Play The Benko Gambit (PDFDrive)

Nicolai V.
Pedersen
play the
Benko gambit
EVERYMAN CHESS
Gloucester Publishers plc www.everymanchess.com
First published in 2011 by Gloucester Publishers plc (formerly Everym an Publishers
plc ) , Northburgh House, 10 Northburgh Street, London EC1V OAT
Copyright © 2011 Nicolai V. Pedersen
The right of Nicolai V. Pedersen 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 m ay be reproduced, stored in a

retrieval system or tran smitted in any form or by any mean s, electronic,
electrostatic, m agnetic tape, photocopying, recording or otherwise, without prior
permission of the publisher.
British Libra ry Cataloguing-in-Publication Data

A catalogue record for this book is available from the British Library.
I S B N : 978 1 8 5 744 634 0
Distributed in North America by The Globe Pequot Press, P.O Box 480,
246 Goose Lane, Guilford, CT 0643 7-0480.
All other sales enquiries should be directed to Everyman Chess, Northburgh House,
10 N orthburgh Street, London EC1V OAT
tel : 020 7 2 5 3 7887 fax: 020 7490 3 708
em ail : info@everymanch ess.com; website: www.everym anchess.com
Everyman is the registered trade m ark of Random House Ine. and is used in this
work under licence from Random House Inc.
Everyman Chess Series

Chief advisor: Byron J acobs
Commissioning editor: John Emms
Assistant editor: Richard Palliser
Typeset and edited by First Rank Publishing, Brighton .

Cover design by Horatio Monteverde.
Printed and bound in the US by Versa Press.
Contents
Bibliography 4
Introduction 5
Benko Accepted
1 The Fianchetto Variation 12
2 The Classical Variation 33
3 The 'Quiet' 5 e3 73
Benko Declined
4 The 5 f3 Variation 85
5 White plays bS-b6 99
6 Other Options for White 1 14
Anti-Benko
7 The Kasparov Gambit 149
8 The Topalov Variation 1 73
Index of Variation s 202

Index of Complete Games 207
Bibliography
Books
1 d4 Volume Two, Boris Avrukh (Quality Chess 2010)
The Benko Gambit, Pal Benko (Batsford 1974)
Da ngerous Weapons: Flank Openings, Richard Palliser et al (Everym an Chess 2008)
The Gambit Guide to the Benko Gambit, Steffen Pedersen (Gambit 1999)
Play the Benko Gambit, Vaidyan athan Ravikum ar (Cadog an Chess 1992)
Squeezing the Gambits, Kiril Georgiev (Chess Stars 2010)
Winning with the Benko, Byron J acobs (Batsford Chess Library 1995)
Websites and periodicals

Chess Inform ants 1-109
New in Chess Yearbooks 1-98
Databases/DVDs
MegaBase 2011
The ABC of the Benko Ga mbit, An drew Martin (Chess Base DVD)
4
I ntrod uction
The Benko Gambit is an aggressive and Gambit h as been my own m ain reply to
sound pawn sacrifice which h as been 1 d4, and with decent results. The
played by such prominent grandmas reader can, for in stance, see my win s
ters as Garry Kasparov, Veselin Topalov, again st grandmasters P. H . Niel sen and
Viswan ath an An and, Alexander Kh alif V. Petrov, as well as the recent game
man, to name but a few. Significantly, again st V. N everov in the Fianchetto
Topalov has used it in m any games, line.
which says something about its Now I can cl aim th at the Benko is
soundness. It was the American sound. Unlike other more dubious
grandmaster, Pal Benko, who found g ambits (I could m ention m any h ere), it
many new ideas in the 1970s and is based on lon g -term compen sation .
pl ayed the gambit successfully, after The old m axim that acceptance i s the
which it increased in popularity. Al critical test i s probably especially true
though there were others who ven for unsound g ambits, however it still
tured it earlier, they were not the main h as truth in it and Black h as to play
in spiration for me; whereas a lot of the with some precision .
ideas in this book are inspired by Benko
and Khalifm an in particular. Some typical ideas
I have faced the Benko myself m any and themes in the Benko
times and feel that it is one of the more To justify the sacrifice of a pawn Black
dangerous replies to 1 d4, as White has aim s to seize the initiative and it is of
to be careful not to cede the initiative ten a feature th at Black h as a slight
to Bl ack. I h ave pl ayed around 10 lead in development, which can lead to
games with tt::lh 3 in the Fianchetto m aterial gain or some weaknesses for
Variation, and maybe 20 games (if one White. In some cases further sacrifices
includes rapid games) with the fash are called for, of which th ere are m any
ionable 10 l:[bl. My experiences al so examples in thi s book. Indeed, the
include the Benko Declined with 5 b6. Benko Gambit i s a good way for Bl ack
Over the last few years the Benko to get the chance to win - and to win in
5
Play t h e B e n k o G a m b i t
style. Furthermore, endgames can •.•C5-C4

sometimes be quite good for Black, too, Although this is a committal move, it
if there is still an initiative on the can be lethal if it supports a knight on
queen side for instance. d3 after ...ltJg4-e5 or ...ltJd7-C5 (or ... ltJd7-
e5). The obvious drawback is that it
Themes for Black weakens the d4-square; in particular
ltJf3-d4-c6 is usually not a manoeuvre to
...e7-e6 allow. So after ltJf3-d2 or ltJh 3, for in
The m ove ... e7-e6 is a comm on theme, stance, ... c5-c4 is more likely to be suc
tryin g to break up the white centre. It is cessful. As examples I could mention
usually more effective if White has Burmakin-5.Kasparov (Game 1), Aseev
played committal m oves such as f2- Ponomariov (Game 7) and Mancini
f3,e2- e 3 or f2-f4. Pal Benko often Degraeve (Game 18), and it is often seen
played ... e7-e6 after some preparation in the analysis too. So ... c5-c4 is a theme
in the middlegame, with both the idea which can be effective under the right
of attacking the centre and sometimes circumstances. It can also be a way to
of a l ater ... d6-d5 starting a rolling cen open files after a white b2-b3. However,
tre of his own . So it is a theme even Black should be careful that White can
when White h asn 't created substantial not simply answer ... c5-c4 by b3-b4, fol
weaknesses. There are m any in stances lowed by a2-a4, with two strong con
of this in the book. It should be noted nected pawns on the queen side.
th at thanks to the missing white c4-
pawn the breaks ... e7-e6, as well as .. .f7- An example is Aseev-Ponomariov
f5, are m ore likely to be effective. (Game 7):
One example, from Comas-Topalov

(Gam e 1 3 ) :
Here the white knight on h 3 can't

reach c6 sufficiently quicKly and there
is no b3-b4.
6
I n t ro d uc t i o n
...tt:Jd7-e5 (or ...tt:Jg4-e5), seen m any times in the g ames and

•••tt:Jd7-C5-d3, or tt:Je8-c7-b5-d4
••• analysis as leading to at least equality
In some case"S the knight on f6 is hin for Black, it should be con sidered very
dering the dark-squared bishop and carefully, with the possible g ain
might seek another post via d7 or g4, weighed against the weaknesses it
while the thematic ... ttJe8-c7-b S (-d4) is leaves around the black king .
al so sometimes effective. Some of the
key squares which can be good for the From N everov-Pedersen (Game 3 ) :
knight are eS and (especially) d4, c4
and d3 .
.. :iVa5, 'iVb6, 'iVC7, or 'iVa7

••. •.. .••
The move ... 1Was is especially thematic,

followed by . ..l::tfb8. Sometimes the
queen might go to b6 where it i s sup
porting the d6-pawn after the ... e7-e6
break, while in some cases f2 might be
weak. The squares c7 and a7 are usu
ally less ideal, but in the Benko De
clined (with bS-b6) the queen often The a2-pawn i s won and there i s
ends up here because of ttJd2-C4. further pressure on the light squares :
Obviously the various pawn m oves ... i.b3 a s well as ... ttJb6-a4 (or ... ttJxdS)
and knight and queen m an oeuvres are possibilities.
should be combined appropriately and
require some timing and precision . ••.f7-f5
The .. .f7-fS break can be quite risky,
.i.xc3
••• weakening the black king and the e6-
This move does indeed seem very square in particular. So in the middle
committal, giving up the good bishop g am e circum stances often h ave to be
on g7. In some cases, however, it en favourable for Black to play .. .f7-fS. In
ables Black to regain the white a-pawn the endg ame, on the other hand, it is
(usually on a2) with further threats. very thematic and i s seen quite often,
Some examples are seen in the g ames especially when Black is pressing for a
and notes to Neverov-Pedersen (Game win . Then the ds-pawn can be weak as
3), Cossin-Vachier (Game 8), Lugovoi it already l acks the support of a pawn
Khalifman (Game 9), and Piket-Topalov on c4.
(Game 10). Essentially, ... i.xc3 is a very An example is Cossin-Vachier
concrete move and, while it can be (Game 8):
7
A classical formation or Themes for White

a more rapid approach
The thematic and slightly slow set-up The blockade: a2-a4, b2-b3
with ... liJbd7, .. .'iWas, ... 0-0 and ... .l:.fb8, and/or a piece on bS
possibly followed by ... liJc7-bs, is often In the Benko Accepted one of the cen
a good plan . H owever, I recommend tral themes is a white blockade on the
th at Black keeps a good degree of flexi queenside. A big part of Black's strat
bility in the Benko. For in stance, egy is counterpl ay on the queen side, so
ag ain st the Fian chetto Variation, one of White's plan s is to play a2-a4
9 ... liJfd7 with the idea of ... liJb6-C4 and and plant a piece on bs, often prepared
.. .'�as (and ... liJd7-b6) is m ore effective by moves like "C2 and b2-b3. Naturally
than the classical form ation. Against Black will try to hinder these plans. For
White's plan with f2-f3 I suggest a in stance, the move b2-b3 weakens the
quick ... e7-e6 in reply, which leads to a a3-square, while ... cs-c4 is sometimes
very sharp and unclear g ame. And more dangerous (this is mostly charac
again st some of the less correct lines, teristic of the Fianchetto Variation).
such as 4 �g s in Grivas- Khalifm an The theme also occurs now and then
(Game 32), Black plays ... liJe4 and ... "as in the Benko Declined, for instance in
and even ... g 7-g s (see fol1owing d ia the f2-f3 and e2-e3 lines if Black plays a
gra m), gaining time, and is already bet later ... a6xbs. On the other hand, in the
ter from the early opening phase: bs-b6 lines I mostly do not recommend
With the g ambit Black i s trying to playing ... a6-as, as the bs-square is then
obtain a lead in development (amon g st a serious weakness.
other thing s), so slow moves by White In general the blockade on bs com
are often a good reason to sharpen the bined with the e4-e5 break (see below)
play. is something Black should make a seri
ous effort to prevent.
8
I n t r o d uc t i o n
M.Gurevich -Z.Azm aiparashvili, Stras

bourg 1994. White h as a comfortable
advantage (see the n otes to Game 10).
i.g5, i.f4
As with .if4 in the previous n ote, .ig 5
often leaves the b2-pawn weak. There
fore .ig 5 m akes particular sen se when
it attacks an undefended e7-pawn, af
ter ... ltJe8 for example. On the other
h an d, under some circumstances Black
An example of a good blockade for can con sider leavin g the pawn en prise
White. This can arise after inaccurate if there's the possibility of .. .f7-f6, trap
play by Black in the notes to G ame 14. ping the bishop on e7, in which case
.ig 5 might just be m et be ... h 7-h 6.
The e4-e5 advance
Another key theme for White is the e4- From Campos-Topalov (Game 25)
e5 advance. Sometimes this needs to be
prepared, as in the Fianchetto Variation,
whereas in the Classical main line it can
often be played sooner. The move e4-e5
does have drawbacks; for instance the
d5-pawn may become weak, while pre
paring it with .icl-f4 can leave the b2-
pawn loose. However, if White has good
control, then e4-e5 can be effective.
White h as just an swered Black's

mistimed ... ltJf6-e8 with .icl-g 5, at
tackin g e7, and Black's best idea m ay
be simply to put the knight back on f6.
Prophylactic moves (such as l:.e2,

�C2, 'ii'c2, h2-h3, b2-b3 or a2-a4)
In the Benko, as in chess in general, a
prophylactic approach i s often worth
This position arose in the g ame considering and l:te2, l:tc2, 'iVc2 (etc) are
9
all m oves which Karpov h as used with Benko nonetheless with 2 ... cS 3 dS bS,
success in the Classical m ain line. How intending 4 c4 g 6 S cxbS a6. If White
ever, I think the recommendations in pl ays 4 �g s instead we have a system
Chapter Two show h ow to deal with which Topalov has used with success.
these ideas, so the Classical is no longer Since Black has played the ambitious
that popular. ... b7-bS, I am n ot in favour of allowing
The modern .l:.bl in the Fianchetto �g sxf6, compromising the pawn
Variation is al so prophylactic in n ature, structure further. So here the recom
as White takes steps (with b2-b 3 and mendation is 4 ... ttJe4, involving ideas
�b2) to prevent a l ater ... cS-c4, ... 'iWa3 like ... 'iYb6 (or .. "iVas) and ... g7-g6,
or ... ttJC4 arising from Black's classical ... �g7. In general the concept remains
formation with ... ttJbd7, ... 0-0, ... 'iWas active pl ay. The struggle is often quite
and ... l:r.fb8. In fact thi s idea h as been positional, but it can quickly turn and
the most serious threat to the Benko in be very tactical . In either case Bl ack is
the l ast decade or so, which i s why in fine as long as he knows wh at he is do
Chapter One I suggest th at Black take ing.
preventative m easures in turn by play Another anti-Benko system is 1 d4
ing 9 ... ttJfd7 and 10 ... ttJb6. ttJf6 2 c4 cS 3 ttJf3, refusing to push the
d-pawn, and tran sposes to a line of the
English Opening. This time Bl ack can
sh arpen the play with the Kasparov
G ambit: 3 ... cxd4 4 ttJxd4 e s ! ? S ttJb S dS
6 cxdS �cS . I think this is a sound gam
bit as well.
One of the critical variations is 7
ttJSC3 0-0 8 g 3 , as played by the new
guy in the 2800+ club, Lev Aronian, and
which he used to beat 5utovsky. In re
spon se I recommend that Black follow
5utovsky with 8 ... "i!Vb6 ! ?, forcing 9 e3,
White avoids the Benko when the follow-up 9 ... Sl.g4 more or
Although there are other m ore or less less prevents the plan of �g 2 and
qualified ways to test the G ambit, my leaves White with weaknesses on the
conclusion i s that Bl ack i s doing fin e light squares (especially c4, d3, f3 and
from a theoretical perspective. So what h 3 ) . Furthermore, Black has a lead in
if White avoids the Benko? development which can be used for the
One approach i s with 1 d4 ttJf6 2 piece sacrifice ... �d4 ! ? as -a way of at
ttJf3 . In thi s case Black can aim for the tackin g the exposed ds-pawn and the
10
I n t ro d uc t i o n
white king, as in Matlakov-Evdokimov ttJf6 2 ttJf3, but these cannot b e re

(Game 41): g arded as within the scope of thi s book,
particularly if White i s aiming more for
a flank approach than direct occupa
tion of the centre.
In conclusion, I hope that readers
n ew to the opening will be inspired to
try the Benko Gambit for themselves,
while existing Benko players will find
some new ideas enabling them to play
the Gambit m ore successfully. The
Anti- Benko chapters might even prove
useful as part of another repertoire,
There are many new ideas in these such as the Benoni, Nimzo-Indian,
lines and I think Black is doing fin e Queen 's Indian, Bogo-Indian or Grun
against g 2 -g 3 . White's other option s fel d, for instance.
are less critical, so Black should be abl e
t o prove compensation there too. Nicolai Pedersen,
Other ways for White to avoid the Aarhus, Denmark,
Benko include 1 c4 and 2 g 3 , or 1 c4 June 2011.
11
Chapter One
The Fianchetto Variation
1 d4 tLlf6 2 c4 cS 3 d s bs 4 cxbs a6 S strong bishop on g 7, and the other

bxa6 g6 6 tLlC3 .1i.xa6 7 g3 d6 8 ..tg2 pieces h ave good potential to be very
..tg7 active on the queen side, particularly on
the h alf-open a- and b-files. Mostly
Black h as followed the standard pl an in
this position : ... 0-0, ... 'iYas, ... l:tfb8,
... tLlg4-es (or ... tLle8-c7-bS), etc. How
ever, 10 l:tbl and the ideas behind it
(which are typical for the variation)
h ave proved quite effective again st this
set-up. And other books which cover
th e same subject h ave failed to show
solution s for Bl ack.
50 the reader might ask, wh at
Thi s is the critical position for the should Bl ack do again st 10 :bl - ? Well,
Fianchetto Variation, which is itself one my suggestion is 9 ... tLlfd7, aiming to
of the most important lines of the g ain counterplay earlier in the middle
Benko Gambit from a theoretical view g ame with moves such as ... tLlb6,
point. Now 9 tLlf3 tLlbd7 10 l1bl brings ... tLl8d7, ... 0-0, ...tLlC4 and .. :iWas. Th e
about perh aps the most serious chal main idea i s to plant a knight on the
lenge to the Benko. Here White i s aim c4-square before White has managed
ing for two m ain strategic pl an s : a to prevent it by 'iVC2, ..td2 and b2-b3.
blockade on the queen side with moves The conclusion is th at Bl ack is
like 'iYC2, b2-b3, a2-a4, and placing a achieving an unclear game here. This
knight or bishop on b S ; and the central line has been played many times with
advance e4-e s . success by GM 5ergey Kasparov, who
From the other side, Bl ack h as a wrote an article on it in New in Chess
12
Th e Fia nc he t to Va ria t i o n
Yearbook #94. In m y recent g am e Benko Gambit, s o much so th at I think

again st Neverov (Game 3 ) I was al so Black should avoid it altogether.
happy to test this idea and obtained,
first, an unclear game from the open
ing, and then the better position, all
without too much effort at the board
(as most of the moves were prepared).
In general, it i s said th at acceptance
is the critical test of a gambit, and with
the Fianchetto Variation th at is some
how confirmed, as this has been the
main theoretical focus in recent years,
and will probably continue to be so.
With this in mind, and as theory will With 9 ... tDfd7 Black intends to pl ay
inevitably continue to evolve, the ... tDb6, from where it can quickly jump
reader is advised to follow develop to c4 (10 ... tDb6, 11 .. tD8d7 i s the typical
.
ments in this variation carefully. At the follow-up). By sending the king's knight
moment, however, it is up to White to on thi s journey, Black uncovers the fi
come up with something new, and anchettoed bishop on g 7, which mean s
that, I think, makes quite a change. th at White's prophyl actic m ove b2-b3
As the variation with 9 tDf3 tDfd7 i s is itself prevented. White will usually
the most important in this ch apter, we try to prepare it anyway with Wc2 and
will examine it straightaway in the first Jt.d2, but as we will see, thi s can be an
four games below, before going back swered by ... tDC4 and ... Was .
and considering earlier moves and pos Sergey Kasparov wrote a n opening
sibilities in Games S-7. survey of thi s variation in New in C hess
Yea rbook #94. I will use a few of his
games in thi s ch apter, while some of
Game 1 the analysi s is also inspired by his arti
V.Bu rmakin-S.Kaspa rov cle.
Deizisa u 2003 10 0-0
The immediate 10 'iYc2 is seen in
1 d4 tDf6 2 tDf3 cS 3 d s d6 4 c4 bs S the next g ame, while other moves are
cxbs a6 6 bxa6 Jt.xa6 7 tDC3 g6 8 g3 covered in the n otes to Gam e 4.
Jt.g7 9 Jt.g2 tDfd7 10 tDb6
...
To reiterate wh at I 've already said, Before castling Bl ack continues to

9. .tDbd7 10 .l:f.bl is one of the most - if
. mobilize his pieces on the queen side.
not the most - critical tests of the 11 'iYC2
13
The queen is usually placed o n c 2 i n still seems to be equalizing . For in

this line. stance: 1 3 ...liJC4 (a standard move in
11 ...liJ Sd7 this variation ; the knight is quite well
placed here) 14 lbcl 'ifas 15 lLld2
liJdb6 16 liJxC4 liJxc4 17 b3 liJa3 (show
ing another difference from the
9 ... liJbd7 10 l:tbl m ain line: White's
plan of b2-b3, a2-a4 and blocking the
queen side is more easily prevented in
the 9 ... liJfd7 line) 18 'ii'd 2 c4 19 .i.h 6
l:tfb8 20 .i.xg 7 'it>xg 7 2 1 bXc4 liJxC4 22
'iWd4+ liJe s 2 3 'ife3 11c8, when Black is
quite active and has enough compen
sation for the pawn .
A basic position for the 9 ... liJfd7 line. Alternatively, 1 3 b3 is covered in the
Here White h as a few different paths. n ext g ame, while 13 .l:t.bl is examined
In the current game Burmakin tries in Games 3 and 4.
going forward in the centre, but I don 't 13 liJ C 4
.••
think this shoul d cause Black any prob

lems.
12 :d1 0-0
14 liJd2
14 .i.f4 can be tried here too,
though Black is again doing fine. A pos
13 e4?1 sible continuation is 14 .. :iVas 15 .l:[acl
This allows Black to get a rather i.. x c3 ! 16 'ifxC3 -.xC3 17 bxC3 liJb2 18
comfortable g am e by simply playin g .l:t.d2 liJd3 19 .l:t.cdl liJxf4 20 gxf4 liJf6,
. . .liJC4. when Black is slightly better and is set
Instead, 13 .i.f4 might be a better to win at least one of the weak pawns
attempt to get an edge, although Black on e4, f4, a2 or ds. So it seem s th at 13
14
Th e Fi a nc h e t t o Va ria t i o n
e4, allowing the knight to jump to c4, Stronger was 2 2 .. :ii'a5 ! and Bl ack
means a comfortable game for Bl ack. keeps a slight advantage - White i s
14 ... tDxd2 quite uncoordinated.
Black is al so fine after 14 ... tDde 5 ! ? 23 Iidbl?
intending 15 tDxc4 tDxc4 and . . .Va5 2 3 l:.xb 8 ! equalizes.
and sometimes ...l1fb8 with pressure 2 3 .. Jlxb4 24 l:ixb4 'iVC5 1 2 5 tDb5
on the dark squares, or just 14 ... tDce 5
15 tDf3 tDC4 with a repetition of m oves.
15 .i.xd2 tDe5
25 ... tDd 3?
Overlookin g a simple combin ation :
2 5 ... .i.xds 2 6 'iVxc5 dxc5, regaining the
Black has solved all his opening pawn with g ood winning ch ances.
problems and it i s White who should Bl ack i s m ore active and the c-pawn is
think about equalizing . Notice that more dangerous th an the white a
... c5-c4 and ... tDd3 is now an effective pawn .
plan . Often in the fianchetto variation 26 Ubi tDxel 27 .l:txel .i.xb5 28 'iix C5
White can respond with tDf3-d4-c6, but lixC5 29 axb5 l:r.xb5 30 .i.f3 l:tb2 Vz-Vz
obviously th at is not a theme h ere.
16 .i.el 'ili'b6 17 l:tabl c41 18 b4 cxb3 19 Game summary
'iVxb3 'iVa7 20 a4 .i.C4 21 'iVc2?1 Burmakin played un ambitiously with
2 1 'iVb7 would m aintain the bal the white pieces and Sergey Kasparov
ance. m anaged to equalize. Later on Black
21 ... .:tfc8 even h ad the slightly better position,
Now Black has the better position , though after inaccuracies the g am e
due to his good activity and well placed ended in a draw. In conclusion, the l i n e
pieces. The bishop on g7 and the rook with .lir. d l and e2-e4 i s n ot a theoretical
on the open c-file are workin g espe problem for Black, whose knight
cially well together. shouldn 't be allowed to go to c4 so eas
22 lib4 l:!ab8?1 ily.
15
played in lVeng-R.5kytte, Copenhagen

