf6 was perfectly okay for Black, H.Wolf-H.Atkins, Hanover 1902 – two old games
showing (by transposition) how to play as Black.
e) 8.Be3 is a more useful move, but Black is still okay after 8...0-0 9.Bc4+ Kh8 10.Ng5 Qe8 11.Ne6
(not 11.f4? Ng4)

482
 11...Qf7! (another option is 11...Na5!? 12.Nxc7 Qd7 13.Nxa8 Nxc4 14.Bg5 Nxb2 15.Qe2 Qc6 –
I’ve drawn twice from here, but finding the right moves is tough; e.g. 16.Bxf6 Bxf6 17.Nd5 Bd8
18.f4 Nc4 19.fxe5 Rg8! 20.exd6 Nxd6 21.Rad1 b5 and so forth, Zoltar-tsmenace, ChessWorld.net
2020) 12.Nxc7 Qxc4 13.Nxa8 Nd4; e.g. 14.Qd3 Qc5! (instead, 14...Be6 15.Qxc4 Bxc4 16.Nc7 Nxc2
17.Bxa7 Nxa1 18.Rxa1 Rc8 19.N7d5 Nxd5 20.Nxd5 Bxd5 21.exd5 left White with an extra pawn in
A.Davies-J.Manley, corr. 2019, though Black drew anyway) 15.Bxd4 (or 15.Nd1 Qa5) 15...exd4
16.Nb5 Bd7 (Stockfish goes for 16...Nh5! 17.g3 Nf6, provoking a weakness, so that after ...Bh3 both
the f1-rook and a8-knight are attacked) 17.Nac7 (or 17.Nbc7 Ng4) 17...Bd8 18.b4 Qb6 19.a4 Bxc7
20.Nxc7 Qxc7 21.Qxd4 Re8 is roughly equal.
f) 8.h3 0-0 9.Be3 (again 9.Bc4+ Kh8 10.Ng5 Qe8 11.Ne6 is met by 11...Qf7! 12.Nxc7 Qxc4
13.Nxa8 Nd4, as in A.Anelli-H.Link, Buenos Aires 1976: 14.Qd3 Qc6 15.Nd5?!, when 15...Nxd5
16.exd5 Qxd5 17.Nc7 Qf7 and ...Bf5 looks very strong for Black) 9...Kh8 10.a3 Qe8 11.Bd3 Nh5
shows typical moves for Black against anodyne play. The obvious ideas are ...Qg6 and ...Nf4; while
after 12.Nd5 Bd8 13.c3, I tried 13...g5!? 14.Nxg5 (declining the pawn with 14.Nd2 Nf4 15.Be2 Qg6
16.Bg4 is preferred by Stockfish; not 14.Bxg5? Bxg5 15.Nxc7 Qg6 16.Nxa8 Nf4 with a very strong
attack) 14...Nf4 15.Bxf4 exf4 16.Nf3 Qg6 with good compensation for the pawn in A.Corish-J.Tait,
corr. 2021.
g) 8.Bc4 attempts to prevent castling, as in line E22 below, but the inclusion of Nc3 and ...Be7
means that 8...Na5! works here.

483
 In fact Stockfish recommends that White go back with 9.Bb5+ Nc6! and have another go. A few
games have ended in a repetition this way. If Black wants to continue, then 9...Bd7 10.Bxd7+ Qxd7 is
playable; e.g. 11.Nd5 Nc6 (avoiding 11...0-0?? 12.Nxf6+ and Qd5+) 12.Re1 0-0 13.c3 Kh8 14.Bg5
Rae8 15.Nxe7 Rxe7 was fine for Black, who did go on to win in G.Sanakoev-V.Leonidov, Tula 2004.
Other moves don’t give White anything:
g1) 9.Be2 0-0

10.b4 (logically gaining space; 10.Nd5 Kh8 11.Nxe7 Qxe7 12.Bg5 Qe8 13.Qd3 Qg6 14.Bxf6 Qxf6
15.Rad1 Nc6 16.c3 Ne7 17.Qc4 Ng6 18.Qxc7 Nf4 19.Bc4 Nxg2 worked out well for Black in
F.Cottarelli-J.Tait, corr. 1998) 10...Nc6 11.b5 Na5 12.Ba3 (or 12.Nd5 Nxd5 13.Qxd5+ Kh8 14.Bd2
c6) 12...c6 13.bxc6 (or 13.Bb4 d5 14.Bxe7 Qxe7 15.exd5 Qc5) 13...bxc6 14.Rb1 Be6 15.Ng5 Bc8!

484
and Stockfish says “0.00”; e.g. 16.Bb4 (if 16.f4 then 16...exf4 17.Rxf4 d5, or 16.Qd3 h6 17.Nf3 Be6
18.Nh4 Nc4 19.Bc1 Rf7) 16...h6 17.Nf3 Be6 18.Ba6 Nc4 19.Qe2 Nb6 20.Rfd1 Qc7 with a draw eight
moves later in H.Galjé-D.Donk, corr. 2019.
g2) 9.Bd3 0-0 10.Nd5 (or 10.b4 Nc6 11.b5 Na5 as in ‘g1’) 10...Kh8 11.Re1 Be6 12.Bd2 Nc6
13.Ng5 Bg8 14.Nxe7 Qxe7 15.Re3!? h6 16.Rh3 Nd4 17.b3 a5 (first solidifying the queenside) 18.a4
b6 19.Be3 c5 20.c4 Nh7 (before looking to the other flank) 21.Nxh7 Bxh7 22.Rb1 Rf6 23.Rg3 Raf8
and Black is ready to take the initiative, schwabu-tsmenace, ChessWorld.net 2020.
g3) 9.Qd3 Nxc4 10.Qxc4 is well met by 10...Qd7 11.Ng5 Rf8 12.Nd5 (or if 12.f4 then 12...Qc6!
13.Qxc6+ bxc6 14.fxe5 dxe5) 12...Bd8 13.f4 Nxd5 14.exd5 Bxg5 15.fxg5 Qg4 16.Rxf8+ Kxf8
17.Qxg4 Bxg4 and Black has whatever chances are going, aperturaf-tsmenace, ChessWorld.net 2018.

8...0-0!

Black is not discouraged!

9.Qc4+

Or 9.Bxc6 bxc6 10.Qc4+ Kh8, transposing. White might as well grab the pawn, since nothing else
offers anything:
a) 9.Nd5 Nxd5?! 10.Qxd5+ Kh8 11.Bxc6 bxc6 12.Qxc6 left White a much safer pawn up in
S.Alapin-H.Trenchard, Vienna 1898. Obviously Black should play 9...Kh8 first and then exchange on
d5.
b) 9.a3, H.Suchting-F.Dus Chotimirsky, Prague 1908, also prompts 9...Kh8, avoiding any tactics
before setting about kingside manoeuvres; e.g. 10.Be3 Qe8 11.Ne2 Nh5 12.Ng3 Nf4 13.Qc3 Bd7
14.Rfe1 a6 15.Bf1 Qg6 with an edge for Black, D.Tucker-J.Tait, corr. 1996.
c) 9.Bc4+ Kh8 10.Ng5 is again an option, and this time Black has to give up the light-squared

485
bishop; i.e. 10...Qe8 11.Ne6 Bxe6 (not 11...Qf7?? since the c4-bishop is defended) 12.Bxe6.
Nonetheless, Black still seems okay; e.g. 12...Nd4 13.Bh3 (retreating this way to defend the kingside;
13.Bb3, giving the bishop back, makes no sense; while Sokolov’s “about equal” after 13.Bc4 c6
might be amended to “at least equal”) 13...Qg6 14.f4 (or 14.Be3 c5 15.Rad1 Rab8 16.Bxd4 cxd4
17.Nd5 Bd8 ½-½ M.Womacka-A.Krastev, Schwäbisch Gmünd 2017) 14...d5 15.f5 dxe4 16.Nxe4
Qh5 17.Ng3 Qe8 (or 17...Qh4!? 18.c3 Nc6, intending 19.Qb5 Bd6 20.Qxb7 Ne7, followed by ...Ned5
and ...e5-e4 with a strong initiative) 18.c3 Nb5 19.Bg5 Nd6 20.Rae1 Rd8 21.Qc2 Nf7 22.Bxf6 gxf6
23.Bg4 Rg8 24.Bf3 Qf8 25.Ne4 (not 25.Bxb7?? Bc5+ 26.Kh1 Rxg3 and wins) 25...Qh6 led to a draw
in M.Calogridis-G.Van Habberney, corr. 2015.

9...Kh8 10.Bxc6 bxc6 11.Qxc6

The first indication (at least to me) that this line might be okay for Black was when Boey used it to
draw with Grigory Sanakoev, the future World CC Champion. GM Ivan Sokolov wrote in 2009 “that
Black does not get enough compensation”, but results and newer engines beg to differ.

11...Rb8

11...Bd7 and 11...Bg4 have also been played successfully, the latter to a win in Nirvana 2-Pedone,
TCEC 16 League 2 2019: 12.Ne1 Rb8 13.Qc4 d5! 14.exd5 Rb4 15.Qd3 e4 16.Qe3 Rc4 17.Bd2 Bc5
18.Qg3 Bc8 19.Be3 Ba6 20.Bxc5 Rxc5 (0-1, 60), but Stockfish 13 rejects a lot of White’s moves in
that sequence.

486
 With the text Black applies pressure to b2 and plans ...Rb4, both attacking e4 and threatening to
trap the queen with ...Bb7. Things can easily go very wrong for White:
a) 12.b3? Rb4 13.Nd5 Nxd5 14.Qxd5 (or if 14.exd5, Mi.Bartel-M.Timmermans, Titled Tuesday
blitz 2021, then 14...e4! 15.Nd2 Bf6 16.Rb1 Bb7 or 15.Qc3 a5 16.Nd4 Rxd4 17.Qxd4 Bf6 wins)
14...Bb7 15.Qa5 Rxe4 16.Qxa7 Qc8 17.Ne1 c5! 18.Be3 Qc6 19.Qa5 d5 20.Nf3 d4 21.Bg5 Rg4
22.Bxe7 Qxf3 0-1 R.Torella-R.Fidel, Mar del Plata 2011.
b) 12.Rd1? Bg4 13.Rd3 Nd7 14.Nd5 Nc5 15.Ra3 Bxf3 16.gxf3? Ne6 was good for Black in
E.Schmittdiel-C.Strugnell, Augsburg 2020. Annotating this game in Chess Monthly, Strugnell gives
16.Rxf3 and 14.Be3 as improvements for White, but also a big earlier one for Black: 12...Rb4!
13.Nxe5 (or 13.Nd5 Nxd5 14.exd5 Bg4) 13...Bb7 14.a3 (the only move) 14...Bxc6 15.Nxc6 Qa8
16.Nxb4 Nxe4 17.Re1 Nxc3 18.Rxe7, and now Stockfish 13 enhances his analysis with 18...Rxf2!
19.Kxf2 Qf8+ 20.Bf4 Qxe7 21.bxc3 Qf6 22.g3 (or 22.Nd5 Qf7) 22...g5 and Black wins.
c) 12.Nd5 Bb7 13.Qc4 (not 13.Qxc7? Qxc7 14.Nxc7 Bxe4) 13...Nxd5 14.exd5 Qc8 15.Rd1?!
(instead, 15.Qe4 c6 16.c4 keeps it level; e.g. 16...cxd5 17.cxd5 Qf5 18.Qxf5 Rxf5 19.Rd1 e4 20.Nd4
Rxd5 21.Be3) 15...c6! 16.dxc6?! Bxc6 17.Ng5?! Qf5! 18.Be3? Bb5 19.g4 Qxf2+ 20.Bxf2 Bxc4 and
Black won, C.Terrazzoni-P.Dubois, Sautron 2013.
d) 12.Qc4 can be answered by 12...c6! and then:

487
 d1) 13.Qxc6?! is not a good idea for White: 13...Rb4 (threatening ...Bb7 again) 14.Nd5 Nxd5!
(better than 14...Bb7 15.Qc3 Rxe4 16.Nxe7 Qxe7 “with compensation” – Sokolov) 15.exd5 (or
15.Qxd5 Bb7) 15...Rxf3! (and better than 15...Bb7 16.Qc3 Rb5 17.Rd1 Bxd5, even though Black won
in J.Ljubarskij-J.Boey, European Seniors Ch., Saint Vincent 2002) 16.gxf3 Bd7 17.Qc3 Rh4,
threatening ...Qe8-g6, when Stockfish gives Black a clear advantage; e.g. 18.Re1 (or 18.f4 Bh3 19.f3
Bxf1 20.Kxf1 Rxh2 21.Kg1 Rh4 22.Be3 Qe8) 18...Qe8 19.Re4 Qg6+ 20.Kh1 Bh3 21.Rg4 Bxg4
22.fxg4 Qxg4.
d2) 13.Qd3

13...Qe8! (improving on 13...Nh5 14.b3 g5!? 15.Ne2 g4 16.Nd2 Bf6 17.Ba3 Rf7 18.Nc4 Be6
19.Qxd6 Bxc4 20.bxc4 Qxd6 21.Bxd6 Rb2 22.Rfc1, when Black shouldn’t really have enough,

488
G.Sanakoev-J.Boey, corr. 1978) 14.b3 (if 14.Nd2, as in G.J.De Boer-W.Bor, Waddinxveen 1979,
then 14...d5 15.exd5 cxd5 16.Nxd5 Nxd5 17.Qxd5 Qg6 with compensation; the mysterious 14...a5!?
is a Stockfish suggestion) 14...d5! (a very thematic move in this variation, hassling White in the centre
and/or opening it up for Black’s pieces) 15.exd5 (after 15.Bg5 d4 16.Bxf6 Rxf6 Black should be okay
with the bishops; while 16.Ne2 Nd7 17.Bxe7 Qxe7 18.Ng3 Nc5 19.Qd2 Rxf3! 20.gxf3 Ne6 gives
excellent play against the white kingside) 15...Bb4! (threatening ...e5-e4) 16.a3 Ba6 17.Qxa6 Bxc3
18.Rb1 e4 19.Ng5?! (19.Qxc6 Qxc6 20.dxc6 exf3 21.b4 is probably alright for White with three
pawns for a knight) 19...cxd5 20.Qxa7 d4 21.Qa6 e3! 22.fxe3 dxe3 23.Qe2 Nd5 24.Rxf8+ Qxf8 and
Black stands very well with the white queenside out of the game. Ares777-tsmenace, ChessWorld.net
2018, continued 25.Nf3 Re8 26.Kh1 Bf6 27.Qe1 Nc3 28.Ra1 e2 29.Bb2 Ne4 30.Bxf6 Qxf6 31.c3 g5
32.Ra2 g4 33.Qxe2 gxf3 34.gxf3 Ng3+ 35.hxg3 Rxe2 36.Rxe2 Qxf3+ 37.Rg2 Qxc3 and Black
eventually won in a technical queen vs. rook and pawn ending.
e) 12.Qa4, “going after another pawn”, is given an exclamation mark by Sokolov. Here too Black
can play 12...c6! (alternatively, 12...Rb7 13.b3 c6 14.Bd2 Qe8 15.a3 Bg4 16.Ng5?! Nh5 17.Qa6 Rc7
was good for Black in H.Gnirk-J.Tait, corr. 2003; while 13.Re1 c6 14.b3 Bg4 15.Re3 Nh5 16.Ne2
Rc7 17.Bd2 Qe8 18.Re1 h6 led to a draw in Shredder 13-Laser, CCC 2 Blitz Battle 2018 – but it
seems there’s no need to defend the a7-pawn) 13.Qxa7 (for 13.Qxc6?! see ‘d1’ above) 13...Bg4
14.Qe3 (recentralizing the queen; both 14.Nd2 d5 and 14.Ne1 d5 get the “0.00” approval; while after
14.Kh1 Bxf3 15.gxf3 Nh5 16.Qe3 Qd7 17.Qd3 Qh3 18.f4 Qxd3 19.cxd3 Nxf4 Black has enough for
the pawn) 14...d5!, creating great activity across the board, which appears sufficient for Black:

e1) 15.exd5 cxd5 16.Qxe5 (after 16.Nxe5 d4 17.Qe1 dxc3 18.Nc6 Qb6 19.Nxb8 Bd6, Black will
pick up a second piece; and 16.Ne2 Rb4 17.Ng3 Bxf3 18.Qxf3 Ng4 19.Qd3 Bc5 20.Nh1 Qh4 21.Qh3
Nxf2 also works out okay) 16...Bd6 17.Qg5 (or 17.Qe3 d4 18.Nxd4 Bc5) 17...h6 18.Qg6 Rb4 and
Black’s attack is worth three pawns; e.g. 19.Ne2 Bxf3 20.gxf3 Rh4 21.Ng3 Qd7 22.Nf5 Bxh2+
23.Kg2 Rh5 24.Rh1 Rxf5 25.Rxh2 Rxf3!? 26.Bxh6 gxh6 27.Kxf3 Ng4+ 28.Ke2 Qe7+ 29.Kd1 Nxf2+
30.Rxf2 Rxf2 31.Qxh6+ Kg8 32.Qg6+ Kf8 33.Qh6+ Ke8 34.Qh5+ with a draw.

489
 e2) 15.Nxe5 d4

16.Qe1 (or 16.Qg3 Bd6 17.Nb1 Rb5! 18.Nxc6 Qc7 19.e5 Qxc6 20.exd6 Be2 21.Re1 Ne4 22.Qh4
g5 23.Qh6 Rf6 and the queen is gone) 16...dxc3 (if 16...Qc7 then 17.Nd3 dxc3 18.Bf4) 17.Nxc6 Qc7
18.Nxb8 Qxb8 and we’re at “0.00” again; e.g. 19.b3 Bc8 (threatening ...Ba6) 20.Qxc3 Bd6 21.Bb2
Re8 22.Qd3 Bxh2+ 23.Kh1 Qf4 24.Bxf6 Qh6 (or 24...gxf6 25.Qb5 Re5 26.Qb8 Rh5 27.Qxc8+ Kg7
with a draw) 25.Bg5 Qh5 26.g4 Bxg4 27.Kg2 Qxg5 28.Kxh2 Re6 29.Rg1 Qh5+ 30.Kg2 Be2
31.Qd8+ Re8 32.Qd6 Qf3+ with a draw anyway, L.Weiss-K.R.Jensen, corr. 2020.
e3) 15.Qd3 Bxf3 16.Qxf3 Nxe4 17.Qe2 Nxc3 18.bxc3 Qa5 19.Bd2 Qa4 20.Qxe5 Bf6 21.Qg3 Qxc2
22.Be3 (or 22.Bf4 Rb2) 22...Qxc3 is just equal, A.Lukásová-L.Milde, corr. 2021.
And that’s the end of established theory and practice. However:
f) 12.a3!? is Stockfish’s choice, preventing ...Rb4, with the option to secure the b-pawn by pushing
it two squares.

490
 In view of the latter, the engine itself regards 12...a5!? as best, and if 13.Qc4 then 13...h6!? – slow
play despite the missing pawn. Are these moves a human would even consider?
I’d probably look at something like 12...Bb7 13.Qc4 (or 13.Qa4 d5 14.Nxe5 Bd6! 15.Nc6 Qe8
16.Nxb8 Bxh2+ 17.Kxh2 Qh5+ 18.Kg3 Qg4+ and draws) 13...c6 14.Re1 (if 14.Bg5 then 14...Nh5, or
14.Qd3 Qc8, intending ...Ba6) 14...Qe8 (or again 14...d5, though after 15.Qd3 d4 16.Ne2 c5 17.Ng3
Qc7 White seems very solid) 15.Qd3 Qg6 16.b4 Bc8!? 17.h3 Nh5 18.Kh2 Bd8 with ideas of ...Rb7-f7
and ...Bb6.
There are no practical examples of 12.a3 to draw on, so at the moment this is all vague conjecture,
but Black seems to have reasonable compensation for the pawn with the usual two bishops and
activity. Citation required.

E22: 7.Bc4

491
 This looks likely to become the new main line, having been recommended in recent works by GMs
Caruana and Swiercz, as well as FM Joshua Doknjas. Although it musters a mere 47 games in the
standard databases, four of those are from 2700+ super-GMs, plus one from AlphaZero, all on the
White side. Other engines have tested it too – in four games from the Top Chess Engine
Championships (TCEC). Remarkably, prior to all this, 7.Bc4 was hardly ever mentioned at all. As for
me, I’ve faced it no less than 15 times in the past three years (my score: W4, D11, L0).
By retreating the bishop to the a2-g8 diagonal, White prevents ...0-0 for the near future and
threatens Ng5 into the bargain.

7...Bg4

Pretty much the only move.

a) 7...Be7?! allows 8.Ng5 and if 8...Rf8 then 9.Nxh7!.
b) 7...h6 is well met by 8.Nh4!, aiming at the weakened light squares; e.g. 8...Bg4 9.f3 Bh5 10.g4
Bf7 11.Bxf7+ Kxf7 12.Qd3 g5 13.Nf5 Ne7 14.c4 with a clear advantage (Doknjas).
c) 7...Na5 doesn’t have the desired effect following 8.Be2!, the point being that after 8...Be7 Black
is still unable to castle due to 9.b4 Nc6 10.b5 Na5 11.Bd2 b6 (or 11...c6? 12.Qe1) 12.Bxa5 bxa5
13.Bc4; while 8...Be6 9.Ng5 Bg8 10.Nc3 Qd7 (or 10...Be7 11.f4!) 11.Nd5 c6, R.Villeneuve-P.Green,
corr. 2020, and then 12.Nxf6+ gxf6 13.Bg4 (Doknjas) is no more appealing.

8.h3

This must be played straight away if White is to reach the critical position below. Instead, 8.Be3
Qd7 9.Nc3 (or 9.h3 Bh5) 9...Be7 10.h3 Bh5 transposes to the lesser 10.Be3 variation; 8.Nc3 would
allow 8...Nd4; and 8.Nbd2 Qd7 9.c3 Be6 10.Re1 Be7 11.Qc2 Nd8 12.Qb3 0-0 was fine for Black in
pionroi-tsmenace, ChessWorld.net.2019.

492
8...Bh5

9.Nc3

Also best.
a) 9.g4?! Bg6 10.Ng5 Qe7 11.Ne6?? is not a good idea due to 11...Bf7 (Caruana); and even after
11.Nc3 h6 (or 11...0-0-0 12.Ne6 Rd7) 12.Ne6 Bf7 13.Nd5 Nxd5 14.exd5 Bxe6 15.dxe6 0-0-0, White
has just weakened the kingside.
b) 9.Qd3 can be met by 9...Nd7! (rather than 9...Qd7, when 10.a3! lands Black in the type of
depressing position that can arise after 7.a3 etc)

493
 and then:
b1) 10.Qc3 Be7 11.b4 Nb6 12.Bb3 a5 13.b5 Nd4! 14.Nxd4 exd4 15.Qd2 (not 15.Qxd4? Bf6 16.e5
Bxe5 17.Re1 Qe7 18.Qd2 0-0-0 19.f4 Bd4+ and Black keeps the piece) 15...a4 16.Be6 Bf6 17.Bb2
Qe7 18.Bd5 Nxd5 19.exd5 0-0 20.Bxd4 Qe4 21.Bxf6 Rxf6 22.f3 Bxf3! 23.Nc3 (if 23.Rxf3 Rxf3
24.gxf3 then 24...Qe5 25.Nc3 Qg3+ 26.Qg2 Qe5 or 25.c3 Rf8) 23...Qh4 24.Rae1 Bh5 25.Rxf6 Qxf6
led to a draw, lysol-tsmenace, ChessWorld.net 2019.
b2) 10.Nc3 Be7 11.Nd5 (after 11.Qe3 Nb6 12.Be2 Bg6 13.Qd3 0-0 14.Be3 Kh8 15.a4 Nd7 16.a5
a6 17.Qc4 Nf6 Black’s lesser board room isn’t really an issue; e.g. 18.Nd2 Bf7 19.Nd5 Nd4 20.Bxd4
exd4 21.Qxd4 Nxd5 22.exd5 Bf6 23.Qc4 c6 24.Qb4 Bxd5 25.Qxb7 Bd4 26.Bf3 Rb8 27.Qxa6 Bxf3
28.Nxf3 Rxf3 29.gxf3 Qg5+ and draws, K.Kitson-J.Tait, corr. 2020) 11...Nb6 12.Bb3 (or 12.Nxb6
axb6, followed by ...Bf6, ...Qe7 and ...Bf7) 12...Nxd5 13.exd5 (or 13.Bxd5 Qd7 and ...Nd8) 13...Bxf3
14.gxf3 (or 14.Qxf3 Nd4) 14...Nb8 15.Ba4+ Nd7 16.Qf5 Qc8! 17.c4 Kd8 and ...Nf6 seems okay for
Black.