Came 2 2010, when 15 ... 'ii'a 5 16 ltJxC4 .txC3+ 17
A.Goganov-S. Kasparov .td2 ii.xd2+ 18 ltJxd2 l::tb 8 19 b3 ltJe 5
Peterhof 2008 gives Bl ack compensation - Skytte)
15 ... 'i1t'a5 16 a3, when Avrukh rates
1 d4 ltJf6 2 c4 cS 3 ds bs 4 cxbs a6 S White's chances as preferable, men
bxa6 g6 6 ltJC3 ii. xa6 7 g3 d6 8 .tg2 tioning the pl an of l:ia2, moving the c3-
.tg7 g ltJf3 ltJfd7 10 'iWc2 knight, followed by b2-b3. But I don 't
Since Gogan ov follows this with 11 agree White is better here ! After
0-0, it just leads to the same position as 16 ... ltJdb6, I think Black has definite
in the previous g ame. H owever, the compen sation whether White plays
current move order does give White Iita2 or any other move:
another option (see the n ext n ote).
10 ltJb6
...
a) 17 :a2 ? ! (this move is dubious, in

my opinion, as the rook is offside here)
11 0-0 17 ... ii.xc3 ! 18 'ii'x c3 (or 18 bxc3 ltJa4 19
In his book 1 d4 Volume Two Avrukh l1al .tc8 20 ltJe3 ltJxe3 21 .txe3 'iix c3
recommends 11 h4 here. Black has to 22 'iWe4 'iWb2 and Bl ack is slightly bet
react precisely as there is some danger ter) 18 .. :iVxc3 19 bxc3 f5 20 ltJe 3 ltJxe3
on the king side with the knight away 21 .txe3 .tC4 and Black regains the
on the other side of the board: 11 ...h 6 ! pawn with a slight advantage.
12 ltJh 2 ltJ8d7 1 3 h S (after 1 3 ltJg4 h 5 b) 17 :bl (a better move than l:.a2)
1 4 ltJe3 .td4 1 5 .td2 ltJf6 " Black has 17 ... .txc3 18 bxc3 .tc8 19 ltJe3 ltJxe3 20
nice play" - Avrukh ; a possible con ii.xe3 ltJc4 21 ii.cl ltJxa3 2 2 ii.xa3 'ikxa3
tinuation is 16 0-0 0-0 17 b3 ltJbd7 with 23 f4 (or 2 3 l:tb6 0-0 24 .:r.fbl 'iWa4 is just
compen sation for the pawn) 13 ... g 5 14 equal) 23 ... 'iIi'a2 24 .l:tb2 'ii'C4 with an
ltJg4 ltJC4 (continuing with the plan of unclear game .
... ltJC4 and .. :tIi'a5) 15 0-0 (15 ltJe3 was c) 17 ltJe3 .txc3 18 bXc3 'iia4 19
16
Th e F i a nc h e t to Va ria t i o n
'iVb1 ttJxe 3 20 i.. x e3 litb8 2 1 liIe1 i.. C 4

with compen sation .
11 ... ttJ8d7 12 �d 1 0-0 13 b3
1 9 'ii'd 2 llfa8
And h ere 19 ... lIa3 ! with the better
chances.
A straightforward attempt to pre 20 a4 c4 2 1 'iid 4 'iid 8 2 2 bxc4 lba4 23
vent the further ... ttJc4. But n ow we see lIxa4 ]:txa4 24 litb1 ttJbd7
one of the reason s for preparing b2-b3 Despite the reduction in forces
with litb1 in these lines. Black's g am e is still favourable, due to
13 ... i..xe2 ! the weak c4-pawn .
A tactical trick, exploitin g White's 2 5 l:Ia1?! 'ii' b 6 26 'iVC3 ':'xa1+ 27 'iix a1
vulnerability on the long diagona1 . ttJe5
Black equalizes quite easily.
14 'ii'x e2 i.. x c3
The position is roughly bal anced.
Bl ack has the superior structure, a
strong dark-squared bishop, and good
chances of activity in the middlegame.
White has the bi shop pair.
15 i.. b 2?!
See the note just above. It is more
logical to play 15 .l:.b1, keeping the
bishop pair; e.g. 1 5 ... i.. g 7 16 a4 lite8 17
i..f4 ttJf6 with chances for both sides. Now Black i s clearly better with his
15 ... i..x b2 16 'ii'x b2 ttJf6 17 ttJg5 'iVd7 18 well-placed and active pieces, while the
h4?! l:la7 white c-pawn can n o longer be de
Stronger was the altern ative fen ded.
18 ... l:tfb8!, when White has a difficult 28 'ii'e 1 ttJfg4 29 'ii'e 2 'ii'd 4 30 i.. h 3 h6
position . 3 1 ttJe4 f5 3 2 i.. xg4 fxe4 3 3 i.. e 6+ 'ii;>g 7
17
3 4 'ii'u ?? strong opponent.

A blunder in a difficult position . 10 0-0 tbb6 11 �C2 tbSd7 12 l::td 1 0-0 13
34 ...'ii'd 1+ 3 5 �g2 tbf3 36 'iVb2+ � h 7 3 7 :b1 tbc4
'iVC3 'iVg1+ 0-1 This position can also be reached via
9 ... tbbd7 10 .l:tbl 0-0 11 0-0 'ii'a s 12 'iVc2
Game summary tbb6 13 .l:tdl tbc4, but this move order
White played the set-up with 'iVc2 and allows White the strong alternative 12
.l:r.dl and we quickly saw why this �d2. Hence I recommend 9 ... tbfd7 .
should be followed by 1 3 l:lb1. The 14 tbd2
g ame continuation 1 3 b3 allowed I don't think this move is critical
13 ... �xe 2 ! , after which Black h as n o from a theoretical standpoint. More
problem s. Once Goganov exch anged interestin g is 14 b3, which is examined
the bishop on b2, he could n o longer in the next game (Gall ardo-Ortega).
count on the bishop pair to balance 14 ... 'iVa s 15 tbxc4 �xC4 16 �d2
Black's superior structure. After a fur
ther few inaccuracies and a final blun
der Sergey Kasparov won the rather
one-sided g ame.
A m ore critical continuation was
seen in the n otes to White's 1 1th m ove:
Avrukh 's recommendation of 1 1 h4 -
though, as shown in the analysi s, Black
should still be able to obtain an unclear
g ame at l east.
H ere the game C.Horvath -M.Pap,

Game 3 Pula 2003, ended abruptly 16 �g s .l:r.fe8
V.Neverov-N.V.Pedersen and a draw was agreed in this unclear
D ieren 2010 position . I had al so looked at the text
move and concluded that it is okay for
1 d4 tbf6 2 c4 cS 3 d s bs 4 cxbs a6 5 Bl ack.
bxa6 g6 6 tbc3 �xa6 7 tbf3 d6 S g3 �g7 16 ... i.xc31
9 �g2 tbfd7 I saw in my preparation th at this is
I was h appy with N everov's choice unclear. It might look odd to give up
of variation as I h ad already m ade the g7-bishop, but it is justified by the
quite some study of 10 .l:tbl ideas and g ain of the a2-pawn and further
9 ... tbfd7, and could n ow show in prac threats again st dS, with' ideas like
tice th at Black is fine, even ag ain st a ... �b3 and ... tba4.
18
Th e Fia nc h e tt o Va ria t i o n
In stead, E.Bacrot-P.Tregubov, Paris position via the 9 . . .ttJbd7 move order

2006, continued 16 .. :iVa6 ! ? 17 a4! i.xe2 given above) saw 19 h4 f6 20 h 5 g 5 2 1
18 ttJxe2 'ii'x e2 19 'uel 'ii'a 6 (19 .. :ji'g4?! ..te4 i.b3, when White elected t o force
20 b4 was good for White in a draw by 22 i.xh 7+ ! ? 'itt x h 7 2 3 'iVd3+
H.5tefansson-R.Lagerman, I celandic �h 8 24 'ii'g 6 i.xdl 2 5 'iVh 6+ 'itt g 8 2 6
Championship, Reykjavik 2007) 20 'iVg 6+ 'itth 8 2 7 'iVh 6+ 'itt g 8 2 8 'iVg 6+ Y2-Y2
I!xe7 ttJe5 2 1 b4 i.f6 2 2 b 5 (more accu A.Yusupov-P.Tregubov, German League
rate was 22 %:tC7 ! i.d8 23 b5 'ii'x a4 24 2006.
'ii'xa4 l:txa4 25 l:tb7, when White i s 19 f6 20 'ii'e 3 I!f7 2 1 b3!? i.xb3 22
•..
slightly better due to the bishop pair lildcl ttJxd S 23 'iVf3

and the strong b-pawn) 22 .. :ii'x a4 2 3 So Neverov seems to be planning to
'ii'xa4 l:txa4 2 4 l:tb7 l:fa8 (now Black win some m aterial based on a weak
has some activity and the position i s ness on the light squares. I was n ot too
unclear) 2 5 i. C 3 lla3 2 6 ..txe5 ..txe 5 2 7 worried about thi s durin g the g ame
b 6 I!al 28 I!xal I!xal+ 2 9 i.fl l:lbl 30 though.
':'b8+ 'itt g 7 3 1 b7 c4 3 2 'itt g 2 i.d4 33 2 3 ... ttJxC3 1 24 'it'xC3 'ub8 2 S 11b2 'ii'a 4 26
..txc4 i.a7 34 I!e8 I!xb7 35 I!e2 Y2-Y2. l:tcbl?!
17 i.xc3 'ii'x a2 18 'iVd2 ttJb6 2 6 i.d5 i.xd5 2 7 l:txb8+ �g 7 was
perhaps a better way to win the ex
change, although White has n o advan
tage.
26 ... C4
The position i s unclear. The bishop

on c3 pointing towards the black king
might seem threatening, but 'iVh 6 can
always be answered by .. .f7-f6. Mean
while Black has queen side counterplay, Now Black is slightly better with
particularly again st the weak pawn on two pawns for the exch an g e (after
d5 and the threat of ... ttJa4. i.d5Xf7), while the white rooks are
19 'ii'h 6 quite passive. In addition Neverov was
An earlier game (which reached thi s in time trouble. So quite a good situa-
19
tion for me; the m ain worry here was a 62 fxg4 'ifxd 5+ 6 3 �g3 'ii'e 5+ 64 <t>f3
draw. 'ii'f6+ 65 �g3 d 5 66 Ve2+ �g7 67 "fie7
27 .lid S libf8 28 Iitd2 �g7 29 .lie6 Iita8 d4 68 "fie5 d 3 ?
30 .lixf7 �xf7 3 1 l:td 5 'ii'e 6 3 2 e4 'ii' b 6 But this was a mistake. I t was nec
33 �g2 :ta2 34 J:tb2 l:txb2 35 Vxb2 e3 essary to keep the pawns together with
3S ...'iVb4 might provide more win 68 . . e s . Then it would be possible in
.
ning chances, but the ri sk i s al so a bit m any cases to pl ace the queen on f4
greater. I saw thi s durin g the game, but and use the eS- and d4-pawn s as shel
thought that the queen endgame of ter from the checks.
fered reasonable winning chances too. 69 'ii'd 5 Ve3 70 �f3 Ve2 71 Vg5+ �f7
36 'ii'x e3 .txd 5 37 exd 5 Vb1 72 'iff5+ �e8 7 3 �e3 'ife2+ 74 �d4 e5+
75 �d 5 'ifU+ 76 �e4 d2 77 'ii'xe5+ �f7
78 'ii'e 7+ �g6 79 Vd6+ �f7 80 "fid7+
�f6 81 �e3 Ve2 82 'ii'xd2 'ilkxd2+ 83
�xd2 �g5 Yz-Yz
Game summary
With 9 ... liJfd7 I had the chance to test
my recommendation against a strong
grandmaster. In the opening phase
Neverov didn 't play the theoretically
most testing line, and the thematic
Unfortunately I didn 't win man age 16 .. .lixc3 ! led to an unclear position in
.
to win thi s endgame, despite being a which Black was not worse. In the mid
clear pawn up. Since the opening dlegame Neverov pl ayed sh arply, trying
ph ase i s the m ain focus here, I will go for an advantage after all, but an inac
quickly through the rest of the g ame. curacy g ave me the slightly better posi
3 8 'ii'd 4 Ve1 39 h4 'ii'e 5 40 'ifd2 �g7 41 tion . Eventually I obtained good win
<t>g1 'ilke4 42 'it'a 5 �f7 43 'it'd2 �g7 44 ning ch ances in a queen endgame, but
'ilka 5 'ii' b 1+ 45 �g2 Vd3 46 Va8 g5 47 had to concede a draw in the end.
hxg5 fxgs 48 g4 Ve2 49 'ifa1+ <M7 50
'ifd4 h5 5 1 gxh 5 'ilkxh 5 52 f3 'it'h7 53
'ilkg4 'ilkg6 54 �g3 'ii'f6 55 �g2 'ifb2+ 56 Game 4
�h3 Ve5 57 'ii' h 5+ <t>g7 58 'ilke8 'ilkf5+ A.Gallardo Ciarcia
59 �g2 'ii'f7 60 Ve6 �g6 61 Ve7? g4 D.Ortega Hermida
N ow Bl ack obtains two connected Monda riz 2007
pawn s and his hopes of winning in
crease significantly. 1 d4 liJf6 2 e4 CS 3 d5 b5 4 exb5 a6 5
20
Th e Fia nc h e t t o Va ria t i o n
bxa6 g6 6 tDc3 .ixa6 7 g3 d6 8 .ig2

.i.g7 9 tDf3 tDfd7 10 0-0
Apart from 10 0-0 and 10 'iWC2,
White has tried two other moves in this
position :
a) 10 .i.d2 is a different way to aim
for the desired set-up with b2-b3, a2-
a4. The drawback i s that the d5-pawn
is temporarily vulnerable, which Black
can exploit by playing the sequence
10 ... tDb6 11 b3 .i.b7 ! , and if White de
fends the pawn with 12 e4, then the all) 17 f3 (defending the pawn)
very thematic return 12 ... .i.a6 hinders 17 ... 'ii'd 7 18 ..t>g 2 h 6 19 tDh 3 e6 20 dxe6
his castling short. tDxe6 21 'ii'b l tDd4 22 'iWd3 g5 and
Black's activity, together with the poor
white knight on h 3 and exposed king
on g2, gives him enough compen sation
for the pawn .
a12) 17 'ii'e 2 'ii'd 7 18 a4 h 6 ! ? (alter
n atively, 18 ... .i.xc 3 ! ? 19 .i.xc3 fxe4 20
'ii'x e4 tDbxd5 tran sposes to line 'a3'
below) 19 as tDbxd5 20 tDxd5 tDxd5 21
exd5 hxg 5 leads to an unclear position .
Now White h as to pl ay rather artifi

cially, returning his fianchettoed
bishop to its original square, simply in
order to get the king to a safe l ocation
away from the middle of the board:
al) 13 .i.fl 0-0 14 .i.xa6 tDxa6 15 0-0
f5 ! (after this White h as some difficulty
defending the d5-pawn i in stead 15 ... c4
16 .l:.bl tDC5 17 'ii'e 2 cxb3 18 axb3 was
slightly better for White in D.Hou For instance, 2 2 1:[a4 1:[fb8 2 3 'ii'C 4
sieaux-V.5olodovnich enko, Condom g4 24 l:tel :a7 and 22 .:ta3 f4! both of
2005) 16 tDg 5 tDC7 with the following fer Black counterplay and/or equality,
possibilities: so White might be tempted by the ex-
21
Play t h e B e n k o G a m b it
chan g e sacrifice 2 2 .i.xg S ! ? i.xal 2 3

':xal, when he h as a strong a-pawn
and possible attacking chances against
the king on g 8 . On the other h an d, after
23 ... l:tfe8, intending ... e7-e6(es), Black
can l ook for counterplay in the centre,
while White also h as a slight problem
on the back rank. A possible continua
tion is 24 a6 eS 25 dxe6 ':xe6 26 'ii a 2
dS 2 7 a7 d4 2 8 'iVas 'iVc6 2 9 h4 and the
g am e remain s unclear.
a13) 17 'ii'C 2 (this gives Black fewer Black should reply 10 ... tiJb6 1 1 'ii'c l
options than 'iVe2) 17 ... 'iVd7 18 a4 fxe4 h6 (weakening, but safeguarding the
(18 ... h 6 ? ! would be dubious h ere, due g 7-bishop) 12 0-0 (as in Y.Yakovich
to 19 as tiJbxds 20 tiJxds tiJxdS 2 1 exds S.Kasparov, Russian Team Ch ampion
hxg s 22 i.c3 and White's chances are ship 2004), and now 12 ... g S ! with com
preferable) 19 "ii'x e4 i.xc3 20 i.xC3 pensation. For instance: 13 i.d2 tiJ8d7
tiJbxds 21 'iVh4 h S ! 22 i.d2 "g4 2 3 14 l:tdl (or 14 'iVC2 tiJC4 15 i.el 'iVb6 16
'iix g4 hxg4 with equal chances in the tiJa4 'iWb7 17 i.c3 tiJf6 18 %bcl 0-0 19
endgame. ':'fdl lIfb8) 14 ... tiJC4 15 i.el 0-0 16 b3
a2) 1 3 litcl (a waitin g move) tiJces 17 'iVc2 tiJxf3+ 18 i.xf3 fS 19
13 ... tiJ8d7 ( Bl ack develops and keeps llacl 'iVe8 20 i.d2 tiJe s 21 i.g2 'iWhs
the bal an ce) 14 a4 (14 i.fl i.xfl 1 5 and Black has a strong initiative.
'iii' xfl 0-0 16 a4 comes t o the sam e 10 ... tiJb6 11 "iVc2 0 - 0
thing) 14 ... 0-0 1 5 i.fl i.xfl 16 'iii' xfl 11 ... tiJ8d7 12 l:tdl 0-0 is the move
was played in V.Epishin-M.lvanov, Arco order I recommend.
2002, and now instead of 16 ... tiJf6 ? ! 12 ':'d1 tiJSd7 13 l1b1 tiJC4
Black should prefer 16 . . .C 4 17 b 4 i.xC3
18 i.xc3 tiJxa4 regaining the pawn.
Fo11owing 19 i.al White's bishop
mig ht l ook powerful, but Black can
block it off with the c-pawn if n eces
sary, so the g am e is just unclear after
19 ... 'iVb6.
b) 10 i.f4 is another possibility,
which h as the idea of continuing "ii'c l
and i.h 6, exchanging the strong
bishop on g 7 .
22
14 b3 l:tdcl l:ta3 19 .i.xe7 l:te8 20 .i.xd6 .i.xd5

As we saw in the previous g ame, 14 2 1 e 3 with a slight plus for White in his
ttJd2 is less critical for Black. New in Chess article. I agree with this
In stead, 14 .i.h 3 was tried in evaluation, but Black might prefer
L.Schan dorff-R.5kytte, Danish Team 17 ... .i.f6 ! 18 .i.xf6 ttJxf6 19 ttJd2 .i.c8 20
Championship 2011. From h3 the .l:1b2 l:ta3 2 1 e4 .i.d7 with a likely draw.
bishop exerts pressure against d7, hin 15 ... Ji.b5
dering Black's intended ... 'iWa5 . The Now Black is very active and White
drawback is that d5 becomes weaker, h as some tactical problems. This is in
so a logical reply is 14 ... ttJce 5 . For in deed something to aim for from a Black
stance, 1 5 ttJxe5 ttJxe 5 16 b3 'iWa5 17 perspective and shows that White of
iob2 ioc8 18 .i.xc8 lIfxc8 19 .i.al c4 and ten doesn 't know to react to an early
Black has compensation with good, ... ttJd7-b6-C4.
active play. 16 ttJd2
The text asks the question of the After 16 a3 l::tfb8 17 .i.fl ttJxa3 ! 18
knight on c4, noting th at its retreat .i.xa3 c4 Black win s the piece back with
would mean the failure of Black's en a slight advantage. Again he is very
tire strategy, and the triumph of active.
White's. Fortunately, the knight does 16 ... ttJcb6 17 ttJxb6 ttJxb6 18 a4 .i.xe2
not have to retreat. 19 l:tel 'iWa6
14 ... 'iWa5!
Black h as regained the pawn with a