9...Qd7

Against other moves 10.g4 is strong, as Caruana indicates: 9...Nd4?! 10.g4 Nxf3+ 11.Qxf3 Bg6,
F.Caruana-L.Aronian, London Classic (blitz) 2018, and 12.g5! Nd7 13.h4; or if 9...Be7, as in J.Van
Foreest-Ma.Carlsen, Chess24.com (blitz) 2020 (by transposition), then 10.g4! Bg6 11.Ng5 Nd4 12.f4
causes persistent problems.
The cautious 9...Qd7 negates the threat, since 10.g4?! now runs into 10...Nxg4! 11.Nd5 (the only
move) 11...Nf6 12.Nxe5! Qxh3!? (or just 12...Bxd1 13.Nxd7 Kxd7) 13.Nxc7+ Kd8 14.Ne6+ Ke7
15.Nxc6+ bxc6 16.Nf4 Bxd1 17.Nxh3 Bxc2 and Black is not worse.

10.Nd5!

494
 GM David Navara’s idea and recommended by GM Swiercz.
Caruana and Doknjas both go for 10.Be3, which is less testing because Black can aim to castle
short after 10...Be7 and then:

a) 11.g4 Nxg4 12.hxg4 Qxg4+ 13.Kh2 Nd4 (not 13...Qxf3?? 14.Be2) 14.Bxd4 Qf4+ 15.Kg2 Qg4+
16.Kh2 is a draw.
b) 11.Be2 Bg6 12.Nh4 Bf7 13.Nf5 is given without comment in the accompanying database to
Caruana’s video. It’d be interesting to know how he proposed to proceed after 13...0-0, as Black can
reorganize with moves like ...Kh8, ...Rhe8, ...Bd8, and ...a7-a6 inserted if required. I’ve played this
out six times (W2 D4 L0) and was never really in any trouble. For example: 14.Qd2 (or 14.Qd3 Kh8
15.Nxe7 Qxe7 16.Rad1 Qd7 17.b3 a6 18.Kh2 Be6 19.Qd2 Rf7 20.f3 Raf8, Capalaskine-tsmenace,
ChessWorld.net 2020) 14...Kh8 15.g4!? Rae8 16.Rad1 Bd8 17.b3 a6 18.Bf3 Be6 (an “old rules”
bishop, moving one square at a time) 19.Bg2 h5 20.f3 Nh7 21.Ng3 g6 22.Nd5 Bh4 23.Kh2 hxg4
24.fxg4 Nd8, Capalaskine-tsmenace, ChessWorld.net 2021. No doubt I’d find it more difficult
against Caruana himself.
c) 11.a4 is what Caruana has actually played, if only a blitz game. Black should still aim to castle
short.

495
 The question is how best to accomplish that:
c1) 11...Rf8?! 12.Be2! (threatening Nxe5 again) 12...Bxf3 13.Bxf3 Kf7 14.Nd5 Kg8 and Black has
castled short, at the cost of the light-squared bishop and time. White now has a dominant position, as
Caruana demonstrated: 15.a5 a6 16.c3 Kh8 17.Qb3 Nd8 18.Rad1 Nxd5 19.Bg4 Qb5 20.Qxd5 Qxb2
21.Qc4 c6 22.Rb1 Qa3 23.Bb6 Nf7 24.Ra1 Qb2 25.Rfb1 d5 26.Qd3 dxe4 27.Qd7 Qxc3 28.Qxe7 and
wins, F.Caruana-Ma.Carlsen, St. Louis (blitz) 2019.
c2) 11...h6! (Swiercz) is much better. This can transpose to ‘c3’ below after, say, 12.a5 a6 13.Re1;
while 12.Nd5 Nxe4 13.Nxe5 Bxd1 14.Nxd7 Bxc2 15.Nxc7+?! Kxd7 16.Nxa8 Nc5 17.Rac1 Bd3
18.Rfd1 Bxc4 19.Rxc4 Ne6 worked out well for Black in Aa.Rodriguez-N.Anastasovski, Montreal
2005.
c3) 11...a6 12.Re1

496
 12...h6! (my own small novelty, preparing ...Bd8, ...Bf7 and ...0-0; not yet 12...Bd8?! because of
13.b4!, intending 13...Nxb4 14.Qb1! c5 15.Ng5 with a strong initiative; and the attempted concrete
solution 12...Nb4?! 13.g4! Bxg4? doesn’t work: 14.hxg4 Qxg4+ 15.Kf1 Nxc2 16.Qxc2 Qxf3 17.Qd1
Qh3+ 18.Ke2 and White won again, R.Hovhannisyan-L.D.Nisipeanu, Dubai 2012) 13.Nd5 (now if
13.b4?! Nxb4, there is no Ng5; while after 13.Rb1 Nisipeanu’s idea of 13...Nb4! is feasible, since
14.g4 fails to 14...Nxg4 15.hxg4 Qxg4+ 16.Kh2 Bg5!; finally, if 13.a5 then 13...Bd8 and ...Bf7 etc)
13...Bf7 and Black will castle next move with a perfectly sound position.
For example: 14.c3 (or 14.Qe2 0-0 15.Ra3!? Kh8 16.Nxf6 Bxf6 17.Bxf7 Qxf7 18.Rd1 a5, followed
by ...b7-b6, ...Nd8-e6, ...Rae8 and so forth, doclolly-tsmenace, ChessWorld.net 2018) 14...0-0 15.Qd3
(or 15.Qb1 Kh8 16.b4 Bg6 17.Bd3 Nd8 18.Nh4 Bh7 19.Nxe7 Qxe7 20.Nf5 Qf7, E4DOG-tsmenace,
ChessWorld.net 2021) 15...Kh8 16.Nh2 Bg6 17.Nxf6 Bxf6 18.b4 Ne7 19.Ng4 Rad8 20.Bb3 d5 21.f3
Qe8 22.Qe2 dxe4 23.fxe4 Nc8 24.Bc5 Be7 25.Be3 and ½-½ DragonLore1994-tsmenace,
ChessWorld.net 2020.
The immediate 10.Be2 (threatening guess what)

497
 10...Bg6 is much the same, and 11.Nh4 Bf7 12.Be3 Be7 just transposes to note ‘b’. Instead,
Ethereal 11-Stockfish, TCEC 16 VSO 2019, saw 12...a6 13.Bg5 Be7 14.Nf5 0-0 15.Qd2 Be6
16.Nxe7+ Nxe7 17.f4 exf4 18.Rxf4 Ng6 19.Rf2 Qc6 20.Raf1 Nd7 21.Nd5 Rxf2 22.Rxf2 Rf8
23.Rxf8+ Kxf8 and Black had no trouble drawing.
Finally, 10.Qd3 is mistimed in view of 10...Nb4! 11.Qe2 Bxf3 12.Qxf3 Nxc2 13.Rb1 Nd4 14.Qd3
c6 with ideas of ...b7-b5 or ...0-0-0 and ...d6-d5, when White will have to justify the sacrifice.

10...0-0-0

Basically forced, as GM Swiercz says – which is annoying because Black would much rather castle

498
short. Unfortunately, 10...Be7? loses material: 11.Nxe5! Bxd1 (or 11...Nxe5 12.Nxf6+ Bxf6
13.Qxh5+ g6 14.Qe2 – Swiercz) 12.Nxd7 Nxd5 13.exd5 Bxc2 14.dxc6 bxc6 15.Be6. And 10...Nxd5?
11.exd5 Nd4 12.Nxe5! Bxd1 13.Nxd7 is little different: 13...Bxc2 14.Nxf8 Rxf8 15.Re1+ Kd7
16.Be3 Nf5 17.Bb5+ c6 18.dxc6+ bxc6 19.Bxc6+ Kxc6 20.Rac1.

11.Qd3

Defending the e4-pawn and breaking the pin. Black responds with useful, and probably necessary,
prophylaxis.

11...Kb8 12.a4

As played in all games so far bar one:

a) 12.Bd2 Be7 13.b4 was seen in D.Navara-S.Azarov, Czech League 2016, where 13...Bxf3
14.Qxf3 Nd4 15.Qd3 Nxd5 16.exd5 (or 16.Bxd5 c6 17.Bb3 d5) 16...g5 17.Qe4 (17.c3 Nf5) 17...h5
18.Bd3 Rdf8 seems okay for Black.
b) 12.Bb5 can also be met by 12...Bxf3, when 13.gxf3 (or 13.Nxf6 gxf6 14.Bxc6 Qxc6 15.Qxf3 d5)
13...a6 (rather than 13...Qxh3?! 14.Bxc6 bxc6 15.Nb4) 14.Ba4 (or 14.a4 Qe8) 14...b5! 15.Bxb5 axb5
16.Qxb5+ Ka7 17.Be3+ (or 17.c3 Nxd5 18.exd5 Nb8 19.Be3+ Ka8) 17...Nd4 18.Nxc7! Qxb5
19.Nxb5+ Kb7 20.Nxd4 exd4 21.Bxd4 Be7 is roughly equal. The kingside weaknesses balance out
White’s four pawns for the knight.

12...Bg6

Swiercz regards this as best: “By pinning the e4-pawn against the queen, Black aims to oblige
White to recapture on d5 with a pawn.” Instead:

499
 a) 12...Bxf3?! is no longer as effective: 13.Qxf3 Nd4 14.Qd3 c6 15.Nxf6 gxf6 16.c3 d5 17.exd5
cxd5 18.Ba2 Nc6 (or 18...Ne6 19.Qf3) 19.Rd1 and White is clearly better (Swiercz).
b) 12...Be7?! 13.Bb5 (Swiercz) is also good for White; e.g. 13...Bxf3 14.gxf3 Qxh3 15.Bxc6 Nxd5
16.exd5 (or 16.Bxd5! c6 17.Ba2 Rdf8 18.f4) 16...bxc6 17.Ra3 Qh5 18.Re1 Qg6+ 19.Qxg6 hxg6
20.dxc6 Ka8 21.Rb3 with an edge that was converted in C.Deakin-L.Milde, corr. 2021.
c) 12...Ne7!? is a possible alternative: 13.Ng5 Bg6 14.Nxf6 gxf6 15.Nf7 Bxf7 16.Bxf7 gives White
“the bishop pair and good attacking prospects on the queenside” (Swiercz), but White has found it
difficult to win against steadfast resistance:

16...h5 (after 16...d5!? 17.exd5?! Nxd5 18.Bxd5 Qxd5 19.Qxd5 Rxd5 20.Be3 Bd6 21.Rfd1 Rxd1+
22.Rxd1 h5 23.c4 b6 Black was fine in canstein-tsmenace, ChessWorld.net 2019; but either 17.Qf3 or
17.Rd1 improves for White) 17.a5 a6 18.Qb3 Nc6 19.Be3 Bh6 20.Be6 Qg7 21.Bd5 Qd7 and while
Stockfish concedes the advantage, White was unable to get past its defences in Alpha-Zero-Stockfish
8, London (match) 2018. The same applied to 17.Qb3 a6 18.Rd1 Bh6 19.Bxh5 Bxc1 20.Bg4 f5
21.exf5 Bg5 22.f6 Nf5 23.Qd3 d5 24.Qxf5 Qxf5 25.Bxf5 Bxf6 26.g3 e4 27.c3 c6 in
W.Krzyzanowski-K.Kögler, corr. 2018.

13.Re1

Not 13.Nh4?! due to 13...Bxe4 14.Nxf6 gxf6 15.Qxe4 d5 and Black is okay (Swiercz).

13...Nxd5 14.Qxd5

Nor 14.Bxd5 Nb4 15.Qb3 Nxd5 16.exd5 Be7 17.Be3 Rhf8 18.Nd2 Bf5 with counterplay (Swiercz).

14...Be7 15.Ra3

500
 Now Black has to... well... have prepared their moves at home.

15...Nb4! 16.Qd2 Rhf8! 17.Qxb4

If 17.Rb3 c5 18.c3 Nc6 19.Bd5 then 19...Bh5! 20.Ng5 (or 20.Nh2 Bh4) 20...Rf6, again with
counterplay.

17...d5 18.Qc3

After 18.Qa5 dxc4 19.Nxe5 Qe6 20.Nxg6 hxg6! (not 20...Qxg6? 21.Rg3) 21.Rae3 b6 22.Qc3 a5
23.Qxg7 g5 24.Qh7 Bc5 25.R3e2 g4 26.hxg4 c3 27.b3 Qxg4 28.Qh3 Qg6, Black had sufficient
activity to draw in C.Schakel-K.Reinhart, corr. 2020.

18...Bxa3 19.Bb5 c6 20.Nxe5 Qd6 21.bxa3 cxb5 22.Nxg6 Qxg6 23.axb5 Qb6

If 23...dxe4 24.Be3, White has “a long-term initiative and basically risk-free position” (Swiercz),
which is no doubt true, though 24...Qf6 25.Qb4 g5 26.Qxe4 Rfe8 27.Qb4 Qf5 led to a draw anyway
in T.Mosconi-G.Ziese, corr. 2020.

24.Qg3+ Ka8 25.exd5

Instead, 25.e5 gave rise to an unexpected repetition: 25...Qa5 26.Rf1 (otherwise, 26.c3 d4 and
26.Qe3 Rc8 and 26.Rd1 Qxb5 are all about equal) 26...Qc7 27.Re1 Qa5 28.Rf1 Qc7 ½-½ A.Faridani-
M.Hrubcík, corr. 2019.

25...Qxb5 26.d6 Qc6 27.Bf4

501
 Here Swiercz concludes: “This looks very good for White. His advantage is cemented by having
two pawns for the exchange, including the powerful and well-supported d6-pawn.”
In response, Stockfish throws out 27...g5! 28.Be5 (not 28.Qxg5?? Rg8 or 28.Bxg5?! Rxd6)
28...Qxc2 29.Qe3 Qc6 and it no longer looks quite so good. White is down to one pawn for the
exchange, which means that Black can, if necessary, await a convenient moment to take on d6 and
reach a drawable endgame; e.g. 30.Rc1 (or 30.Rd1 Rfe8 31.Qd4 Re6) 30...Qd5 31.Rc5 Qd1+ 32.Kh2
Rxd6 33.Bxd6 Qxd6+ 34.Qe5 Qxe5+ 35.Rxe5 h6 36.Re6 Rh8, when counterplay with the queenside
pawns should be sufficient.
All the same, Black is having to work for the draw after 10.Nd5, so it can certainly be
recommended as a practical weapon for White. From the other side: fair enough, sometimes that’s
just how it is.
while 6.Nxe4 d5 is the same as 4.Nc3 fxe4 5.Nxe4 Nf6 6.d3 d5 and also fine; e.g.

502
 Chapter Eleven
Jaenisch Gambit with 4.Nc3

1.e4 e5 2.Nf3 Nc6 3.Bb5 f5 4.Nc3

The traditional main line, reacting to Black’s aggression by accelerating development.

4...fxe4

Removing the e-pawn was the point after all. I’ve only played 4...Nf6 and 4...Nd4 in thematic
tournaments – with mixed success as Black, but a massive plus score from the White side.

5.Nxe4

One practical bonus of this variation is that White sometimes only half-remembers the moves and
plays them in the wrong place or order. For example, knowing that Nxe5 is the standard idea can lead
to things like 5.Bxc6 (or even 5.Nxe5?? Nxe5 6.Qh5+ Ng6 and Black has won a piece, A.Donatti-
F.Vivas Zamora, Titled Tuesday blitz 2020) 5.Bxc6 bxc6 6.Nxe5? (for 6.Nxe4 d5 see below) 6...Qg5!
7.Ng4 d5 8.h3 (otherwise 8.Ne3 d4 wins a piece) 8...h5 9.d4 Qg6 10.Ne3, N.Oyvind-Gerd.Meyer,
Bad Ems 2004, and Black has a superb position (centre, bishops, activity, initiative) for nothing.

5...d5

Back in 1982/83 S&S recommended 5...Nf6 “!”, giving 5...d5 a dismissive “?!”. I could never see
what was wrong with it, certainly not at my level of play. Then came the game Timman-Speelman,
Candidates 1989 (0-1, 30). Someone asked me what the improvement was; I replied that Black didn’t

503
need one.
Nonetheless, grandmasters mostly go along with 5...Nf6, and if 6.Nxf6+ Qxf6 7.Qe2 Be7 8.Bxc6
bxc6 9.Qxe5 (or 9.Nxe5 0-0, or 9.d4 Qg6) 9...d6 10.Qxf6 Bxf6, analysis and praxis has shown that
Black has very decent drawing chances, if you like that sort of thing. Bulgarian GM Ventzislav
Inkiov is a notable exception, a dedicated adherent of 5...d5.
After the text, White has two main replies:
A: 6.Ng3
B: 6.Nxe5
White has two minor options:
a) 6.Bxc6+ bxc6 7.Ng3 Bg4 8.h3 Bxf3 9.Qxf3 transposes to 7.Bxc6+ in line A.
Another possibility is 7...e4!? 8.Ne5 Bd6 9.Nxc6 (or 9.Qh5+ g6 10.Nxg6 Nf6 11.Qh6 Rg8,
N.Checa-B.Nikel, corr. 2009) 9...Qf6!? (rather than 9...Qh4 10.d3!) 10.Qh5+ Qf7 11.Qxf7+ Kxf7,
when Stockfish thinks Black is okay even with queens off.
b) 6.Nc3 is inferior since d5 easily defended.

After 6...Bg4 7.h3 (instead, 7.Qe2 Bxf3 8.Qxf3 is much the same; while 8.gxf3!? can be met by
8...Kf7! 9.Nxd5 Nd4 10.Qc4 c6 or 9.Bxc6 bxc6 10.Qxe5 Bd6 11.Qf5+ Nf6 – I&K; or if 8.Qxe5+ Kf7
9.Bxc6 bxc6 10.Qf4+, J.Möller-F.Englund, Christiania 1903, then 10...Qf6! 11.Qxc7+ Be7 12.gxf3
Nh6 or 11.Qxf3 Qxf3 12.gxf3 Re8+ with more than enough counterplay) 7...Bxf3 8.Qxf3 Nf6 9.0-0
a6 (or 9...Bc5 10.Qg3 Qe7 11.d3 0-0 12.Bg5, E.Rozentalis-P.Kolosowski, Polish League 2003, when
Sokolov gives 12...Nd4! 13.Nxd5 Nxd5 14.Bxe7 Ne2+ 15.Kh2 Nxg3 16.Bxc5 Nxf1+ 17.Rxf1 Rf6
and “with an exchange up, Black is obviously better”, though White does have two bishops)
10.Bxc6+ bxc6 11.Qg3 (or 11.Qf5 Bd6 12.d4 exd4 13.Re1+, Wei Sin Tan-J.Tait, Sheffield League
1998, and why I played 13...Kf8? rather than 13...Kf7, I’ve really no idea) 11...Bd6 12.d4 (here

504
12.Qxg7?! Rg8 13.Qh6 Rg6 14.Qh4 Kd7 would be very greedy) 12...Qe7 13.dxe5 Bxe5 14.Qd3 0-0
Black is fine, T.Ratcu-A.Souza, Sao Paulo 1996.

A: 6.Ng3

This simple retreat is a reasonable choice. White generally obtains the two bishops and can play
against the centre. Indeed, the e5-pawn needs to be secured and Black must be precise in order to
equalize.

6...Bg4

The strongest reply, developing with a threat (...e5-e4) while preparing to castle queenside as
quickly as possible. Note that 6...Bd6? fails to 7.Nxe5! Bxe5 8.Bxc6+ bxc6 9.Qh5+ etc; while
6...Qd6?! is a suspect move order because of 7.0-0 Bg4 8.Re1! (rather than 8.h3 etc) 8...Nge7 9.d4,
and if 9...0-0-0?, P.Sciacca-J.Friedel, New Hampshire 1998, then 10.dxe5! Nxe5 11.Rxe5 Bxf3
12.Qd4! wins. Finally, if 6...e4!? 7.Nd4 Qf6 (not 7...Qd6?! 8.d3) 8.Qh5+ Ke7, Stockfish proposes
9.Nb3!? and claims an edge, albeit in a complicated position.

505
7.h3

Duly gaining the two bishops. Instead:

a) 7.Bxc6+ bxc6 8.h3 Bxf3 9.Qxf3 is not an improvement because Black can now go short: 9...Bd6
10.d3 (if 10.Qc3?! here then 10...Ne7!, or 10.Nf5 Qd7) 10...Nf6 11.Bg5 0-0 12.0-0 (or 12.Nh5? Nxh5
13.Qxh5 Qb8!, unklesubz-jatait47, Chess.com 2020) 12...Qe8 and Black already stands well, with the
threat of ...Ne4, H.Reussner-P.Leisebein, corr. 2001.
b) 7.d4 e4 8.h3 can be answered by 8...exf3 (or just 8...Bxf3 9.gxf3 exf3 10.Qxf3 Qe7+) 9.hxg4
fxg2 10.Rg1 (or 10.Qe2+ Kd7) 10...Qd7 (rather than 10...Qd6 11.Qe2+ Nge7, L.Shusterman-
V.Agzamov, USSR 1966, when 12.c3 0-0-0 13.Rxg2 Ng6 prompts 14.Nf5; while 10...Qe7+ 11.Qe2
Qxe2+ 12.Kxe2 0-0-0 13.Be3 gives White an edge with the two bishops, V.Kupreichik-A.Barreras,
Plovdiv 1981) 11.Qe2+ Nge7, and now if 12.c3 (or 12.Be3 0-0-0) 12...0-0-0 13.Rxg2 then 13...Ng6
and ...Bd6 targets the f4-square.
c) 7.Qe2 allows 7...Bd6, when 8.d4 (or 8.c4 a6) 8...e4 (not 8...Qe7? 9.dxe5 Bxf3, B.Gurgenidze-
A.Hennings, Kislovodsk 1972, in view of 10.exd6! Bxe2 11.Bxc6+ bxc6 12.dxe7 and White is clearly
better – I&K) 9.h3 (not 9.c4? Ne7) 9...Bd7 10.Bxc6 (not 10.Ne5? Nxd4) 10...bxc6 11.Ne5 Nf6 is
equal (I&K).
d) 7.0-0 sets White on the back foot after 7...Qf6! (not 7...e4?! 8.Re1 Nge7 9.h3 which turns out
well for White, and 8...Kf7? 9.Nxe4! just wins)

506
 8.d3 (White already needs to think about equalizing; other moves are worse: 8.Re1 0-0-0 9.Bxc6
bxc6 10.Qe2?! e4 11.h3 Bxf3 12.gxf3 exf3 13.Qa6+ Kd7 14.c4 d4, V.Varavin-V.L.Ivanov,
Chelyabinsk 1991; if 8.Bxc6+ bxc6 9.d3 Ne7 10.Re1, E.Geller-Mi.Tseitlin, Moscow 1992, then
10...Ng6! intending 11.Bg5 Bxf3 12.Bxf6 Bxd1 13.Bxe5 Bxc2 14.Bxg7+ Kd7! 15.Bxh8 Nxh8
16.Re3 Bh6 – I&K; or 8.c4 Bxf3 9.Qxf3 Qxf3 10.gxf3 0-0-0 11.Bxc6 bxc6 12.d3 Ne7, M.Smith
Hansen-L.Dobrovolsky, Tastrup 1992; or 8.Be2 e4 9.Ne1 Bxe2 10.Qxe2 h5 11.d3 h4 12.Nh5 Qf5
13.Nf4 g5, K.Richardson-R.Mallee, corr. 1993; White scored just ½/4 in these games) 8...Bxf3
9.Qxf3 Qxf3 10.gxf3 Bd6 11.Nf5 Nge7 12.Nxd6+ (12.Nxg7+?! Kd7 13.Kh1 Raf8 is not good for
White) 12...cxd6 13.f4 0-0 14.fxe5 dxe5 15.Be3 Rf6 16.Ba4 Nf5 17.c3 h6 18.Kh1 g5 and Black is
fine with the knights and stronger pawns, F.Colombo Berra-H.D.Vötter, corr. 2018.

7...Bxf3 8.Qxf3

507
8...Qd6!

IM Josef Boey’s improvement. Hitherto 8...Nf6 was routinely played, but then 9.Bxc6+ (either 9.0-
0 Bd6 10.Nh5 e4 or 9.Nh5 Qd6 10.Nxf6+ gxf6 11.Qh5+ Kd7 is easier on Black) 9...bxc6 10.Qc3 Qd6
11.0-0 is trickier than it looks, even if White slow plays the position: the a7-pawn is a permanent
weakness and the black king won’t be wholly happy wherever he goes; e.g. if 11...Kd7 12.d3!? (12.d4
e4 13.Qa5 Qb4! is no problem) 12...g6 13.Bg5 Nh5, K.Rjabzev-V.Knyazev, Samara 2002, then
14.Nxh5 gxh5 15.Rae1 Re8 16.f4 Qc5+ 17.Kh2 Qxc3 18.bxc3 exf4 19.Rxe8 Kxe8 20.Bxf4 and
White’s edge continues into the endgame.

508
9.0-0

White has also tried:

a) 9.c3 0-0-0 (there’s no need to get involved in 9...e4?! 10.Qh5+ g6 11.Bxc6+ bxc6 12.Qe2 Ne7
13.0-0 Bg7 14.d3 exd3 15.Qxd3, where White is at least a little better, W.Smiley-J.Ryska, corr. 2000)
10.0-0 transposes to the main line.
b) 9.Bxc6+ can be answered by 9...Qxc6 here, and if 10.Qh5+ then 10...Kd7; e.g. 11.Qf5+ (or 11.0-
0 Re8) 11...Qe6 12.Qxe6+ Kxe6 13.d3 Bd6 14.Ne2 Ne7 15.Bd2 Rhf8 and Black is well set up behind
the strong centre, M.Hogarth-J.Tait, Scarborough 1994.
c) 9.Qh5+!? g6 10.Bxc6+ bxc6 11.Qe2 shouldn’t trouble Black, though 11...Ne7 12.b3 Bg7 13.a4
(13.Bb2 0-0 14.0-0-0? is just asking for it; i.e. 14...a5 15.a4 c5) 13...e4 14.Ba3 Qe6 15.Rd1 Nf5
(Stockfish offers 15...h5!? as an improvement) 16.0-0 Nd4 17.Qa6 Nxc2 18.Bc5 Nd4 19.d3 worked
out well for White in V.Kupreichik-I.Yagupov, Groningen 1995.
d) 9.c4 is overexcitable: 9...0-0-0 10.Bxc6 (not 10.cxd5?? Nd4, or 10.Nf5?! Qf6 11.cxd5 Nb4!)
10...Qxc6 11.cxd5 Qxd5 12.Qxd5 Rxd5 13.Ke2 Ne7 14.d3 Nc6 15.Be3 Bc5 and Black is fine,
V.Akopian-A.Kuzmin, USSR Team Ch. 1990.
e) 9.Nf5 doesn’t get anywhere: 9...Qe6 10.Qg4 (if 10.c4 then 10...Nf6, or 10.0-0 0-0-0) 10...Kf7
11.Bxc6 (now if 11.c4 then 11...d4, while 11.Ne3 Nf6 12.Qxe6+ Kxe6 13.0-0? Nd4 14.Ba4 b5
15.Bb3 a5 16.c3 Nxb3 17.axb3 d4 gave Black a strong initiative in H.Camilleri-J.Boey, Nice
Olympiad 1974) 11...Qxc6 12.d4 Nf6 13.Qe2 Bb4+ 14.c3 (not 14.Kf1?! Rhe8 15.dxe5? Rxe5
16.Qxe5 Re8 and wins, R.Agg-J.Tait, corr. 2006) 14...exd4 15.cxb4 Rhe8 16.Be3 dxe3 17.0-0 and the
best to be said is that White may be okay.

9...0-0-0

509
10.c3

Preventing ...Nd4. White can then continue with d2-d4 or d2-d3 as desired.
The alternative 10.Bxc6 Qxc6 11.Qf5+ Kb8 12.Qxe5 Qxc2 is fine for Black; e.g. 13.d4 (or 13.Nf5
Nf6 ½-½ A.Serbinenko-J.Tait, corr. 1995 – premature, as after 14.d3 Qxd3 15.Bf4, B.Abramovic-
S.Tatai, Italy 1981, and 15...Rd7 16.Rad1 Qe4 it’s up to White to prove the draw) 13...Nf6 14.Qf5
Qxf5 (or 14...Qc4!? 15.Bg5 Qxd4 16.Nh5, M.Smith Hansen-E.Bentzen, Danish League 1990, as long
as 16...h6! follows) 15.Nxf5, G.Sigurjonsson-S.Tatai, Barcelona 1980, when 15...Rd7 16.Bf4 Kc8
17.Rac1 c6 is solid enough.

10...e4

Having tried nearly everything here, I’ve decided this direct response is best.
Instead, 10...Nf6 11.d4 exd4 (or 11...e4 12.Qf5+ Qd7 13.Bg5 – Sokolov) 12.Nf5 Qc5 13.Bxc6
Qxc6 14.Nxd4, I.Smirin-V.Tukmakov, Lvov Zonal 1990; and 10...g6 11.d4 e4 (not 11...exd4? 12.Bf4
Qc5 13.Bxc6 Qxc6 14.Be5 – I&K) 12.Qg4+ Qd7 13.Qxd7+ Rxd7 14.f3 Bd6 15.Ne2, V.Palciauskas-
J.Boey, corr. 1984, are both good for White.
The final option 10...Kb8!? pre-empts a check, giving Black more flexibility after 11.d4 e4; e.g.
12.Qf4 Qe6 13.Qf5 Qe8 14.Bf4 was a quick draw in D.Svetushkin-L.D.Nisipeanu, French League
2011. White might therefore prefer 11.d3 and attack on the queenside; e.g. 11...g6 12.Be3 Bg7 13.b4
Nge7 14.Bc5 Qf6 15.Qd1 Nc8 16.Qc2 N6e7 17.c4 c6 18.Ba4 Nf5 19.Rab1 with an ongoing initiative,
afms-tsmenace, ChessWorld.net 2010.

11.Qf5+

Queen retreats give Black time for ...a7-a6, putting an awkward question to the bishop. Playing Be2

510
renders it at least temporarily passive, Ba4 relinquishes control of d3 (which S&S thought was an
issue for Black), and Bxc6 gives up the two bishops. For example:
a) 11.Qe2 a6 12.Bxc6 (or 12.Ba4 Ne5!) 12...Qxc6 13.d3 exd3 14.Qxd3 Nf6 15.Bg5 (or 15.Re1 Bc5
16.Be3 Rhe8, H.Kotz-L.Dobrovolsky, Oberwart 1991; while if 15.Be3 then 15...g6) 15...Bc5 16.b4
Bb6 17.Nf5 Rd7 18.a4 Qc4 19.Rad1 Qxd3 20.Rxd3 Ne4 was roughly equal in Y.Balashov-
A.Kuzmin, Moscow 1989.
b) 11.Qe3 a6 12.Be2 g6 13.d3 Bh6 14.f4 exf3 15.Qxf3 Bxc1 16.Raxc1, E.Janev-J.M.Degraeve,
Saint Affrique 2008, and now 16...Re8 looks safest, clearing the way for ...Nd8 and ...c7-c6 if needs
be.
c) 11.Qd1 Ne5? (11...a6) 12.d4! exd3 13.Bf4 shows why Black should insert ...a7-a6. White is
clearly better here, as in V.Tsarev-A.Karpachev, Moscow 1988.

11...Kb8

12.d3

If 12.d4 then 12...Nce7 (or 12...Nh6 13.Qg5 Nf7 14.Qf5 Nh6, repeating in bobagem-tsmenace,
ChessWorld.net 2010) 13.Qf7 Nf6 14.c4 dxc4 15.Bxc4 Ned5 16.Bxd5 Qxd5 17.Qxd5 with a draw in
F.Garscha-H.D.Vötter, corr. 2005.

12...Nce7!

Improving on 12...g6 13.Qf4 exd3 14.Qxd6 Bxd6 15.Bxd3, where GM Ivan Sokolov writes that
“thanks to his bishop pair White had a small, long-term advantage and ultimately won” in
An.Sokolov-Mi.Tseitlin, Moscow 1983.

511
 After 12...Nce7! Sokolov says “the position is messy, but Black is not worse. On 13.Qf4 there is
13...Qb6.” This had indeed been demonstrated in several games prior to his book. (I’m almost the
only person to lose as Black.)
a) 13.Qf4 Qb6 (even 13...Qxf4 14.Bxf4 a6 15.Ba4 exd3 16.Rad1, B.Lepelletier-G.Grimberg,
French Ch., Toulouse 1995, and 16...b5 17.Bb3 Ng6 18.Bg5 Be7 19.Bxe7 N8xe7 20.Rxd3 c5 looks
okay) 14.Ba4 Ng6 15.Qe3 Bc5 16.Qg5 Be7 17.Qe3 Bc5 18.Qg5 offered a repetition (which Black
declined) in B.Hanison-F.Titzhoff, corr. 2002.
b) 13.Qg5 Qb6 14.Ba4 (if 14.a4 a6 15.Be3 Qg6 16.Qe5 Qd6 17.Qxd6 Rxd6 18.dxe4 axb5 19.axb5
Nc8 20.Ra4 Nge7 21.Rfa1 b6, White is struggling to prove full compensation, Y.Dothan-H.D.Vötter,
corr. 2005; in this line 15.dxe4!? axb5 16.axb5 h6 17.Qf4 Nc8 18.exd5 Nf6 may give White enough
for the piece, but no more than that) 14...exd3 15.Qe3 Qa6 16.Bb3 Nf6 17.Rd1 Ng6 (not 17...Nd7?
18.Qxd3 Qxd3 19.Rxd3 Nc5 20.Rd1 Nxb3 21.axb3 a6 22.Be3 with a plus as it’s hard for Black to
mobilize effectively, F.Lunek-J.Tait, corr. 2003) 18.Qxd3 Qb6 (or 18...Qxd3 19.Rxd3 Bc5 20.Bg5 ½-
½ A.Junker-B.Hanison, corr. 2001) 19.Nf5 Bc5 20.Be3 Ne4 21.Bxc5 Nxc5 22.Qe3 Rde8 23.Qd4
Rhf8 24.Re1 Rd8 25.Ne3 Nxb3 26.Qxb6 axb6 27.axb3 d4 led to a draw in H.Behling-F.Lunek, corr.
2003.

B: 6.Nxe5

512
 White does right to take up the challenge, thus forcing Black to justify the provocation.

6...dxe4

Not 6...Qe7?? (as Marshall once played) 7.d4! Bd7 (or 7...dxe4 8.Nxc6) 8.Bxc6 Bxc6 9.Ng5 Nh6
10.0-0 and White is virtually winning already, J.H.Alvarez-M.Paglino, corr. 1996.

7.Nxc6

White is occasionally tempted by 7.Qh5+?!, which gains a rook and two pawns for the knights.
Black’s two pieces are superior: 7...g6 8.Nxg6 hxg6 9.Qxh8 (not 9.Qxg6+? Kd7 10.Bxc6+ bxc6
11.Qf5+ Kd6 12.Qf4+ Ke6 13.Qxe4+ Kf7 14.d3 Nf6 15.Qxc6 Bd7 and Black won, G.Weber-
G.Möhring, Leipzig 1963) 9...Qd5 (defending g8 and hitting the bishop; I’ve won with both 9...Be6
and 9...Qf6!? as well) 10.Bxc6+ (if 10.Ba4 then 10...Be6 and ...0-0-0, or 10.c4 Qf7) 10...bxc6 11.0-0
Be7 12.Qc3 (or 12.d3 exd3 13.Re1 Be6 14.cxd3 Kd7 15.Qc3 Bf6 16.Qb4 Ne7 – Tay) 12...Ba6 13.d3
exd3 14.cxd3 0-0-0 15.Be3 Bf6 16.Qa3 Qxd3 17.Qxd3 Bxd3 and Black is clearly better, raymac-
tsmenace, ChessWorld.net 2020, in view of 18.Rfe1 Bxb2 19.Rad1 Rd5 20.Bxa7 Nf6, followed by
pushing the c-pawn.

7...Qg5

513
 The viability of 5...d5 relies upon this move. Instead:
a) 7...bxc6? 8.Bxc6+ Bd7 fails to 9.Qh5+ Ke7 10.Qe5+ Be6 11.Bxa8 Qxa8 12.Qxc7+, dating back
to M.Yudovich-N.Krogius, USSR 1959, when White has a rook and three pawns for bishop and
knight, with the black king stuck in the middle too.
b) 7...Qd5?! is also dubious though requires precise play to prove it: 8.c4 Qd6 9.Nxa7+! (rightful
greed; 9.Qh5+ g6 10.Qe5+ Qxe5 11.Nxe5+ c6 isn’t so bad for Black) 9...Bd7 10.Qh5+ g6 11.Bxd7+
Qxd7 12.Qe5+ Kf7 13.Nb5 c6 14.Qd4! Qe7 15.Qxh8 Nf6 16.b3 Rd8 17.Bb2! Bg7 18.Ba3! Qd7
19.Nd6+ Ke6 20.Qxd8 Qxd8 21.Nxb7 and White has collected two rooks and three pawns for the
queen. All the same, White does have to get this far, and containing Black’s counterplay can still be
tricky over the board; e.g. 21...Qb6 22.Nc5+ Kf7 23.0-0-0? Bf8 24.b4 Bxc5 25.bxc5 Qa6 26.Bb2
Qxc4+ 27.Kb1 Nd5 and Black was very much in the game, P.Mercs-J.Tait, Notts Championship
1998.
B1: 8.Nd4+
B2: 8.Qe2
Everything else is bad:
a) 8.Ne5+?? is the right move at the wrong time again: 8...c6 9.d4 (or similarly 9.f4 Qxg2 10.Bf1
Qxh1 11.Qh5+ g6 12.Nxg6 hxg6 13.Qxh8 Kf7, R.Britton-J.Tait, Sheffield 1991) 9...Qxg2 10.Qh5+
g6 11.Nxg6 Qxh1+ 12.Bf1 hxg6 13.Qxh8 Kf7 and White has nothing for the piece, A.Butterworth-
J.Tait, Sheffield League 2011.
b) 8.Nxa7+? fails to 8...Kd8! (not 8...c6? 9.Nxc8 Qxb5 10.Qg4! Kd8 11.Qf4 and White escapes
with the pawns) 9.d3 (now 9.Nxc8? Qxb5 just wins) 9...Qxg2 10.dxe4+ Bd6 11.Rf1 Rxa7 12.Be3
Ra5 13.Qd3 (certainly better than 13.c4? Bh3 14.Qh5 Nf6 0-1 M.J.Turner-J.Tait, Warrington rapid
1991) 13...Nf6 14.0-0-0 (threatening Bb6 or e4-e5) 14...Bd7! 15.Bxd7 Kxd7 16.b4 Rh5 (the rook
must stay on the fifth rank) 17.f3 (or 17.Rg1 Qxe4 18.Rxg7+ Ke6! 19.Qb3+ Nd5) 17...Rd8 (so that

514
the king can go either way) 18.Rg1 (or 18.Bf4 Kc8 19.e5 Be7) 18...Qxf3 19.Rdf1 (or 19.Rxg7+ Kc8
20.Qa3 Qxd1+! etc) 19...Qxe4 20.Rxf6 Qxb4 21.Rxd6+ Qxd6 22.Qb3 Rd5 23.Qa4+ b5 24.Rxg7+
Ke8 25.Rg8+ (or 25.Qe4+ Kf8) 25...Kf7 26.Rxd8 bxa4 27.Rxd6 cxd6 0-1 pionroi-tsmenace,
ChessWorld.net 2020, since Black has a winning endgame.
c) 8.c4?!

(hoping for 8...Qxg2?? 9.Qh5+ etc) 8...Bd7 (this is good for Black but very sharp; Stockfish prefers
8...h5!?, preventing Qh5+ and hence making ...Qxg2 a serious threat) 9.d3 (note that 9.d4? is worse
because, after 9...Qxg2 10.Qh5+ g6 11.Qe5+ Kf7 12.Qd5+ Be6 13.Ne5+ Kf6, the white queen can’t
use the d4-square) 9...Qxg2 10.Qh5+ g6 11.Qe5+ Kf7 12.Qd5+, K.Suwara-Jar.Pech, corr. 2013, and
now 12...Be6! 13.Ne5+ Kf6 14.Qd4 (or 14.Nd7+ Bxd7 15.Qd4+ Kf7 16.Qxd7+ Ne7 17.Rf1 c6)
14...c5! 15.Qc3 (or 15.Nd7+ Ke7 16.Bg5+ Qxg5 17.Qxh8 Bh6 18.Rd1 e3!) 15...Bg7! 16.Nf3+ Kf7
17.Ng5+ Qxg5 18.Qxg7+ Kxg7 19.Bxg5 a6 20.Ba4 exd3 would see Black emerge with a clear
advantage.

B1: 8.Nd4+!?

515
 In The Schliemann Defence: Move by Move, Junior Tay writes: “This ‘time-wasting’ move has been
ignored by most Schliemann/Jaenisch texts”. That’s true. On the other hand, it was the first thought of
IM (soon to be GM) Péter Székely when I confronted him with this line while he was staying locally
(with János) in 1991.

8...c6 9.Bf1

Tay: “In case you are thinking about pouring scorn on this retreat, consider that in the Ulvestad
Variation (...) the retreat home with 6.Bf1 is considered the best line for White” – as shown in
Chapter Nine.

9...Qg6!

I think this prophylactic counter-retreat is best. Black doesn’t commit to anything, waiting to see
what White does next. It’s “a move in the spirit of the Latvian Gambit” (Tay), but I disagree that “it
doesn’t work well”. I think Black has fully adequate compensation.
Note that some of the games below were reached via 9...Bg4 10.d3 Qg6. Shifting the queen first is
more practical since 9...Bg4 allows White significant alternatives, not least 10.h4!? (Hergert); e.g.
10...Qh5 11.Be2 Bxe2 12.Qxe2 Qd5 13.c3 Bc5 14.d3 Bxd4 15.cxd4 Nf6 16.dxe4 Qxe4 17.Qxe4+
Nxe4 and White kept the extra pawn for the ending, V.Kotronias-M.Ghannoum, Budapest 2018.

516
10.d3!

The only testing move, as well as virtually the only tested move:
a) 10.Qe2 doesn’t help White: 10...Nf6 11.f3 (not 11.d3? Bg4) 11...Bd6 12.d3 (not 12.fxe4? 0-0)
12...Bxh2!? 13.dxe4 (not 13.Rxh2? Qg3+) 13...Bg3+ 14.Kd1 Ng4! 15.Be3 (not 15.fxg4?? Bxg4
16.Nf3 0-0-0+ 17.Bd2 Bf4) 15...Nxe3+ 16.Qxe3 Qd6 and Black has quite enough for the pawn.
b) 10.c3 Bg4 11.Qb3 0-0-0 12.h3 (here 12.Bc4 Nh6 13.Be6+ can be met by 13...Rd7!, and if
14.Bxd7+ Bxd7 15.0-0 then 15...Bh3 16.g3 Bc5 17.d3 Bxf1 18.Kxf1 Re8) 12...Bd7 13.d3 Nf6
14.dxe4 (or 14.Be3 Nd5 15.0-0-0 Nxe3 16.fxe3 Be7) 14...Qxe4+ 15.Be2 Bd6 (15...Qxg2?! 16.Bf3
Re8+ 17.Be3 is better for White) 16.0-0 Rhe8 with a strong initiative.
c) 10.Ne2 (intending Nf4) 10...Bd6 11.d4 exd3 12.Qxd3 Bf5 13.Qe3+?! (13.Qc4 Bxc2 14.Nf4 Bxf4
15.Bxf4 Ne7 16.Bg3 is roughly equal) 13...Ne7 14.Nf4 Bxf4 15.Qxf4 Nd5 16.Qe5+?! (instead,
16.Qg5 Qe6+ 17.Be2 0-0 18.Be3 h6 19.Qg3 Nxe3 20.Qxe3 Qxe3 21.fxe3 Bxc2 restricts Black to a
slight edge) 16...Kf7 17.Qg3 Rhe8+ 18.Be3 (not 18.Be2? Nb4) 18...Nxe3 19.fxe3 Qxg3+ 20.hxg3
Rad8 was good for Black, DanielMena-tsmenace, ChessWorld.net 2019.

10...Bg4

517
 Attacking the queen, accelerating development, and (perhaps) preparing to castle long.

11.f3

Others give White nothing:

a) 11.Be2 Bxe2 12.Qxe2?! (overambitious; 12.Nxe2 Qxg2 13.Rg1 Qxh2 14.Bf4 Qh5 15.dxe4 Rd8
16.Nd4 Qf7 17.Bg5 Be7 18.c3 Bxg5 19.Rxg5 Ne7 and ...0-0 is level) 12...0-0-0 13.Be3 (not 13.Nb3?
exd3 14.cxd3 Bb4+ 15.Kf1 Nf6 and White is in trouble, C.Cabrera Trujillo-J.Ryska, corr. 2013)
13...Bc5 14.c3 (here 14.Nb3 exd3 15.cxd3 Bxe3 16.fxe3 Qxd3 17.Qxd3 Rxd3 18.Ke2 Rd5 should be
a favourable endgame for Black, J.Kozma-J.Fichtl, Czech Ch., Prague 1954) 14...exd3 15.Qf3 Bxd4
16.cxd4 (or 16.Bxd4 d2+! – I&K) 16...Ne7 17.0-0? (either 17.0-0-0 Nd5 18.h4 or 17.Qh3+ Rd7
18.Rd1 keeps White in the game) 17...Nf5 18.d5 Nxe3 19.Qxe3 Rxd5 20.Qxa7 Rhd8 21.a4 d2
22.Rfd1, V.Tsarev-V.L.Ivanov, Moscow 1990, and now Stockfish assesses 22...Re5! as approaching
“+6” for Black; e.g. 23.a5 (if 23.h3 then 23...Rg5, or 23.g3 Qe6) 23...Qe6 24.h3 Re1+ 25.Kh2 Qe5+
26.g3 Qe4 wins.
b) 11.Qd2 0-0-0 12.h3! (definitely best; not 12.Qc3?! Ne7 13.h3 Nd5 14.Qd2 Bc5 15.Qg5 Bxd4
and White is in difficulties, V.Hergert-P.Leisebein, corr. 1997; or if 12.Nb3? then 12...Re8 13.d4 e3!
14.fxe3 Bd6 is one way for Black, e.g. 15.Be2 Nf6 16.0-0? Qh5 and wins) 12...Bh5 (if 12...Rxd4 at
once, then 13.hxg4 Nf6 14.Qg5 Qxg5 15.Bxg5 Nxg4 16.g3 and Bh3 is annoying) 13.g4 (delaying this
doesn’t strengthen White’s hand; e.g. 13.Nb3 Nf6 14.g4 Qe8! 15.gxh5? exd3+ 16.Qe3 Qxh5, or
13.Qf4 Bb4+ 14.c3 Rxd4 15.cxb4 Ne7 16.g4 Qe6) 13...Rxd4 14.gxh5 Qxh5 15.Qg5 Qxg5 16.Bxg5
exd3 17.Bxd3 Rd5 is roughly equal. White has the two bishops but also a slightly inferior structure.

518
11...exf3

Stockfish considers 11...0-0-0!? to be viable too; e.g. 12.fxg4 (or 12.dxe4 Nf6 13.c3 c5) 12...Rxd4
13.c3 (or 13.Be3 Rd7) 13...Rd8 14.d4 Bd6 or 14...h5!?.

12.gxf3

White needs the gain of tempo. After 12.Nxf3?! 0-0-0 Black has good play; e.g. 13.Be2 Bb4+ 14.c3
Bxf3 15.Bxf3 Rxd3 16.Bg4+ Kb8 17.Bf4+? (17.Qe2 Nf6 18.cxb4 Nxg4 19.Bf4+ Ka8 20.0-0 escapes
the worst) 17...Bd6 18.Bxd6+ Qxd6 19.Qe2 Nf6 20.0-0 h5! 21.Bf3 Ng4 22.Bxg4 hxg4 23.g3 Rd2 and
Black won, K.Reed-P.Goffin, corr. 2012.

12...Bd7

12...0-0-0? is worse than on the previous turn: 13.fxg4 Rxd4 14.Be3 Rd5 15.Qf3 Be7 16.0-0-0 and
White is a clear pawn up, P.Doberitz-A.Kropman, Uelzen 2007.