Now White's plan of .i.d2, a2-a4, etc good g ame. It is clear that White h as
is prevented. failed in the opening phase.
15 ttJa4?! 20 ttJC4 .i.xC4 21 bxc4 'ili'xC4 2 2 'iWxC4
White should follow through ; i.e. 1 5 ttJxc4 23 l:txe7 l::tx a4 24 .i.fl ttJe5
bXC4! "tIi'xC3 16 'i!VxC3 .i.xc3, when Ser N ow Black i s a pawn up and stands
gey Kasparov gives 17 .i.g 5 ! .i.xC4 18 clearly better. Again, I will just give the
23
rest in brief a s i t is not relevant t o the H avin g now dealt with the most
opening discussion . critical line in the Fianchetto, we will
2 5 �f4 11d4 turn our attention to the earlier moves
Or 25 ... g 5 ! 26 �xe 5 �xe 5 . and examine other White possibilities.
26 �xe5 �xe5 27 f4 �f6 28 J::t d 7 l::tx d 5
29 l:tbb7 J::t d 1 30 'it>g2 d 5 3 1 �C4 dXC41?
3 2 ':xd1 c3 Game S
S.Pavlovic-T.Giorgadze
luga no 1985
1 d4 ttJf6 2 c4 cS 3 d 5 b5 4 cxb5 a6 5
bxa6 g6
White faces a difficult defen sive

task, as one of his rooks rendered quite
passive by the need to take care of the
advanced black c-pawn .
3 3 �f3 .l::[ e 8 34 .l::[c 1 �d4 3 5 lic2 .l::[ e 3+
36 �g4 litd 3 37 lia7 h 5+ 3 8 �h3 �g7 This move requires a little explan a
39 :e2 �e3 40 l1a6 �xf4 41 lIa4 lId2 tion . Very often Bl ack plays 5 ... �xa6
42 lie1 l:id4 43 :a 5 �d2 44 .:tf1 ':'C4 45 here, when 6 ttJc3 d6 7 g3 g6 8 �g2
l:ta7 c2 0 - 1 �g7 reaches the initial tabiya for this
ch apter. H owever, via that move order
Game summary there is a serious alternative in 6 g3 d6
In the early part of the g am e Gallardo 7 �g 2 g6 8 b3 �g7 9 �b2, and White
pl ayed in accurately. In stead of 1 5 has scored well from that position.
ttJa4 ? ! , the logical 1 5 bXc4 would have By playing 5 ... g6 first and delaying
been a stern er test of Black's concept, the capture on a6, Black neutralizes the
although Black should be doing alright possible double fi an ch etto, since if
here too. After White's inaccuracy Or White goes ahead with it anyway, the
tega was already slightly better, and he pawns on d5 and a2 can find them
kept the advantage and converted it to selves weak. This is becau se Black still
a win without m ajor problems. has the option of ...ttJxa6, and can then
24
Th e Fi a nc h e t to Va ria t i o n
apply pressure with ... ltJb4 and possibly Again Black i s preparing for . . .ltJb4,
... �b7 or ... �f5 (as in the current game). followed by ... �b7 or ... �f5, highlight
In fact, after Tb3 ? ! it is White who must ing White's weaknesses on the light
think about equalizin g ! For this reason, squares. Both the pawn s on a2 and d5
the double fianchetto is rarely seen after are vuln erable, while the C2-square has
5 ... g6 in modem practice. al so to be defended somehow after
6 g3 . .. �f5.
After 6 b 3 ? ! ltJxa6 Black is already 9 �g2 0 - 0 10 'iVd2?!
slightly better; e.g. 7 �b2 �g 7 8 g 3 10 ltJc3 �f5 al so gives Black the ini
ltJb4 9 �g 2 l:txa2, and the b 3-pawn, tiative (11 e4? fails to 1 1 ... ltJxe4!). In
too, will be weak in the middlegame. A stead, 10 e4, trying to build a strong
logical continuation could be 10 li[xa2 centre, seem s like a m ore logical pl an
ltJxa2 11 'iWd2 ltJb4 12 ltJh 3 0-0 13 ltJC3 to me. A possible continuation i s
d6 14 0-0 'iWb6 1 5 ltJf4 �d7 16 ltJe4 10 ... �b7 (10 ... c4! ?, intending ... ltJc5, i s
ltJxe4 17 �xe4 �xb2 18 'iWxb2 l'1a8 19 another good m ove) 1 1 ltJ e 2 e 6 , and i f
l:[al 'iWb7 and Black has good winning 12 dxe6 fxe6 1 3 0-0 �xe4 14 �xf6
chances. 'iWxf6 15 �xe4, then 1 5 ... d5 16 �xd5
6 d6 7 b3?1
... l::tfd8 17 ltJbc3 exd5 and Black i s slightly
Black gets a lead in development better.
and good play on the light squares af 10 �f5 ! l1 ltJf3?
•..
ter this move, so it seems that the plan White h ad to play 1 1 ltJc3 ltJb4! and
of b2-b3 and �b2 is a luxury White then :
cannot afford; whereas if Black h ad a} 12 e4? still fails to 1 2 ... ltJxe4! 1 3
played 5 ... �xa6, instead of 5 ... �g7, the �xe4 (or 1 3 ltJxe4 �xb2 14 l:tbl ':'xa2
double fianchetto would have been with a decisive attack) 13 ... �xe4 and
more challenging. Black i s winning, based on the simple
7 ... �g7 8 �b2 ltJxa61 tactic 14 ltJxe4 �xb2 15 'iWxb 2 ? ? ltJd3+.
b} 1 2 'iitf 1 ! looks a bit artificial but it
i s the best m ove in the circum stances
(i.e. followin g the dubious m oves 7 b3
and 10 'iVd2). Nevertheless, after
12 ... 'iVa5 13 ltJf3 �e4 Bl ack was better
in p.spacek-R.Mainka, Luxembourg
1988. He is about to regain his pawn
(on d5) and h as a good lead in devel
opment. Black went on to win in 43
m oves.
11...ltJb4!
25
White resigned without waiting for

a reply, since his position is hopeless
after 21 ...li'lxd5, and 21 i.xf3 li'lxd5 !
was winning for Black as well.
Game summary
The idea of the double fianchetto sim
ply does not work after 5 ... g6; indeed
Benko player should be very happy to
face it, as White is struggling to stay in
the game right from the opening
Pinpointing the light-squared weak phase. This is m ain reason why 5 ... g6 is
ness on the c2-square. Now Black is the most accurate move in practice. In
clearly better. the game White could have kept his
12 0 - 0 disadvantage to a minimum with 10
The attempt to defend tactically by e4. Instead, 10 'ii'd 2 ? ! lost more time
12 li'ld4 ! ? is m et by 12 ... i.e4! 13 0-0 and after 10 ... i.f5 ! it was no longer
i.xg 2 14 �xg 2 li'le4 etc. possible for White to guard against the
12 ...li'lc2 13 li'lh4li'lxal 14li'lxfS?! threat of ... li'lb4-C2 in any satisfactory
14 i.xal might h ave been a better way.
try.
14 ...li'lxb3
14 ... l:txa 2 ! looks even stronger than Game 6
the text. S.Furman-E.Cieller
lS axb3 gxfs 16 'ii'g s h6 17 'ii'xfs 'iid 7 USSR Tea m
18 'ii'f 3 .l:ta2 1 9 i.C3 Cha m pions h i p 1975
1 d4 li'lf6 2 c4 cS 3 d s bs 4 cxbs a6 S
bxa6 i.xa6
To repeat the lesson from the previ
ous g ame, I recommend 5 ... g6 6 li'lc3
i.xa6 as a m ore accurate move order.
6li'lC3 g6 7li'lf3 i.g7 8 g3 0 - 0
This is also inaccurate, as White
could now play 9 i.g 2 d6 10 l:tbl !
reaching his favourable set-up. Correct
was 8 ... d6, and if 9 i.h 3 , lhen 9 ... li'lbd7
19 ... fl'g4! 20 ttJd2 "iVxf3 21 ttJxf3 0 - 1 (or just 9 ... 0-0) 10 0-0 li'lb6 (or again
26
10 ... 0-0) 11 l':.el 0-0 transposes below. 11...�C4

While on the subject of m ove orders, it As the bishop doesn't cover dS
should be noted that 9 e4 �xfl 10 pawn from h 3, it i s quite logical for
'it>xfl 0-0 11 'it>g 2 lLlbd7 here would Black to target the d-pawn directly. Of
tran spose to the Classical m ain line in course with the bishop on g 2 thi s m ove
the next ch apter. could just be an swered by 12 lLld2 �a6
9 �h3 13 b3.
12 11dl :t a 7 13 e 4 'ii'a 8
Preventing ... lLlg4-eS which, as we

will see in the next ch apter, is a com The set-up with ...l:1a7 (or ... l:1a6),
mon manoeuvre in the Benko. Also, if ...'iWa8 and ...l:1b8 i s one of the standard
White is going to play e2-e4 the bishop plans the Benko player shoul d con sider.
looks more active on h3 than ob Although it is rarely seen in the Fi
structed by the pawn chain . On the anchetto Variation, here it m akes sense
other hand it doesn 't support the ad as White h as less control over the cen
vance e4-e s so well. tral light squares with the bishop on
9...d6 10 0 - 0 lLlbd7 11 Wc2 h3. For instance, 14 b 3 would be m et by
After 11 .l:.el lLlb6 Black i s doing 14 ... lLlxe4! 1 5 lLlxe4 (or 1 5 'ii'x e4 �xc3 )
fine. For instance, 12 e4 lLlC4 1 3 'ii'c 2 l S ... �xdS 16 l:txdS WxdS 17 l:tbl fS
'ii'a s 14 eS dxes 1 5 lLlxe s lLlxdS ! 16 lLlc6 with good play for Bl ack. In the n ext
lLlb4 17 lLlxe7+ 'it>h 8 18 'ii'e 4 l:[ad8 with g ame we will see this set-up being ef
good activity for the pawn, or 12 �f4 fective for another reason .
lLlC4 with an unclear g ame. Notice that 14 a 3
the c4-square is an excellent square for White i s usually cautious about
the knight, while �f4 is n ot usually playing a2-a3 in these positions, as the
played in thi s line as it weakens b2. b 3 -square i s n ow weak, but h ere it i s
White should con sider defendin g the h ard to avoid i t . 14 b 3 was given in the
c3-knight, preparing b2-b 3 . previous n ote, while if 14 �f4 l:tb8 1 5
27
lLld2 ..ta6 1 6 l1abl, then 1 6 ... lLlh S 17 24 �g2 f5?! 2 5 ..tb2 Itb4 26 Wc2 fxe4
..te3 lLles probes the vulnerable light 27 fxe4 'iVe2+ 2s lLlf2 ..te3?!
squares, and after 18 ..tfl lLlg4 19 ..txa6 With this move Black is overplaying
.l:lxa6 20 a3 lLlxe3 21 fxe 3 lLlf6 Black is his hand; the simple 28 .....txb2 followed
doin g fine. White h as num erous weak by ... lLle s would still have led to an un
nesses, and the likely exch ange of the clear position.
f6-knight for the knight on d2 (arising 29 ..tc3 lIfS 30 lLlC4 'ii'f3+ 31 �gl
from lLlC4 and ... lLlg4-e s) will leave ..txf2+ 3 2 1:txf2 'iVxf2+ 3 3 'iVxf2 :xf2 34
Bl ack with active pl ay on the b-file: the �xf2 1:txb3
heavy pieces are workin g well with the
dark-squared bishop.
14 1ibS 1 5 lLld2 ..ta6 16 ..tfl l1ab7 17
.•.
:a2 ..txfl lS l:txfl lLleS

Another standard m ove in the
Benko, the knight going on to c7 and
sometimes further to b S and d4.
19 b3 Wa6
With the idea of ... lLles-d3.
20 lLldl 'ife2 21 WC4 Wh 5
A string of inaccurate moves have

led Geller from an advantageous posi
tion (after the move 22 f3) to one in
which he is clearly worse, and he now
h ad to fight for the draw - a task in
which he was ultim ately successful .
3 5 lIa3 litbl 36 �e3 �7 3 7 ]:tal .l:r.b3 3S
�d2 lLla6 39 �C2 :bS 40 lLla 5 lla8 41
lLlc6 �e8 42 a 5 lLldbS 43 lLlxbS l:txbS 44
�d3 �d7 45 �C4 e6 46 :fl exd 5+ 47
2 213?! exd 5 �e7 4S h4 lLlc7 49 h5 gxh 5 50 llf5
2 2 ..tb2 was a better way to keep .l:tgS 51 ..tf6+ �d7 52 ..tg5 lIg7 53 ..tf4
the bal ance. h4 54 gxh4 litg4 5 5 :f7+ �cS 56 h5 h6
2 2 lLlc7 2 3 a4 ..td4+?!
•.. 57 a6 lLlxa6 58 �b5 lLlb4 59 ..txd6
H ere Geller starts to g o astray. After lLlxd 5 60 lIh7 ':g5 61 l1xh6 �d7 62
the thematic 23 ... e 6 ! 24 dxe6 lLlxe6 ..txc5 �eS 63 �C4 �f7 64 ..td4 lLle7 65
Black is very active with a strong initia �d3 lLlf5 66 lIh7+ �e6 61 ..tb2 :g4 6S
tive for the pawn . ..tCl :h4 69 ..tg5 lIh3+ 70 �d2 Yz-Yz
28
Th e Fia nc h e tt o Va ria t i o n
Game summary White plans t o blockade the queenside,

The line with 9 .i.h 3 is largely forgotten with moves like .i.d2, 'iVC2, l::tb l, b2-b3,
nowadays. The idea is prophylactic and a2-a4, etc; the well-timed break with e4-
is directed again st ... tZJg4-e s, while eS will also come into consideration,
White reason s that the d-pawn i s well while the knight itself will often go to f4
defended anyway. However, with to reinforce the ds-pawn.
11 ... .i.c4 followed by .. J:ta7 and .. .'iVa8
Geller showed that th e absence of a
defender of the ds-pawn does h ave
some relevance. Thi s gave him some
initiative from the opening, especially
after Furm an let slip the chance to
equalize. But later Geller went wrong
in turn, seeming somehow to miss
White's defen sive resources, and in the
resulting endgame he had to fight for a
draw.
Nevertheless, the conclusion must This system had some popul arity
be th at 9 .i.h 3 is hardly a serious test of around the millennium when it was
the Benko from a th eoretical stand used by GMs Aseev, Aleksandrov, lonov
point, and that Black is doing fine as and Se.lvanov. I, too, h ave pl ayed it
long as he knows what he's doing. For with White in about ten g ames, and it
White the main advantage of thi s line can be quite dangerous if Black doesn't
may be that the opponent has forgot watch out. However, the tZJh 3 line is
ten about it, or else that it is so old they currently unfashionable, as Black has
never knew about it at all . mostly sorted out how to neutralize it.
Ironically, the knight's absence from f3
i s a significant factor, as we will see.
Game 7 9 0 - 0 10 0 - 0 tZJbd7 11 'ikC2
.••
K.Aseev-R.Ponomariov ll .i.d2 l:ta7 leads to similar play.

Eu ropea n Cha m pionship, 11 :a7
•..
Oh rid 2001 I like the idea of ... l:ta7, ... 'iVa8, and
... .l:r.b8 here. The reason i s that Bl ack can
1 d4 tZJf6 2 c4 cS 3 d 5 b5 4 cxb5 a6 5 l ater play ... cS -c4 without worryin g
bxa6 g6 6 tZJC3 .i.xa6 7 g3 d6 8 .i.g2 about the possibility o f tZJf3-d4-c6,
.i.g7 9 tZJh3 since the white knight is n ot on f3. It i s
The ideas behind this move are quite thi s crucial difference th at justifies
similar to those in the 9 tZJf3 main lines: Black's play in thi s line.
29
12 b3 "a8 13 l:tb1 l:tb8 1 4 .l:.d1!? �xe4 2 3 i.. x e4 i.. xh 3 with equal

14 i.. d 2 .l:r.ab7 15 l:tfcl would saved chances.
White a tempo on the g ame, but I think 17 bXc4 i.. x C4 18 �f4
Black h as compensation n evertheless.
14 l:tab7 1 5 i.. d 2 c4
•..
18 .....a6?1
And h ere 18 ... �fd7 ! was more accu
This thrust is well timed and effec rate, supporting the knight on cS and
tive. N ow m ost of Black's pieces will opening for the bishop on g7, so that
become active, givin g him good com 19 ':xb7 ':xb7 20 �a4 (as in the next
pensation for the pawn. As already n ote) can be m et by 20 ... i.. x a2.
m entioned there i s no disturbing �f3- 19 ':'xb7 .l:txb7 20 :b1?1
d4-c6 for him to worry about, and n o N ow 20 �a4! would have been
b 3 -b4 closing t h e queen side either. slightly better for White, since
16 l:tdc1 20 ... i.. x a2 ? fail s to 2 1 �xcS, while after
As it h appens I h ad reached this po 20 .. :ifxa4 (or 20 ... �xa4 2 1 "xc4) 21
sition myself as White in an earlier "xa4 �xc4 2 2 ':xc4 White would have
g ame. N .V.Pedersen-J.Carrasco Marti more chan ces of achieving something
nez, World Junior Championship, Za with his a-pawn and bishop pair.
g an 1997, continued 16 i.. e 3 cxb3 17 20 ':xb1+ 21 "xb1 �fd7 22 "C2 'i'a3
.••
axb3 l1C7 18 i.. d4 'iWb7 ? ! (18 .. J:�b4 was Rather than regain the pawn by
correct) 19 "d2 (here 19 b4! would 2 2 ... i.. x C3 2 3 i.. x c3 "xa2, which would
h ave given White an edge) 19 .. :ifb4, m ean ceding the bishop pair, Black
when Black had reasonable compensa keeps things fluid and is soon re
tion. warded with a mistake.
16 ...�C5?1 23 h4 �e5 24 i..c 1 "a7 25 i.. h 3 "a 5 26
Better was 16 ... cxb3 17 l:.xb3 %bb3 i.. d 2 "a3 27 'iii>f 1? �f3!
18 axb3 �CS 19 l:tbl i.. c 8, and if 20 b4, Oops! Now Black gain-s the bishop
then 20 ... i..f S ! 21 e4 �cxe4 22 �xe4 pair and is clearly better.
30
m an ages t o free his pieces, so that Po

n om ariov only m anages to win the
g ame due to a rather big mistake from
Aseev at the end.
45 .ig2 .ic8 46 ltJe4 .ig7 47 ltJxC3
.ixc3 48 ltJC1 �7 49 .ie4 h6 50 ltJd3 g5
51 hxg 5 hxg 5 52 �1 �e8 53 �e2 �d8
54 f4 g4 55 �e3 �C7 56 .ig6 �b6 57
.ie8 .if5 58 .if7 �b5 59 .ie6 .ixe6 60
dxe6 �c4
28 ltJd1 ltJxd2+ 29 'iWxd2 .ixa2 30 .ig2

.ib3 31 ltJe3 .ia4 3 2 'iWa 5 'iWa1+ 3 3 'iWe1
'iWb2 34 .if3 ltJb3 35 �g2 ltJd2 36 ltJd 3
'ii'C 3 37 'iWc1 'iWxc1 38 ltJxC1 .ib5 39 ltJf1
ltJb1 40 ltJb3 ltJC3 41 e4 .id3 42 ltJfd2
.ih6 43 �gl f5
61 f51
H ere 61 �e4 would h ave h eld; e.g.
6 1 ... dS+ (not 6 1.. . .id4? 6 2 ltJb4! ) 6 2 �e3
.id4+ 6 3 �e2 �bS 64 fS and Black can
m ake no progress. Whereas n ow the
black king invades.
61 ... .id4+ 62 'ifole2 �d5 63 ltJf4+1
Thematic in a Benko endgame, 6 3 ltJb4+ was a better try.
Black pl ays .. .f7-fS to undermine the 63 ... �e5 0 - 1
centre. Note that in the middlegame, Since 64 ltJg6+ 'ifolf6 6 S �d3 .if2 i s
on the other hand, thi s move i s usually winning easily for Black.
a bit risky, and circum stances have to
be favourable for it to work. Game summary
44 exf5 .ixf5?! The line with ltJh 3 is n ot so popular
More accurate was 44 ... gxfS ! , with these days. Many of the ideas for White
good chances of winning the ds-pawn are the same as in the 9 ltJf3 m ain
and the game. After the text White lines. H owever, by choosin g the plan
31
with ... l:ta7, . . -'.a8, ...l:.b8 and a l ater in Gam e 1), the knight goes to c4 at
... c5-c4, Black can exploit the fact that once and Black is doing fine .
the knight can't go from f3 to d4 and White's most consistent plan is to
c6. From the opening phase the g am e continue with moves like 'ii'c 2, l:tdl, l:tbl
i s unclear. and b2-b3 anyway, though I don't think
In the middlegam e Aseev m ade a Black has a theoretical problem here
tactical blunder, which allowed Black either. Avrukh's suggestion of 'iWc2 and
to g ain an edge. His technical realiza h2-h4 is covered in the notes to Game 2;
tion of this in the endgame was less as long as Black responds with ...h7-h 6,
than convincing , however, and it was this leads to an unclear game. The con
only thanks to a fin al mistake from his clusion is that Black seems to be doing
opponent that Ponomariov m an aged fine after 9 ... liJfd7, and it is up to White
to win . to try and find something new.
The conclusion is that liJh 3 is not Regarding White's other possibili
too difficult to neutralize. It i s charac ties: 9 liJh 3 can be met by ...l:Ia7, ... 'ii'a 8,
teristic for the variation that it is Black .. JIb8 and ... c5-c4 (as Ponom ariov
who is pressing and h as the winning played in Game 7). 9 .i.h 3 is a forgotten
chances. move, which prevents ... liJg4 and puts
some pressure on d7, but in Game 6
Summary Geller showed that Black is doing fine
This chapter examined one of the m ost after 1 1 ....i.c4. Finally, the proposed
critical and fashionable ways of playing move order with 5 ... g6 (rather than
against the Benko, where White aim s 5 ... .i.xa6) is specifically designed to
to neutralize Black's queenside initia counter the double fianchetto with g2-
tive, startin g with 9 liJf3 liJbd7 10 :tbl. g 3 and b2-b3. Should White proceed
To counter this, my recommendation is nonetheless, Black responds with the
9 ... liJfd7 ! , by which Black goes for m ore active ... liJxa6 ! , ... liJb4, and possibly
rapid counterplay with ... liJb6-C4 and ... .i.f5, and even comes out of the open
.. :iVa5 . If White essays a quick e2-e4 (as ing with some advantage.
32
Cha pter Two
The Classical Variation
1 d4 t2Jf6 2 c4 cS 3 d5 b5 4 cxb5 a6 5 bxa6 This i s one of the sounder ways to

g6 6 t2JC3 .i.xa6 7 e4 .i.xfl 8 'it>xfl d6 meet the Benko Gambit. Karpov played
it as White with quite good results.
Typical m oves for White include .:tel,
followed by 'iYc2 or 'iVe2, or l:te2 and
'iVel, or :e2-c2 and 'ii'e 2 . In any case
White defends the second rank and the
knight on c3 while con sidering a well
timed e4-e S . Sometimes .i.g s is pl ayed,
especially if the bl ack queen and f6-
knight h ave moved away.
Typical plan s for Bl ack include
... t2Jg4-e s or ... t2Je8-c7-b S, ... 'iVb6/'iVas
The starting point for the Classical and ... :fb8, or ... .l:ta6/.l:ta7, ... Was and
Variation, the main line continuing ... ... e7-e6. In some cases the committal
9 g3 .i.g7 10 ..ti>g2 0-0 11 t2Jf3 t2Jbd7 move ... cS-c4, intending ... t2Jcs-d3, al so
comes into con sideration .
The first question here is: should
White play h 2-h 3 or not? I s ... t2Jg4 th at
dangerous a threat th at White should
use a tempo preventing it? The conclu
sion used to be "yes", but n ow White
often prefers 12 :el, since if Black opts
then for the plan of ... .l:ta6, .. :ii' a 8, ... e7-
e6, White m ay save a tempo. Therefore
again st 12 l:1el I recommend the the
m atic 12 ... t2Jg4, when Bl ack i s doing
33
fine; while against 12 h 3 the plan with So White allows 1 2 ... liJg4. The alter
12 .. .lla6 and ... Va8 is seen to be effec n ative 12 h3 is covered in Games 10-13.
tive, and 12 ... 'i¥b6 is possible too. Either 12 ...liJg4
way Black h as sufficient compen sation
for the pawn .
The Classical h as declined in popu
l arity somewhat, probably because, as
shown by the ideas and analysi s in thi s
chapter, Black seem s t o be doing fine.
Incidentally, there are m any n ovelties
in thi s book and in this chapter in par
ticul ar. I don 't point out each tim e a
n ovelty appears as I only found around
10%-30% of them ; the rest were found
by the computer. H owever, the con se This is the principled reply and the
quence i s th at some of the lines which one I recommend. The knight will often
were previously regarded as problem go to e S , and from there, if Black man
atic for Black are in fact seen to be at ages to justify it tactically with precise
least equal. pl ay, to the good squares d3 and c4.
The first six games below examine Alternatively, it m ay retreat back to d7
the m ain line with g2-g3 and 'itt g 2 . Af after the other knight goes to b6,
ter that, in G ames 14-16, we look at should White prepare f2-f4 with liJf3 -
variations in which the white king d2.
takes another route, via h 2-h 3 and 13 'iWe2
c;t>g l-h 2 . The fin al two g ames in this One of the critical tests of 12 ... liJg4.
chapter see White pl ayin g m ore ag Others are are 13 h 3 , 13 l:!e2 and 13
gressively on the king side, with 9 g4 liJd2. I will cover the first two of these
and 7 f4 respectively. here, and 1 3 liJd2 in the next game.
a) 1 3 h3 (by chasing the knight at
once with h 2-h 3 and f2-f4, White will
Game 8 g ain some space in addition to the
S.Cossin-M.Vachier Lagrave pawn ; but moving forwards can al so
Chalons (ra pid) 2008 backfire for White in the Benko)
13 ... liJge s 14 liJxe s liJxe s 15 f4 liJd7 16
1 d4 liJf6 2 c4 cS 3 ds bs 4 cxbs a6 5 VC2 1:.a6 17 a4 'ii'a 8 18 liJb s 1:.c8 19
bxa6 g6 6 liJC3 .i.xa6 7 e4 .i.xfl 8 c;t>xfl i.. d 2 c4 (clearing the path for the
d6 9 g3 .i.g7 10 'ittg 2 0 - 0 11 liJf3 liJbd7 knight to cS and d3, b3 - or a4) 20 .i.C3
12 l:.el (better was 20 .i.e3 ! liJcS 21 i.. x cs l:txcS
34
Th e C la s s ic a l Va ria t i o n
with a balanced position : it is difficult which mean s Black can usually meet
for either side to m ake any progress) .i.g 5 with ...h 7-h 6; it al so prevents .i.d2,
20 ... .i.xc3 2 1 ttJ)(c3 ttJC5 22 lle2 llb8 2 3 as the b2-pawn would be h anging .
.l:[a3 .l:[b4 2 4 'ifilh 2 "ili'b7 2 5 as ttJd3 and
Black had enough compensation for
the pawn in V.Neverov-V. Bologan, Ni
kolaev 199 5.
b) 13 Iite2 covers the second rank.
White has a good score with this move,
though Black is doing fine, if he plays
correctly:
For example:
b21) 14 .i.g 5 h6
bl) 13 ... 'ii'a 5 14 .i.g 5 (attackin g the

pawn on e7; after 14 ':'c2 .l:lfb8 Black
has compen sation with the idea of
... 'ii'a6 and ... ttJe 5-d3) 14 ... .i.xc3 ! ?
(14 ....l:.fb8? 1 5 .i.xe7 i s clearly better for
White; while 14 ... .l:tfe 8? ! is a bit slow 1 5 .i.cl (after 1 5 .i.xe 7 ? ! l:.fe8 16
and al so offers White a slight advan .i.h4 g 5 17 ttJxg 5 hxg 5 18 .i.xg 5 ttJge5
tage in my opinion) 1 5 bxC3 f6 16 .i.d2 Black is doing fine; 1 5 h3?! ttJxf2 16
'ii'a4 17 .i.f4 g5 (V.Gavrikov-Y.Meshkov, l:txf2 hxg 5 17 ttJxg 5 ttJe5 18 'ii'f l c4
USSR 1981) is given as unclear by a few gives Black a very comfortable position
sources. It is certainly playable for too) 15 ... l:.fb8 16 ':c2 'ii'a 6 17 'ii'e 2
Black, but I would say it is rather ttJge5 (heading for d3; Black h as some
slightly better for White. pressure in the endgame) 18 'ii'x a6
b2) 13 .. :iVb6! is the move I would ':xa6 19 ttJel c4 20 .i.e3 ttJd3 21 ttJxd3
prefer. Although it doesn 't attack c3, cxd3 22 l:td2 ':'xb2 23 Ibb2 .i.xC3 24
from b6 the queen covers the d6-pawn, litabl .i.xb2 25 l:txb2 ttJe5 and Black
35
was slightly better in F.Fleury a) 14 h 3 (not the most critical move)

S.Andrianoho, correspondence 1997. 14 ... tDge5 15 tDxe 5 .txe5 16 f4? ! .txC3
b2 2) 14 .tf4 'i¥a6 1 5 I!c2 (or 1 5 'i¥c2 17 bXc3 'ili'xC3 18 .tb2 'iWb4.
1:[fb8 16 h3 tDge 5 17 tDxe5 tDxe5 18
.txe 5 .txe5 with compensation)
15 .. Jitfb8 16 litacl :b4 17 'iVe2 .txc3 18
bXc3 'i¥xe2 19 ':'xe2 :ba4 20 litcc2 tDgf6
2 1 c4 (or 2 1 .tg 5 1:[c4 2 2 tDd2 ':ca4 2 3
tDf3 11C4 with a repetition) 2 1...tDb6 2 2
e 5 tDh 5 and Black regains the pawn
with an unclear g ame.
13 'iWa 5
•.•
Given the unopposed bishop on b2

one might think this position is good
for White. But in fact Bl ack is slightly
better here with the a- and b-files for
the rooks, while the king on g 2 is not
too safe either. S. Bil sel-F.Buettner, cor
respondence 1998, continued 19 a3
'iWb3 20 lIabl l:tfb8! 21 1i'd2 ? ! (but if 21
.tf6, then the queen sacrifice 2 1...tDxf6 !
Attackin g the knight on c3, so that 2 2 l:txb3 l:txb3 shows the point of
Black can just win the pawn back in 20 ... I:tfb8 - Bl ack will soon have rook,
some lines. Otherwise .. JIfb8 and knight and pawn for the queen, and
.. :�a6 (as in the g ame) is an interestin g plan s ... tDh 5, ... llaxa3 and Itxg 3 with a
plan . A s in m any Benko positions, Black clear advantage; e.g. 2 3 e5 tDh 5 24
is doing alright in the endgame despite exd6 ':'xg 3+ 2 5 �h 2 exd6 26 'fie7
bein g a pawn down. The bishop on g 7 Itaxa3 27 'i¥xd6 l:txh 3+ 28 �g l l:tag 3+
is strong , Black h as the nice a - and b 2 9 �fl l:tf3+ followed by ... tDxf4 and
files for the rooks, while a knight can wins) 2 1...'iVa2 2 2 'iVCl ':xa3 2 3 l:1e2
go to an active square such as d3, c4 or 'iWb3 24 .txa3 'i¥xbl 25 'i¥xbl ':'xbl
a4. with a decisive advantage for Bl ack,
14 .td2 who went on to win in 34 moves.
White h as tried several other moves b) 14 .tg 5 .txc 3 ! ( Black doesn't
here, of which 14 tDd2 (line 'c') is the want to spend time covering the e7-
m ost significant: pawn) 15 bxc3 f6 16 .td2 l:tfb8 with
36
compensation . The position i s similar indeed sufficient t o draw in the g ame

to th at in Game 10 (Piket-Topalov). after 29 lldl 'ili'a5 (the immediate
c) 14 liJd2 - �xc3 ! (the immediate 29 ... e5 m ay improve) 30 h4 e5 31 l:tfl
14 ... liJge 5 ? ? would lose a piece to 15 f4, 'ika3 3 2 l:te2 'fixc3 3 3 J:[e3 'iVC2 34 llf3
while 14 ... liJgf6 ? would also be a hu liJC5 3 5 llel 'iix a2 3 6 11ee3 'ii a 8 3 7 l:tf6
miliating retreat, leaving White clearly l:tb3 3 8 %le2 l:tb8 3 9 lIe3 Yz-Yz.
better after 15 liJc4) and n ow: c2) 1 5 liJC4 'ii a 6 16 bxc3 liJge5 and
now White h as two sound choices, 17
liJb2 and 17 liJxe5, though in both
cases Bl ack will obtain good compen sa
tion for the pawn :
Cl) 15 liJb3 'ii'a 6 16 bxC3 liJge5 is

unclear. Again there are no worries in
the endgame: with the open lines for
the rooks and the squares d3 and c4 for
the knights, Black has good compensa 17 liJxe 5 (or if 17 liJb2, intending f2-
tion for the pawn . One example: 17 f4, a2-a4 and liJc4 again, then 17 ... c4!
�h 6 l:tfb8 18 'iiC 2 ? ! (or 18 'iix a6 l:txa6 leads to an unclear position) 17 ... liJxe 5
19 l:tebl l::t a 3 with sufficient compen 18 'ii'x a6 (more accurate was 18 �h 6 !
sation) 18 ... liJd3 19 lle2 liJ7e5 20 lldl l:tfb8 19 'iix a6 1txa6 20 llebl l:tba8,
c4 (Bl ack has the initiative so White M. Krallm ann-M. Kopylov, Dortmund
gives up the queen for rook, knight and 1999, though Black should be fin e
pawn) 2 1 liJd4 litb2 2 2 'ii'x b2 liJxb2 2 3 anyway; e . g . 2 1 l:tb7 liJ g 4 2 2 �g 5 f6 or
l:txb2 liJd3 2 4 1:1C2 l:te8 2 5 llbl e6 2 6 22 �f4 �f8) 18 ... 11xa6 19 i..h 6 (or 19
dxe6 fxe6 27 liJf3 'ili'a8 2 8 liJg 5 l:1b8 and �f4 liJd3 20 .l:.ebl f6 with compen sa
Black had good winning ch ances in tion) 19 .. JUa8, when 20 l:te2 i s just
K.Frey Beckh am-D.Chan, correspon equal ; whereas after 20 l:teb l ? ! (as in
dence 2000. Nevertheless, White h as C. Becker-M. Burschowsky, Austrian
some hopes of making a fortress, while Team Champion ship 1997) 20 .. .f6 !
Black should watch out for a TOok or would even give Black the edge in the
two appearing on the f-file, which was endg ame. Some ideas are ... g 6-g5,
37
. . .cj;;f7, .. J!a4, . . .lla3, and . . .l:txa2 (as long Black wants to keep things fluid, but
as there is n o m ate on f8). the more concrete l s ... lLlges 16 lLlxe s
c3) 1 5 'iVxg4 .i.xd2 16 .i.xd2 'ii'x d2 .i.xe s 17 l:tecl .i.xC3 18 .i.xc3 'iVxa2 was
17 'iWxd7 lilfe8 18 b3 (or 18 a4 'iVxb2 19 correct, with an equal position. Black
as c4 20 a6 Y2-Y2 H . Horta-A.Tsvetkov, can neutralize the bishop with .. .f7-f6,
correspondence 1999, since the posi if n ecessary, or sometimes with ... lLlb6-
tion is equal ; e.g. 20 ... c3 21 l:tebl 'iVe2 a4, and target the pawns on b2 and
22 l1el 'iVb2 or 21 'iiC 7 (2) 18 ... cj;;f8 ! (de possibly dS (if White plays e4-e s).
fen ding e7 and e8 and thus threaten 16 b3 .i.XC3? 17 .i.xc3 'iWxa2 18 h3?
ing ...l:.xa2, as well as enabling Black Here 18 lLld2 ! , with the threats of
n ext m ove) 19 a4 (19 'iVh 3 ? ! cj;; g 8 20 'ii'x g4 and l:.al, would have given
'iVd7 cj;;f8 led to another draw in White a clear advantage; e.g. 18 ... lLlgf6
S. B.Jensen-R.5kytte, Dani sh Champion (unfortunately 18 ... lLlge s is no longer
ship, Odense 2011, though here possible, due to 19 f4) 19 .:tal 'iic 2 20
19 .. J:txa2 20 liba2 'iVxel 2 1 'iVxh 7 l:.ecl 'ilkxc3 2 1 l:.xC3 :'xal 2 2 lLlc4.
'iVxe4+ 2 2 f3 'iVes would h ave been 18 ... 'iVxe2 19 :'xe2 lLlgf6 20 lLld2 lLle8
slightly better for Black) 19 ...l:.eb8 20 21 lLlc4
l:tadl 'iWc2 21 e S l:.d8 2 2 'iWC7 llac8 23 21 li:teel followed by li:tal is okay for
'ii'b 6 llb8 24 'ii'C 7 l:.bc8 Y2-Y2 R. Ruether White.
F.Buettner, correspondence 1998. 21 ... lLlc7 22 ':'eb2 f5 2 3 f3
White might h ave tried 24 'iVc6, hopin g Or 23 .i.as lLlbS 24 exfs gxfs 25 l:[el
for 24 ... l:tbc8? ! 2 5 exd6 ! exd6 2 6 'iVb7 with an unclear g ame.
with the advantage, but 24 .. :iVxb3 2 5 2 3 ...lLlb5 24 .i.d2 fxe4 2 5 fxe4
exd6 exd6 2 6 a s li:tbc8 would still be
equal .
14 ... l:tfb8 1 5 nabl
Black has no problem s here; indeed

after 2s ... lLla3 26 lLlxa3 .uxa3 27 b4
cxb4 28 li:txb4 li:txb4 29 .i.xb4 li:td3 he is
15 'iWa6?1
••. even on the favourable side of a draw,
38
as his rook is more active, while the 34 .. .'�g7 3 5 lta7+ � h 6 36 11ee7 ttJf5 3 7
white pawn s e4 and dS are slightly .l:txh7+ �g5 38 l:thd7?1
more vulnerable. In stead, Black plays And h ere 3 8 �2 lld3 3 9 lla4 l:txdS
for more and should h ave gotten less. 40 h4+ �f6 41 hS offered more chances
25 ...ttJf6 26 11e1 ttJa 3 2 7 ttJa 5 1 l':txa 5? of saving the g ame.
A faulty combination . No better is 38 ...l1d 3 39 l:ta4 Iixd 5
27 .. J:tbS ? 28 l:ta2 l:tbxas 2 9 �xas l:lxas
30 e s ! ttJxdS 31 l:tea1, when White re
gains the piece with the advantage.
Black should admit his mistake and
play 27 .. J:tb6 28 �C3 (not 2 8 ttJc6 ? !
ttJc4!) 28 . . .ttJbS still with a fairly level
position.
28 �xa 5 ttJC4 29 11a2?
Missing his chance with 29 �c7 ! , af
ter which White is clearly better; e.g.
29 ...lk7 30 l:tc2 ttJa3 3 1 lIa2 Iitxb3 32
es!. Now White h as a difficult task, de
29 ... lixb3 30 � d 2 ttJxd2 3 1 ':'xd2 ttJ h 5 fen ding against the connected c- and
d-pawns.
40 �3 �6 41 �2 ltd2+ 42 �3 l:td3+
43 �g2 c4 44 l:tC7 d5 45 lla6+ �g5 46
':'cc6 �4 47 :C5 c3 48 �2 l:tf3+ 49
'ii;>e 2 d4 50 llxg6 d 3+ 0 - 1
Game summary
When White tries to do without play
ing h 2-h 3, Black should exploit thi s by
playing ... ttJg4, when the knight is
ready to jump to e S . With some knowl
Now Black gains another pawn and edge of the position, Black should
it is White who must try to draw. I will equalize without m ajor problems.
give the rest with just light comments Some of the ideas are .. ..l:Ub8, ... 'ifb6/as
as it is not relevant for this opening 'ilfa6, ... ttJeS-d3/c4 and ... ttJb6-c4/a4.
book. Another recurrin g theme is ... �xc3
32 e5 ttJxg3 33 exd6 exd6 34 lla2?1 with the idea of winning the a-pawn,
Taking the seventh rank with 34 which often works surprisingly well
l:te7 was stronger. despite its risky appearance. Here it i s
39
suggested a s a n improvement for Black precision not to get in trouble. White's

on move 1 5 ; we saw another example l ast move offers to return the pawn
in Game 3 (Neverov-Pedersen). (after ... 1.xc3, etc). Here it is con sidered
In the g am e Cossin had two oppor unfavourable for Black to accept it, but
tunities to obtain the better position : I think Black is doing alright after
when Vachier didn 't play l s ... 1.xc3, ... 1.xc 3 .
and l ater when he sacrificed the ex Another option is 16 tDf3, aiming
change. Once White h ad spurned those for e4-es, when play can continue
chances (in his defence, it was a rapid 16 ... tDa4 17 tDdl (or 17 'ii'c 2 tDxC3 18
g ame) he found him self in a difficult bxc3 Was 19 1.d2 'iVa4 with compensa
endgame and was outplayed by his tion) and now:
higher rated opponent.
Game 9
A.LuIOVoi-A.Khalifman
St Petersbu rg 1996
1 d4 tDf6 2 c4 cS 3 d s bs 4 cxbs a6 S
bxa6 g6 6 tDC3 1.xa6 7 e4 1.xfl 8 <io>xfl
d6 9 tDf3 tDbd7 10 g3 1.g7 11 <io>g2 0 - 0
1 2 l:tel tDg4 1 3 tDd2 tDges 14 'ii'e 2
a) 17 .. :fVb6 18 .ubl (for 18 es 'i'b7
see line 'b') 18 ... 'ii'a 6 (them atic: thanks
to his active pieces Black is not worse in
the endgame) 19 J.d2 l:tfb8 (19 ... c4 is
al so fine) 20 b3 c4 21 tDf2 tDacs 22 bXc4
'it'xa2 23 ':'xb8+ ':'xb8 with compensa
tion for the pawn in M.Gurevich
H . Banikas, Corinth (1st g ame, rapid
m atch) 2002.
b) 17 .. :ifc8 ! ? (similar to ... 1i'b6 in
that Black can pl ay either ... 'ii'a 6 or
Controlling d3 and preparing f2-f4, .. :jib7, while in some cases the queen
so Black must m ake room for the e s might also support ... cs-c4) 18 es 'i!Vb7
knight to retreat. (attacking the ds-pawn, made vulner
14 ... tDb6 lS f4 tDed7 16 a4 able by White's last move) 19 'fWe4 (or
In this position Black must pl ay with 19 e6 tDf6 20 exf7+ llxf7 21 tDg s
40
'iVxdS+ 22 'iif3 lilff8 2 3 ':'xe7 lUe8 with 2 1 �xa3 'fIC7 with a bal anced position .
compensation) 19 .. 5�h 8 20 b3 ttJab6 2 1 b) 18 'iid 3 liJab6 19 �b2 'ii'C 7 20 c4
ttJC3 fS 2 2 1Vd3 was A.Yusupov liJa4 2 1 �C3 was A.5trunsky-M.Vasilev,
V. Baklan, German League 1998, when Neuhausen 2008, and n ow 21 ... ttJxc3
22 ... dxes 23 fxes e6 24 �f4 ttJxdS 2 S 2 2 'iix c3 lilxal 23 lilxal lilb8 would lead
liJxdS exds would have given Bl ack a to an equal game.
slight edge according to Baklan . c) 18 c4 liJC3 ! (Black is tryin g to use
16....l:ta7?1 the slight lead in development, in some
There has been some debate about cases playing ... ttJa2 and ... �fb8, re
this position over the years, as it is strictin g the bishop on cl; instead
quite logical to chase the black knights. 18 ... e 6? ! 19 dxe6 fxe6 20 liJf3 was
Indeed, while the game is very unclear slightly better for White in Y. Pelletier
after 16 ... lila7? ! 17 as, it seem s that l . Leventic, Buekfuerdo 1995) 19 lilxa8
White can get a slight advantage. 'iWxa8 20 'ii'e 3 ttJa2 21 �b2 (21 i.. a 3
In stead, Black should prefer liJb4 is unclear) 2 1..Jlb8 2 2 litbl 'ii'a4 23
16 ... �xc3 ! 17 bXC3. This used to re e S 'iWc2 24 exd6 exd6 2S �f6 ! (a tactical
garded as good for White. However, I trick based on the back rank m ate)
think - as I hope the annotation s will 2s ... l:ta8 with equality.
show - th at Black is not worse after 17 as
17 ... liJxa4! (17 ... lilxa4?! 18 ':xa4 liJxa4
19 c4 is slightly better for White) and
then :
17...'fIa8
17 ... liJa8? is way too passive and 18
ttJC4 is clearly better for White, as Black
a) 18 J:ta3 ttJab6 19 c4 (19 'iWh s l:txa3 lacks any compen sation for the pawn .
20 �xa3 'i'a8 21 .:tal 'iWa4 22 'iWd3 c4 18 'iWbs
23 'iib l lila8 24 liJf3 'iWas 2 S 'iWel ttJf6 With the threat of 19 axb6 J:[xal 20
and 19 lIb3 lila2 20 c4 'iWa8 2 1 'ifd3 .:tb8 'iWxd7; the an swer is another develop
are both unclear) 19 .. .f6 20 �b2 lilxa3 ing move. Alternatively:
41
a ) 18 liJb3 ? i s a tactical mistake: i.d2 l:ta4 27 'itt e 2 l::t x c4 28 'itt d 3 l:txd2+