13.Qe2+ Be7

519
 White can now develop the dark-squared bishop and castle to safety on the queenside.

14.Bd2

If 14.Bf4 Qf7 15.Be5 0-0-0 16.f4 Nf6 17.Bg2 Rhe8 18.0-0 (not 18.0-0-0? Qxa2) 18...Bc5 19.c3
Bf5, Black can count on sufficient play in the centre. For example: 20.b4 (after 20.Rae1 Qg6 21.Qf3
Nd7 22.Kh1 Bxd3 23.Rg1 Bc4, Black had safely regained the pawn, The_Thinker95-tsmenace,
ChessWorld.net 2020) 20...Rxd4! 21.cxd4 (not 21.bxc5? Rxd3) 21...Bxd4+ 22.Kh1 Bxa1 23.Rxa1
Ng4 24.b5 Qh5 25.Bf3 Qh4 26.bxc6 Rxe5! 27.cxb7+ Kb8 28.fxe5 Nf2+ 29.Kg2 (not 29.Kg1 Nh3+
30.Kf1?? Bxd3 and wins) 29...Bh3+ 30.Kg1 Qd4 31.Re1 Nxd3+ 32.Kh1 Nf2+ with a draw.

14...Qf7

Overprotecting the e7-bishop in preparation for castling, and eyeing the f3- and a2-pawns.

520
15.Nb3

Preparing to castle long at the cost of rendering the knight rather passive.
Instead, 15.Qe5 Nf6 16.Nf5 (more logical than 16.0-0-0?! 0-0 17.Nb3, as in H.Faber-P.Leisebein,
corr. 2014, where Stockfish likes 17...a5! for Black) 16...Bxf5 17.Qxf5 0-0 looks as if it might be
good for White with the two bishops and still the extra pawn. Computer says “no”. White’s minuses
are the king (which needs to castle soon), the a2-pawn (if the king goes long), the h2-pawn (if the
king goes short) and the isolated f-pawn.
For example: 18.Bh3 (18.c4 creates more weaknesses, so even 18...Nd7!? 19.Qxf7+ Rxf7 is
possible) 18...Bd6 19.0-0 (or 19.0-0-0 Qxa2) 19...Qc7 20.Kh1 (or 20.f4 Rae8) 20...Rae8 (not yet
20...Bxh2? 21.f4 Rae8 22.Bb4) 21.Rae1 (or 21.f4 Nd5) 21...Bxh2 22.f4 Bg3 23.Rxe8 Rxe8 24.Bc3
Bh4 and Black is fine.

15...Nf6!

Improving on 15...0-0-0?! 16.0-0-0, when “White is a pawn up without any issues” (Tay).
S.Abramov-M.Shchebenyuk, corr. 2012, continued 16...Nf6 17.Qf2 b6 18.Bg2 Rhe8 19.Kb1 Bd6
20.f4 and White was clearly better.

16.0-0-0

I was looking forward to exploring this properly in a recent ChessWorld.net game. Unfortunately,
my opponent had an aberration, played 16.Qe4?? instead, and resigned on the next turn. So all I have
is a few variations from analysing with the engines.
Note that 16.Bb4 is well met by 16...0-0 17.Bxe7 Rfe8 18.0-0-0 Rxe7 with ...Nd5 to follow, when
White has just relinquished the dark squares.

521
16...0-0

Never mind the half-open g-file – the king is secure on the kingside; the f8-rook also adds pressure
on the f-file; and there are now options of pushing the queenside pawns. Meanwhile, without a single
well-placed piece, White will take some time to develop active play. For example:
a) 17.d4 Bf5 (or 17...Rae8) 18.Rg1 Bd6 19.Qg2 (or 19.Nc5 b5) 19...b5 (securing the d5 outpost)
20.Kb1 Nd5 21.Nc1 Qg6! 22.Bd3 Qxg2 23.Rxg2 Bxd3 24.Nxd3 Rxf3 is equal.
b) 17.Rg1 a5 (or 17...Nd5) 18.d4 b6 (not 18...a4?! 19.Nc5) 19.a4 (or 19.Kb1 Nd5) 19...Bf5 20.Qg2
(or 20.Bh6 Nh5) 20...Bd6 21.Bc4 Nd5 22.Rde1 (or 22.h4 Ra7) 22...Bg6 23.f4 Kh8 and Black is fine.
c) 17.Kb1 Nd5 (or 17...Rae8) 18.Bg2 (or 18.c4 Nf6 19.Bc3 Nh5 20.Rg1 Rae8) 18...Bh4 19.Nc5
Rae8 20.Qf1 Ne3 21.Bxe3 Rxe3 22.Nxd7 Qxd7 and with control of the dark squares and much
greater activity, I think Black should be perfectly okay.

B2: 8.Qe2

522
 The main line. From here the white queen defends the b5-bishop, attacks e4, and is ready to cover
g2 sideways after f2-f4. Note that 8...Qxg2?? would be a blunder due to 9.Qh5+ g6 (or 9...Kd7
10.Ne5+) 10.Qe5+ and wins.

8...Nf6

The only move, developing and defending the e-pawn. Now White has:
B21: 9.Nxa7+
B22: 9.f4
Instead:
a) 9.Ne5+ c6 10.f4 Qxf4 transposes to line B223.
b) 9.d4?! allows 9...Qxg2 10.Ne5+ (not 10.Rf1?? a6 11.Ba4 Bd7) 10...c6 11.Qf1 (the only way to
save the piece) 11...Qxf1+ 12.Bxf1 Be6 and Black is fine; e.g. 13.Bg2?! (here 13.Bc4, as in
K.Slechta-A.Stojkovic, Prague 2007, looks safer and equal; indeed a draw was agreed one move
later) 13...0-0-0 14.Be3 Bd6 15.0-0-0 Rhe8 16.Rhe1?, O.Bashirinia-H.Toomanian, Iranian League
2005, and now 16...Bxa2! was possible, since 17.b3?? loses to 17...Ba3+ 18.Kd2 Rxe5 19.Ra1 Ra5.

B21: 9.Nxa7+

523
 White grabs a pawn with check. Sokolov writes that it leads to “a labyrinth of variations” and “may
well become one of the main lines in the Jaenisch Gambit.” Certainly, the timely capture is more
testing than 9.f4 Qxf4 10.Nxa7+ (see line B221 below).

9...Bd7

The correct way to answer the discovery. After 9...Kd8?! (9...c6? 10.Nxc6 is no good at all)
10.Nxc8 Qxg2 11.Qf1 Qxf1+ (or 11...Qf3?! 12.Nb6 cxb6 13.Be2!? Qf5 14.d4 exd3 15.Bxd3 Qe5+
16.Qe2 Qxe2+ 17.Bxe2, F.Fleury-M.Preussner, corr. 2013) 12.Kxf1! Kxc8 13.d4 exd3 14.Bxd3 Bd6
15.a4 Ra5 16.Rg1 g6 17.Rg5, S.Schubert-A.Osipov, corr. 2011, Black doesn’t have a deal for
White’s pawn and two bishops, even if winning is not entirely simple.

10.Bxd7+ Nxd7

Also practically forced. After 10...Kxd7? 11.Qb5+ Qxb5 12.Nxb5 c6 13.Nc3, Joh.Anderson-G.Van
Dijk, St. Albans 1998, Black has very little for two pawns.

524
11.f4!

Other moves are at least equal for Black:

a) 11.Qxe4+?! Kd8 12.Qxb7 (after 12.d4 Qa5+ 13.Bd2 Bb4 14.0-0-0 Bxd2+ 15.Rxd2, V.Hergert-
P.Leisebein, corr. 1997, and 15...Re8 16.Qxb7 Rxa7 17.Qb3 Qxa2; or 12.d3 Qa5+ 13.Bd2 Qxa7 14.0-
0 Nf6 15.Qe6 Qc5 16.Rae1 Qd5 17.Qxd5+ Nxd5, it’s White who has to try and draw – which he did
manage in D.Jakovenko-A.Karpatchev, Tomsk 1998) 12...Rxa7 13.Qxa7? (13.Qe4 keeps White in the
game) 13...Qxg2 14.Rf1 Bc5 15.Qa6 Re8+ and Black won, D.Thomas-M.Boskovic, USA 1975.
b) 11.Nb5?! Qxg2

525
 and now:
b1) 12.Rf1?! 0-0-0 13.Nc3 (relatively best; not 13.d4?? exd3 14.Qxd3 Bb4+ and wins; or 13.f4?
exf3 14.Qxg2 fxg2 15.Rg1 Re8+ 16.Kd1, J.Fiser-V.Hora, Prague 1965, and now 16...Re5! 17.Nc3
Bc5 18.Rxg2 Rf8 with a very strong attack; or if 13.Qc4, G.Monnisha-M.Mitali, Bhayandar 2015,
then 13...c6! 14.Na7+ Kb8 15.Nxc6+ bxc6 16.Qxc6 Qg6 and White has problems) 13...Nf6 (or
13...Ne5 14.f4 Nf3+ 15.Kd1 Qxe2+ 16.Kxe2 Re8) 14.b3 Bc5 15.Bb2 Rhe8 16.0-0-0 Qxh2 is fine for
Black.
b2) 12.Qf1 Qxf1+ 13.Rxf1 0-0-0 (the cramping e4-pawn and lead in development offer Black
ample compensation) 14.b3 (if 14.Nc3 Re8 15.b3 Ne5 16.Bb2 Nf3+ 17.Ke2, N.Djokic-D.Sekulic,
Serbian League 2003, then 17...Bc5 intending ...Bd4) 14...Ne5 15.Nc3 Bb4 16.a3?? (either 16.Nxe4
Rhe8 17.f3 Nxf3+ 18.Rxf3 Rxe4+, or 16.Ke2!? Nc6, threatening ...Nd4+, would maintain the
balance) 16...Nf3+ 17.Ke2 Nd4+ 18.Ke3 (or 18.Kd1 Bxc3 19.dxc3 Nxb3+) 18...Nxc2+ 19.Kxe4
Rhe8+ 20.Kf3 Bxc3 21.dxc3 Nxa1 0-1 R.Wynarczyk-J.Tait, Scarborough 1993.

In my copy of Spanish: Schliemann (Jaenisch) there’s a terse “11.f4!?” scrawled in the margin,
showing I’d noticed this move as a potential problem. The point is that 11...Qxf4? lands Black in a
bad 9.f4 Qxf4 line, having answered 10.Nxa7+ Bd7 11.Bxd7+ with 11...Nxd7? (rather than
11...Kxd7). So, what to do instead?

11...Qf5!

GM Ivan Sokolov first brought attention to this move, analysing it in his book The Ruy Lopez
Revisited. Subsequent praxis has refined the theory, confirming 11...Qf5 as the best option for Black.
This includes an important correspondence game by GM Liviu-Dieter Nisipeanu and a Stockfish-
Leela encounter from TCEC 17, both drawn. As for me, I’ve tested it in twelve online games so far
(all referenced below), for a score of W5, D7, L0.

526
 Previously, Black rushed to regain a pawn with 11...Qc5 12.Nb5 Qxc2 (12...0-0-0?! 13.Nc3 is just
good for White), but then 13.d4! more or less forces the queens off and Black isn’t quite active
enough to make up for the other pawn; e.g. 13...Bb4+ 14.Kf2 Qxe2+ 15.Kxe2 0-0-0 (15...Kd8 16.Bd2
Bxd2 17.Kxd2 Ra4 18.a3 is much the same) 16.Be3 Nf6 17.a3 Ba5 18.Nc3 Rhe8 19.Rhd1 c6 20.Rac1
Bc7 21.d5 Nxd5 22.Rxd5 cxd5 23.Nb5 Rd7 24.Bb6 Ree7 25.Bxc7 1-0 Joh.Anderson-B.Hanison,
corr. 1995, since 25...Rxc7 26.Nxc7 Rxc7 27.Rxc7+ Kxc7 28.Ke3 is winning for White who can
create a passed pawns on both wings.
And 11...Qxf4? 12.d4 is the aforementioned bad line; e.g. 12...Qf5 13.Nb5 0-0-0 14.Rf1 Qe6
15.Bf4 Bb4+ 16.c3 Ba5 17.b4 Bb6 18.a4 c6 19.Nd6+ 1-0 J.Tait-W.Goedhart, corr. 1998.

12.Nb5

Here I initially wrote that no one has yet tried 12.g4!?, probably because 12...Qc5 13.Nb5 0-0-0!
gives Black good play given White’s weakened kingside; e.g. 14.Kd1 Nf6 or 14.Nc3 Be7 15.g5 h6.
But now someone has, continuing 14.b3!? Qxc2 15.Nd4 Qd3! 16.Qxd3 exd3 17.Bb2 h5 18.0-0 hxg4
19.Ne6 Re8 20.Nxf8 Nxf8 21.Bxg7 Rh3 22.Rae1 Rxe1 23.Rxe1 Ng6 24.Re4 Nh4 25.Bc3 (or 25.f5
g3) 25...Nf3+ 26.Kf2 Rxh2+ 27.Ke3 Nh4 28.Re8+ Kd7 29.Rg8 g3 30.Kxd3 g2 31.Bd4 Rh1 and
Black held, pionroi-tsmenace, ChessWorld.net 2021.

12...0-0-0

527
13.a4

The most popular reply for White, overprotecting the knight with options to push the pawn further
and/or develop the queen’s rook via a3.
Five other moves have been tried:
a) 13.0-0 Nf6 (not 13...Bc5+?! 14.Kh1 c6 15.b4! Bxb4, E.Powierski-S.Tennert, Oberhof 2012, due
to 16.Ba3! Bc5 17.Nd6+ Bxd6 18.Bxd6 – Tay; nor 13...Qc5+?! 14.Kh1 Qxc2 – Sokolov, because of
15.Nd4! Qd3 16.Qxd3 exd3 17.b3 – Bologan) 14.Kh1 (sensible prophylaxis; 14.b3?! Qc5+! 15.Kh1
Qxc2 is now good for Black, drink1966-tsmenace, ChessWorld.net 2020; or if 14.Nc3 then 14...Bc5+
15.Kh1 h5 16.Qb5 g6!, intending 17.b4 Bd4 18.Qxf5+ gxf5 19.Rb1 Bxc3 20.dxc3 b5 with a total
clamp on the light squares, and not 16.Na4?? Bd4 17.c3 Ba7, followed by ...h5-h4, ...Nh5-g3+ and
wins, PoulErik1962-tsmenace, ChessWorld.net 2018) 14...c6 15.Nc3 h5 16.Nd1 h4 (preparing ...Nh5-
g3+ again) 17.Ne3 Qg6 18.f5 Qg5 gives Black sufficient play for the pawns; e.g. 19.Nd1 (or 19.Kg1
Nh5 20.Qg4 Qxg4 21.Nxg4 Bc5+ 22.Ne3 Nf6 23.a4 h3, “followed by ...Rh5 and White is suffering,
unless he is a cyborg” – Tay; and not 19.b3? Bc5 20.Qc4? Bd4 21.Rb1 Nh5 22.Re1 Ng3+ 23.Kg1
Nxf5 with a decisive attack, tripoduk-jatait47, Chess.com 2019) 19...Nh5 20.Nf2 Bd6 21.Nxe4
Ng3+! 22.Nxg3 (not 22.hxg3?? hxg3+ 23.Kg1 Rh1+ etc) 22...hxg3 23.h3 Rxh3+! 24.gxh3 Rh8
25.Kg2 Rxh3 26.Qe8+ Kc7 ½-½ D.Korze-K.Kögler, corr. 2013.
b) 13.g3 just encourages 13...h5 (rather than 13...c6 14.Nc3 Bb4?! 15.Nd1 e3? 16.Nxe3 Qe4 17.0-0
and Black soon resigned in T.Burg-M.Timmermans, Vlissingen 2013) 14.a4 (or 14.Qc4 c6 15.Nd4
Qf6 16.Ne6 h4! 17.Nxd8 hxg3) 14...Bb4 15.Qc4 (not 15.c3? Nc5 16.cxb4 Nd3+ 17.Kd1 e3!,
intending 18.Qxe3? Qd5 and wins) 15...Qc5 16.Qxc5 Nxc5 17.Nc3 h4 “and Black is exerting strong
pressure” (Tay).
c) 13.Nc3 Bc5 “followed by ...Rhe8, ...Nf6 gives Black nice compensation” (Sokolov),

528
 as proved the case in docjan-tsmenace, ChessWorld.net 2018: 14.a4 Rhe8 15.a5 Nf6 16.a6 b6
17.Rf1 (or 17.a7 Kb7) 17...Kb8 18.Ra4 Qd7! 19.b4 e3! (with everything developed Black is ready to
take action) 20.a7+ (not 20.bxc5?? exd2+, or 20.dxe3? Bxe3 21.Bxe3 Ng4, or 20.d3? Qd4! 21.a7+
Kb7 22.bxc5 Qxc3+ 23.Kd1 Qxc5) 20...Ka8 21.Qf3+ c6 22.d3 b5 23.Nxb5? (taking the pawn is a
mistake; 23.bxc5 bxa4 24.Nxa4 Kxa7 25.Bb2 Ng4 26.Nb6 Nxh2 27.Nxd7 Nxf3+ 28.Rxf3 Rxd7
29.Be5 should be a draw) 23...Bf8! 24.Nc3 Nd5 25.Na2 (or 25.Nxd5 cxd5 26.Ra1 Qc6) 25...g5!
26.fxg5 Bg7 (threatening ...Rf8) 27.Ra3? (here 27.b5 cxb5 28.Re4 was the best try) 27...Rf8 28.Qh3
Rxf1+ 29.Kxf1 Rf8+ 30.Ke1 Qd6 (threatening ...Qf4) 31.g3 (or 31.Qh4 Be5) 31...Nxb4 32.Nxb4
Qxb4+ 33.Kd1 e2+! 34.Kxe2 Re8+ 35.Kf3 (35.Be3, hoping for 35...Qxa3?? 36.Qd7, fails to
35...Qe7! 36.g4 and then 36...Qxa3) 35...Qe1 36.Ra5 h5! 0-1, in view of 37.gxh6 (or 37.Qxh5 Qe2+)
37...Qd1+ 38.Kf4 Bxh6+! 39.Qxh6 Qxc1+ etc.
d) 13.Qc4 c6

529
 14.Nd4 (after 14.Nc3 Bc5 15.b4 b5! 16.Nxb5 cxb5 17.Qc3 Qxf4 18.bxc5 e3 19.dxe3 Qh4+ 20.g3
Qe4 21.0-0 Ne5 22.Qa5 Nf3+ 23.Rxf3 Qxf3, White needs to take the draw) 14...Qxf4 15.Ne6 Qf7
16.Rf1 is answered by 16...Qg8! (improving on 16...Nf6?! 17.b3 – Sokolov) 17.Qb3 (White has to
break the pin) 17...Nc5 18.Nxc5 Bxc5 and Black has sufficient compensation; e.g. 19.Rf7 (or
19.Qxg8 Rhxg8 20.Rf5 Rd5 21.Rf7 Rf8 22.Rxf8+ Bxf8 – Tay) 19...Rd7 20.Rxd7 Qxb3 21.axb3
Kxd7 22.Ra4 Re8, which led to a draw in J.De Waard-H.D.Vötter, corr. 2011.
e) 13.b3 (hoping to go long) 13...Qf6! (interfering with that plan) 14.Rb1 (in Sokolov’s line 14.Nc3
e3! 15.0-0 Bc5 16.Kh1 Rhe8 17.Bb2 exd2 18.Qxd2, Tay’s 18...Nf8! gives Black good play; e.g.
19.Na4 Rxd2 20.Bxf6 gxf6 21.Nxc5 b6 22.Nd3 Rxc2, or 19.Qc1 Ng6 20.f5 Nh4 21.Qf4 Nxg2!
22.Kxg2 Qc6+ 23.Qf3 Rd2+ 24.Kh1 Qh6 “and Black has at least a draw” – Tay) 14...Qxf4 15.d4 was
tested at the very highest level: in Stockfish-Lc0, TCEC 17 Superfinal 2020.

530
 Leela came up with an impressive defence: 15...e3! (jettisoning the just-regained pawn to allow
pressure on the e-file and light squares) 16.Bxe3 Qf5 17.Rf1 Qg6 18.b4 (not 18.Kd2 Nf6 19.Rbe1? c6
20.Nc3 Bb4; while 18.Rf3 c6 19.Na7+!? Kb8 20.d5 Bc5! is just a mess, e.g. 21.dxc6 bxc6 22.Rg3
Qf6 23.Bg5 Bb4+! 24.c3 Bxc3+ 25.Kd1 Nc5+ 26.Kc2 Qf5+ 27.Kxc3 Ne4+ 28.Kb2 Kxa7) 18...Bd6
19.Nxd6+ (later 19.Qd3 Qxg2 20.Nxd6+ cxd6 21.Qe2 Qe4 22.Kd2 Nb6 23.Qd3 Rhe8 24.Rbe1 Qg2+
25.Rf2 Qd5 26.Qf5+ Rd7 27.Qxd5 Nxd5 28.c4 Nxb4 29.Kc3 Rde7 led to a drawn rook endgame in
Reg-UK-tsmenace, ChessWorld.net 2020) 19...Qxd6 20.h3 Nf6 21.Rb3 Rhe8 22.Rf3 Rd7 23.Kd1 (not
23.Kf2?!, since 23...Qh2! keeps the king in the middle and threatens ...Ne4+) 23...Nd5 24.Qd3 b5!
(gaining more light square control) 25.Kc1 (presumably after 25.Qxb5 Nxe3+ 26.Rfxe3 Rxe3
27.Rxe3 Qxd4+ 28.Rd3 Qg1+ 29.Kd2 Rxd3+ 30.Qxd3 Qxg2+ the engines judged that Black should
hold the queen ending – the white king can only hide on the queenside, when Black can prepare the
advance of the kingside pawns and blackmail White with queen exchanges; e.g. 31.Kc3 Qc6+ 32.Kb2
Qf6+ 33.Qc3 Qf5 34.a4 Qe4 35.Kb3 h5 36.a5 Qe6+ 37.Ka3 h4 38.Qd3 Qd7 and suchlike) 25...Qc6
26.Qf1 Re4 27.Rf8+ Kb7 28.Ra3 Rde7 29.Bd2 Nb6 30.Qf3 Qd5 31.Re3 Qxa2! (not 31...Qxd4? 32.c3
Qd5 33.Rxe4 Qxe4 34.Qxe4+ Rxe4 35.Rf7 and Black is in trouble again) 32.Kd1 (obviously
32.Rxe4?? Qa1 is mate) 32...Qd5 33.Rxe4 Rxe4 34.Bc3 g6 35.Rd8 Qe6 36.Kd2 c6 and White could
make no progress (½-½, 50).

531
13...Bb4!

Both the most aggressive and best move.

Sokolov’s main line was 13...Bc5, preventing White from castling. Then 14.a5 c6 15.Nc3 Rhe8 (or
15...Nf6 16.Rf1 Rhe8) 16.Rf1 (there’s no rush to insert 16.a6 b5, as in Sokolov’s analysis) 16...Nf6
(if 16...Nb8!? 17.Ra4 Bd4 18.Nd1 Na6 19.Ne3 Qf6 20.g3 Kb8 21.b3 Nc7, M.Jaworski-G.Vetter, corr.
2011, Stockfish wants to sacrifice the exchange with 22.Ng4 Qe6 23.Rxd4 Rxd4 24.Bb2, accounting
White’s bishop and pawns clearly superior to Black’s extra rook) 17.Nd1! Nd5 18.Ne3 Qf7 (or
18...Nxe3 19.dxe3 Bb4+ 20.Kf2 Bxa5 21.Ra4 Kb8 22.Kg1, as in D.Mielczarek-W.Kozlowski, corr.
2011) 19.g3 Kb8 20.Ra4 Nb4 21.b3 Qf6 22.Kf2 Nxc2 23.Qc4 Nxe3 24.dxe3 Qf5 25.b4 Be7 26.Ra2
and White consolidated, A.Serazeev-D.Wilhelmi, corr. 2013.
Another option is 14.Rb1!? Be7 15.b4 Nb6 16.0-0 Bxb4 17.Kh1 Ba5 18.Rd1 Rhe8 19.Bb2 Qxf4
20.Bd4 Kb8 21.Be3 Qf7 22.Nd4 and, despite regaining both pawns, Black was clearly worse in
Komodo-Komodo MCTS, TCEC 16 VSO 2019.
Note that 13...c6 14.Nc3 Bc5 15.a5 reaches the same position as above (though Sokolov actually
analysed it twice).

532
14.c3

A natural response, given that Black was already threatening ...e4-e3.

Instead:
a) 14.a5 is duly answered by 14...e3! 15.c3 (or 15.0-0 exd2 16.Bxd2 Bxd2 17.Nd4 Qf6 18.Qxd2
Nb8!) 15...exd2+ 16.Bxd2 Rhe8 17.Be3 Bc5 18.Nd4 Qe4 and Black is okay; e.g. 19.Kf2 (or 19.a6
Qxe3 20.Qxe3 Rxe3+ 21.Kd2 Bxd4 22.cxd4 Re4 23.a7 Nb6 – Sokolov) 19...g5! 20.Qf3 gxf4
21.Qxe4 Rxe4 22.Bd2 Bxd4+ 23.cxd4 Nc5!? 24.dxc5 Rxd2+ 25.Kf3 Rb4.
b) 14.0-0 Qc5+ 15.Rf2 (after 15.Kh1 Qxc2 Black is fine; e.g. 16.a5? Nc5 – Sokolov, or 16.Nd4
Qd3 17.Qxd3 exd3 18.b3 Rhe8 – Tay) 15...Qxc2 16.d4 is much easier on Black than the 11...Qc5
line; e.g. 16...Qxe2 17.Rxe2 c6 18.Nc3 Nf6 19.Be3 (or 19.Nxe4 Rhe8 20.Ng3 Rxe2 21.Nxe2 c5!)
19...Ba5 20.g3 (or 20.h3 h5 21.Rb1 Bb6 22.Rd1 Rd7 23.b4 Rhd8 24.Red2 Nd5 25.Nxd5 Rxd5 “and it
is really tough for White to make use of his extra pawn” – Tay) 20...h5 21.Rb1 h4 22.Bf2 hxg3
23.hxg3 Bxc3 24.bxc3 Rh5 25.Kg2 Ra5 26.Ra1 Kc7 27.c4 Rda8 28.Rb1 Rxa4 29.Reb2 Ra2
30.Rxb7+ Kd6 31.Kg1 Ng4 32.Rxg7 Nxf2 33.Rbb7 ½-½ pionroi-tsmenace, ChessWorld.net 2018, in
view of 33...Nh3+ 34.Kh1 Nf2+ 35.Kg1 (not 35.Kh2?! Ng4+ 36.Kh3 Nf6) 35...Nh3+ etc.
c) 14.Nc3 Qxf4 15.Nd5 Qd6 16.Nxb4 Qxb4 is also acceptable for Black; e.g. 17.0-0 Qc5+ 18.Qe3
Qxc2 is now level on pawns, while 17.a5 Rhe8 18.a6 b6 19.Ra3 Nf6 20.0-0 Kb8 hasn’t solved
White’s problems.
d) 14.Qc4 is the engines’ top choice (once they’ve discarded 14.c3). After 14...Qc5 15.Qxc5 Nxc5
White has forced the queens off and is still two pawns up – and yet it’s surprisingly difficult to
consolidate.

533
 I’ve defended this successfully four times out of four:
d1) 16.0-0 Bxd2 17.Bxd2 Rxd2 regains one pawn with sufficient play for the other; e.g. 18.Rf2
Rhd8 19.Re1 Nxa4 20.b3 Nc5 21.Nc3 Ne6! (going after both f4 and c2) 22.Rxe4 Rxf2 23.Kxf2 Nd4
24.Re5 Nxc2 ½-½ Nimzovik-tsmenace, ChessWorld.net 2020.
d2) 16.Nc3 Ne6 17.a5 (or 17.Ne2 Nc5 18.Kf2 Bxd2 19.b3 Rhe8 20.Rd1 e3+ 21.Kf3 g5 22.f5 h5
23.h3 h4 and it’s White who has to be careful: 24.Bxd2 exd2 25.g4 hxg3 26.Nxg3 Rd4 27.Kg2 Ne4
28.Nxe4 Rexe4 29.c4 Re2+ 30.Kf3 Rde4 31.f6 R2e3+ ½-½ alexkhesin-tsmenace, ChessWorld.net
2020) 17...Rhe8 18.a6 bxa6 19.Rxa6 Kb7 20.Ra4 Rd4 21.f5 (an attempted improvement on 21.Ra2
Nxf4 22.g3 Ne6 23.Kd1 Nc5 and Black held easily in R.Sherwood-R.Burridge, corr. 2019) 21...Nf4
22.Rxb4+!? Rxb4 23.0-0 Nh5 24.b3 Nf6 25.Nd1 Rd4 26.Kf2 with the initiative and two pawns for
the exchange. But it’s hard to achieve anything concrete if Black sits tight, and 26...Rd7 27.Ke2 Ra8
28.Bb2 Rad8 29.Bc3 Nd5 30.Ne3 Nxc3+ 31.dxc3 Rd2+ 32.Ke1 R2d6 33.Rf4 Re8 34.Ke2 g6 35.fxg6
Rxg6 led to another draw in Vishnuduta-tsmenace, ChessWorld.net 2021.
d3) 16.Ke2 c6 17.Nc3 Ne6 18.Nxe4 (White is now three pawns up) 18...Rhe8

534
 19.Kd1 (after 19.d3?! Nd4+ 20.Kd1 Nb3! 21.cxb3 Rxe4 “Black has enough play for three pawns” –
Tay) 19...Nxf4 20.Ng3 Nxg2 21.d3 Ne3+ 22.Bxe3 Rxe3 23.a5 Kb8! (refusing to commit to anything;
23...Rde8?! 24.Ne4 Rd8 25.Kc1 Re2 26.Ra4! is worse for Black) 24.Rf1 Rde8 25.Ne4 Re5 26.a6
Rf5! 27.Rg1 (not 27.Rxf5?? Re1 mate) 27...Rh3 28.Rh1 Ra5!? (it looks strange to swap an active
rook for one sitting on its home square, but White’s rooks wouldn’t be passive for long; e.g. 28...b6
29.Ra4 Be7 30.Re1 and if 30...Rxh2?? then 31.a7+ Ka8 32.Nd6! wins) 29.Rxa5 Bxa5 30.axb7 Bb6
31.Ng5 Rh5 32.Ne4 Rh3 33.Ng5 Rh5 34.Ne6 g5! (Black holds the kingside, so White tries to make
something of his own majority) 35.Rf1 h6 36.c4!? Rxh2 37.c5 (now Black must defend precisely)
37...Ba5 38.Nd4 Bc7! 39.Rf8+ Kxb7 40.Rf7 Kc8 41.b4 g4! 42.b5 (or 42.Ne6 Bd8 43.Rf8 Kd7
44.Nxd8 g3, leading to a drawn rook ending) 42...cxb5 43.Nxb5 Ba5 44.c6 (or 44.Ra7 g3 45.c6 Kb8
46.Rxa5 g2 47.c7+ Kb7 48.Ra7+ Kb6 49.Ra6+ Kb7 etc) 44...Rd2+ 45.Kc1 Rxd3 46.Ra7 Rc3+! (this
needed to be seen in advance) 47.Nxc3 Bxc3 led to a draw once more in ianl-tsmenace,
ChessWorld.net 2020. The black pawns are irrelevant since the queenside is a fortress, dating back to
E.Del Rio in 1750.

14...Nc5!

Obviously not 14...Be7? 15.0-0 (Sokolov).

15.cxb4

Refusing the piece is no good: 15.0-0? Nb3 16.cxb4 Nxa1 17.b3 Rd3.

15...Nd3+ 16.Kd1 g5!

535
 Nisipeanu’s novelty. Sokolov had analysed 16...e3? 17.Qxe3! Rhe8 18.Qf3 Re4 19.Ra3! Rxf4
20.Rxd3 to a win for White.

17.g4!

Pushing the black queen back is the only serious try: 17.fxg5? runs into 17...e3! (even stronger than
17...Nf2+) 18.Rf1 Nf2+ 19.Ke1 Rhf8, threatening ...Nd3+ and wins. White gets nowhere with 17.Ra3
gxf4 18.Rc3 c6 19.Kc2 either, in view of 19...Kb8 20.Na3 Nxb4+ (or 20...f3) 21.Kb1 f3!.

17...Qe6

Threatening ...Qb3 mate!

18.Kc2

Bringing the rook forward with 18.Ra3 still doesn’t succeed: 18...gxf4 19.Rc3 (or 19.Rxd3 Rxd3
20.Kc2 Qc4+ 21.Kb1 f3) 19...c6 20.b3 (or 20.Rf1?! Rhe8 21.Na3 Qa2!) 20...Kb8 21.Na3 f3 22.Qf1
e3 23.Qxf3 Rhf8 24.Qg3+ Ka8, when White will have to try and draw after 25.Qxe3 Qxe3 26.dxe3
Nf2+ 27.Kc2 Nxh1 28.b5.

18...e3!

Now threatening ...Qc6+, as well as ...Nxb4+. Not 18...gxf4? 19.Kb1 Rhe8 20.Ra3 and Black’s
initiative dissipates.

19.Nc3 Qc6

536
20.Rf1

If 20.Rd1 gxf4!, the new passed f-pawn provides Black with sufficient counterplay; e.g. 21.dxe3 (or
21.Kb1 f3 22.Qxe3 f2, threatening ...Rhe8) 21...f3 22.Qf1 (after 22.Qxd3 Rxd3 23.Rxd3 f2 24.Rd1
Qc4 25.b3 Qxg4 26.Rf1 Qg2 27.Bd2 Qxh2, Black gets a passed h-pawn too) 22...Nxb4+ 23.Kb3
Rxd1 24.Qxd1 (or 24.Nxd1 Rd8 25.Nf2 Nc2 26.Rb1 Rd4!) 24...Rd8 25.Bd2 (not 25.Qf1? Nd3 26.e4
f2 27.Be3 because of 27...Qe6+ 28.Kc2 Nb4+ 29.Kb1 Qb3 30.Ra3 Qc2+ 31.Ka1 Rd3! 32.Bxf2 Rd2
33.Qb1 Rxf2 and White is in trouble) 25...Nd3 26.Kc2 (or 26.Qc2 Qe6+ 27.Ka3 Qd6+ and draws)
26...f2 27.Qe2 Qc4 28.Rf1 Nb4+ 29.Kb1 Qxe2 30.Nxe2 Rxd2 31.Nf4 Kd7 with a level endgame.

20...exd2 21.Bxd2 Rhe8

537
22.Qxd3

White has extra material and so returns some in the hope of seeing out the storm.
The alternative was 22.Qf3 Nxb4+ 23.Kc1 (if 23.Kb3 then 23...Qe6+ 24.Kxb4 Rxd2 25.Rfb1 Qb6+
26.Ka3 Re3 27.Qf1 and 27...Qc5+ 28.Kb3 Rxb2+ 29.Rxb2 Rxc3+ 30.Ka2 Qd5+ 31.Kb1 Qe4+
32.Ka2 Qd5+ draws) 23...Qc4 24.fxg5 Kb8! (hiding the king eliminates any counter-tactics; Black
now threatens ...Qb3 and mates) 25.Ra3 Nd3+ 26.Kb1 Ne5 27.Qd1 (or 27.Qf5 Rxd2 28.Re1 Qb4
29.Ra2 Rxh2 with a draw if White takes the knight) 27...Rxd2 28.Qxd2 Qxf1+ 29.Qd1 (or 29.Qc1
Qg2) 29...Qc4 “and Black will recoup all the material; e.g. 30.Na2 Qe4+ 31.Ka1 Qxg4 32.Qxg4
Nxg4 33.Rc3 Nxh2 with an equal game” (Tay).

22...Rxd3 23.Kxd3 Rd8+ 24.Kc2 Qg2 25.Rfd1

Surrendering the f-file since 25.Rad1 drops a piece to 25...Rxd2+ 26.Rxd2 Qxf1.

25...gxf4

Once again this pawn, supported by the queen, gives Black enough for the technical material
deficit.

26.a5 f3 27.b5

538
27...Qxh2!?

A small novelty. A.Nekhaev-L.D.Nisipeanu, corr. 2012, concluded 27...f2 28.Rf1 Qf3 29.Rad1 b6
30.a6 Kb8 31.b3 Ka7 ½-½, as White can do nothing constructive. This was the first real chance for
Black to deviate, so I took it in Diadoco-tsmenace, ChessWorld.net 2019, just to prevent the game
from being a carbon copy. I wasn’t expecting any other result.

28.a6?!

Here 28.b6 f2 29.bxc7 Rd4 30.Rf1 Qg3 31.Rad1 Qd3+ 32.Kc1 Qf3 33.Kc2 Qd3+ is one way to
draw.

28...Kb8 29.a7+?

And now 29.Ra3 f2 30.Bf4! Rxd1 31.Bxh2 f1=Q 32.Bxc7+ Kxc7 33.Nxd1 Qe2+ 34.Kc1 Kb8
35.Ne3 Qxb5 36.axb7 Kxb7 37.Rc3 is a fortress.

29...Ka8

White perhaps had fantasies of mating the black king in the corner. Unfortunately, without the
above defences, White’s own king is in far more trouble. The game ended 30.b6 (otherwise 30.Ra4
b6) 30...cxb6 31.Nb5 Rc8+ 32.Kb3 Qe2 33.Nc3 Qc4+ 34.Kc2 b5 35.b3 Qxg4 36.Kb2 b4 37.Na4
Qd4+ 38.Ka2 Rc2+ 39.Kb1 Rc6 (threatening ...Qd3+ and mates) 40.Ra2 (or 40.Ka2 b5 41.Nb2 Ra6+
42.Kb1 Qe4+ etc) 40...b5 41.Nb2 f2 42.Bg5 Qxd1+ 43.Nxd1 f1=Q 0-1.
The point is that 11...Qxf4? lands Black in a bad 9.f4 Qxf4 line, having answered 10.Nxa7+ Bd7
11.Bxd7+ with 11...Nxd7? (rather than 11...Kxd7).

539
B22: 9.f4 Qxf4

The main line of the main line. By returning the pawn White covers g2 with tempo and draws the
enemy queen into the line of fire, enabling a discovery with d2-d4.
Black has to acquiesce as 9...Qh4+? is unsound: 10.g3 Qh3 11.Ne5+ c6 12.Bc4 Bc5 13.d3!, and if
13...Ng4 then 14.Qxe4 Nf2 15.Bf7+ leads to a decisive advantage; e.g. 15...Kd8 (or 15...Ke7 16.Qc4
Bb6 17.f5! Qxf5 18.Qb4+ c5 19.Qh4+ Kf8 20.0-0 Qxe5 21.Bc4, W.Goedhart-F.Espinosa Pabon, corr.
1998) 16.Qc4 Bb6 and here Stockfish gives 17.f5! (improving on the previously obligatory 17.Rf1,
which is also very good) 17...Nxh1 18.Bg5+ Kc7 19.Bf4 Be3 (or 19...Kd8 20.Be6) 20.Qb4! c5 (or
20...Bxf4 21.Qe7+ Bd7 22.Qxd7+ Kb6 23.Qd4+ Kc7 24.Qxf4) 21.Qa5+ b6 22.Qa4 Qg2 23.Bxe3 and
White wins (apparently).
B221: 10.Nxa7+
B222: 10.d4
B223: 10.Ne5+
All other moves are mistakes:
a) 10.Nd4+? (the wrong square: the knight is vulnerable here, blocks the d-pawn, and leaves the b8-
h2 diagonal as a tactical weakness) 10...c6 11.Bc4 Bd6

540
 12.Nf3 (or 12.d3 Qh4+) 12...Kd7 13.Nd4 Qh4+ 14.Qf2 Qxf2+ 15.Kxf2 Ng4+ 16.Ke2 Bxh2 and
Black was a pawn up, M.Seris Granier-J.Moreira Lopes, corr. 2003.
b) 10.d3? Qd6 sees the white knight in trouble: 11.Na5+ (11.Nxa7+ Bd7 12.Bxd7+ Qxd7, and
11.Ne5+ c6 12.Nc4 Qb4+ 13.Bd2 Qxb5, P.Gorbea-R.Hernando, Barcelona 1996, are no better)
11...c6 12.Bc4 Qe5 13.Bd2, V.Koutecky-M.Vesely, Klatovy 1997, and now 13...Bg4 forces White to
give up a pawn for nothing with 14.d4 (or 14.Qe3 Bc5) 14...Qxd4.
c) 10.g3? Qd6 11.Nxa7+ Bd7 12.Bxd7+ Qxd7 13.Nb5 0-0-0 is a much superior version of line B21
for Black. F.Da Silva-L.Mezera, corr. 2015, continued 14.a4 Bc5 15.b4 Bxb4 16.c3 Ba5 17.Nd4 e3!
18.0-0 Rhe8 19.Qb5 e2 20.Qxd7+ Rxd7 21.Re1 Bb6 22.a5 Bxd4+ 23.cxd4 Rxd4 and Black was again
a pawn up.
d) 10.Rf1? drops the h2-pawn for no reason: 10...Qxh2 11.Nxa7+ (or 11.Rxf6? gxf6 12.Qxe4+ Kf7
13.Bc4+ Kg7 14.Qd5 Qg3+ 15.Kf1 Qg6 16.Nd4 c6 17.Qd8 Qg3 18.Qe8 Qf4+ 19.Ke1 Qe5+ 20.Qxe5
fxe5 and Black was winning, C.Gorka-J.Tait, Nottingham 1996) 11...Bd7! (stronger than 11...c6
12.Nxc8 Rxc8 13.Bc4, or 11...Kd8 12.Nxc8 Kxc8 13.d4, K.Seegert-M.Strange, Kolding 2016, though
Black won anyway) 12.Bxd7+ Nxd7! 13.Nb5 (or 13.Qxe4+ Be7 14.Qxb7 Qe5+ 15.Kd1 Qh5+ 16.Rf3
Qh1+) 13...0-0-0 (I&K) and Black is clearly better, not even a pawn down.

B221: 10.Nxa7+

541
 Despite the engines often favouring this move, it shouldn’t give White anything. I’ve faced it
fifteen times in the last couple of years, scoring W7, D8, L0. In four games White didn’t even try,
electing to force a repetition in the opening. That’s the only thing wrong with this for Black. If your
opponent wants to take a draw there’s nothing you can do about it.

10...Bd7

So that White will have to spend time extracting the offside knight. Not 10...Kd8? 11.Nxc8 Kxc8
12.d4 and Black is a pawn down for nothing.

11.Bxd7+

Not 11.d3? Qd6 as in note ‘b’ above; while 11.d4 allows 11...Qh4+ 12.g3 Qh3 13.Bxd7+ Qxd7
14.Nb5, F.Vidalina-M.Tritt, corr. 2004, and now 14...Bb4+! 15.c3 (or 15.Nc3 Qxd4) 15...Be7 (or
even 15...c6!? 16.cxb4 cxb5) 16.Bf4 (or 16.a4 c6 17.Na3 Rxa4) 16...Nd5 17.0-0 c6 18.Na3 Bxa3
19.bxa3 0-0 with equality.

11...Kxd7!

Not 11...Nxd7? 12.d4, which is what we had to avoid in line B21.

542
12.Qb5+

Coming in for either another pawn or the draw. Instead:

a) 12.d4?! is well met by 12...e3! (improving on 12...Qf5 13.Qb5+ Qxb5 14.Nxb5 c6 15.Nc3 Bb4,
Mi.Adams-J.Lautier, Terrassa 1991, where Black was only equal) 13.Bxe3 (or 13.Nb5 c6 14.Bxe3
Qh4+ 15.g3 Qe4 16.0-0 cxb5 17.Qxb5+ Kc8 18.Rae1 Bd6 19.c3 Qc6, U.Jansen-M.Preussner, corr.
2012) 13...Qe4 14.Nb5 Re8 15.Kd2 Nd5 16.Rhe1 (not 16.Rae1?? Nxe3 17.Qxe3 Qxg2+ 18.Re2 Bb4+
19.c3 Qxh1 and wins, J.Van Cappellen-S.Beukema, Houffalize 2011) 16...Be7 17.Nc3 (not 17.Qd3?!,
P.Reichard-R.Brachtel, corr. 1996, because of 17...Qg4!, threatening ...Nxe3, which prompts 18.h3
Qxg2+ 19.Re2 Qxh3, picking up both kingside pawns) 17...Qxe3+ 18.Qxe3 Nxe3 19.Rxe3 Bg5 and
Black has the better chances.
b) 12.Nb5 “should not worry Black. For only one pawn, Black has plenty of play here” (Sokolov);
e.g. 12...Re8 (or 12...c6 13.Nc3 Bd6 – S&S) 13.b3 (not 13.d4?! e3!, as in C.Monzon-R.Bergquist,
corr. 2003) 13...Qe5! (rather than 13...Bb4?! 14.c3 Bc5 15.Ba3, V.Lukov-V.Inkiov, Bulgarian Ch.,
Sofia 1984) 14.c3 (or 14.Nc3?! e3! – Sokolov) 14...c6 15.Nd4 Bc5 16.Bb2 h5 17.0-0-0 Nd5 and
Black indeed has quite enough.
c) 12.g3!? seems never to have been tried, perhaps because it creates a hook for Black’s
counterplay; e.g. 12...Qe5 13.Qb5+ (or 13.Nb5 c6 14.Nc3 Bc5 – I&K) 13...Qxb5 14.Nxb5 Ra5!?
15.a4 c6 16.Nc3 h5, intending ...h5-h4 and so on, as in J.Clancey-M.Mujunen, corr. 2015 (by
transposition).

12...Ke6

543
13.Qxb7

Consistent – and “Courageous!” (Flear). White’s other options are to push the d-pawn or throw in
further queen checks, either for the draw or to bring about subtle changes in the position:
a) 13.d4 is best met by 13...e3! 14.Rf1 (not 14.Qe5+? Qxe5 15.dxe5 Nd5 and Black is better,
E.Ritter-J.Heesen, corr. 1995) 14...Qh4+ 15.g3 Qxh2 16.Bxe3 Qxg3+ 17.Kd2 Qg2+ with a draw,
since 18.Bf2?? Rxa7 and 18.Rf2?? Ne4+ are clearly unplayable.
b) 13.d3 is a reasonable try.