18 ... i.xc3 19 bxc3 liJxd5 and Black is 29 'itt x d2 and White is slightly better. In
better, since the b 3 -knight is h anging the rook endg ame there are ideas of
after 20 exd5 ? ! 'iWxd5+. e4-e 5, as well as 'itt d 3 and litbl-b7. The
b) 18 a6 (advancing the pawn, with pawn endgame after 29 ... :a4 30 �d3
the threat of liJb5, is another logical ':a2 31 l:!e2 :'xe2 32 'itt x e2 is al so prob
m ove) 18 ... l:tb8 ! (18 .. :.c8? ! h as al so l ematic for Black, since the white king
been played, but leads to a worse posi is far m ore active.
tion for Black) 19 liJb5 l:txa6 20 liJc7
':'xal 21 liJxa8 l:tbxa8 2 2 liJf3 c4 and
Black h ad some compensation for the
queen in C.Olausson-R.Skytte, Stock
h olm 2000, though objectively White
must be slightly better here.
lS .. JlbS! 19 axb6
The only sen sible m ove.
19 Jlxal 20 'iWxd7
.•
21 liJdbl??
Missing his chance. Lugovoi later
found the correct line 21 'iVxe7! iLxc3
22 bxc3 l:!xcl 23 l:[xcl llb2 24 'iWxd6
l:lxd2+ 2 5 'itth 3 and White is winning;
e.g . 25 ...'iWa2 (or 25 .. :ifc8+ 26 fS ! gxfs
2 7 l:.bl) 2 6 ':'h l ! (stronger than Lugo
voi's 26 'itth 4, which allows 26 ... h 6 ! )
26 .. :ifc4 2 7 Wb8+ (or just 27 'iVe s)
20 ':xb6??
•.. 27 ... 'itt g 7 28 'fie 5+ 'itth 6 29 'fig 5+ 'itt g 7 30
This should lose. Black h ad to play f5 'iWxe4 3 1 f6+ 'itt g 8 3 2 'iVxd2 'ifxh l 33
20 ... iLxc 3 ! 2 1 bxC3 'iWa2 2 2 'iWa7 ! l:txb6 'iWd3 with a decisive advantage.
23 'iVxa2 ( 2 3 'ii'x b6? ! l:txcl l eads to 21 'iWbS 2 2 e5 .l:ta7 !
•••
equality, as after 24 l:te2 l:tC2 Black win s Better than 2 2 . . .l:tb7 ? ! 2 3 'iWg4 dxe5
the knight o n d 2 ; e . g . 2 5 'iVd8+ 'itt g 7 2 6 24 fxe 5 i.xe 5 2 5 l:txe s 'iWxe s 26 'ii'c 8+
'fixe7 ':'xd2 2 7 ':'xd2 'iVxd2+ 2 8 'itth 3 'itt g 7 27 'iVxb7 l:txbl 28 i.d2 ! (not 28
'iWe2 with a draw i n S.Bekker Jen sen liJxb l ? ! 'iVe2+ with perpetual check)
S. Petrosian, N orderstedt 2003) 28 ... .l:.al 29 'ilfb5, wh en the minor
2 3 ... .l:txa2 24 'ittfl :c2 2 5 liJC4 l:!a6 2 6 pieces are somewhat superior to
42
Black's rook and pawn. errors, with White missing a clear win
23 "iVg4 dxe5 24 fxe5 .i.xe5 2 5 l:txe5 and Black obtaining some chances in
'ii'xe5 26 'ifc8+ 'iti>g7 27 "ii'x c5 l:.ab7 28 the ending.
'ikf2 I don 't think we will see that much
of 16 a4 in the future, as Black h as a
relatively easy g ame, and White should
search for a theoretical advantage in
other lines.
Game 10
J.Piket-V. Topaloy
Monte Ca rlo (rapid) 1997
1 d4 liJf6 2 c4 cS 3 d5 b5 4 cxb5 a6 5
28 .. .'itg8 bxa6 g6 6 liJC3 .i.xa6 7 e4 .i.xfl 8 'iti>xfl
28 ... e6! might give Black a slight d6 9 g3 .i.g7 10 'iti>g2 0 - 0 11 liJf3 liJbd7
edge. 12 h 3
29 'ife2 'ii'x e2+ 30 liJxe2 'iti>g7 Here, and in the n ext three g ames,
And here 30 ... l:ta6 3 1 liJbc3 1:tal 3 2 White opts for thi s prophylactic m ove,
'it>f3 l:tb4 perhaps offered more win preventin g the knight m an oeuvre to
ning chances, though it i s still drawish. g4.
31 liJbc3 'iti>f6 3 2 'iti>f3 Itxb2 3 3 .i.xb2 12 "ii' b 6
.•.
.!:[xb2 34 'iti>f4 g5+ 35 'iti>e4 h6 36 g4 'iti>g6

37 h3 h5 38 liJd4 hxg4 39 hxg4 'it>f6 40
liJde2 Yz-Yz
Game summary
The plan of chasing the knights on g4
and b6 used to be reg arded as prob
lematic for Black. However, with the
thematic 16 ... .i.xc3 followed by
17 . liJxa4, Black seem s to equalize,
..
which I think is new. When Khalifm an

didn't play this idea, Lugovoi could In my opinion thi s is the soundest
have gained a nice advantage. In stead, choice for Black in the Classical m ain
the game continuation was very un line. From b6 the queen prevents the
clear and ended in a draw after mutual cl-bishop m oving for the time being,
43
and Black h as a reasonable chance to a) 1 5 e S i s a good choice if White is

g ain good counterplay with typical tryin g for a draw. Then ls ... liJxe s 16
moves such as .. JUb8, ... Wa6, ... liJe8-C7- liJxe s dxe s 17 ':xe s 'ii'b 7 tran sposes to
b S and ... liJb6-C4. The drawback is that 14 e S (see the next game), albeit with
White can adopt a plan with l:.el and each side havin g made an extra move.
e4-eS which leads to a drawish posi b) 15 :bl (planning b2-b3 and �b2)
tion. If the Benko player wants to avoid l s ... liJe8 (the logical response, after
this and is ready to take some risks, the which White has to spend a tempo
alternative idea of 12 ... l:.a6 followed by covering the knight on c3) 16 �d2 and:
... Was can be considered, for which see
G ames 12 and 1 3 .
13 l:!el .l:tfb8 14 l:!e2
The immediate 14 eS is examined in
the next g ame.
14 'ii'a 6
...
bl) 16 ... C4? ! was played in M.Gure

vich-Z.Azmaiparashvili, European Cup,
Strasbourg 1994, but it's not a good
move. After 17 �e3 l:.b7 18 �d4 White
had a comfortable edge. The game con
tinued 18 ... �h6 (18 ... �xd4 19 'ii'xd4
A good square for the queen . Now litab8 20 eS dxes 21 liJxes liJxes 22 litxes
one of the m ain priorities is to find is no better) 19 e S dxes 20 liJxes liJxes,
good posts for the knights, such as d3, and now 2 1 �xes :d8 22 :e4! liJd6 (or
c4 0r a4. 22 .. .f6 23 �d4) 23 �xd6 followed by 24
15 'ii'e l We2 would have maintained White's
Mirrorin g the similar idea of ... :ta6, plus, according to Gurevich .
... Was as Black can play in this line. The b2) 16 ... �xc 3 ! is preferable. It might
threat is now e4-e S with pressure on look dangerous for Black to take the a2-
the e-file. pawn with e4-eS coming, but the dS
White h as quite a few alternatives pawn is weak after that, so basically it
here, though Bl ack is doin g fin e against leads to an unclear position . For exam
all of them : ple: 17 �xc3 Wxa2 18 b3 (the immedi-
44
Th e C l a s s ic a l Va ria t i o n
ate 18 eS could be met by 18 .. :iVb3) t o C 4 o r sometimes a4) 17 1:[ c l (White

18 ... 'iWa6 19 es?! (Gurevich's intention ; can 't play 17 b 3 ? due to 17 ... ttJfxdS ! 18
instead 1 9 ttJd2 'iVd3 2 0 1:[e3 1Wbs is un exds i.xC3 19 i.xh 6 ttJxdS and Black
clear after 21 ttJC4 ttJb6 or 21 b4 %:ta2) was clearly better in M.Gurevich-J . Koch,
19 ...ttJC7 ! (targeting the now vulnerable European Team Championship, H aifa
ds-pawn) 20 e6 (20 l::td 2 "iVb7 2 1 b4 c4 1989) 17 ... ttJC4 18 b3 (otherwise Black
22 lIcl ttJb6 is also fine for Black, or 20 will get at least a slight edge)
exd6 'ii'xd6 21 .l:.d2 f6 ! 22 1:[cl 1:[c8 2 3 18 ... ttJxe3+ 19 ':xe3 ttJd7 and Black h ad
.Ilb2 ttJb6 and Black wins the ds-pawn excellent compensation for the pawn
with a good position) 20 .. .fxe6 21 dxe6 in J . Hjartarson- L.Alburt, Reykjavik 1986.
ttJf8 22 'iWel (or if 22 ttJg s, then C2) 16 i.f4 gs 17 i.e3 ttJh 7 (an unor
22..."ir'c6+ 23 'it;>g l 1i'ds with a clear ad thodox move with ideas of ... ttJhf8-g6,
vantage) 22 ... 1i'c6 23 i.al (now the ... .l:txb2 and sometimes ... i.x(3); e.g. 18
threat of "ir'C3 looks dangerous, but...) :cl i.xC 3 ! 19 l:txC3 (19 bXc3 ttJhf6 20
23 ... 'iWds (a cool reaction - the ideas are h4 g4 2 1 ttJel ttJxe4 intending .. :.C4
...ttJcxe6(-d4) and ...l:.xb3) 24 'iVC3 ttJcxe6 and Black is doing fine) 19 ... ttJhf6 20
25 1i'h8+ 'it;>f7 26 1:[bel ':xb3 27 i.C3 l:ta3 'iWbs with compensation.
I1aa3 28 1:[e3 ':xc3 29 ':'xC3 ':'xc3 30 C3) 16 i.xf6 i.xf6 17 :'Cl and n ow
'ii'x C3 ttJd4 and Black is better with two Black h as three possibilities. As they all
strong pawns for the exchange. lead to a bal anced position, it's a m at
c) 15 i.g s is a logical move with the ter of style which one to play:
b2-pawn defended, and the bishop is
well placed on gS. So Black should pl ay
lS ... h6, when White has three sound
choices, though none is sufficient for
more than an unclear game:
c 3 1) 17 ... i.xC3 18 bxc3 (18 l:txc 3 ? !

'ii'x a2, reg aining the pawn, is fin e for
Black) 18 ...'iVC4 19 l:tcc2 lIa3 20 "iVd2
'it;>g7 2 1 1i'e3 1:[ba8 22 ttJd2 "iVbs 23 ttJb3
1:[3a4 24 1:[b2 ttJes 2 5 f4 ttJC4 2 6 'ii'd 3
Cl) 16 i.e3 ttJb6 (preparin g to jump 'ii'd 7 with compensation for the pawn .
45
C32) 1 7 ... lLles 18 lLlxes i.xes 1 9 'iVd2 in B.5passky-1Koch, French Champion

i.g 7 is roughly equal, since any at ship, Montpellier 1991) 18 bXc3 lLlXa4
tempt by White to m ake progress will 19 eS (Stohl) 19 ... lLlxes 20 lLlxe s dxe s
allow Black counterplay. L.Sch andorff 2 1 l:[xes 'iVC4 2 2 .l:txe7 'iVxdS+ 2 3 'i¥i>h 2
A. Khalifman, Aarhus 1997, ended: 20 (better than 2 3 f3 ? ! lLlb6 with compen
b3 c4 21 bxc4 'i!VxC4 2 2 lLldl Y2-Y2. sation) 2 3 ... lLlb6 24 l:txa8 l:txa8 2 S 'iWe2
C 3 3 ) 17 ... l:tb4 (the sharpest choice, with some initiative for White.
leadin g to an unclear g ame) 18 l1cc2 (a d2) ls ... lLle8 ! is correct. Then 16 a4
thematic m ove in this line, this was the :b4 m aintains Black's queenside coun
continuation in G . Danner-R.Mainka, terplay, while either 16 i.g s or 16 i.f4
Dortmund 1993) 18 ... .l:tab8 (Mainka run s into 16 ... l:!.xb2; i.e. 16 i.f4?! ':'xb2
played 18 ... 'i¥i>h 7 here) 19 'iVd2 and n ow 17 'ii'x b2 i.xC3 18 'ivxc3 'iVxe2 and Black
White might soon play h 2 -h4, so I has a slight advantage.
would suggest 19 ... i.g7, tryin g to 1s lLle8
...
minimize the slight weakness on h 6 ; ls ... lLlb6 is a simple way to get com
e.g. 20 h4 lLlb6 (thi s was obviously the pen sation for the pawn, and I would
m ain plan here) 21 'ii'c l lLlC4 22 hS g s recommend this move active move; e.g.
with a sharp, unclear g ame, though I 16 i.g s h 6 17 i.xf6 i.xf6 and now we
would find it easier for Black to play. h ave:
d) 1 5 'ii'C 2 is a little tricky:
a) 18 e S (one should always watch

dl) l s ... lLlb6 ? ! is now met by 16 a4! out for this move in the Classical main
lLlfd7 17 lita2, when White h as a good line, but Black is okay here, thanks to
chance of obtaining an advantage; e.g. the pressure against the ds-pawn)
17 ... i.xc 3 ? ! (but n ot 17 ... lLlc4? 18 b3 18 ... dxe s 19 lLlxes (19 l:.dl lLld7 is un
lLlces 19 lLlxe s lLlxe s 20 lLlbS ! c4 2 1 clear) 19 ... 'iVb7 (notice the unfortu
i.f4! :c8 2 2 i.xe s cxb3 2 3 'iVxb3 i.xe s n ately placed king on g2) 20 lldl :d8
24 l:[ec2 and White was a safe pawn up 2 1 lLlg4 lLlxdS and Black is doing fine.
46
b) 18 :cl lbd7 19 b3 (or 19 .:tcc2 White cannot m ake progress, s o the

.Jlxc3 and if White captures with rook position is drawn .
or queen, then - 20 .. :ifxa2; otherwise 20 30 �f3 'it;e8 3 1 lbf1 'it;d7 3 2 lbg3 l1b8
bxc3 'iVc4 is unclear) 19 ... c4 20 bXc4 33 lbfS lbxfS 34 exfs ,Ub3 35 l:ta3 .ll b 8
'ii'xC4 21 lbbl 'iWa6 22 .:tec2 was 36 e4 'it;d6 37 1i13a2 1Ib3 38 :a3 'it;cs 39
M. Bertok-D.Gliksan, Zagreb 1971, and <ifi>e3 .l:.b8 40 .:td1 'uba8 41 lIb1 l:txa4 42
now 22 ... lbcs would be fine for Black; lIxa4 l:txa4 43 :b7 lita 3 44 ':'C7+ 'it;bS
e.g. 2 3 'iWe3 'ili'd3 . 45 :b7+ 'ifi>cs 46 :c7+ 'it;bS 47 .l:lb7+
16 .Jlf4 .Jlxc3 1 7 bxc3 f6 18 a 4 lbes 19 �cS Yz-Yz
.Jlxes dxes
Game summary
After White played h 2-h 3, n ot allowing
... lbg4, Topalov adopted the plan with
... .l:.fb8, ... 'i!i'b6-a6 and ... lbb6-C4 and
equalized quite easily. He might al so
h ave tried ls ... lbb6 with an unclear
g ame.
Overall, while the Classical m ain
line is still pl ayed by some GMs n owa
days, I don't think it is a problem for
Black. On the other h an d, it is probably
Black has good compensation, and sensible for White to prefer 12 h 3 , since
the game is close to equal . whatever ch ances he does h ave of
20 .:tea2 'ii'd 3 21 'ii'e 3 'ifxe3 2 2 fxe3 fig hting for an opening advantage look
lbd6 23 lbd2 l:ta s 24 g4 'it;f7 2 5 'it;f3 hS to be in this line, rather than after 12
26 'it;e2 c4 27 'it;f3 .l:.ba8 28 'it;g3 hxg4 lIel.
29 hxg4 gs
Game 1 1
V.Kachar-V.Arbakoy
Russian C u p, Moscow 1999
1 d4 lbf6 2 c4 cS 3 ds bs 4 cxbs a6 5
bxa6 g6 6 lbC3 .Jlxa6 7 lbf3 d6 8 e4
i.. xf1 9 'it;xf1 lbbd7 10 h 3
If White intends t o play the king to
g2 it is more usual to play 10 g3 first,
although there is nothing wrong with
47
the current move order. trap - see the note to White's next
m ove.
17 'iVf3
Instead:
a) 17 'iVe2 ? fall s into the trap with
17 ... li:JxdS ! ,
10 ..tg7 11 g3 0 - 0 12 �g2 'ilkb6 13 l:[e1

.••
':fb8 14 e5
A m ore straightforward approach
than the 14 .l:r.e2 of the previous g ame.
14 dxe5 15 li:Jxe5 li:Jxe5 16 l:[xe5
•..
a s occurred in G . Rechlis-J . Plachetka,

Ostrava 1981. Probably GM Rechlis
simply overlooked this move, since the
position after 18 lIxdS e6 19 'iVf3 exds
20 li:JxdS doesn 't look like it offers
White sufficient compensation for the
exchange. Indeed, 20 ... l:ta7 21 a4 (not
2 1 ..tg S ? 'iVxb2) 2 1 ... .l:r.d8 would lead to
a clear plus for Black.
b) 17 'iit g l removes the king from
the long diagonal, but is a rather pas
16 'ilkb71
..• sive move, and White must be careful
Inferior is 1 6 ... l:tb7 ? ! 17 'iVf3 li:Je8 18 n ot to get the worse position . For ex
l:te2 li:Jd6 (the knight would be great ample, 17 .. JId8 18 'iVe2 (not now 18
after ... li:JfS -d4, but this is easily pre 'ilkf3 ? ! e6 and Black is better, as 19
vented) 19 g4 li:JC4 20 ..tf4 with the bet ..tg S ? ! can be met by 19 ... li:JxdS)
ter position for White in N . Diedam 18 ... li:Jxds 19 l:.xdS l:txdS 20 li:Jxds
R.Mainka, Liechten stein 1993, a g ame 'iVxds 21 'iVxe7 'iVdl+ 22 'iit g 2 'ilkdS+
which Diedam quite surprisingly won when Black is certainly equal and
ag ain st his GM opponent. might try for more with -... ..td4, ... l:.xa2
The text involves a rather simple or .. :iVf3 (if allowed), hoping for play on
48
the dark squares, although 2 3 �g l . . .s.. x c3 and . . .'ii'x d5, without all owin g
s.. d4 24 'ii'e 2 (or 24 a4 'ii'f3 2 5 s.. e 3 l::t x e7) h as the drawback that, com
i.xb2) 24. . .:xa2 2 5 ':xa2 'ii'x a2 is still pared with the text, it takes an extra
drawish . tempo. After 20 l:te2 f5 21 J:[ael White
17 .. Jld8 is slightly better, since 21 ... s.. x c3 2 2
bXc3 'ili'xd 5 ? ? would n o w l o s e t o 2 3
l::t e 8+.
20 %:te2 s..X C3 21 bXc3 'ili'xd 5 2 2 'ili'xd 5
':'xd 5 2 3 lIxe7 ':'U
18 s..g 5
18 a4 is more ambitious, since the
a-pawn can be annoying if Bl ack
doesn 't pl ay precisely: 18 ... l:ta5 ! (from
as the rook can attack the d5-pawn Black is winning the pawn back, af
after ... c5-c4 in some lines; 18 ... l:.a6? ! 19 ter which the rook endgame is very
lIe2 left White with a slight edge in likely to end in a draw.
R.Akesson -S.Pinchuk, Pardubice 2008) 24 l:te3
19 s.. g 5 h 6 20 s.. xf6 s.. xf6 2 1 l:te2 c4 22 24 C4 l:td4 25 lIbl lIa6 also ended
tLle4 (after 22 'ii'e 4 'ii'b 4 or 2 2 l::t d 2 'ii'b 4 peacefully in W.lclicki-1 Bellon Lopez,
Black should al so keep the bal an ce) Brussel s 1987.
22 ... l:tdxd5 with equality in I.Saaed 24 ...lId2 2 5 :f3 lIaxa2 26 lIel lIe2 27
K.5eegert, World Junior Champion l::t b l .l:tab2 2 8 l::t d l 1/2-1/2
ships, Kemer 2008.
18 ... h6 Game summary
18 ... 'ii'x b2 ?? is a huge blunder, due The e4-e5 line is known as a way for
to 19 l:tbl �C2 (or 19 ... 'ii'a 3 20 lIxe7) 20 White to try and force a draw. Black
l:[cl ! 'iVb2 2 1 l:lxe7 and White is win m ay obtain some chances again st inac
ning. curate play, but the g ame was a logical
19 i.xf6 s.. xf6 con sequence of White's peaceful atti
The attempt to create some compli tude, and is perh aps the m ain draw
cations by 19 ... exf6 (planning .. .f6-f5, back to 12 ... 'ii'b 6.
49
Therefore, i f Black is i n a must win a6 al so covers the d6-pawn . So this can