Black should again reply 13...e3! 14.Rf1 Qh4+ 15.g3 Qxh2 16.Bxe3 Qxg3+ 17.Kd2 Qg2+,

544
although this time, with e4 covered and the knight defended, White can block the check. From the
other side, Black has regained the pawns, so the only issue is to secure the king, to which end an
exchange of queens is called for: 18.Rf2 (or 18.Bf2 Qd5 19.Rae1+ Kf7 20.Qxd5+ Nxd5 21.Nb5 c6
22.Bc5+ Nf6 23.Bxf8 Rhxf8 24.Nd6+ Kg8 25.a3 Rab8, which led to a draw in M.Höppenstein-
R.Sabel, corr. 2017) 18...Qd5 19.Qxd5+ (if 19.Qa4 h5 20.Re1 Kf7 21.Ref1 ½-½ H.Faber-
P.Leisebein, corr. 2014, Black might play on with 21...c6!, as White’s knight is more of a problem
than Black’s king) 19...Kxd5 20.Nb5 Kc6 and Black is fine. I even won after 21.c4 Ng4 22.Rf7 Bb4+
23.Nc3?! (here 23.Ke2 Nxe3 24.Rxc7+ Kb6 25.Kxe3 Rhf8 26.Rxg7 Rae8+ 27.Kd4 Rf4+ 28.Kd5
Rf5+ 29.Kd4 would be a draw) 23...Rhf8 24.Raf1?? Rxf7 25.Rxf7 Bxc3+ 26.bxc3 Rxa2+ 0-1
Pawn1957-tsmenace, ChessWorld.net 2018.
c) 13.Qc4+ Kd7 and now:

c1) 14.d4?, as in D.Velimirovic-Z.Nikolic, Yugoslav Ch., Novi Sad 1995, again runs into 14...e3!
(Sokolov) 15.Rf1 (not 15.Qb5+? Kd8 16.Rf1 Qxd4, or 15.Nb5? c6 16.Nc3 Qf2+ 17.Kd1 Qxg2 –
Sokolov; e.g. 18.Re1 Re8 19.Bxe3 Bd6 and Black wins according to Stockfish) 15...Qh4+ 16.g3 Qxh2
17.Qb5+ Kd8! 18.Qxb7 Qxg3+ 19.Kd1 Qg4+ 20.Ke1 Rxa7! 21.Qxa7 Bb4+ 22.c3 Qg3+ 23.Ke2
Qg2+ 24.Ke1 Bd6 (Sokolov), and if 25.Qb8+ (not 25.Bxe3? Bg3+ 26.Bf2 Qe4+ 27.Kd2 Bf4+ and
mates) 25...Ke7 26.Qb5 Re8 27.Qe2, then 27...Qg6 28.Bxe3 Kd8 29.Rg1 Bg3+ 30.Kd2 Nd5 31.Qd3
Qxd3+ 32.Kxd3 Rxe3+ 33.Kc4 Nf6 with a clear advantage in the endgame.
c2) 14.Nb5 c6 15.Qd4+?! (15.Nd4 Bd6 is still probably equal) 15...Kc8 16.g3?! (or 16.Nc3 Qg5! –
Sokolov) 16...Qf5! 17.Nd6+ Bxd6 18.Qxd6, E.Schmittdiel-V.Inkiov, Gausdal 1989, and now
Stockfish likes 18...Ra5!, when White is already in serious trouble with the queenside sitting idle.
c3) 14.d3 Qe5! (more accurate than 14...Qh4+ 15.g3 Qh5 16.Qb5+ Qxb5 17.Nxb5 exd3 18.cxd3
Bc5, J.C.Howell-J.Tait, Nottingham 1991, since the extra g2-g3 is not unhelpful to White, even if
Black should still be fine) 15.Qb5+ (not 15.Be3? Rxa7 16.Bxa7 Qa5+ 17.c3 Qxa7 with advantage,
Sa.Campbell-M.Kern, corr. 1997, and 16...exd3+! 17.Kf1 Qe2+ 18.Kg1 b6 looks even better)

545
15...Qxb5 16.Nxb5 exd3 17.cxd3 Bc5 18.Nc3 Rhf8 with excellent compensation for the pawn.
d) 13.Qb3+ Kd7 and now:

d1) 14.d4?, A.Drei-M.Vivo, Bologna 1992, again allows 14...e3! 15.Rf1 Qh4+ 16.g3 Qxh2
17.Qb5+ Kd8! 18.Qxb7 Qxg3+ as in note ‘c1’ above.
d2) 14.Nb5 can be met by 14...Bc5! (14...Qh4+?! 15.g3 Qh3 16.d4! is good for White) 15.d4 (not
15.Qf7+?! Be7 16.Nd4 Qe5 17.Qc4 Nd5 and Black is better; while if 15.Qh3+?!, M.Ulibin-
G.Timmerman, Cappelle la Grande 1998, then 15...Ng4! gives Black good play, e.g. 16.Rf1 Bf2+
17.Kd1 Rae8 18.Rxf2 Qxf2 19.Qxg4+ Kc6) 15...e3 16.Rf1 (not 16.Bxe3? Rae8 17.Kd2 Bb4+!
18.Qxb4 Qxe3+ 19.Kd1 Ne4 and Black won, K.Shoup-R.Brachtel, corr. 1996) 16...Qh4+ 17.Kd1
Qg4+ 18.Ke1 and Black has at least a draw with 18...Qh4+.
d3) 14.g3 Qe5 (14...Qg4 is okay too) 15.Qxb7 (15.Qb5+ returns to 12.g3 above) 15...Bc5! 16.Qc6+
Ke7 17.Nb5 Rhb8 18.c3 (not 18.c4? Rxb5! 19.cxb5 Rd8 20.b4 Bd4 21.Rb1 e3 22.dxe3 Bc3+ 23.Ke2
Qh5+ with a decisive attack, R.Marconi-J.Clancey, corr. 2015) 18...Ra5 19.b4 Raxb5 20.bxc5 Kf7 ½-
½ W.Class-P.Leisebein, corr. 2013, presumably in view of 21.0-0 Rxc5 22.Qd7+ Kg6 23.Qd4 Qxd4+
24.cxd4 Rd5.
d4) 14.Qxb7 Bd6 looks very similar to the main line, but there are some differences when g2-g3 is
inserted:

546
 d41) 15.g3 Qf3 16.Qb5+ c6! (stronger here than 16...Ke6 17.Rf1 Bxg3+ 18.hxg3 Qxg3+ 19.Kd1
Qg4+ 20.Qe2 Qxe2+ 21.Kxe2 Rxa7 22.b3, when White can play for a win at no risk, samurai-
tsmenace, ChessWorld.net 2011) 17.Qxc6+ Ke7 18.Qb7+ (or 18.Rf1 Bxg3+ etc) 18...Ke6 19.Rg1
Rhb8 20.Qxg7 Ng4 21.Rf1 Bxg3+ (finally) 22.hxg3 Qxg3+ 23.Ke2 Qg2+ ½-½ G.Plume-M.Mujunen,
corr. 2016.
d42) 15.Qb5+ Ke6 16.Qc4+ Kd7 17.g3 (17.Qf7+ is the main line) 17...Qg4! (17...Qf3 is similar to
samurai-tsmenace above) 18.Qf7+ (or 18.Qa4+ Ke6 19.Qb3+ Nd5 20.Nb5 e3! 21.Nxd6 cxd6 22.c4
exd2+ 23.Kxd2 Qg5+ 24.Ke2 Qg4+ 25.Qf3 Qxc4+ 26.Qd3 Qxd3+ 27.Kxd3 Nb4+ 28.Ke2 Rxa2 ½-½
loper10-tsmenace, ChessWorld.net 2020) 18...Be7 19.Nb5 Ra4! 20.c3 c6 21.Nd4 Rxd4!? 22.cxd4
Nd5 23.Qf2 Rf8, intending 24.Qe2 Rf3 and ...Nb4 or 25.a3 Bg5 and ...e4-e3. Instead, I got 24.Qg2??
Nf4 25.Qxe4 Re8 0-1 bulent_2010-tsmenace, ChessWorld.net 2018.

13...Bd6

547
 Black is two pawns down, with the king in the middle of the board too, and yet has fully adequate
compensation in terms of activity and time. While White is trying to develop, return the offside
knight and secure their own king, Black has attacking ideas with ...Qh4+ (and ...Bxg3), ...Ng4 and
...e4-e3, and can bring the rooks into action via f8, e8, and/or a6.

14.Qb3+

The only good move. Having grabbed the pawn, the queen needs to relocate pronto.
Not 14.d4?? Qh4+ 15.Kf1 (or 15.g3 Bxg3+) 15...Nd5 16.Nb5 Rhf8+ 17.Ke2 Rf2+ 18.Kd1 Qg4+ 0-
1 A.Stacey-W.Goedhart, corr. 1998. Or if 14.g3?!, A.Kuhlmann-G.Staf, corr. 1994, then 14...Qf3!
15.Rf1 Rhb8! 16.Qa6 Qh5 17.Qc4+ Kd7 18.Nc6 Rb6 19.Nd4 Qxh2 with advantage.

14...Kd7

Not 14...Ke7? as that gives White a tempo with the knight: 15.Nc6+ Kd7 16.Nd4 Qh4+ 17.g3 Qh3,
when Flear writes: “I get the impression that White is doing well about here”, which is certainly true;
e.g. 18.Qf7+ Be7 (or 18...Kc8 19.Qe6+) 19.Rf1 Rae8? (but if 19...Raf8 20.Qc4 Bd6 then 21.b3 Nd5
22.Bb2, intending Rxf8 and 0-0-0) 20.Rf5! Kd8 21.Qe6 Nd7 22.Nc6+ Kc8 23.Ne5 1-0 S.Zielinski-
F.Colombo Berra, corr. 1994.

548
15.Qf7+

Assuming White isn’t going to take the draw after all (with 15.Qb5+ etc) – because why would
you?! – this check is as good a try as any. Others:
a) 15.Nb5? is met by 15...e3! (not 15...Bc5?, as I’ve played twice, because of simply 16.Qf7+! Kc6
17.d4 e3 18.Qxc7+ or 16...Be7 17.Qc4 and Black’s pieces are misplaced) 16.Qd3 (not 16.Rf1??
exd2+ 0-1 S.Cullip-R.Ackermann, Bern 1991; or 16.dxe3? Qh4+ 17.Kf1 Rhf8 18.Nd4 Ne4+ 19.Ke2
Rf2+ 20.Kd1 Qg4+ 21.Ke1 Raf8 0-1 Abacus-tsmenace, ChessWorld.net 2019) 16...Qf2+ 17.Kd1
Qxg2 18.Rf1 (or if 18.Qf5+ Kd8 19.Re1, H.Wiesmann-G.Timmerman, Dutch League 1998, then
19...Ra5! is again strong) 18...Qg4+ 19.Qe2, M.Dabija-M.Auzins, corr. 2003, and now 19...Bxh2!
20.Qxg4+ (or 20.dxe3 h5!) 20...Nxg4 21.Rf7+ Ke6 22.Rxg7 (or 22.Rxc7 Be5!) 22...h5 leaves Black
with a very dangerous passed h-pawn.
b) 15.Qh3+?! sidelines the queen after 15...Ng4! 16.Rf1 (not 16.Nb5? Qf2+ 17.Kd1 Qf5 – I&K,
and Black is winning; e.g. 18.Nxd6 cxd6 19.Ke2 Rhf8 20.Qg3 Nf2 21.Rf1 Qh5+ 22.Ke3 Ng4+ etc)
16...Qg5 17.d3 (similarly 17.d4 e3 18.Nb5 Rhf8; and not 17.Nb5? Raf8 18.Rh1 Bxh2 19.d3 Bf4
20.Bxf4 Rxf4 21.Nc3 Re8 0-1 T.Stock-A.Curran, corr. 1999) 17...e3 18.Nb5 Rhf8 19.Rf3 h5! and
White is struggling; e.g. 20.Nxd6 cxd6 21.a4 Qc5! 22.c3 Qd5 23.Bxe3 Rae8 24.0-0-0 Qa2 25.Rxf8
Rxe3! 26.Qh4 Qa1+ 27.Kc2 Qxa4+ 28.Kc1 Kc6! 29.h3 g5 30.Qxg5 Re2 31.Rc8+ Kb7 32.Qd5+ Kxc8
33.Qc4+ Qxc4 34.dxc4 Ne3 35.Rd3 Nxc4 and Black won the endgame, S.Lucki-L.Ellis, corr. 2010.
c) 15.d4!? prompts the usual 15...e3!

549
 16.Bxe3, when Black is fine after any of 16...Bb4+ (M.Alcock), 16...Qe4 (An.Dyce), or 16...Qh4+
(J.Wagenbach). I opted instead for 16...Rae8? 17.Qb5+ Kd8 in An.Dyce-J.Tait, Notts League 1995,
when Stockfish proposes 18.Nc6+ Kc8 19.0-0-0! Qxe3+ 20.Kb1 as good for White.
d) 15.g3 should be met by 15...Qf3!.

Black has no need to fear the exchange of queens as the new f-pawn would be very dangerous; in
many cases even Rf1xf3 is not a threat. We’ll see this idea again in the main line. The exclamation
mark was the result of hard analysis by me and Volker Hergert, so it’s interesting to look at it now
with Stockfish. The engine finds the positions to be completely trivial:
d1) 16.Qxf3? exf3 17.Nb5 Rae8+ 18.Kf1 (if 18.Kf2 then 18...Ne4+ 19.Kxf3 Rhf8+ wins; e.g.
20.Kg2 Rf2+ 21.Kh3 Ng5+ 22.Kh4 Re4+ 23.Kxg5 Be7+ 24.Kh5 Rf5 mate) 18...Re2! (we’d given

550
18...f2 as best, and then 19.Kxf2 Ng4+ 20.Kg2 Rhf8 21.Nc3 Rf2+ 22.Kh3 h5 with good play, as later
occurred in F.Hoffmann-B.Hanison, corr. 1995) 19.Nd4 Rhe8 and Stockfish says that Black is
winning; e.g. 20.Nxe2 (or 20.Nxf3 Ng4) 20...Rxe2 21.d3 Ng4 22.Bg5 (or 22.Bf4 Bxf4 23.gxf4 Ne3+
24.Kg1 Rg2 mate) 22...Nxh2+ 23.Rxh2 Rxh2 24.Kg1 Bxg3 25.Be3 f2+ etc.
d2) 16.Qf7+ Be7!? (for 16...Kd8 17.Nc6+ Kc8 see 17.g3 in the main line) 17.Rf1 Qg2 18.Nc6 (if
18.Qc4 Rxa7 19.Qd4+ Bd6 20.Qxa7 Qxh2 21.Qa4+ c6 22.Qc4 Qxg3+ 23.Kd1 h5 and 24.a4? h4
25.Qf7+ Kd8 26.c3 h3 27.Rxf6, Steve9834-PatzerBernie, ChessWorld.net 2003, then 27...h2! 28.Rf1
e3! 29.dxe3 Qg2 wins, now that Kc2 has been ruled out) 18...Kxc6 19.Qxe7 Raf8 (intending ...Nd7 –
Hergert) 20.Rf2? (White should bail out with 20.Qe6+ Kb7 21.Qb3+ as in H.Hausmann-P.Leisebein,
corr. 2000) 20...Qh1+ 21.Ke2 Kb7! (improving on 21...Nd7 22.Qe6+ Kb7 23.Qxd7 Rxf2+ 24.Kxf2
Qf3+ and draws – Tait & Hergert 1995) 22.Qb4+ (if 22.Qe6 then 22...Ne8!, or 22.b4 Nd5!) 22...Kc8
23.Qa5 Nd7 24.Qa6+ Kb8 25.Qb5+ Nb6 26.Qc5 Rxf2+ 27.Qxf2 e3 28.dxe3 Rd8 with a decisive
attack, PeterRs-tsmenace, ChessWorld.net 2009.

15...Kd8!

In an article written in 1998 for the BCCA magazine I concluded that blocking with the bishop was
best. I’ve changed my mind about that, though 15...Be7 is probably okay: 16.Nc6 Kxc6 17.Qxe7
Rae8 18.Qa3 (18.Qb4 can be answered by 18...Ng4! 19.Qc4+, J.Assunçao-P.Goffin, corr. 2013, and
now 19...Kb7! 20.Qf1 Qg5 21.a4 c6 22.Ra3 Rhf8 with typical compensation) 18...Rhf8 19.g3
(swapping queens with 19.Qg3 Qxg3+ 20.hxg3 Nd5 21.c3 e3 22.Rxh7 Rf2 is fine for Black, M.De
Oliveira-N.Patrici, corr. 2002; while 19.Qa6+ Kd7 20.Qe2?, which I’d given as “more testing”, is
well met by 20...Nd5!, e.g. 21.a4 e3 22.dxe3 Qh4+ 23.g3 Qe4 24.Rg1 Nb4 25.Rg2 Rf3 26.Ra3 Ref8)
19...Qf3! 20.Qa6+ (or 20.Rf1 Ng4!; not 20.Qxf3? exf3+ 21.Kf1 f2 22.Kxf2 Ng4+ 23.Kg2 Re2+
24.Kh3 Nf2+ and wins) 20...Kd7 21.Rf1 Ng4 22.Qc4 h5 23.Qd5+ Ke7 24.Rxf3? (taking the queen is
still a mistake) 24...exf3 25.Kf1 Nxh2+ 26.Kf2 Ng4+ 27.Kf1 f2 28.Qe4+ Kf7 29.d3 Rxe4 30.dxe4

551
Ke6 31.Bf4 g5 32.Bxg5 Nh2+ and Black won, W.Goedhart-J.Tait, corr. 2003.

16.Nc6+

The only other option is 16.g3 Qf3 17.Nc6+ Kc8, which transposes below.

16...Kc8

17.Ne7+

Instead:
a) 17.Qe6+?! Kb7 18.Nd4 Qh4+ 19.Ke2 Ra6 20.Qf5? (20.Qh3 Qg5 21.Rf1 is more tenacious)
20...Re8 21.Kd1 Re5 22.g3 Qh6 23.Qf2 Bc5 24.c3 Qg6 (looking at White’s rooks, bishop and king,
it’s unsurprising that serious trouble is looming) 25.Rf1 Rc6 26.Qe2 Bxd4 27.cxd4 Rd5 28.b3 Rxd4
29.Bb2 Rd3 30.Rc1 e3 31.Rxc6 Kxc6 32.Bc3 Ne4 33.Qxd3 Nf2+ 34.Rxf2 Qxd3 35.Rf4 g5 36.Rc4+
Kb7 37.Ke1 h5 and there is no fortress draw, draco69-tsmenace, ChessWorld.net 2018.
b) 17.g3 is again met by 17...Qf3! 18.Rf1 Bxg3+ 19.hxg3 Qxg3+ 20.Rf2 (20.Ke2 Qg4+ 21.Ke1
Qg3+ repeats) 20...Kb7!? (trying for the win) 21.Nd4 (or if 21.Ne7 Qg1+ 22.Ke2 Qg4+ 23.Kf1 Qd1+
24.Kg2 Qg4+ 25.Kh2 Ra5!? then 26.Qb3+ Ka8 27.Rf5! Rxf5 28.Qa4+ and White takes the draw –
Tait & Hergert) 21...Qg1+ 22.Rf1? (22.Ke2 Qg4+ is still a draw) 22...Qxd4 23.Qxg7? (but 23.Qb3+
Kc8 24.Qe6+ Kd8 is good for Black anyway) 23...Raf8 24.Qg1 Qd6 25.Qe3 Nd5 26.Qb3+ Nb6
27.Qb5 Rhg8 28.Rxf8 Qxf8 29.Qf1 Qc5 and White cannot survive, C.Dunn-B.Hanison, corr. 2000.

17...Bxe7 18.Qxe7 Re8

552
19.Qc5

From here the queen can retreat further to defend on the a7-g1 diagonal.
Instead:
a) 19.Qb4 Ra6! (rather than 19...Ng4 20.Rf1) 20.Qb5 Rb6 21.Qe2 Rc6 gives Black the better
chances; e.g. 22.c3 Rce6! (threatening ...e4-e3) 23.Qe3 Qe5 24.0-0 Ng4 25.Qf4 Qc5+ 26.Kh1 h5,
threatening ...Rf6.
b) 19.Qf7 allows 19...e3 20.dxe3 Rxe3+ 21.Bxe3 Qxe3+, when White has to be careful: 22.Kd1
(not 22.Kf1?? Ra4!) 22...Ra4 (or 22...Ra6 23.Qf8+ Kb7 24.Qb4+ Rb6 25.Re1! Qf2 26.Qc3 Rc6
27.Qb4+ Rb6 28.Qc3 and draws) 23.Qf8+?? (after 23.c3! Ra6 24.Qc4 Rd6+ 25.Kc2 White should
survive) 23...Kb7 24.c3 Ne4 25.Qd8 Nd6! 0-1 J.Kusmierek-S.Kudela, corr. 2000, since there is no
good defence to ...Rf4-f2 etc.
c) 19.Qxg7?? is far too greedy: 19...Rg8 20.Qe7 Rxg2 21.Qf8+ Kd7! 22.Qf7+ Kd6 and Black wins.

19...Ra6!

553
 This is the critical improvement, discovered by Peter Leisebein. I’d previously rejected 15...Kd8 on
account of 19...e3 (“!”) 20.dxe3 Ra6 (or 20...Ra4 21.g3 Qf3 22.Rf1, B.Gentinetta-F.Pezzi, corr. 1998,
where “White converted his advantage into the full point” – Flear) 21.Bd2! Re5 22.Qf8+ Kb7
23.Qb4+ Qxb4 24.Bxb4 Rxe3+ 25.Kd2 Re4 26.c3 and White is clearly better, W.Goedhart-J.Tait,
corr. 1998, even if I did manage to draw.
The threat is stronger than the execution. After 19...Ra6 something must be done about 20...e3!
21.dxe3 Re5 and wins. White has tried three moves.

20.g3

The others:
a) 20.Qf2 Qxf2+ 21.Kxf2 Rc6!

554
 (I like this idea from CCSIM Hans-Dieter Vötter; the obvious intention is to create light square
holes in the white queenside – in particular, the d3-square becomes a lovely outpost for the knight)
22.c3 Ng4+ 23.Ke1 (or 23.Ke2 Rf6 24.h3 Ne5 25.Rf1 Rg6 26.g4 Rh6 27.b3 Rxh3 28.Ba3 Nd3 and
Black is fine, A.Solari-H.D.Vötter, corr. 2015) 23...Rf6 24.b3 Nf2 25.Rg1 Nd3+ 26.Ke2 Rd8 27.Ke3
g5 28.c4 c5 (I’ve reached this position twice in online games) 29.Rb1 (my other game ended quickly:
29.a4 Nf4 30.g3 Nd3 31.Ba3 Rf3+ 32.Ke2 Rf2+ 33.Ke3 Rf3+ ½-½ bulent_2010-tsmenace,
ChessWorld.net 2019; note that White can’t take the e-pawn due to 34.Kxe4?? g4 and ...Rd4 mate)
29...Nf4 30.Bb2?! (instead, 30.g3 Nd3 31.Bb2 Rf3+ 32.Ke2 Rf2+ 33.Ke3 Rf3+ is another draw)
30...Re6 31.Bc1? (giving Black two tempi is too much; 31.Rbd1 Rd3+ 32.Kf2 e3+ 33.dxe3 Rdxe3
34.Bc1 Nd3+ 35.Rxd3 Rxd3 36.Bxg5 is still okay for White) 31...Rd3+ 32.Kf2 Rh6! 33.Rh1 (or
33.h3 Rf6 34.Ke1 h5) 33...Rh4 (threatening to push the g-pawn) 34.Kg1 e3! (or the e-pawn) 35.g3 e2
36.Bb2 Rxd2 37.Re1 Nd3 38.Bc3 Rd1 39.gxh4 Nxe1 40.Kf2 Nd3+ 41.Kxe2 Rxh1 42.Kxd3 Rxh2
43.hxg5 Rxa2 and Black won the endgame, Bahnfahrer0629-tsmenace, ChessWorld.net 2018.
b) 20.Qe3 Qf5 21.Rf1

555
 21...Qd7 (keeping the queens on; 21...Qg6 22.Qg3 Qxg3+ 23.hxg3 Rc6 looks fine too) 22.Qh3
(offering a queen swap which is of course declined; instead, playing to prevent ...Ng4 with 22.h3?!
allows Black a strong initiative after 22...Rc6 23.c3 Rd6, planning ...Rd3 and ...e4-e3 or ...Qd5
depending where the white queen goes, e.g. 24.b3? Rd3 25.Qa7 Qd5 26.Qa6+ Kd8 27.Kf2 Re5 0-1
Honeybunch-tsmenace, ChessWorld.net 2018, since 28.Kg1 Rg5 29.Rf2 e3! 30.dxe3 Rd1+ 31.Kh2
Rxg2+! 32.Rxg2 Nh5 and ...Qe5+ wins; rather than that, it would be better to let the black knight
come in, e.g. 22.a4 Ng4 23.Qe2 Rc6 24.Rf4 Nxh2 25.Kd1 e3 26.Ra3 Rce6 27.Rd3 exd2, S.Lucki-
G.Ziese, corr. 2011, when 28.Rxd7 Rxe2 29.Rff7 draws) 22...Rae6 23.Rf5 (or 23.a4 Nd5 24.Ra3 g5,
intending ...Nf4, T.Hagen-M.Preussner, corr. 2009) 23...Qc6 (if 23...e3 24.dxe3 Nd5, K.Boldysh-
G.Ziese, corr. 2011, then 25.Kf2 g6 26.Rf3 Nb4 27.Kg1 is another draw after 27...Nxc2 28.Rb1 Ne1
29.Rf1 Qd3 30.Ra1 Nc2 etc) 24.Qc3 e3 25.Qxc6 exd2+ 26.Kxd2 Rxc6 and again Black has typical
counterplay; e.g. 27.c3 Rd6+ 28.Kc2 Re2+ 29.Kb3 Rd1 30.a4 ½-½ C.Matei-G.Ziese, corr. 2009, in
view of 30...Rg1 31.a5 Ree1 32.Kc2 Rxg2+ 33.Kb3 Rgg1 34.Kc2.
c) 20.a4!? is a fourth, unplayed option, intending 20...e3 21.dxe3 Rae6 22.Ra3, when 22...Ng4
23.Rf1 Qxh2 24.Rc3 Qg3+ 25.Kd1 Rd8+ 26.Bd2 Nxe3+ 27.Kc1 Rxd2 28.Rxe3 Qxe3 29.Qxe3
Rxc2+ 30.Kxc2 Rxe3 reaches a drawn rook endgame.

556
20...Qf3

Preventing castling seems most logical, though 20...Qh6 21.0-0 Nd7 also offers good play; e.g.
22.Qc3 g5 23.Qe3 Qh5 24.d3 Rf6 ½-½ M.Wurschner-R.Sikorsky, corr. 2008, in view of 25.Rxf6
Nxf6 26.Bd2 Ng4 27.Qe2 exd3 28.Qg2 dxc2 29.Qa8+ Kd7 30.Qd5+ Kc8 etc.

21.Rf1 Rae6!

This move was the reason I made 20.g3 the main line. Black can just leave the queen en prise to the
rook – and does so for six moves, meanwhile bringing the knight up and pushing the kingside pawns.

557
22.Qb5

Taking the queen is a mistake: 22.Rxf3? was ½-½ J.Fuchs-G.Vetter, corr. 2011, but agreeing a
draw was premature. After 22...exf3+ 23.Kf1 Ng4! 24.Qf5 h5 25.Kg1 f2+ 26.Kg2 g6 27.Qb5 Re1
Black may well be winning, as work with Stockfish shows no perpetual check; e.g. 28.Qa6+ Kd7
29.Qd3+ Ke6 30.Qxg6+ Ke7 31.Qg5+ Kd6 32.Qg6+ R8e6 33.Qd3+ Kc6 34.Qc4+ Kb7 35.Qd5+ Rc6
36.Qb3+ Ka7 37.Qa4+ Kb8! 38.Qb4+ Kc8 39.Qf8+ Kb7 40.Qb4+ Rb6 and wins, now that there’s no
check on the h1-a8 diagonal.

22...Ng4 23.Qe2

Not 23.Rxf3?? exf3+ and ...Re1 mate.

23...h5!?

Black is okay after 23...Qxe2+ 24.Kxe2 Nxh2, but there’s no need to swap while the queen remains
immune.

24.a4

Still not 24.Rxf3?? exf3 25.Qe3 Nxh2, or 24.Qxf3?? exf3+ 25.Kd1 Re1+ 26.Rxe1 Nf2 mate.

24...h4 25.Ra3

Here 25.gxh4! would force Black’s hand: 25...Qh3 26.Ra3 e3 27.dxe3 Nxe3 28.Bxe3 Rxe3 29.Kd1!
Rd8+ 30.Kc1 Rxa3 31.Qa6+ (or 31.bxa3 Qxa3+ 32.Kb1) 31...Kb8 32.Qb5+ with a draw.

25...h3!

558
 Leaving the black queen en prise to both white rooks because she can’t be taken by either:
26.Rfxf3?? exf3, or 26.Raxf3?? exf3 27.Qe3 Rxe3+ 28.dxe3 Nxh2 29.Rh1 (or 29.Rf2 Ng4 30.Rxf3
h2 31.Rf1 Nf2!) 29...Ng4 30.Rxh3 f2+ 31.Ke2 (or 31.Kf1 Rd8 32.Ke2 Rd1) 31...Rf8 32.Rh1 Nh2!
33.Rf1 Nxf1 34.Kxf1 Rd8 35.Kxf2 Rd1 and wins.

26.b4 g5 27.a5

Now there’s nothing else to be done, so Black finally exchanges.

27...Qxe2+ 28.Kxe2 Nxh2 29.Rh1 Nf3 30.Bb2

Not 30.Rxh3?? Ng1+.

30...g4 31.Ke3 Rd8 32.Bc3 Rd5

559
 Black has control of the position and can aim to penetrate via ...Rf5, ...Ng5 and ...Rf3+. The game
tripoduk-jatait47, Chess.com 2018, continued 33.a6 (Stockfish suggests the mysterious 33.Ra4!? as a
possible defence) 33...Kb8 34.a7+ Ka8 35.Raa1 (not 35.Kf4? e3! 36.dxe3 Ng5!, threatening ...Re4
mate, or 37.Kxg4 Re4+ 38.Kh5 Nf3+ 39.Kg6 Re6+ and wins; e.g. 40.Kf7 Ng5+ 41.Kf8 Rd8+ 42.Kg7
Rd7+ 43.Kf8 Rf7+ 44.Kg8 Re8 mate) 35...Rf5 (proceeding as planned) 36.Ra4 Ng5 37.Ra5 Rf3+
38.Ke2 Nf7 39.Bd4 (intending Rc5 and Rxc7, or if 39...Rxg3 then 40.Rf1) 39...Re8 40.Be3 (now if
40.Rc5 then 40...Nd6 41.Rxc7 Nf5 42.Bf2 Rxg3! wins) 40...Nd6 (not yet 40...Rxg3? 41.Bf4 Rg2+
42.Ke3) 41.Rg5 (if 41.Bf4 then 41...e3! 42.Bxe3 Nf5 or 42.dxe3 Ne4 wins) 41...Nf5 42.Rxg4 Rxg3
43.Rxg3? (but 43.Rf4 Rxe3+ 44.Kf2 Rf3+ 45.Rxf3 exf3 46.Rxh3 Nd4 47.c3 Re2+ 48.Kf1 Nb5
49.Rxf3 Rxd2 would also win) 43...Nxg3+ 44.Kf2 Nxh1+ 45.Kg1 Ng3 and White resigned.

B222: 10.d4

560
 Straightforward and consistent. Having lured the target to f4, White pushes the d-pawn with tempo.
The drawback – if it is one – of playing d2-d4 straight away is that the black queen is able to return
homewards.

10...Qd6

Which is fortunate because 10...Qh4+? 11.g3 Qh3 fails to 12.Bg5! (rather than 12.Ne5+ c6,
transposing to line B223; or 12.Nxa7+ Bd7 13.Bxd7+ Qxd7 14.Nb5 c6 15.Nc3 Bb4, when Black has
excellent compensation) 12...Bd6 (nothing else is any better: 12...a6 13.Ba4 Bd7 14.Bxf6 gxf6
15.Qxe4+ Kf7 16.Ne5+! fxe5, R.J.Fischer-M.Matulovic, Herceg Novi blitz 1970, and now 17.Bb3+!
Be6 18.Rf1+ Kg8 19.d5! Qg4 20.Qe3 or 19...Bf7 20.d6! Bxb3 21.axb3 Qe6 22.Qxb7 Rc8 23.dxc7
and Rxa6 appears to win for White; while 12...Bd7 13.Ne5 c6 14.Bc4, J.Tait-W.Goedhart, corr. 1998,
is a much improved version of line B223 – the black bishop should be on e6) 13.Ne5+ (instead, 13.0-
0-0? a6 14.Bxf6 gxf6 15.Ba4 Bd7 16.Qxe4+ Kf8 17.Ne5 Bxa4 18.Rhf1 Kg7 19.Rf4 Qh6 20.Kb1
turned out well for Black in C.Thomson-J.Tait, corr. 1993 – or would have if I’d continued 20...Rhf8!
21.Rg4+ Kh8 22.Ng6+ hxg6 23.Rh4 Qxh4 24.gxh4 Be8 25.Qxb7 Bf7; whereas I made a series of
sub-optimal moves and resigned in a position Stockfish regards as still likely drawn) 13...c6
14.Bxc6+! (not 14.Nxc6? bxc6 15.Bxc6+ Kf7 16.Bxa8 Bg4 17.Qc4+ Be6) 14...bxc6 15.Qc4 Bxe5 (or
15...Be6 16.Qxc6+ Ke7 17.Qb7+ Bd7 18.g4! Rad8 19.0-0-0 with a crushing attack) 16.dxe5 Qe6
17.Qxe6+ Bxe6 18.exf6 gxf6 19.Bxf6 0-0 20.Rf1 and White is probably winning even with the
opposite-coloured bishops.
The variations arising from Fischer’s game have been analysed numerous times over the years.
Stockfish destroys the whole thing in a few seconds.

561
11.Ne5+

Not 11.Nxa7+? c6 12.Nxc8 (two of my opponents opted hopelessly for 12.0-0 Rxa7 and resigned in
due course, as in C.Duncan-J.Tait, Leics vs. Notts 1991) 12...Qb4+ and Black wins a piece, as the
knight is trapped after 13.c3 (or 13.Bd2 Qxb5 14.c4 Qxb2 15.0-0 Rxc8 and wins, L.Spassov-
A.Hennings, Zinnowitz 1965 – the first time this trap was sprung) 13...Qxb5 14.Qxb5 cxb5 15.Nb6
Ra6 16.Nc8 Kd7 0-1 A.Beresford-J.Tait, corr. 1998.

11...c6 12.Bc4 Be6

Grabbing the d-pawn is far too greedy. I’ll just mention a little trap which has caught out three
more opponents in rapid/ blitz games: 12...Qxd4? 13.Bf4 (simply 13.Bf7+! Ke7 14.Bf4 is terrible for
Black) 13...Bb4+?! (here 13...Nd5 is correct) 14.c3 Bxc3+ 15.bxc3 Qxc3+ 16.Kf2 g5!? 17.Rhc1
Qd4+ 18.Qe3 Qb2+ 19.Qd2??, when 19...e3+! 20.Bxe3 Ne4+ wins. With a little more time they
might all have found 19.Kg1! gxf4 20.Bf7+ Kf8 21.Qxf4, when Black is completely lost.

562
13.c3!

Defending the d4-pawn is simplest and strongest. White hopes to trade pieces on favourable terms
and leave Black with a weak e-pawn.
Instead:
a) 13.Bg5?! sacrifices the d-pawn for not very much: 13...Qxd4 14.Bxf6 gxf6 15.Bxe6 Qxe5
16.Qg4 Rd8 17.0-0 Bd6 18.Qh3 e3 19.Bc4, M.Kozakov-A.Karpatchev, Guingamp 1999, and now
19...Rd7!, making way for the king, is very good for Black.
b) 13.Bxe6 Qxe6 14.Bg5 (or 14.0-0 Bc5!? 15.Be3 Bb6) 14...0-0-0 15.c3 c5 already fights for the
initiative; e.g. 16.Qe3? h6 17.Bxf6 gxf6 18.Ng6 cxd4 19.cxd4? (19.Nxf8 Rhxf8 is still good for
Black) 19...Bb4+ 20.Kf2 Qf5+ 21.Nf4 Kb8 22.Rhd1 Bd6 23.g3 h5 and Black soon won, J.Kleine-
P.Leisebein, corr. 2004.
c) 13.Bf4 Bxc4 14.Qxc4 Qd5

563
 15.Qb3 (or 15.0-0-0 Bd6 and Black is fine) 15...Bd6! solves all problems by offering the b-pawn:
16.Qxb7!? (other moves lead to equality: 16.c4 Qa5+ 17.Bd2 Qc7; or 16.0-0-0 Qxb3 17.axb3 0-0
18.Rhe1 Rae8, A.Cela-P.Frendzas, Athens 1994; or 16.0-0 0-0 17.Ng6 Qxb3 18.axb3 Bxf4 19.Nxf4
½-½ D.Minic-M.Matulovic, Yugoslav Ch., Bor 1976) 16...0-0 17.c4 (not 17.Qxc6?? Qxc6 18.Nxc6
Bxf4, or 17.Qb3? Rab8 18.Qxd5+ Nxd5 and Black is clearly better, K.Lewandowski-H.Chomicki,
corr. 1996) 17...Qxd4 18.Nxc6 Qc5 19.Bxd6 Qe3+, when Sokolov writes that “White should consider
himself very lucky if Black does not have more than perpetual check.” Black does not have more:
20.Kd1 (not 20.Kf1?? Nd7+) 20...Qd3+ 21.Ke1 (not 21.Kc1?? Qxd6 22.Rd1 Qf4+ 23.Kb1 e3)
21...Qe3+ ½-½ gluv-tsmenace, ChessWorld.net 2011 (since 21...Qxd6?? 22.Ne7+ Kh8 23.Nf5 wins
for White).

564
13...Be7

With the queen in the way, the bishop has to occupy a lesser post.
This time 13...Bxc4 14.Qxc4 (for 14.Nxc4 Qe6 15.0-0 Be7 16.Bg5 0-0 see 15.Bg5 in the next note)
14...Qd5 15.Qb3! (15.0-0 Bd6 and 15.Bg5 Nd7!? are fairly level) 15...Bd6!? is more speculative:
16.Qxb7 0-0 17.Qxc6 (not 17.Nxc6?? Kh8, when the pin will cost White a piece) 17...Qxc6 18.Nxc6
Ng4! 19.Ne5 (GM Dorian Rogozenco mentions 19.h3! Bg3+ 20.Ke2 Rf2+ 21.Kd1 as an engine
tendency, which Stockfish continues 21...Nf6 22.c4 Rxg2 23.a4, followed by Ra3 with an edge)
19...Bxe5 20.dxe5 e3! 21.h3 Nxe5 22.Bxe3 Nc4 23.Bc5 (here 23.Bd4 Rae8+ 24.Kd1 Nxb2+ 25.Kd2
Nc4+ 26.Kc2 Ne3+ 27.Bxe3 Rxe3 is a draw – Rogozenco) 23...Rae8+ 24.Kd1 Nxb2+ 25.Kd2 Rf5
26.Bxa7 was S.Karjakin-L.D.Nisipeanu, Medias 2011. Black was running low on the clock or would
doubtless have found 26...Rfe5!, when “White can hardly avoid the repetition of moves” –
Rogozenco; e.g. 27.Kc2 (not 27.a4?? Re2+ 28.Kc1 Nd3+ and Black wins) 27...Nc4 28.Kd3 (and not
28.Bd4 Re2+ 29.Kb3?? Rc8) 28...Nb2+ 29.Kc2 Nc4 30.Kb3 Nd2+ etc.

14.Bf4

Threatening Ng6, so Black is virtually forced to simplify. I’ve made this the main line because the
options for White at move 16 typify the entire variation. Double castling with 14.0-0 0-0 is the more
common alternative:

a) 15.Bg5 Bxc4 16.Nxc4 (or 16.Qxc4+ Qd5 17.Qb3 Qxb3 18.axb3 c5!) 16...Qe6 17.Rae1 Rae8
18.Nd2 Qxa2!? (Kamsky’s suggestion 18...Qd5 is safer; e.g. 19.Bh4 Bd8 20.Nb3 Bc7 21.Bxf6 Rxf6
22.Rxf6 gxf6 23.Nc5 f5 24.Nxb7 Qxa2 and Black was okay, N.Ferreira-J.Deforel, corr. 1999; and not
19.Bxf6 Bxf6 20.Nxe4? since 20...Bxd4+ 21.cxd4 Qxd4+ wins a pawn) 19.Bxf6 Bxf6 20.Nxe4 Bh4!
(an old improvement on 20...Re6?! 21.Qg4 with an edge for White, G.Kamsky-J.Piket, Groningen
1995; not 20...Be7?? 21.Nd6 and White wins) 21.g3 Rxf1+ 22.Rxf1 Qe6 23.Re1 Be7! (blocking the

565
e-file against knight tricks) 24.Nc5 Qxe2 25.Rxe2 b6 26.Nd7 c5 led to a draw in tarby-tsmenace,
ChessWorld.net 2009.
b) 15.Bf4 Bxc4 16.Nxc4 (for 16.Qxc4+ Qd5 see 16.0-0 in the main line) 16...Qe6 17.Be5 Nd7
18.Bg3 c5!? (swapping rooks first with 18...Rxf1+ 19.Rxf1 c5 might be more accurate) 19.Rfe1!
(rather more testing than 19.Nd2 cxd4 20.Qxe4 Qxe4 ½-½ A.Southall-J.Tait, Derby 1996) 19...Rae8
20.d5! Qxd5 21.Rad1 Qe6 22.Qxe4 Qxe4 23.Rxe4 Nf6 24.Ree1, R.V.Hall-J.Tait, corr. 1995, where
Black now needs to be careful: 24...b5! (I played 24...b6? and ended up in a bind after 25.Ne5 Nh5
26.Nd7 Rf7 27.Bd6) 25.Nd6 Bxd6 26.Bxd6 Rxe1+ 27.Rxe1 Rd8! 28.Bxc5 Rd2, which Stockfish
assesses to be a drawn endgame.
c) 15.Bb3!? has only been played once to my knowledge.

Mi.Adams-J.Tait, Blackpool 1996, saw 15...Bxb3 (15...c5! is a logical response; e.g. 16.Be3 Bxb3
17.axb3 Qe6 or 16.Bf4 Bxb3 17.Ng6 Qd8 18.Nxe7+ Qxe7 19.axb3 cxd4 20.cxd4 Nd5 21.Be5 e3 with
equality) 16.axb3 Qd5?! (still 16...c5!) 17.Qc4 a6?! (not now 17...c5?? 18.Rxf6 and wins; 17...Bd6
was necessary) 18.Bg5 Rad8 19.b4 Bd6 20.Ra5 Qxc4 21.Nxc4 Bc7 22.Raf5 and White was clearly
better, if not as played: 22...Rd5 23.Ne5? Nd7 24.Rxf8+ Nxf8 25.Bf4 Ne6 26.Bg3 c5! 27.bxc5 Nxc5
28.dxc5 Bxe5 29.Bxe5 Rxe5 and I’d escaped into a drawn rook endgame – which GM Adams of
course won anyway.

14...Bxc4 15.Qxc4 Qd5

566
 White has tried four moves here:
a) 16.0-0-0 0-0 17.Rde1 Bd6 18.Qb3 Ng4 (or even 18...a5!? now White has gone long) 19.Rhf1
Bxe5 20.dxe5 (or 20.Bxe5 Rxf1 21.Rxf1 Nxe5 22.dxe5 Re8) 20...Qxb3 21.axb3 Rf5 22.h3 Nxe5
23.Rxe4 Nd7 was level in T.Gubaydulin-V.Nozdrachev, Dubna 2007.
b) 16.0-0 0-0 17.Rae1 (or 17.Bg5 Rae8 18.Qb3 c5, and if 19.Rxf6!? then 19...Qxb3 20.Rxf8+ Rxf8
21.axb3 Bxg5 22.dxc5 Be3+ 23.Kh1 Re8 and ...Bxc5) 17...Rae8 18.Bg5 Bd8! 19.Qb3 Qxb3 20.axb3
c5 21.Bxf6 Rxf6 22.Rxe4 Rxf1+ 23.Kxf1 cxd4 24.cxd4 Bb6 25.Kf2 ½-½ J.Plaskett-M.J.Turner,
Hastings 2000/01, as either 25...Bxd4+ or 25...Rxe5 follows.
c) 16.Rf1!? 0-0 17.Ke2 (getting ready for the endgame) 17...Rae8 (I might prefer 17...Bd6,
intending 18.Qxd5+ cxd5) 18.Qxd5+ Nxd5?! (18...cxd5 here too) 19.Bg3 Nf6?! 20.Nc4! c5 21.Nd6
cxd4 22.Nxe8 d3+ 23.Kd2 Rxe8 24.Rae1 and the immobile pawns were not worth the exchange,
loper10-Alekhine1962, ChessWorld.net 2013.
d) 16.Qb3 should be answered by 16...0-0! (after 16...Qxb3?! 17.axb3 Black had to work hard to
draw in J.Dourerassou-V.Inkiov, Guingamp 2004) 17.Qxb7 (or 17.0-0 c5) 17...Bd6. Notice that
White has the extra move Bf4 on Karjakin-Nisipeanu above, but this in fact helps Black since the
bishop is vulnerable there. In particular 18.Qxc6?? just loses a piece, while after 18.0-0 c5! 19.Qxd5+
Nxd5 Black looks fine; e.g. 20.g3 (or 20.Bg3 Rxf1+ 21.Rxf1 cxd4 22.cxd4 Rc8) 20...Nxf4 21.Rxf4
(or 21.gxf4 cxd4 22.cxd4 Bb4) 21...Rxf4 22.gxf4 cxd4 23.cxd4 Rf8 24.Rf1 Bb4, intending ...Bd2.

B223: 10.Ne5+

567
 S&S give this an exclamation mark, while Tseitlin & Glazkov call it “the critical test of Black’s
variation.” Certainly it’s the best scoring move for White in MegaBase, so Black must be prepared to
answer whatever comes their way. Since there’s a lot to get through, I’ll mostly be passing over
alternative options for Black. Regard the first few moves as forced.

10...c6 11.d4

Note – in case you hadn’t – that 11.Nxc6?? loses to 11...a6! 12.Ba4 Bd7, since the knight is now
pinned; while 12.Rf1!? Qxh2 13.Rxf6 gxf6 14.Qxe4+ Kf7 15.Bc4+ Kg7 16.Ne7 Qg3+ 17.Kf1 Qe5
came to nothing in J.Muno-B.Hanison, corr. 1994.

11...Qh4+ 12.g3 Qh3

568
13.Bc4

Still not 13.Nxc6?? a6 14.Ba4 Bd7 and wins, unless you’re stupid: 15.Bg5 Qg4?! 16.Qxg4 Nxg4
17.d5 h6 18.Bf4 Bc5? 19.b4 Bb6? 20.c4 Nf2? 21.c5 Nd3+ 22.Kd2 Nxf4 23.gxf4, when Black was
more at risk of losing, An.Dyce-J.Tait, Derby 1994.
And let 13.Ba4 stand as a warning against routine play: 13...Be6 14.Bg5 0-0-0? 15.0-0-0 Bd6?
16.Nxc6! bxc6 17.Qa6+ and White wins. With the white bishop missing from the a2-g8 diagonal,
Black can happily go short: 13...Bd6 14.Bf4 0-0 15.Qf1 Nd5 16.Qxh3 Bxh3 17.Bb3 Be6 and already
stands well, J.Gilland-B.Hanison, corr. 1999.

13...Be6

569
 I like the geometry of this move, with the queen unusually supporting the bishop backwards down
the diagonal. There’s no merit in allowing Bf7+ in any case.

14.Bg5

By far the most popular continuation. Of the rest, 14.Bf4 is the most significant:
a) 14.Be3 creates no pressure, so Black can castle short: 14...Bd6 15.0-0-0 0-0 16.Qf1 Nd5
17.Qxh3 Bxh3 18.Rhe1 Rae8 led to a quick draw in A.C.Hunt-J.Tait, Nottingham 1994.
b) 14.Bxe6 Qxe6 15.Qc4 is neutralized by 15...Qxc4 16.Nxc4 b5! (not 16...0-0-0?! 17.c3 c5 18.Be3
Ng4 19.Ke2!, when the black e-pawn is looking vulnerable, C.Tippleston-J.Tait, Birmingham 2000)
17.Ne5 c5, blowing up the centre and equalizing.
c) 14.Qf1 is a move to be wary of, but not yet: 14...Qxf1+ (or 14...Nd5) 15.Rxf1 Nd5 16.Bb3 Bd6
17.Nc4 Bc7 18.Ne3 0-0-0 19.Bd2 h5 and Black is fine, G.Toro Solis de Ovando-N.Patrici, corr. 2002.
d) 14.Bf4 can also be answered by 14...Bd6 15.0-0-0 0-0! (15...0-0-0 is nearly always played, or
else 14...0-0-0 first – indeed I’ve nearly always played this myself – but if it’s possible to go short in
the Jaenisch, I think Black nearly always should)

570
 16.Ng6 (instead: if 16.Rhf1 or 16.Rde1 then 16...b5 17.Bb3 Bxb3 18.axb3 Qe6 – Sokolov; 16.Bg5
b5 17.Bb3, F.Salamero Pelay-M.Royo Moya, Saragossa 1998, and 17...Bxb3 18.axb3 Bxe5 19.dxe5
Qg4 20.Qxg4 Nxg4 is level; 16.Kb1 Rad8 17.Bg5? Bxe5 18.dxe5 Rxd1+ 19.Rxd1 Bxc4 20.Qxc4+
Nd5 21.Qxe4 Qh5 had White resign in M.Horras-F.Torweihe, corr. 2008; while if 16.Qf1 then just
16...Nd5 17.Qxh3 Bxh3) 16...hxg6 17.Bxd6 Rf7, when Sokolov writes: “Black has his usual
counterplay related to the passed e-pawn, the open e- and f-files, and the well-placed knight on g4 or
d5, while Black’s doubled pawns do not have to be bad at all – often the contrary is true!” Stockfish
sees nothing wrong whatsoever with Black’s position; e.g. 18.Rde1 (not 18.Be5?! Bxc4 19.Qxc4 Nd5
20.Rhe1?, V.Nevostrujev-S.Starkov, Kemerovo 2010, as 20...Qh6+! 21.Kb1 Ne3 wins the exchange;
18.Kb1 Re8 19.Be5 Bxc4 20.Qxc4 Nd5 21.Rde1 e3 looks very good for Black too, S.P.Sethuraman-
D.Leygue, Marseilles 2006) 18...b5 19.Bb3 a5 20.Qf1 Qxf1 ½-½ D.Rivera Kuzawka-C.Ferron
Garcia, Mondariz 1995, since 21.Rhxf1 Bd5 is quite equal.
Incidentally, Stockfish wants to throw in 14...Bb4+!? 15.c3 before 15...Bd6 16.0-0-0 0-0 and says
it’s even more equal: resolutely “0.00”. This is yet to receive a practical test.