situation, or facing a much lower rated be a quite a dynamic plan . For White, a
opponent, he might try the alternative set-up with lbb 5 and a2-a4 looks to be
plan of 12 .. Jla6 and ... 'ita8, which we the most effective response.
will examine in the n ext two g ames. 13 liel 'ii'a 8
Objectively White is slightly better but, In addition to preparing the ... e7-e6
as we will see, it is still interestin g from break, Black's 'iVa8/l1a6 form ation al so
a practical viewpoint. exerts some pressure on the queenside.
14 'tiC2
Game 12
J.Aagaard,.L.D.Nisipeanu
Berli n 1997
1 d4 lbf6 2 c4 cS 3 ds bs 4 cxbs a6 S
bxa6 iLxa6
Please remember that 5 ... g6 is our
preferred move order, after which 6
lbc3 .i.xa6 7 e4 .i.xfl 8 'ifolxfl d6 tran s
poses below.
6 lbc3 d6 7 e4 .i.xfl 8 'ifolxfl g6 9 lbf3 White has several ideas here: .i.g s,
iLg7 10 g3 lbbd7 11 h3 0 - 0 12 'ifolg2 lia6 preventing ... lbe8, while if Black plays
... lbb6 White might consider .i.xf6. An
other m ove is l:te2 covering the second
rank, followed sometimes by :ec2 or
:cl-c2. The pl an of a2-a4 and lbb S
should al so be mentioned, threatening
lbC7 as well as lbxd6 in some cases.
14 ':'e2 is examined in th e next
g ame. After 14 .i.g 5 Bl ack should
probably pl ay 14 ... h 6 (even though this
weaken s the g6-pawn somewh at) and
if 1 5 iLd2 then 1 5 .. J:tb8, intending
A m ove pl ayed by Topalov and ...lbe8-c7 with compen sation . Instead,
Kh alifm an . It is usually followed by 15 ... e 6 ? ! 16 dxe6 fxe6 17 'iVc2 is slightly
... 'i!i'a8, puttin g pressure on the long better for White.
diagon al after the them atic ... e7-e6 14 ... e6?!
break, when we n otice th at the TOok on Here 14 .. :ij'b7 ! is a better try, avoid-
50
ing the idea of d5xe6 and liJb5 (see the a 2 } 1 5 . . .e6 is another option ; the
note to 15 i..f4 in the game). idea of .. .f7xe6, d6-d5 and play in the
centre is interestin g from a practical
perspective, but I think the position
slightly favours White here: 16 dxe6
fxe6 17 i.. g s d5 18 i.. xf6 liJxf6 19 exd5
exd5 20 l:tdl 'it>h 8 2 1 liJa4 liJe4 22 liJd2
l:txa4 ! ? 23 'ii'x a4 'ii'x b2 24 liJxe4?? (a
blunder; 24 :'xe4! dxe4 25 liJxe4 i.. d4
2 6 l:td2 'iVb7 27 f3 would still be a little
better for White) 24 ... 'ii'x e2 2 5 l:Ixd5 c4
0-1 D.Fuhrm an n - H . Ellers, German
league 1999, since White cannot de
Now White has tried: fen d both e4 and f2.
a} 15 11e2 ? ! and then : b} 1 5 b 3 ! e6 16 dxe6 fxe6 and n ow:
al} 1 5 ... 'it'b4! I like this move.
-
bl} 17 i..b 2 d5 with a further split:

The idea is ... l:tfb8 and ... liJe8-(c7- bll} 18 exd5 ? ! exds (now the posi
b5}, while, as we have already seen a tion is unclear, though White must be
few times, ... i.. x c3 is al so a possibility careful not to g et in trouble since Black
with an unclear g ame; for example, 16 has a good centre and active play) 19
i..d 2 (or 16 i.. g 5 l:[b8 17 J:tdl h 6 18 i.. c l 'it>g l ? ! (tactics en sure that this neither
liJe8 with compensation) 16 .. JUa8 17 win s nor loses m aterial, but it cedes the
.l:!.dl liJe8 18 i..f4 i.. x c3 19 bxc3 'ii C4 20 initiative) 19 ... d4 20 liJdl 'it'xf3 2 1 'it'C4+
'ii'd 3 'it'xd3 2 1 llxd3 f6 2 2 litdd2 g5 23 (White's point) 22 ... 'iid 5 22 'iix a6 liJe5
.i.e3 liJb6 24 ':c2 liJC4 25 i.. c l liJC7 26 (the counter-point: Black win s the ex
liJd2 liJe5 27 liJb3 l::t a4 Yz-Yz M.Marin chan g e back, and there is the threat of
N.Nedobora, Saragossa 1994. ... liJe4 among st others) 2 3 l:[xe5 'ii'x e5
51
24 tiJe3 tiJe4 2 5 f4 'ii e 7 2 6 tiJfl tiJC3 ! 2 7 Black has excellent compensation ac
'ii'c 6 d 3 2 8 .1i.xC3 .1i.xC3 2 9 1:[cl .1i.d4+ 3 0 cording to West; for example, 22 ':'acl
�h l 'iVe2, when Black h ad a strong ini (or 22 .1i.xd4 ! ? cxd4 23 ':'acl tiJcS 24
tiative for the pawn and went on to l:tedl .1i.h6 25 l:[xcS 'iVxcs 26 'iVxe6+
win in J . Dorn er-G.West, correspon �h 8 with roughly equal chances)
dence 1997. 22 ... tiJxes 23 ':xe s .1i.xes 24 l:!txcS "ili'xcs
b12) 18 tiJd2 (West g ave thi s an ex 2 5 'i!Vxe6+ �g 7 26 tiJxe s ':'xf2+! 27 �xf2
clamation m ark without further analy "ili'c2+ with a draw.
sis) b2) 17 tiJg S ! ? dS is very unclear too.
18 ... C4! (the threat is n ow ... dS-d4 Black has a strong centre, while the
and c4-c3) 19 bXc4 tiJxe4 (this is the king on g 2 is in some danger. Therefore
point; the bishop on b2 is h anging in White must play with some precision :
some lines) 20 tiJdxe4 dxe4 2 1 l:[e2 e3+ 18 .l::r.b l ! (West shows that other moves
( 2 1 ... tiJcS is also possible) 2 2 'ii e 4 11c6 are good for Black; e.g. 18 e s ? ! d4+ 19
with an unclear g ame, the threat is f3 tiJds or 18 f4? ! h6 19 eS hxg s 20 exf6
n ow ... tiJcS . ':'xf6 with a clear advantage) 19 ... d4 19
b 13 ) 18 e S tiJh s 19 tiJa4 (or 19 a 4 d4 tiJa4 'ii'b s 20 tiJb2 (or 20 .1i.a3 ':'c8 21 f4
20 tiJe4 'iid s with compensation : Black e s ) 20 ... 1txa2 2 1 tiJxe6 ':'b8 22 tiJxg 7
h as a strong centre and pressure on the �xg 7 2 3 eS tiJds 24 e6 tiJ7f6 2 5 .1i.g s
light squares and ag ain st the e s-pawn) tiJC3 26 J:[al J:[xal 2 7 1:[xal 'iWc6+ and
19 ... .:xa4 ! ? (19 .. JUS is al so possible; e.g. Black regains the pawn with an unclear
20 .l::r. a cl c4 2 1 .1i.d4 na8 intending g ame .
... .:.af8, or 20 tiJxcs tiJxcs 2 1 tiJd4 .1i.f8 15 .1i.f4?1
2 2 tiJxfs exfs with an unclear g ame, The critical test of Black's play is 15
and not 21 'iix cS ? ? because of 2 1 ... d4 dxe6! fxe6 16 tiJbs, when White looks to
winning the knight on f3 and the have a slight advantage, - although the
g ame) 20 bxa4 d4 21 'iib 3 'iic 6 and game remains very complicated:
52
16 ... 'iVb7! (if 16 ... "iYc6?! 17 a4 dS 18 exds 2 2 "iYe3) 2 2 ttJxdS ttJxdS (or 2 2 . . .l:txdS 2 3
ttJxdS, as in Y.Yakovich-P.Cramling, 'ike1) 2 3 l:td7, when the reduced m ate
Stockholm 1999, then 19 'i'e4 is good rial gives him g ood hopes of a draw.
for White) 17 a4 dS 18 exds exds (again 20 ... ttJfxd S 21 l::t e 2? ttJxc3
18 ... ttJxdS ? ! 19 "iYe4 gives White a com
fortable advantage) 19 ttJg s d4+, and
now White should play 20 <itg l ! ttJdS 2 1
�d2 with the better game. Instead,
M.Medic-D.5ermek, Pula 1999, contin
ued 20 f3 ?! ttJdS 2 1 l::t a 3?! (21 "i¥b3 ttJes
22 ttJxd4! was correct, when 2 2 ...'i'a8 2 3
ttJbS c4 2 4 'i'C2 remains very unclear)
21 ... ttJes 22 ttJe4?! (and here 2 2 �d2
improves, although White's position is
more difficult to play) 2 2 ... d3 2 3 l:lxd3
ttJb4 24 "i¥b3+ c4 2S 'ikxb4 cxd3 and And n ow Black is just winning.
Black was somewhat better, while 2 2 bXc3 lid S 0 - 1
22 ... c4! 23 ttJcS d3 24 ttJxb7 dxc2 might
have been even stronger. Game summary
ls ... exd s 16 exd s :c8 17 l::t a dl c4 The idea of ... l:!a6 and ... "iYa8 followed
by ... e7-e6 i s tempting, even if it is ob
jectively inferior to ... 'i'b6, ... .l:.fb8 and
... "iYa6 (as seen in Piket-Topalov). H ow
ever, I h ave included thi s (and the n ext)
g ame to illustrate some of the ideas
and dynamic options for Black in the
Classical m ain line. Indeed, for
Nisipeanu it all went swimmingly; but
it shoul d be n oted that, with correct
play, White can probably achieve a
slight advantage.
Now Bl ack can target the ds-pawn
and the position is unclear.
18 l:te7 ttJb6 19 'iVd2 l:ta s 20 �h6? Game 1 3
This loses material . White should L.Comas Fabrego-V.Topalov
have tried 20 b4! ? cxb3 (20 ... ttJfxdS 2 1 Pa m plona 1994
bxas .txc3 2 2 'i'e2 is unclear) 2 1 axb3
ttJfxdS (if 2 1 ... .:tacs 2 2 ttJa4, or 2 1 ... ttJh S 1 d4 ttJf6 2 c4 cS 3 ds bs 4 cxbs a6 5
53
bxa6 g 6 6 ltJC3 i.. x a6 7 e 4 i.. xf1 8 �xf1 llxd3 ltJds 27 l1adl i.. d4 with an un
d6 9 g3 i.. g 7 10 �g2 ltJbd7 11 ltJf3 0 - 0 clear position.
12 h 3 l:ta6 13 l:le1 'ita8 14 l:.te2?1 18 f4
A thematic m ove in the Classical : Quite ambitious, but also weaken-

the rook m akes room for the queen to ing.
double up behind it on el, while cover 18 c41
.••
ing the second rank and enabling the This move is especially effective
cl-bishop to develop g s or f4 if desired. when ... ltJcs-d3 is a possibility. The
In this precise position, however, Benko player should always be aware
White's plan is a bit slow, and I would of this theme, and it can be found quite
recommend the previous g ame's 14 a few times in this book. Nevertheless,
'iWc2 as a m ore serious test of Black's other moves were good too, such as
set-up. 18 ... h6 19 ltJf3 c4, or 18 ... dS, now th at
14 ... l:tb8 the white king has been further ex
The immedi ate 14 ... e6 is al so play posed by f2-f4.
able. With the text Black m aintains all
his options, such as ... ltJb6-a4, ... ltJe8-
c7-cS or ... .:.b4, depending on what
White does next.
15 'iVe1 e6
With the white queen h avin g left
the d-file thi s move is even stronger.
16 dxe6 fxe6 17 ltJg5 lite8
17 ... dS ! ? is also interesting; for in
stance, 18 exds exds 19 i..f4 d4+ 20
�h 2 :bb6 21 i.. C 7 .l:r.b4 2 2 ltJce4 ltJxe4
2 3 %:txe4 ltJf6 24 ':'e2 d3 2 5 ':'e3 ':'xb2 2 6 19 i.. e 3
54
Preventing the black knight ma (not 30 <it>xf3 ? 'ii'xh 3+ and m ates)
noeuvre. 30 ... <it>h 8 31 'ii'h 4 'ili'd5 ! and Black h as
19 e5 20 litdl116 21 liJf3 exf4?!
... the strong threats of ... l:tf2+ or ... litxh 3+,
Correct was 2 1...liJxe4 2 2 liJxe4 whereas White's attack is stymied.
'iVxe4 2 3 �C5 'i!Vb7 24 �xd6 e4 25 liJe5 28 ...liJxd6 29 "C3 'ili'f6 30 "xf6 liJxf6
ttJxe 5 26 �xe5 (or 26 fxe 5 litxa2)
26 ... ..txe 5 27 fxe5 .l:Ixa2 with a level po
sition .
22 gxf4?
Returning the favour. 2 2 �xf4
would have been good for White; e.g.
22 ...liJC5 2 3 l:td4 liJh 5 24 litxC4 liJd3 25
'ii'd 2 liJdxf4+ 26 gxf4 �xC3 27 bXc3 d5
28 l:.d4 dxe4 29 11exe4.
22 ... liJxe4 23 liJd 5?
Of course not 2 3 liJxe4 'iVxe4 24
�C5?? "xe2+ (this was the point 3 1 liJe5?
21...exf4), but 23 �d4 would only fa 3 1 l:tb6 would offer more chances of
vour Black slightly. holding .
23 ..... b7 24 �d4 "xd 5 3 1 ... liJfe4 3 2 h4 c3 3 3 liIc2 liJf5
The twin threats of ... liJxh6 and
... liJe3+ are decisive.
34 <it>f3 liJd2+ 35 <it>f2
3 5 <it>e2 is m et by 3 5 ... liJC4! 3 6 �g 5
liJxe5 3 7 fxe 5 liJd4+ 3 8 <it>d3 liJxc2 3 9
<it>xc2 J:txe5 with a winning endgame.
3 5 liJxh6 36 �e2 liJc4 0 - 1
.•.
Game summary
The plan with 14 l:le2 and 'ili'e1 is not
effective here. Topalov g ets an unclear
Now Black is clearly better, due to position and win s a nice (if imperfect)
the strong knight on e4. g ame, illustratin g m any of the themes
25 ..txg7 'ii'f 5?! for Black: such as the ... e7-e6 and ... c5-
25 ... 'iib 7! was more accurate. c4 breaks and play again st the exposed
26 �xh6 l:lxa2 27 ':xd6 l:txb2 28 l:txb2 white king on g 2 . N evertheless, I would
Or 28 l:1dd2 l:tb3 with the initiative; only recommend ... l:ta6 and .. :�a8 for
for example, 29 ':xd7 ':'xf3 ! 30 1:[g7+ the Benko player who is desperate to
55
avoid a draw. The theoretical status After <t>gl-h 2, White's ideas are
n owadays is m ore in White's favour much the same as before: developing
than when this game was pl ayed. the king's rook to el; reinforcing the
second rank and the queen 's knight
with l:te2 and 1:[c2 or 'iic 2; the cl
Game 14 bishop is developed to g s, f4, d2, or
A.Blees-A.Greenfeld sometimes on b2 after b2-b3. Later in
Te l Aviv 1988 the middlegame White might pl ay e4-
eS anyway, while the familiar plan of
1 d4 lDf6 2 c4 cS 3 d s bs 4 cxbs a6 S blockading the queen side with a2-a4
bxa6 g6 6 lDC3 �xa6 7 e4 �xfl 8 'itxfl and lDbS remain s an option .
d6 g lDf3 �g7 10 h 3 10 lDbd7
•••
My recommendation is to pl ay the
cl assical set-up with ... lDbd7, .. 0-0,
.
.. "as and ... l:tfbB, after which when

.
Black's various ideas include ... l:tb4,

... lIabB, ... lDb6-c4/a4, and sometimes
... c S-c4 and ... lDcS-d3.
However, 10 ...0-0 11 'itgl 'irb6 12
'ith 2 lDa6 is an interesting altern ative.
This is playable and can be quite dan
gerous, so I will show a few ideas with
this line; but as White can probably
In the first s i x g ames o f this chapter achieve a small advantage, I think the
White con solidated his king with 10 g 3 game continuation of ... lDbd7 should
and 'itg 2. With the text White prepares be preferred.
a different bolthole for his m ajesty.
While the m anoeuvre h 2-h 3 and 'itg l
h 2 takes one move lon g er, on this
square the king is a bit m ore secure;
and as we have seen, White often pl ays
h 2-h 3 in any case in order to prevent
... lDg4, so the tempo is n ot wasted. Fur
thermore, Bl ack's plan with ... e7-e6 is
less effective without the white king on
the long diagon al . On the other han d,
the thematic e4-e S is n ow more diffi
cult for White to achieve. From this position White has two
56
main moves: 2004 - a good illustration o f how

a) 1 3 J:tel ttJb4 (a direct approach, things can go wrong for White after a
Black is pl anning to place the knight on few inaccuracies.
d3 after ... 'ii'a 6 - it works well here) 14 b) 1 3 'ike2 ! ttJb4 14 a4 e6! (the the
:e2 ? ! (a them atic move in the iLxfl m atic 14 ... 'iVa6 ? ! i s dubious here be
lines, but ineffective here; then again, cause of 15 ttJb5 ! .l:r.fb 8 ? ! 16 I1a3 and
after 14 Ve2 'iVa6 ! the endgame i s fine White is better - Black should be care
for Black) 14... 'iVa6 1 5 cJilg l ttJd7 16 iLg 5 ful to avoid a blockade with a2-a4 and
lUe8 17 ttJel ttJe 5 18 a4 h 6 19 iLh4?! ttJb5, especially after the light-squared
(preferable was 19 iLd2, though Black bishops h ave been traded; 15 . . . ttJd7 16
still has good compensation) 19 ... l:teb8 iLg 5 is also good for White) 15 lldl
(now Black has a slight advantage) 20 exd5 16 exd5 lUe8 17 'iVC4 'iVa6 18
l:ta3 (20 iLxe7 run s into 20 ... g5 and Vxa6 :txa6 19 %:tbl ttJe4 20 ttJxe4 ':'xe4
21 ... ttJg6, picking up the trapped 21 b3 with a slight advantage to White.
bishop) 20 ... :b7 (or 20 ... ttJc4!) 21 b3 11 cJilg1
%:tab8 22 :e3 c4 It is n ot too l ate for White to revert
to the m ain lines with 1 1 g 3 and 1 2
cJil g 2 , should he wish t o d o s o . Indeed,
we saw thi s m ove order used in G am e
11.
11 0 - 0 12 cJilh2 Va s 13 l:[ e 1 l:tfb8
..•
23 'iVe2 ? (more resilient was 2 3

ttJbl, when 2 3 . . . cxb3 2 4 l:taxb3 ttJC4 i s
only clearly better for Bl ack, whereas
after the text Black is winning)
2 3 ... ttJed3 (the black pieces are much
more active than their white counter This position can be compared with
parts) 24 'iVd2 ttJxel 25 :xel ttJd3 2 6 the m ain line 9 g 3 iLg 7 10 cJilg 2 0-0 1 1
%:tdl cxb3 (or 26 ... .l:1xb 3 ! 2 7 l:txb3 ':'xb3 ttJf3 ttJbd7 12 h 3 'iVb6 1 3 : e l J:tfb8,
28 ttJe2 l:tb2 29 'ife3 'iVxa4 with a deci seen in Games 10 and 11. The only
sive advantage) 27 ttJb5 ttJC5 28 'iVb4 difference is that the black queen is on
%:ta8 0-1 V.Jianu-M.Ghaem, I stanbul as rather th an b6, while White h as
57
played �g l-h 2 in stead of g 2-g 3 and