571
14...0-0-0

It’s necessary to castle long here. I had hoped to propose 14...Bd6 and ...0-0 again, which is fine for
Black in every instance – except the one given by GM Sokolov: 15.0-0-0 0-0?! (15...0-0-0! returns to
the main line) 16.Bxf6! gxf6 (not 16...Rxf6? 17.Qxe4) 17.Ng4! (the only way to show up Black’s
problems on the light squares) 17...Bxc4 (after 17...Rae8? 18.Bxe6+ Rxe6 19.Rhf1, Gild.Garcia-
G.Hernandez Guerrero, Matanzas 1994, White is close to winning because the desired 19...f5 loses to
20.Qc4 Rfe8 21.Nf6+ etc) 18.Qxc4+ Kg7 (or if 18...Kh8?! 19.Qe6 Rae8, A.Tamashiro-T.Van
Bommel, corr. 2000, then 20.Qf5! h5 21.Ne3 Qxf5 22.Nxf5 “and White’s dominant knight on f5 is a
monster” – Sokolov) 19.Qe6 h5 (or 19...Rae8 20.Qf5) 20.Qxd6 Qxg4 21.Rhf1 (Wedberg) with a
definite advantage, since Black’s position is rather loose. The best Stockfish can come up with is
21...Rae8 22.Rf4 Qe6 23.Qc5 Kg6 24.Qxa7 Rf7, intending ...f7-f5 and ...e4-e3, when it considers
Black to be only slightly worse, despite the pawn deficit. But that’s Stockfish. I wouldn’t fancy trying
to defend this over the board.

15.0-0-0 Bd6

572
 Perhaps the key position of the 9.f4 variation, where White has numerous tries.

16.Nf7

Forcing Black to surrender the light-squared bishop. GM Roeland Pruijssers says this would be his
choice here.
a) 16.Bxf6 gxf6 17.Bxe6+ Qxe6 is harmless. In effect White has relinquished the bishops to
strengthen Black’s pawns. Taking on f6 only makes sense if Black is unable to follow with ...f6-f5,
which is not the case here: 18.Qc4 (or 18.Nc4 Bc7 19.Kb1 f5 20.Ne3 Rhf8 21.c4 f4, V.Lukov-
V.Inkiov, Bulgarian Ch., Sofia 1982) 18...Qxc4 (18...Qg8!? is an option) 19.Nxc4 f5 20.Rhf1 Rhf8
21.Nxd6+ Rxd6 22.Rf4 Rh6 23.Rd2 Rh3 24.Rdf2 Kd7 and Black should hold without difficulty,
Cs.Balogh-M.J.Turner, Zürich 2013.
b) 16.Bxe6+ Qxe6 17.Qc4 (17.Bxf6 gxf6 is note ‘a’; Black can even spend a tempo to induce the
exchange, e.g. 17.Nc4 Bc7 18.Kb1 h6! 19.Bxf6 gxf6 20.Rhf1 f5 21.Ne3 Rhf8 22.Ng2 Rd7 and Black
was fine in the seminal game, C.Von Bardeleben-R.Spielmann, Berlin 1909) 17...Rhe8 18.Qxe6+
Rxe6 is no more impressive. The e-pawn is not such a weakness that White can win by such simple
means: 19.Nc4 (or 19.Nf7 Rd7 20.Nxd6+ Rdxd6 21.c3 Rd5 22.Bxf6 gxf6 23.Kd2, D.Kunzelmann-
G.Vetter, corr. 1996, and then, say, 23...Rh5 24.Ke3 f5 25.Kf4 Rg6 26.h3 Rhg5 27.Rhg1 Kd7, when
White cannot make progress) 19...Be7 20.Rhe1 (after 20.Bxf6?! Bxf6 21.c3 b5 22.Ne3 Bg5 23.Rhe1
Rf8 Black is certainly not worse, E.Güroff-J.Tait, corr. 1994) 20...Ng4! 21.Bxe7 Rxe7 22.Re2 Kc7
23.c3 b5 24.Ne3 Nxe3 25.Rxe3, when Black can even give the pawn away: 25...Rf8!? 26.Rde1 Kd6
27.Rxe4 Rxe4 28.Rxe4 Rf2 with a drawn rook endgame, G.Schuchardt-J.Tait, corr. 1995.
c) 16.Rhf1 Rhf8 (I’ve always put the rook here; it can go to e8 too, e.g. 16...Rhe8 17.Bxf6 gxf6
18.Rxf6 Bxe5 19.Rxe6 Rxe6 20.Bxe6+ Qxe6 21.dxe5 Qh6+! 22.Rd2 Rxd2 23.Qxd2 e3 24.Qe2 Qg5
25.Kd1 Kc7! led to a very high-level draw in Ma.Carlsen-L.D.Nisipeanu, Medias 2010)

573
 17.g4 (supposedly critical; instead, 17.Bxe6+ Qxe6 18.Nc4 Bc7 is much the same as ‘b’, or if
18.Qc4 then 18...Qe8! – I&K; while 17.Kb1 Bxe5 leads to equality: 18.dxe5 Rxd1+ 19.Rxd1 Bxc4
20.Qxc4 Nd5 21.Qc5 Rf1 22.Rc1 Rxc1+ 23.Bxc1 Qf5, juliangon-tsmenace, ChessWorld.net 2020,
intending 24.Qxa7 Qf1 and ...e4-e3 etc) 17...Bxc4 (not 17...Rde8?! 18.Bxe6+ Rxe6 19.Rg1! Bxe5
20.dxe5 Rxe5 21.Rg3 Qxg3 22.hxg3 Rxg5 23.Qe3, when 23...Rxg4? 24.Qxa7 Re8 loses to 25.Rd4!
Rg5 26.Rb4 b5 27.Rb3 Rh5 28.Ra3, so Black has to play 23...Re5 24.Qxa7 Rfe8 and doesn’t have
quite enough for the queen, Glawurtz89-tsmenace, ChessWorld.net 2018) 18.Qxc4 h6! 19.Bxf6 Rxf6
20.Rxf6 gxf6 21.Qe6+ Kc7 (I&K) should be a draw, though I’ve won this as Black: 22.Qf7+ Kb8
23.Qxf6 Qe3+ 24.Kb1 Bc7 25.c3 Qe2 26.Rf1? (26.Qf1 Qxh2 27.Qh1 Bxe5 is equal) 26...e3 27.Qf3?
Qxf3 28.Nxf3 Rf8, followed by ...e3-e2 etc, TicklyTim-tsmenace, ChessWorld.net 2019.
d) 16.Rdf1 Rhf8 is much the same as ‘c’: 17.g4 Bxc4 18.Qxc4 h6!, or 17.Bxe6+ Qxe6 18.Qc4
Qe8!, or lazily 17.Rf2 ½-½ M.Matulovic-V.Inkiov, Nis 1983.
e) 16.g4

574
 16...Bxc4 (Black can even allow the threat: 16...Rhe8!? 17.Bxe6+ Rxe6 18.Rdg1 Bxe5 19.dxe5
Rxe5 20.Rg3 Qxg3 21.hxg3 Rxg5, T.Marinsek-B.Markun, Slovenian League 2009, and is probably
okay as the rooks are placed more favourably than in note ‘c’) 17.Qxc4 Bxe5 18.dxe5 (not 18.Qe6+?!
Rd7 19.dxe5 Qxg4 – Tay, or 19.Qxe5 Re8 20.Qg3 Qxg4 21.Bxf6 Qxg3 22.hxg3 gxf6 and Black is
better, J.Klovans-E.Levi, Adelaide 1990) 18...Rxd1+ 19.Rxd1 Qxg4 is a simple equalizer; e.g. 20.Be3
(not 20.h4? e3 21.Qd3 e2! 22.Re1 Ne4 23.Qxe2 Qxe2 24.Rxe2 Nxg5 25.hxg5 Kd7, when Black has a
very good rook endgame, M.Dabija-H.D.Vötter, corr. 2003) 20...Nd7 21.a4 (if 21.Qf7 Nxe5 22.Qe7,
D.Donchev-V.Inkiov, Bulgarian Ch., Sofia 1989, then 22...Qf5 23.Qxg7 Rd8 – Sokolov; or 21.e6
Nb6 22.Qb3 Re8 – Tay) 21...Qf5 22.a5, S.Brynell-S.Lindemann, Swedish Ch., Borlänge 1995, and
now after 22...Nxe5 23.Qc5 a6 24.Qa7 Nd7 White must think about drawing, as with 25.Bf2 Qf4+
26.Kb1 Qxh2 27.Rxd7! Kxd7 28.Qd4+ Kc8 29.Qxg7 Qxf2 30.Qxh8+ Kc7 31.Qxh7+ and Qxe4.
f) 16.Qf1 Rhf8! (the exclamation mark is for consistency; I don’t know why I’ve generally gone for
16...Rhe8 here – as Sokolov says, the rook is better placed on the f-file in case White exchanges on
f6) 17.Nf7 (a showy move; 17.Qxh3 Bxh3 18.Nf7 Rd7 transposes) 17...Rd7 18.Qxh3 (or 18.Nxd6+
Rxd6 19.Qxh3 Bxh3; while 18.Bxf6 gxf6 19.Nxd6+ Rxd6 20.Qf4 Rd7 21.Be2 f5 was equal, at least
before 22.Qe5? Qh6+ 23.Rd2? f4! 24.gxf4 Rf5 25.Qxe4 Rxf4 0-1 R.Di Paolo-V.Inkiov, Clichy 2001)
18...Bxh3 19.Nxd6+ (or 19.Bxf6 gxf6 20.Nxd6+ Rxd6 and ...f6-f5) 19...Rxd6 20.Bf4 (or 20.Rhe1 b5
21.Bb3 Bg4) 20...Rd7 21.Rhe1 b5 22.Bb3 Bg4 23.Rd2 Nd5! 24.Bxd5 ½-½ F.Nijboer-V.Inkiov,
Utrecht 2005 (not 24.Rxe4?? because 24...g5 wins a piece – Sokolov).

16...Bxf7 17.Bxf7 Rhf8

575
18.Bc4

The standard move, albeit for no discernable reason.

Tay suggests that 18.Bb3 “may well be better, keeping the option of c2-c4 for an opportune
moment, as well as c2-c3 followed by Bc2.” Black should continue the same way in any case; i.e.
with ...Rde8, ...Kb8 and ...h7-h6. For example, after 18...Rde8 19.Rhf1 (not yet 19.c4?! Qg4! 20.Qe3,
D.Velimirovic-J.Klinger, Palma de Mallorca 1989, when “simply 20...h6 21.Bxf6 Rxf6 can only be
better for Black” – Sokolov) 19...Kb8 20.Rde1 h6 21.Bf4 (not 21.Bxf6 gxf6 22.d5?! c5 23.Rf2 f5
24.Kb1 h5 and Black took over the initiative in J.Merlicek-M.Hala, corr. 2003) 21...Bxf4+ 22.Rxf4
Re7 23.c3 Rfe8, Black is as solid as you like. The only issue is winning, to which end I tried 24.Ref1
Qd7 25.R4f2 Ka8 26.Bc2 b5!? 27.Rf5 Qd6 28.Kb1 Rb7 29.Ka1 b4 in HonzaCervenka-tsmenace,
ChessWorld.net.2018. It came to nothing: 30.Rc5 Nd7 31.Rc4 bxc3 32.Rxc3 Qxd4 33.Re1 Nf6
34.Rd1 Qb4 35.Bb3 Rb6 36.Qc2 Kb7 37.Rc1 Qd6 38.Ba4 Re6 39.Bb3 ½-½.
A third option 18.Qc4 (spot the threat) 18...Kb8 shouldn’t bother Black. I’ve lost twice by throwing
the b-pawn forward again and playing feebly, but you don’t have to do that: 19.Rdf1 (or 19.Rde1 Bc7
20.c3 Qf5 21.Bxf6?! Qxf6 22.Rhf1 Qh6+ 23.Kb1 Qxh2 24.Rxe4 Qxg3 and Black has the chances,
M.Motycakova-K.Novosadova, Ruzomberok 2014) 19...Qg4 (or 19...Qg2 20.Rfg1 Qh3 ½-½
M.Brodsky-I.Yagupov, St. Petersburg 1993) 20.Bxf6 (or 20.Bd2 Qd7 21.Be6 Qe7 and “Black is fine”
– Sokolov; I’ve won a game from here: 22.Bg5 h6 23.Be3?! Rde8 24.d5? cxd5 25.Bxd5 Rc8 26.Qb3
Bc5 27.Bxc5 Rxc5 28.Be6 Re5 29.Bh3? e3 30.Re1 e2 and the e-pawn proved decisively strong,
G.Coleman-J.Tait, corr. 1997) 20...gxf6 21.Qe6 Qg7 22.Qxf6 (or 22.Rxf6?! Bc7) 22...Qxf6 23.Rxf6
Be7 24.Rf5 Rxd4 “and a draw is the most likely outcome here” (Sokolov).

18...Rde8 19.Rhf1

GM Jan Timman will be fed up of seeing this game in print: 19.d5 c5 (Pruijssers mentions that

576
19...Nxd5!? 20.Bxd5 cxd5 21.Rxd5 Bxg3! is possible too) 20.Rhf1 Kb8 21.Bf4 Rd8 22.Bg5 a6!
(perhaps he expected Black to repeat with 22...Rde8) 23.Bxf6 gxf6 24.Qxe4 Qxh2 25.Rh1? (25.g4 is
still equal; or else 25.Rf3 intending 25...Bxg3 26.Bxa6 – Speelman) 25...Qxg3 26.Rxh7? (and this
loses) 26...Rfe8 27.Qf5 (if 27.Qd3 then 27...Re3 28.Qf1 Rde8, or 27.Qh4 Qf3!, threatening ...Bf4+ or
...Re4 and wins – I&K) 27...b5 28.Bf1 Re1 29.Qh5 Qf4+ 30.Kb1 Qxf1 0-1 J.Timman-J.Speelman,
Candidates semi-final, London 1989. After this dreadful shambles, GM Timman showed his mettle
by winning the next game and the match.

19...h6

20.Bf4

Trading dark-squared bishops doesn’t do much for White, as there is then no way to challenge the
f6-knight. Instead:
a) 20.Bxf6 gxf6 21.Rf2 f5 gives Black the standard solid set-up; e.g. 22.Rdf1 Kb8 23.Rf4!?
(threatening to trap the queen) 23...Be7 24.Kb1 Bg5 25.c3 h5 and Black is at least equal, R.G.Evans-
J.Tait, Notts League 1995.
b) 20.Be3 is Pruijssers’ preference, so that the bishop can hide on g1 if necessary; e.g. 20...Kb8
21.c3 Nd5! (Tay’s suggested improvement in his own book) 22.Bg1 and here Black needs to be
careful: 22...Rxf1 23.Rxf1 a6! (guarding against ideas of Rf7, Bxd5 and Qb5) 24.Rf7 g6!? (Tay’s line
continued 24...Be7 25.Rxg7 Bf6 26.Rf7 Qe6 27.Qh5 Bg5+ 28.Kc2 e3 29.Rf1 Nb4+ 30.Kb3 Nd5
31.Qf3 e2 32.Re1 b5 33.Bxd5 cxd5, or 31.Qe2 b5 32.Bxd5 Qxd5+ 33.Kc2 Qe4+ 34.Kc1 h5 “and
neither side can improve their position, so a draw is a fair result”; I thought it seemed simpler to keep
the g-pawn and push the h-pawn) 25.Bb3 h5 26.c4 Nb4 27.c5 Bc7 28.Be3 h4 29.Rh7 g5 30.a3 Nd3+
and here I was rewarded with 31.Kb1?? (31.Kc2 Rf8 32.Bc4 is roughly equal, intending 32...Rf1?!
33.Bd2 Rf2 34.Qxe4) 31...Rf8 32.Kc2 (the loss of tempo is fatal) 32...Rf1 33.Bd2 Rf2 34.Qxe4 Ba5

577
(White cannot now play b2-b4 because the bishop is still in the way) 35.Bc4 Rxd2+ 36.Kb3 Rxb2+
37.Ka4 Nxc5+ 38.dxc5 Qxh2 39.Qe5+ Ka7 0-1 drink1966-tsmenace, ChessWorld.net 2020.

20...Bxf4+ 21.Rxf4 Re7

An improvement given in all the books. Actually, 21...Rd8, from M.Yudovich-J.Boey, corr. 1972,
wasn’t so bad. The next move was the mistake: 22.c3 Rfe8? (if the other rook goes back again, Black
is probably still okay; e.g. 22...Rde8 23.Rdf1 Qd7 24.g4 Re7 and if 25.Rxe4 then 25...Rfe8! 26.Rxe7
Rxe7 and ...Qxg4) 23.Bf7 Re7 24.Bg6 Rd5 25.c4 Rg5 26.Bxe4 Kb8 27.Qc2 and, having nabbed the
e-pawn, White went on to win.

22.d5

Given as potentially strong by S&S in 1983. By 1996 they’d changed their minds, though nothing
else is any better for White; e.g. 22.c3 Rd8 23.Rdf1 Kb8 24.Bb3 Rde8 25.Bc2 Nd5 ½-½
Y.Dzhasmagambetov-H.D.Vötter, corr. 1998, in view of 26.Rxe4 Rxe4 27.Bxe4 Qe6 28.Re1 Nf6
29.Bd3 Qxa2. It seems that Tatai and Zinser’s original (1978) estimation of 21...Re7 as equal was
correct.

22...cxd5 23.Rxd5

Or 23.Bxd5 Rfe8 “with a reasonable position” – Tseitlin & Glazkov (1991); e.g. 24.Qb5 Nxd5
25.Qxd5 Qe6 26.Rd4 e3 27.Qxe6+ ½-½ Joseph K-tsmenace, ChessWorld.net 2006.

23...Kb8

578
 Here I&K (1994) say “=” and the engines concur. I’ve reached this position twice:
a) 24.Qd2 Qc8 25.Rd4 Re5 26.Qb4 Rd8 27.Rxd8 Qxd8 28.Rf1 Qe7 29.Qb3 Qc5 30.Ba6 Re7
31.Rd1 Kc8 32.Be2 a6 33.Kb1 Kc7 34.c3 Rd7 35.Rxd7+ Nxd7 36.Bd1 Ne5 37.Qg8 Qe7 38.Be2 Qf7
39.Qxf7+ Nxf7 40.Bc4 Ne5 41.Bd5 Ng4 42.Bxe4 Nxh2 43.Bd3 Ng4 44.Kc2 Kd6 45.Be4 b6 46.Kd2
g5 47.Kd3 Nf2+ 48.Kd4 Nxe4 49.Kxe4 h5 50.Kf5 h4 51.gxh4 gxh4, and now a clerical error
52.Kf4?? gave me the tempo required to win the pawn ending, M.Micklethwaite-J.Tait, corr. 1991.
b) 24.Rd1 Qc8 25.Bb3 Qc7 26.Qe3 Re5 27.h3 Qc5 28.Rd4 a6 29.Rf1 Ka7 30.Rfd1? (30.Re1 was
still equal) 30...Rf5 31.Qe1 Rf3 32.h4 Re8 33.Bd5 Rd3! 34.R1xd3 exd3 35.Qc3 Re1+ 36.Kd2 Re2+
37.Kd1 dxc2+ 38.Kc1 Qxc3 39.bxc3, R.Ward-J.Tait, corr. 1995. Stockfish likes my play so far but
now says 39...Re3 was winning. Instead, I swapped into a seemingly drawn rook endgame (which I
won anyway).

579
 Index of Variations

Chapter 1: Centre Game (and other second moves)

1.e4 e5 (9) 2.d4 exd4 3.Qxd4 15

3.c3 13

3...Nc6 4.Qe3 Nf6 17

5.Nc3 Bb4 6.Bd2 0-0 7.0-0-0 Re8

8.Qg3 20
8.Bc4 22

Chapter 2: Calabrese Counter-Gambit

1.e4 e5 2.Bc4 f5 26

580
3.d3

3.exf5 27
3.Bxg8 28
3.Nc3 29
3.Nf3 30
3.d4 (35) 3...exd4
4.e5 36
4.Nh3 37

3...Nf6 41

4.Nc3 Nc6 – Chapter 3

4.f4 42
4.Nf3 44

Chapter 3: Vienna Game

1.e4 e5 2.Nc3 Nc6 49

581
3.f4

3.Nf3 – Chapter 6
3.g3 50
3.Bc4 f5 (52) 4.d3 Nf6
5.Nf3 54
5.a3 57

3...exf4 60

582
4.Nf3

4.d4 60

4...g5 64

5.d4 64
5.h4 66
5.g3 69

Chapter 4: Bishop’s Gambit (and other King’s Gambits)

1.e4 e5 2.f4 exf4 72

3.Bc4

3.Be2 72
3.Qf3 73
3.d4 74
3.Nc3 74

3...f5 77

583
 4.exf5 78
4.Nc3 80
4.Qe2 82

Chapter 5: Wagenbach Defence

1.e4 e5 2.f4 exf4 3.Nf3 h5 87

4.d4

584
 4.h4 89
4.Nc3 92
4.Bc4 94

4...g5 5.Bc4

5.h4 99

5...h4 108

6.Nc3 121

6.Ne5 109
6.0-0 d6 112
7.c3 113
7.Nc3 116

6...d6

7.h3 122
7.Qe2 128
7.e5 130

Chapter 6: Three Knights Game (and other third moves)

585
1.e4 e5 2.Nf3 Nc6 135

3.Nc3

3.c3 d5 137

4.Bb5 138
4.Qa4 f6 5.d3 140
5.Bb5 141
3.d4 exd4 145
4.Nxd4 – Chapter 7; 4.Bc4 Nf6 – Chapter 8
4.c3 145

3...Bc5 (149) 4.Nxe5 Nxe5 5.d4 Bd6 6.dxe5 Bxe5 151

586
 7.Qd3 154
7.f4 156
7.Bd3 160

Chapter 7: Scotch Game

1.e4 e5 2.Nf3 Nc6 3.d4 exd4 4.Nxd4 Qh4 164

5.Nc3

587
 5.Qd3 167
5.Nb5 169

5...Bb4 (175) 6.Be2 Qxe4 7.Ndb5 Bxc3+ 8.bxc3 Kd8 9.0-0 Nf6 178

10.Nd4 180
10.Bg5 182
10.Be3 184
10.Re1 188

Chapter 8: Two Knights Defence

1.e4 e5 2.Nf3 Nc6 3.Bc4 Nf6 194

4.d4

4.Ng5 – Chapter 9
4.d3 d5 (222) 5.exd5 Nxd5 6.0-0 Bc5 224

7.c3 226
7.Re1 0-0 233
8.Nbd2 235
8.Nxe5 238

4...exd4 197

588
5.0-0

5.e5 Ng4 199

6.Qe2 200
6.0-0 203

5...Bc5 (208) 6.e5

6.c3 209

6...d5 7.exf6

7.Bb5 211

7...dxc4 213

8.Re1+ 214
8.fxg7 218

Chapter 9: Ulvestad Variation

1.e4 e5 2.Nf3 Nc6 3.Bc4 Nf6 4.Ng5 d5 5.exd5 b5 242

589
6.Bf1

6.Bxb5 Qxd5 (244) 7.Bxc6+ Qxc6

8.Qf3 245
8.0-0 247
6.dxc6 bxc4 249
7.Qe2 Qd5 250
7...h6 251
7.Nc3 253

6...Nxd5 (256) 7.Bxb5 Bb7

8.d3 259
8.d4 262

Chapter 10: Jaenisch Gambit

1.e4 e5 2.Nf3 Nc6 3.Bb5 f5 269

590
4.d3

4.exf5 270
4.Qe2 272
4.Bxc6 274
4.d4 (277) 4...fxe4 5.Bxc6 278
5.Nxe5 280

4...Nf6 (285) 5.0-0

5.exf5 287

5...fxe4 6.dxe4 d6 292

591
 7.Nc3 296
7.Bc4 303

Chapter 11: Jaenisch Gambit with 4.Nc3

1.e4 e5 2.Nf3 Nc6 3.Bb5 f5 4.Nc3 fxe4 5.Nxe4 d5 309

6.Nxe5

6.Ng3 310

592
6...dxe4 7.Nxc6 Qg5 (315) 8.Qe2

8.Nd4+ 317

8...Nf6 (320) 9.f4

9.Nxa7+ 321

9...Qxf4 330

10.Nxa7+ 331
10.d4 341
10.Ne5+ 345

593
Table of Contents
Title page 3
About the Author 5
Introduction 6
1) Centre Game (and other Second Moves) 9
2) Calabrese Counter-Gambit 35
3) Vienna Game 73
4) Bishop’s Gambit (and other King’s Gambits) 112
5) Wagenbach Defence 136
Part 2 171
6) Three Knights Game (and other Third Moves) 217
7) Scotch Game 263
8) Two Knights Defence 310
9) Ulvestad Variation 391
10) Jaenisch Gambit 436
11) Jaenisch Gambit with 4.Nc3 503
Index of Variations 580

594