�g 2 .
14 lie2
As played by Piket in the similar po
sition from Game 10. The m ain alterna
tive, 14 'ifc2, is examined in the next
g ame. But n ote that 14 e s ? ! (as in
Game 11) would be a mistake here due
to 14 ... dxe s l S ttJxe s ? ? ttJxe s 16 lIxe S
ttJg4+ ! , winning m aterial .
14 ...ttJe8
14 ...�b4 is another option, prevent 16 'ife2
ing lS :C2 (as the e-pawn would be Black is okay after the alternatives
h anging) and leading to an unclear as well.
game. For example: l S 'ifel (after l S a) 16 i.g S ? ! leaves White weak on
i.g s the logical follow-up is ls . . .llab8, b2 and c3, so th at 16 ... ttJC4 is then good
creatin g pressure on the b-file, and if 16 for Black.
b3 then 16 ...'ifa6, intendin g ... cS-c4 with b) 16 a4? ! is met by 16 ... 'i!Vb4 and
compensation) ls ... ttJe8 (giving the g7- Black regains the pawn with a slight
bishop more scope, and preparing advantage .
...ttJc7-bS) 16 ttJdl (notice that the oth c) 16 l:tbl ttJc4 leads to an unclear
erwise thematic 16 i.g S ? ! is dubious g ame; e.g. 17 'iVe2 ttJe s 18 a3 (or 18
here due to 16 . .J:lxb 2 ! 17 l:txb2 'ifxc3 ttJel ttJd7) 18 ... ttJxf3+ 19 Wxf3 l:1b3 with
and Black regains the exchange with the compensation, since the move a3 a2-
better position) 16 ...'iVa7 17 i.d2 i.xb2 h as created a substantial weakness on
(17 .. Jia4?! 18 a3 was good for White in b3.
C.Garcia Palerrno-J . Nogueiras Santiago, 1 6 .. .'iVa6 17 'ifxa6 .l:txa6
Cienfuegos 1984) 18 ttJxb2 :lxb2 19 a4
'ifa6 20 as ttJb6 21 'ifdl ttJf6 22 �g l
ttJC4 with compensation.
15 liC2 ttJb61
lS ... ttJC7 ? ! , preparing ... ttJbS , i s n ow
a little slow and allows White to g ain
an edge after 16 .td2 'ifa6 17 �acl
i.xc3 (somewhat risky, but other
moves are good for White too) 18 .txC3
ttJbS 19 'ilfd2 'ifxa2 20 b4 'ifa7 21 i.al
and White is slightly better.
58
This endgame is quite thematic for 2 2 l2Jxe5 2 3 .l:!.el l2Jxf3+ 24 gxf3 f5?!
..•
the Classical main line. Thanks to the Giving away most of the advantage.
activity of his pieces, Black is perfectly Instead, after 24 ... .ieS + 25 Wg 2 ':xa2
fine, and often has greater winning or 25 l:[xe s dxes 26 d6 l2Jg 7 27 d7 l2Je6
chances in practical play. 2 S dS'iV+ l2JxdS 2 9 .ixdS l:txa2 Black h as
18 .ig5 l2JC4 19 .ixe7?! good winning chances with his strong
Safer was 19 ...l2Jxb2 20 %lb1 (but not passed c-pawn .
20 l2Jxb2 .ixb2 2 1 l:tb1 ? .ies+) 2 5 l:te6 l:txa2 26 'iStg2 .id4 2 7 .ih4 cM7
20 ... .l:Iab6 2 1 l2Jxb2 .ixb2 2 2 .ixe7 :6b7 28 f4 l:ta6 29 1;:[e7+ 'iStf8
(now if 2 2 ... .ieS+, then 23 l2Jxes l:lxb1
24 l2Jc4) 23 a4 l:lxe7 24 1kxb2 lIxb2 2 5
.u.xb2 lIxe4 and a draw is the m ost
likely outcome; e.g. 26 as l:.a4 27 l:tb7
':'xas 2S l2Jg s f6 29 l2Jf7 l:ta2 30 l2Jh 6+
with perpetual check.
19....l:!.xb2 20 l:txb2 l2Jxb2 21 e5?
Correct was 2 1 l2JbS l2Jd3 2 2 J:.b1 f6
23 l2Jd2 cM7 24 .idS .l:txa2 2 5 l2JC4 .ifS
26 f3 :a4, when the g ame is in the bal
ance, even if Black has some pressure.
2 1...l2Jd3 30 1:tb7?!
After 30 :xeS+ WxeS 3 1 l2JC7+ 'iStd7
32 l2Jxa6 'iStcS 33 .ie7 Wb7 34 .ixd6
'iStxa6 35 .ies , the g am e would proba
bly be drawn .
30 .if6 3 1 .ixf6?
.•.
A definite mistake, after which

Black i s clearly better again, whereas
31 .ig s would keep his advantage to a
minimum. The rest of the g ame was
probably played in time trouble, as it
was n ot conducted with any great ac
Now Bl ack is clearly better. The a2- curacy, but Black triumphed in the end.
and eS-pawns, as well as the pieces on 3 1 l2Jxf6 3 2 l2JC7 l:tu 33 l2Jb5 l2Je4 34
..•
a1 and c3, are all vulnerable. l:txh7 l2Jxf2 3 5 Wf3 l2Je4 36 l2JC7 'iStg8 3 7
22 l2Jb5?! l:te7 l:ta3+ 3 8 'iStg2 l:. a 7 3 9 'iStf3 c 4 40
22 l:.b1 ! l2Jxe s 2 3 l:tbS l2Jxf3+ 24 gxf3 We3 c3 41 Wd3 .:tb7 42 l:1e8+ Wf7 43
i.xc3 2 5 'iStg3 was a stronger defence. .l:!.c8 l:tbl 0 - 1
59
Game summary would be thematic, but then White can

After White pl ayed the prophylactic gain an advantage with 15 ttJdl, pre
rook moves J:[el-e2-c2, covering c3, paring .itd2. For example, lS ... C4! 7 16
Greenfeld equalized with the them atic a4! (16 .ite3 ttJfd7 17 l:[e2 ttJc S 18 ttJd2
... 'Was , ... l1fb8, ...ttJb6 and ... 'Wa6. Black ttJb3 19 ttJxb3 cxb3 20 'Wxb3 ttJa4 was
even got some pressure in the end unclear in P. H aba-M.Prusikin, German
g am e and won after mistakes by both League 2003) 16 ... ttJxa4 (or 16 ... ttJfd7 17
sides. The knight going to b6 and c4 .itd2 intending a4-as) 17 ttJC3 ltb4 18
was quite thematic. So ag ain th e con .itd2 with an annoying pin on the a-file
clusion is that Black is doing fine. and a clear advantage for White.
15 .itd2 ttJC7 16 ttJdl
If 16 a4 'ifb6 17 l:[ebl, then 17 .. :iVb3
Game 15 18 'Wxb3 .l:.xb3 19 as .l:.ab8 20 ttJdl ttJe s
C.Gabriel-D.Sermek 2 1 ttJxe s .itxes+ 2 2 c;t>g l .itxb2 2 3 ttJxb2
Pu la 2000 l:[xb2 24 l:[xb2 l:[xb2 and the endgame
is level .
1 d4 ttJf6 2 c4 cS 3 d5 b5 4 cxb5 a6 5 16 .. :iia 4
bxa6 g6 6 ttJC3 .itxa6 7 e4 .itxfl S c;t>xfl
d6 9 ttJf3 .itg7 10 h 3 0 - 0 11 c;t>gl ttJbd7
12 'ifolh2 'ii'a 5 13 .l:[el l:.fbS 14 'ii'C 2
Black has sufficient compensation .

As usual the endgame is perfectly ac
ceptable for Black with his active
The second of White's two usual pieces, especially the rooks on the a
choices here. The alternative, 14 l:[e2 and b-files supported by the bishop on
intending ':'C2, was covered in the pre g7.
vious g ame. 17 'WCl
14 ... ttJeS! Or 17 'Wxa4 l:[xa4 18 .itC3 ttJb S 19
The m ost accurate m ove here. In .itxg 7 c;t>xg 7 20 e S ttJd4 and Bl ack is do
stead, 14 ... ttJb6, headin g for c4 or a4, ing fine.
60
17 ... tt:JbS 18 tt:Jc3 'iVa6 19 a4 has a dangerous initiative) 2 6 tt:Jxe3

19 tt:JxbS 'iVxb s 20 �C3 �xC3 2 1 l:.lb2 with a drawish position .
bxc3 'iVb2 2 2 - 'ifxb2 l:Ixb2 shows an 24 c 4 2 S .l:r.a2
•..
other drawback to ..to>g l-h 2 : the f2- Here 2S tt:Jd2 ! tt:JcS 26 'ii'e 2 would be
pawn is en prise. slightly better for White due to the
19 ...tt:JC7 weak c4-pawn .
Black is ready to double the rooks on 2 S I1bl 26 l:r.d2 'ii'x e3 2 7 tt:Jxe3 .l:r.xel
.••
the b-file and, by retreatin g his knight 28 tt:Jxel .l:r.a8 Vz-Vz

to c7 again, prevents White from set
ting up a blockade on b S . More ri sky is Game summary
19 ... tt:Jd4 20 tt:Jxd4 �xd4 2 1 tt:JbS �xf2 With 14 1WC2, followed by �d2 and
22 l:tf1 �d4 23 tt:Jxd4 cxd4 24 b4, when tt:Jd1, White aimed to secure his queen
White might hope to make something side ag ain st opposing counterplay.
of his passed pawn s. However, after Black's them atic and
20 l:ta3 l:tb6 21 tt:Jdl l:tab8 22 �C3 �xC3 accurate ... tt:Je8-c7, White found it diffi
23 bxc3 cult to undertake anything active ei
ther. In the g am e Black equalized.
Game 1 6
Y.Sh u l man-A.Kha lifm a n
FIDE World C u p (blitz playoff).,
Kha nty Ma nsiysk 2005
1 d4 tt:Jf6 2 c4 cS 3 d s bs 4 cxb5 a6 S
bxa6 g6 6 tt:JC3 �xa6 7 e4 �xfl 8 ..to>xfl
d6 9 tt:Jge2
23 ... 'ifd 3?!
With his knight no longer required
on the queenside, Black might prefer
23 ... tt:Je8 and ... tt:Jef6, when the game
remains unclear.
24 'iVe3
If 24 c4, then 24 ... l:.b1 2 S 'ilt'e3 'ii'x e3
(2S ... 'ii'C 2?! is quite risky; e.g. 2 6 tt:JC3
l:.xe1 27 tt:Jxe1 'iVb2 28 tt:JbS tt:Jxb S 29
axb s tt:Jb6 30 1;k3 J:.a8 31 l:tc2 'ili'b1 3 2
es l:ta4 3 3 exd6 exd6 34 .l:te2 and White
61
A solid move; the knight supports its 2 5 �h l �d4 26 l:tffl lDe2 0-1 R.Gross
colleague on c3 and therefore hinders P.Benko, US Open, Aspen 1965.
Black's counterplay somewhat. On the b) 1 3 a3 ! is more interesting. This
other hand, the advance e4-e5 is now move restricts the knight on a6, since if
not really a theme. The classical plan of it continues to c7 White can play b2-b4
... lDbd7, .. :iVa5, .. .l::tfb S , followed by showing the black queen to be less for
... lDb6-a4 (as played in the g ame) gives tunately placed on b6, while 13 ... c4 al
Black enough compensation for the lows 14 �e3 lDc5 1 5 lDd4 and lDc6.
pawn. Although Shulman has scored Therefore Black return s to the original
quite well with 9 lDge2, in my opinion knight m anoeuvre seen in Games S
the main lines with 9 lDf3 offer White and 9 of this chapter: 1 3 ... lDg4 14 h 3
more chances to play for an advantage. lD e 5 1 5 f4 (or 1 5 Wc2 lUcS with com
9 �g7 10 h 3
.•• pensation) 1 5 ... lDc4 16 'fia4 lDb4 17
1 0 g 3 is less effective with the Wh3 Wa6 lS axb4 cxb4 19 lDdl lDd2
knight on e2, as the white king is then (19 ... lDa3 ! ? m ay be even better) 20
even m ore exposed on g 2 . Benko him �xd2 'iVxe2+ 21 lDf2 'iVxd2 22 l:thdl
self won a nice g am e as Black in this We2 23 'iix b4 litfbS and Black has com
line: 10 ...0-0 1 1 �g 2 'iVb6 12 :bl lDa6 pen sation for the pawn .
(intending ... lDC7 and ... e7-e6, attacking
the centre; in the case of d5xe6 Black
can play ... d6-d5, which is not without
venom ) :
10 0-0 11 �gl lDbd7 12 �h2 Wa s

..•
Probably the most common set-up

for Black in the Benko. As in the previ
ous two g ames (where White played
a ) 1 3 b 3 ? ! lD C 7 14 f 3 e6 1 5 dxe6 fxe6 �g l-h 2 in conjunction with lDf3), the
16 �e3 'iVc6 17 'iVd2 ? ! d5! (Black is al plan of .. Jita6 and ... WaS is less effective
ready clearly better) lS exd5 exd5 19 with the white king off the long light
�f4 d4 20 lDa4 lDfd5 21 lthfl lDe6 22 squared diagonal .
.l:r.f2 d3 2 3 lDg l lDexf4+ 24 gxf4 lDxf4+ 13 1i'c2 lDb6
62
the immediate 16 . . .ttJd7 was m ore ac

curate, so as to create the possibility of
... i.xc3 at once.
17 Iihel?1
H ere White could h ave played 17 b3,
and if 17 ...ttJd7 then 18 a4 and 19 ttJb5
with a solid advantage.
17 ...ttJd7 18 i.g5 'iitf8 19 'iii'd 2 i.xC3
This idea occurs quite a few times in
this book. Here it is well timed: Black
gets the a2-pawn, and the king is safe
On its way to a4, seekin g to under enough on f8.
mine White's solid defence of c3. 20 bXc3 ':'xb1 2 1 l:txb1 'ili'xa2 2 2 llb2
14 l:tbl 'iVC4
Deviating from the two players' N ow the position is unclear.
rapid playoff game: 14 l:td1 ttJa4 1 5 23 l::t b 7 'iit e 8 24 'it'f4
i.d2 ttJxc3 16 ttJxc3 l:tfb8 17 l:t e l ttJd7 There was n o need to give up the c
18 b3 'ita6 19 l:tabl ttJe 5 20 lle3 c4 2 1 pawn, but Black was okay in any case.
'iitg l, which Shulman i n fact went o n to For in stance, 24 'ifb2 (planning l:txd7) i s
win, although Khalifm an would n atu easily met b y 24. . .'iWa6 and 2 5 . . ..:.a7.
rally have had an improvement ready 24 ...'i!t'xC3
somewhere.
14 ... ttJa4 15 i.d2 ttJXC3 16 ttJxC3 l:tfb8
25 'iWh4??
25 i.xe7 'iit x e7 26 .l:txd7+ 'iit x d7 2 7
After the exchange on c3 Black is 'iVxf7+ 'iit c 8 2 8 'iVe8+ 'iitb 7 2 9 'iVc6+ with
ready to send his other knight round to a perpetual was called for.
the queenside via ... ttJd7-e5 (or b6) and 25 h5
•..
... ttJd3 or sometimes ... ttJC4. However, I t seems Shulman didn't see this
63
possibility, which I suppose is forgiv bxa6 g6 6 ttJc3 .txa6 7 e4 iLxf1 8 �xf1

able in a blitz game. But Black is n ow d6 9 g4
just a pawn up with the initiative,
whereas the white pieces are n ot play
ing together. The pawn on e7 can n o
l on g er be captured because o f 2 6 iLxe7
'it'eS +, winning a piece.
26 iLe3?1
This and White's n ext m ove only
m ake things worse.
26 ...'it'es+ 27 g3 l:.a4 28 llC7 �d8 29
l:tc6 :xe4 30 .td2 f6 31 .tf4 'iVxd s 3 2
l:ta6 l:te1 0-1
To his credit Shulman recovered In stead of the standard position al
from thi s defeat to win the second blitz g 2-g 3, White uses the g-pawn far more
g am e (as Black), and then progress to aggressively. This idea was semi
the n ext round in sudden death . popularized by Yasser Seirawan, who
won some nice games with it in the
Game summary 1980s (he later switched to S b6). Al
Shulman h as played 9 ttJge2 numerous though 9 g4 is a rare continuation
times with good results, though it is nowadays, it can still be dangerous for
otherwise rel atively uncommon. For the unprepared Benko player. One of
Black the plan with ... 'iIi'as and ... ttJb6- the m ain ideas is to continue with g4-
a4 i s quite effective and should guar g s, attacking the knight on f6, and then
antee sufficient compen sation for h 2-h4-h S in order to attack the king
equal chances. In the g ame Benko ex after Black has castled kingside. Typi
pert Kh alifm an played the thematic cally White will play 'it>g2, ttJe2-g 3 and
... iLxc3, regaining the pawn satisfacto aim for a m ate on the h-file.
rily. Then, probably due to the fast time The best replies, in my opinion, are
limit, Shulman blundered and Kh alif 9 ... 'iYc8 and 9 ... h S ! , when Black
m an won in a rather simple way. shouldn 't have any problems achieving
equality. After 9 ...'iVc8 Black has good
potential counterplay on the queenside
Game 1 7 too. However, I would recommend
I.Sa kovich-Z.La n ka 9 ... h S , because the white king on f1 can
Riga 1980 then be in some danger in the middle
game.
1 d4 ttJf6 2 c4 cS 3 ds bs 4 cxbs a6 S 9 ... h S I
64
In the spirit of the Benko Gambit, 1 2 . . .0-0 1 3 �f4 tiJa6 1 4 'iVe2 (or 1 4 'iVd2
Black simply attacks the pawn . After c4! intending ... tiJacs with an unclear
g4-g s, there is no h4-h s anymore, g ame) 14 .. :ii'as 1 5 :hCl :fb8 (the clas
whereas Black will have chances for sical way of playin g the Benko) 16 l:tabl
good play on the queen side. tiJCl (heading for b s ) 17 e s e 6 ! (a cool
10 g5 reply) 18 a3 (or 18 dxe6 tiJxe6 19 �g 3
10 gxh s ? ! is not the most logical tiJxe s) 18 ... dxe s 19 tiJxes tiJxes 20
move: opening the file for the black �xe s �xes 2 1 'ii'x es C4 2 2 tiJe4 (after
rook on h8 and creatin g weaknesses on 22 l:tdl l:td8 23 tiJe4 tiJe8 Black i s ready
the king side seems a little unwise with for ... l:txds, so White shoul d probably
the white king sittin g on fl, so Black opt for 24 tiJC3 tiJC7 2 5 tiJe4 with a
has good chances to gain an advantage repetition) 2 2 ... 'ii'x ds 2 3 'ii'f4 tiJe8 and
in practice. Nevertheless, it seem s the g am e is roughly equal .
White can keep the balance with pre 12 tiJa6
..•
cise pl ay: 10 ... tiJxh s 11 �g 2 �g7 12 On the way to Cl and bs, or some
tiJf3 ! (after tiJge2 tiJd7 Bl ack i s slightly times d3 after ... cs-c4 and ... tiJcs .
better, since the king on g2 is none too 13 tiJf3 0 - 0 1 4 'ii'e 2 'ii'C 7
safe) 12 ... tiJd7 13 a4 tiJes 14 h3 and Preventing e4-es at the cost of a
Black has sufficient compensation for tempo. Black might al so consider the
the pawn : there are fair prospects for direct 14 ... 'iVas ! ? 1 5 e s tiJb4 which
an active game, whereas White h as leads to a sharp, unclear g ame, though
several weaknesses. it still requires some testing in practice:
10... tiJfd7 11 �g2 �g7
16 :dl ! (the most dangerous m ove;

12 f4 16 e6 tiJb6 17 exf7+ l:txf7 18 'iVe6 forces
Consistent, but al so risky. 12 tiJf3 is Black to defend accurately, but
a quieter alternative, though Bl ack i s 18 ... 'ilia6 ! 19 fs 'ii'd 3 20 tiJh4 'iWc2+ 2 1
doing alright here too. For example: �h 3 'ii'd 3+ draws by perpetual check)
65
16 . . .c4 17 e6 ttJC5 ! ( 1 7. . .ttJb6 ? 18 f 5 gives 18 a3?! ttJa2 !

White a dangerous initiative; e.g.
18 ... g xf5 19 exf7+ .l:txf7 20 g 6 ':'f6 21
ttJh4 ttJ6xd5 2 2 ttJxd5 ttJxd5 2 3 'ikxh 5 e6
24 'ilVh 7+ �f8 25 �h 6 �xh 6 2 6 'iix h 6+
�e8 27 ttJf3 and White is clearly better)
18 exf7+ :xf7 19 'ikxC4 ttJC2 20 .:tbl
�xC3 21 bxc3 Wxa2 22 'ii x a2 1:[xa2 2 3
� 1 ttJe4 24 l:td3 ttJ C 5 2 5 .:tdl ttJe4 with
a repetition of moves.
1 5 �e3 .l:lfb8
A nice trick, probably overlooked by

Sakovich . White is now clearly worse
and exacerbates this with his next.
19 ttJxa2?
19 1:[c2 was n ecessary.
19 ...,Uxb2 20 :C2 ':XC2 21 "ii'x c2 "iWxa3
The point: the knight on a2 and the
bishop on e3 are hanging.
22 �2 'ikxa2
Bl ack has enough compen sation for

the pawn .
16 .ll. a dl?!
16 l:ihcl would be m ore solid.
16 ...'ii a 5
With White's bishop obstructing h i s
pl ay o n the e-file, the black queen goes
to as after all . Actually, White ought to
consider 17 e5 anyway; e.g. 17 ... dxe 5 18
�Cl exf4 19 'iVxe7 l:.b7 with equality,
though I would prefer Black over the White is lost.
board as his pieces are quite active. 2 3 :cl �b2
17 l:tcl ttJb4 23 .. .f5 ! is even more convincing.
Now Black is slightly better, primar 24 lidl ttJb6 25 ttJd2 lira3 26 :bl ttJa4
ily because White's reply does n ot 27 'iYC4 ttJC3 28 'iYxa2 :xa2 29 .l:.el �a 3
work. 30 �f3 �b4 3 1 :cl? ttJxd 5 0-1
66
Game summary For example, 7 dxe6 (7 ttJC3 ? ! ttJxdS

The variation with 9 g4 can be quite 8 ttJxdS exds 9 'ilfxds ttJc6 gives Black
dangerous if White manages to build good compensation, due to the lead in
up an initiative with h 2 -h4, ttJge2-g 3 development, the white weaknesses on
and h4-h S. However, if Bl ack responds e4, f2 and c4, and the queen on dS
with 9 ... 'iVc8 or 9 ... hS (my recommenda which can be chased) 7 ...fxe6 8 ttJC3
tion) he should equalize comfortably; c4! ? (alternatives are okay too, but this
indeed, the latter neutralizes the threat move seem s the most principled to me)
of h4-h S altogether. In the middlegame 9 e3 (9 e4? ! is risky, weakening the dark
Sakovich played inaccurately and al squares and the kin g on el; Black must
lowed Lanka a nice win . For White I be slightly better after 9 ... i.b4, and can
think g 2-g4 should probably only be follow with ... 0-0, ... ttJc6, ... 'ilb6, or pos-
used as a surpri se weapon . sibly ... d7-dS) 9 ... i.cS 10 ttJf3 0-0 11
ttJd4 �6 12 ttJa4 'iVas+ 13 i.d2 i.b4 14
ttJC3, when Black can choose between
Game 1 8 14 ... e S , 14 ... 'ilfb6, or 14 ... 'ilfc6 followed
M.Mancini-J.M.Degraeve by ... ttJc6 and ... d7-dS, with compensa-
French Tea m tion for the pawn in all cases.
Cham pionship 2004 But note that this is only possible if
Black plays an early ... ii.xa6, whereas I
1 d4 ttJf6 2 c4 cS 3 ds bs 4 cxbs a6 5 am recommending s ... g 6 as the m ost
bxa6 i.xa6 6 f4 accurate m ove order (for the explana
The usual move order is 6 ttJC3 and tion see the n otes to G am e S in the
then f2-f4. previous chapter). Then 6 ttJc3 i.xa6 7
6 ... g6 f4 reaches the position below.
Here 6 ... e6 ! ? is an interestin g at 7 ttJc3
tempt to exploit White's early f2-f4.
White's idea is to continue with
67
tiJf3, e2-e4 and J::t xfl, most likely fol tinued 9 ... 0-0 (in stead, 9 ... tiJg4+! is
lowed by e4-e S, l ooking for attacking quite an annoying check: Black must be
ch ances. This direct approach was slightly better after 10 'iit g 3 i.xc3 11
played n ow and then in the l ate 1980s, 'iit x g4 i.g7 with the exposed king now
in particular by GM J acob Murey, Now on g4, or 11 bxc3 tiJf6 I would be very
-
it is out of fashion, and for g ood rea h appy if anyone played like that
son : Black should h ave few problems if against me in the Benko) 10 h 3 (pre
he pl ays correctly, whereas the white ventin g ... tiJg4+) 10 ... e6! (opening the
king will often feel quite vulnerable; centre; a logical strategy with the white
and Black will generally have his king on f2) 11 e4? ! (11 dxe6 followed by
queenside counterplay as well. e2-e3 was less ambitious, but superior
7 :i!Va s l
•• to the move played) 11 ... exds 12 eS
Preventing the immediate e2-e4 by tiJe4+ 13 tiJxe4 dxe4 14 tiJg s c4 15
pinning the knight, and enabling Bl ack tiJxe4 dS!
to recapture with the queen after a
l ater exchange on a6. The straightfor
ward 7 ... i.g7 8 tiJf3 0-0 9 e4 gives
White some chances to prove an open
ing advantage.
16 tiJg s (no better is 16 exd6 :te8 17

tiJc3 tiJd7 18 i.e2 tiJcS, when Black's
activity gives him huge compen sation
for the pawn s; or 16 tiJf6+ i.xf6 17 exf6
'iVb6+ and Black is very much on top of
8 i.d 2 things after 18 'iit g 3 d4 19 i.e2 tiJd7 or
The m ost sensible move, preparing 18 i.e3 'ilVxb2+ 19 i.e2 (3) 16 .. .f6 17
e2-e4 again. White h as a few other op exf6 i.xf6 18 tiJf3 tiJc6 19 'iit g 3 �c S 20
tion s, but none of them should pose l:tbl ? ! (20 'iith 2 ) 20 ... i.c8 2 1 i.e3 ? (a
any problems for Black: tactical trick which doesn't work)
a) 8 tiJf3 i.g7 9 <itf2 ? ! was a daring, 2 1 .. :iWxe3 2 2 'iVxdS+ i.e6 23 'iVxc6
though unsuccessful choice by Murey i.h4+ ! 24 'iit x h4 'iif2+ 0-1 lMurey
again st Fedorowicz. Their g ame con- J . Fedorowicz, Paris 1989; it is mate in
68
four after 2 5 g 3 l:r.xf4+. Quite a nice 0-0 1 4 .i. e 3 liJC5 1 5 0-0 liJg4 and Black
game by Fedorowicz. is at least slightly better, with good
b) 8 'ii'd 2 also prepares e2-e4 but prospects of increasin g his advantage.
looks a bit artificial, while the queen 10 e4 d6 11 .i.xa6 'iVxa6 12 <;i;>f2?!
doesn 't support e2-e4-e 5 so well from Stronger was 12 'i'e2 ! liJbd7 1 3
d2, so Bl ack should equalize quite eas 'ii'x a6 l:txa6, although Black h as suffi
ily. For example, 8 ... .i.g7 9 liJf3 (allow cient compen sation for the pawn .
ing White to recapture on fl with the
rook; after 9 e4 .i.xfl 10 <;i;>xfl d6 Black
has the usual long-term compen sation
for the sacrificed pawn) 9 ... d6 10 e4
.ixf1 11 1'bfl 0-0 12 e5 liJfd7 and Black
is doing fine.
8 i..g 7 9 liJf3
...
In m any Benko endgames, as we

have seen several times already, Black's
pieces are so active that he is h ardly
ever worse, despite being a pawn
down . Indeed, from a practical view
point, I would generally prefer Black.
For instance:
9 0-0
... a) 14 a-a-a? (as in E . B ayer
Black can al so try 9 . . .d 6 and answer lFedorowicz, Porz 1988) is just a mis
e2-e4 with ... c5-c4. This is a sharper take: after 14 ... liJg4 Black threatens
option than the text and I would say is ... .i.xc3, ... liJf2 and ... l:.xa2 and is clearly
a matter of taste which one to play. better.
After 10 e4 c4 Black has excellent play b) 14 <;i;>e2 ! l:r.b8 (bringing the rook to
on the dark squares and a good lead in the open file seems the correct m ove
development for the pawn ; e.g. 11 'iVc2 here) 15 b3 (or 15 l:r.ab l liJb6 16 b4
(or 11 .i.e2 0-0 12 0-0 �6+ 13 <;i;>h l liJfd7 17 bXc5 liJxC5 18 e5 liJcd7 19
liJbd7 14 l:tbl liJg4 1 5 'ili'el liJe3 16 J:thcl l:td8 20 .i.e3 liJc4 21 exd6 exd6 2 2
.i.xe3 'i'xe3 17 'i'd2 'i'a7 with compen i.. d4 liJC5 with compen sation; one idea
sation) 11 ... liJbd7 12 .i.e2 �6 (pre is ... l:r.e8+ and ... liJd3xf4) 15 ... h 6 ! (the
venting castling for a while) 13 liJdl immediate 1 5 ... liJxe4 is quite possible,
69
but a s White's best m ove i s t o develop tage) 20 ... l:txbl+ 2 1 �xbl l:txa4 22 �C2
the h l-rook anyway, Black plays a use fs 23 tDeg s h6 24 tDe6 .i.f6, when Black
ful waitin g m ove, coverin g gs in case of is slightly better with active, well
a l ater .. .f7-fS) 16 l:thcl tDxe4 17 tDxe4 coordinated pieces and three strong
.i.xal 18 ':xal nxb3 with an unclear pawn s for the knight.
g ame. 12 Wb6
••.
c) 14 �dl? ! was played in 12 ... tDbd7 ! 13 l:tel c4 seem s more

G .Andruet-lFedorowicz, Wijk aan Zee precise. After the text White might play
1989. Now rather than Fedorowicz's 1 3 'ifh3 ! ? 'ii'a 7 14 'iWC4 tDbd7 15 .l:I.hbl
14 ... tDg4, I think 14 ... l:tb8 ! again looks and the d7-knight cannot reach d3,
best, intending 15 �C2 tDxe4! 16 tDxe4 though Black should still have enough
':xb2+ and then : compensation .
13 lIb1?! c4+ 14 �g3
Or 14 �1 tDa6 1 5 'ii'e 2 tDb4 and
Bl ack is slightly better, one idea being
16 'ii'x c4 tDxa2 ! 17 tDxa2 lUc8 and
Black's activity has increased.
14 ... tDa6
Cl) 17 �d3 ? c4+! 18 �e2 (or 18 �e3

tDb6, followed by ... tDxds+ and ... c4-c3)
18 ... C3 and Black win s the piece back
with a big advantage.
C2) 17 �dl? ! tDb6 18 a4 tDxds 19 as
fs 20 tDg 3 tDC3+ 2 1 �el tDbl 2 2 tDe2
ds and Black i s clearly in the lead. Heading for d3.
c3) 17 �Cl (the best m ove) 17 ... tDb6 15 We2 tDC5 16 �h3
18 l:tel tDxds 19 a4 J:tb8 20 l:tbl (20 16 'ii'x c4? run s into 16 .. :iVa6 ! 17
l:ta2 ? ! is m et by 20 ... C4! , when the Wxa6 tDfxe4+ 18 tDxe4 tDxe4+ 19 �h 3
threat of ... c4-c3 is quite serious; e.g. 2 1 tDf2+ 20 �h4 .i.f6+ and White loses
�dl .i.b2 ! 2 2 �e2 ? ! c 3 2 3 liitb l liitb 4 2 4 m aterial . Or if 16 es, then 16 ... tDh s+ 17
tDf2 liit a xa4 o r 2 2 �C2 c3 2 3 .i.xc3 .i.xc3 �h 3 tDd3 18 g3 .i.h 6 and Black is bet
24 tDxC3 ':'c8 25 l:ta3 l:.ac6 and Black ter; the threats are ... .i.xf4! and ... tDf2+.
regains the piece with a clear advan- 16 ... tDd3 17 .i.e3?1
70
17 l:thfl followed by 18 ttJel, ch al and . . .c5-c4, and obtained compensa

lenging the d3-knight, was more resil tion for the pawn without too much
ient. effort. On move 12, Mancini should
17 .. :iWb7 18 g3 e6! probably have preferred 12 'ife2, aim
A thematic attack on the white cen ing for the endgame in which he at
tre. least h as equal chances. Instead, after
19 dxe6 ttJxe4 20 ttJxe4 'ili'xe4 21 ttJg5 12 'iitf2 ? ! the white king became rather
'iYf5+ 22 g4 'iVd 5 2 3 exf7+ 'iit h 8 24 b3 h6 exposed and Degraeve won n icely.
Summary
The Classical has lost some of its popu
l arity nowadays. White's long -term
plan is simply to n eutralize Black's ini
tiative and then play e4-e5, or perh aps
set up a blockade on the queenside.
However, as we h ave seen in this chap
ter, Bl ack can equalize quite comforta
bly, and I don 't think too m any theo
retical discussions will occur here in
Here 2 5 'iVf3 ! was forced, when the foreseeable future.
Black is at least a little better after If White allows 12 ... ttJg4, Black h as
25 ... 'ii'xf3 26 ttJxf3 llxf7i e.g. 27 bXc4 ideas like ... .ixC3, ... ttJg4-e 5 (-c4), ... .:tfb8
.:te7 ! 28 .id2 ttJf2+ or 28 .ib6 ttJxf4+. and ... 'iVa6, after which White h as failed
Instead, White blundered with : to demon strate any advantage. Even if
25 l:r.hdl?? hxg5 he plays 1 3 'ife2 and delays chasing the
Now Black just win s a piece. knight, Black is still doing fine after
26 bxc4 'ii'e 4 27 .l:f.xd3 13 .. :iWa5, as shown by Vachier Lagrave
Unfortunately 27 'iVxd3 fail s to in Game 8.
27 .. :iIi'f3 mate. In the m ain line White in serts the
27 ...gxf4 28 l:tbb3 lIxf7 29 'ii'f2 .if6 30 prophyl actic 12 h 3 , preventing ... ttJg4,
'ili'xf4 l:1h7+ 0 - 1 but then 12 ... 'iVb6 1 3 l:te1 l:tfb8 gives
Black a relatively easy g ame, the only
Game summary drawback being that 14 e5 l eads to a
The idea of 6 ttJC3 and 7 f4 (or vice very drawish position (as seen in Gam e
versa) is not a critical line again st the 11). If Black wants t o play for a win, he
Benko. Degraeve responded with the might con sider the dynamic 1 2 ... .l:f.a6,
accurate 7 ... 'iVa5 ! , followed by the followed by ... 'ifa8 and ... e7-e6 (as in
standard moves ... .ig7, ... d7-d6, ... ttJbd7 G ames 12 and 13), although in this
71
case White probably can achieve a usual compen sation for the pawn,
slight advantage. which continues into the endgame.
With h 2-h 3 the white king can also This set-up is also appropriate again st
shuffle over to h2 (instead of the m ore Shulman's line with 9 lLlge2.
common g 2), when there are some Finally, the aggressive options of 9
slight differen ces in the play. H ere g4 or 7 f4 are not to be feared either.
Black should opt for the standard Indeed, these can often backfire on
... lLlbd7, .. :ii' a s, .. J:tfb8 , and possibly White, whose own kin g position is then
... lLle8-c7-b S , when he obtain s the more vulnerable in the middlegame.
72
C h a pter Three
The 'Ouiet' 5 e 3
-
1 d4 tDf6 2 C4 cS 3 d 5 b5 4 cxb5 a6 5 e3
Game 1 9
A.Ka rpov-A.Miles
Ti l b u rg 1986
1 d4 tDf6 2 c4 cS 3 d 5 b5 4 cxb5 a6 5 e3
i.. b 7!
5 ... axb5 6 i.. x b5 'iWa5+ 7 tDC3 i..b 7 8
i.. d 2 i s a more common move order,
but vi a thi s route White al so has 8
tDge2 tDxd5 (or 8 ... i.. x d5) 9 0-0 which i s
both critical and, in my opinion, fre
This system doesn't have an estab quently quite difficult for Bl ack. For
lished name. Despite its quiet appear th at reason I prefer the altern ative
ance 5 e3 is actually a sh arp line, as 5 ... i..b 7, which h as the advantage of
White sacrifices a pawn for th e initia avoidin g thi s line.
tive in many variation s; and I am not 6 tDC 3
suggesting a quiet approach for Bl ack 6 bxa6 i.. x d5 7 tDc3 i.. c 6 is roughly
either. equal . For example, 8 tDf3 tDxa6 9 tDe5
After 5 e3 Black has three m ain (aiming for the bishop pair; in return
choices: th e classical 5 ... g 6 (which is Black h as an extra central pawn)
probably okay), 5 ... axb 5 (see the first 9 .. . tDb4! ? (sh arper th an 9 ... tDb8 which,
note below), and 5 ... i..b 7, which is the however, is just as good since White
move I am recommending. However, will mostly take on c6 anyway; e.g. 10
some of the gambit lines can be quite i.. C4 e6 1 1 tDxc6 tDxc6, tran sposing to
dangerous, so Black has to pl ay with line 'c') and now we h ave the following
some precision . possibilities:
73
In comparison to s ... axbs, White

g ains the extra option here of taking
the a-pawn : 7 bxa6 ! ? is examined in
Game 21.
7 ...axbs 8 i.. x bs 'ilVb6
a) 1 0 ttJxc6 ttJxc6 1 1 .i.b s 'ilb6 with

an equal g ame.
b) 10 a3 'ilb8 1 1 f4 (after 11 ttJxc6
ttJxc6 12 .i.bs 'ilb6 the position is basi
cally the same as in line 'a') 1 1 ... 'ilb7 12
ttJxc6 ttJxc6 1 3 i.. d 3 dS, followed by As n oted above, this position often
14 ... e6 and Bl ack is doin g fine. arises via s ... axbs, or sometimes
c) 10 .i.C4 e6 1 1 ttJxc6 ttJxc6 12 e4 s .. :ii'as+. Whichever way, one of the
i.. e 7 is al so TOughly equal . One theme m ain ideas behind ... .i.b7 is to attack
for Black is a well timed ... d7(d6)-dS ; the ds-pawn, and hopefully win it, in
e.g. 1 3 .i.f4 0-0 14 0-0 d6 1 5 a4 'ili'as 16 return for the b- or a-pawn .
i.. d 2 'ifb6 17 b3 J::tfd8 18 l:[cl l1ac8 (or 9 'i'b3
18 ... dS) 19 l':tel ttJxe4 20 ttJxe4 dS with Defending the pawn . Alternatives to
bal anced chances. this move, in particular 9 .i.c4, will be
6 'iWa s
••• discussed in the next game.
9 e6
..•
Tryin g to open the position for the

light-squared bishop.
10 e4
10 i.. C 4 al so tran sposes to Game 20.
The text allows Black to achieve his ob
jective by means of a sm all combina
tion .
10 ... ttJxe41 11 ttJxe4 .i.xd s 12 'ilVd 3 'i!i'b7
The safest move. 12 .. .fS is al so pos
sible and leads to sh arp pl ay after 13
7 .i.d 2 ttJg 3 .i.xg 2 14 ttJle2 .i.xh l 15 ttJxh l
74
Th e 'Q u i e t ' 5 e 3
i.. e 7 16 tDhg 3 (or 16 .i.c3 O-O ! , since 17 13 ... c41 14 Ji.xc4 .i.xc4 15 'iVxC4 d 5 16
.i.xd7?? litd8 18 'ifg 3 fails to 18 ... .i.f8, 'ifc2 dxe4
when the d7-btshop was unpinned but
still trapped in O.Almeida Quintana
M. Leon HOYos, Havana 2011) 16 ... 0-0
17 'iVxe4
Unsurprisingly, Karpov aims for the
clarity of a simplified position, though it
17 a4 (17 tDh 5 ? ! was played in m ay not be not the best move. others:
I.Koga-Dunning, USA 1985, but then a) 17 1:[cl tDd7 18 'iVxe4 'iVxe4+ 19
17 ...'ifb7! is clearly better for Black) fxe4 tDC5 20 'iii> e 2 l::t x a2 21 tDf3 l::t x b2 2 2
17 ... tDc6 18 .i.C3 d5 19 'iVe3 tDd8 20 llc4 Ji. e 7 2 3 ':'al 0-0 24 11a7 .i.d6 2 5 e 5
.i.xg 7 (or 20 tDh S ? ! d4 21 tDxd4 cxd4 22 Ji.b8 was obviously fin e for Black in
�xd4 'ifb7, D.Gurevich-J. Nicholson, J . Litvin chuk-P.Wolff, U SA 1985.
Lugano 1985, when 23 'iii> e 2 'iVg 2 24 b) 17 fxe4 leads to a more unbal
tDxg7 �g 5 2 5 'iff3 'ifxf3+ 26 'iii> xf3 .i.f6 anced position, in which Black's active
27 .l:tg l <M7 28 .i.xf6 'iii> xf6 leaves Black pieces give him the usual compensa
with a slight advantage) 20 ...'iii> x g7 2 1 tion for the pawn ; e.g. 17 ... tDc6 18 tDf3
tDh 5+ 'iii>f7 22 'ifh6 (with the serious tDb4 19 .i.xb4 .i.xb4+ 20 'iii>f2 0-0 2 1
threat of 23 'ifg7 mate!) 22 .. :ii'xb5 ! llhdl l:tfc8 2 2 'ii'e 2 l:ta4 2 3 'iii>fl was un
(Plaskett) 2 3 axbs l:[xa1+ 2 4 'iii>d 2 'iii> e 8 2 5 clear in M.Dlugy-D.Gurevich, U S Ch am
tDef4 'iii>d7 26 'ifxh7 with an unclear po pionship, Estes Park 1985, and
sition. Black has two rooks for a queen, 2 3 ... .i.c5 ! ? 24 nel .i.b4 2 5 l:tedl .i.C5
but his forces are a bit uncoordinated. would h ave been one way to m ake a
13 f3 draw.
Now 13 tDg 3 ? ! .i.xg 2 14 f3 .i.xh l 1 5 17 ...'ii'x e4+
tDxh l Ji.e7 is just good for Black, whose Miles responds with the safest con
rook and two central pawns are much tinuation . H ad h e n ot been playing Kar
better than White's uncoordinated ex pov, he might h ave gone for the m ore
tra bishop and knight. ambitious 17 .. :ilxb 2 ! .
75
Now Black cannot be prevented

from regaining his pawn with equality.
19 liJe2
Worse is 19 .i.e3 ? ! �a4 20 liJf3 l:txe4
21 'itd2 .i.e7 22 a3 0-0 23 b4 .l:td8 24
'ite2 liJb6 and Black has a slight edge,
which led to a win after his opponent
l ater blundered in H . Kummer-L.Milov,
Oberwart 1994.
19 ... liJcS 20 liJc3 liJd3+ 2 1 'ite2 liJxb2 22
liJbS .l::[ a 4 2 3 liJC3 .l:ta8 24 liJbS l:!.a4 2 5
For example, 1 8 �bl (18 'il'xa8 liJc3 ]::t a 8 26 liJbs Yz-Yz
'iVxal+ 19 'ite2 .i.d6 20 1Ii'a4+ 'ite7 is
worse for White, who has more diffi Ciame summary
culty developin g his kingside) 18 ...1Ii'xa2 The S e 3 line can be quite dangerous if
19 liJe2 (19 .i.c3 is met by 19 ... .i.cS ! 20 Bl ack is not careful . With S ... .i.b7 Bl ack
.i.xg7 fS ! 21 'iVb7 'ii'f2+ 22 'itdl 'ii'f l+ attacks the centre straight away, rather
and ... �a2+ with advantage) 19 ... liJd7 (or th an tryin g to win a pawn with the
19 .. .fS ! ?) 20 liJC3 'iVa6 21 liJbs ':'c8 chance of White seizing the initiative.
(Wolff), when Black has slightly the bet In the resulting position after 6 liJC3
ter of things with his extra pawn, al 'iVas 7 .i.d2 axbs 8 .i.xb S 'iVb6, Karpov
though Nguyen Anh Dung-Z.Medvegy, defended the ds-pawn with 9 'ifb3 e6
Budapest 1999, was soon drawn after 10 e4, which allowed Miles to eliminate
22 'il'd3 'iVc6 2 3 'ite2 liJes (23 ... .i.cS ! ?) 24 the centre by mean s of a combination .
'iVd4 'iVds 2 S 'iVxds exds 26 .l:.hcl ':'xcl With White having failed to gain any
27 �xCl liJC4 %-V2. advantage, Bl ack even had two oppor
18 fxe4 liJd7 tunities to sh arpen the play (12 .. .fS and
17 ... 'iIi'xb2). However, Miles was content
to play safely as Black, after which a
draw was th e logical outcome.
Game 20
D.G u revich-J.Benja min
Ch icago 1986
1 d4 liJf6 2 c4 cS 3 d s bs 4 cxbs a6 5 e3
.i.b7 6 liJC3 axbs
76
Th e 'Q u i e t ' 5 e 3
This move order allows White the 17 .Jtg s .Jte7 Black i s slightly better; e.g.
possibility of 8 tiJge2 ag ain. In stead, 18 l:.dl h 6 19 .i.xf6 .Jtxf6 20 'ii'g 3 hS 21
Bl ack should prefer 6 ... 'iVas, when 7 e s h4 2 2 'iVf4 .Jte7 2 3 tiJe4 h3 24 g 3
.Jtd2 axbs 8 .Jtxb S reaches the game ifb8) 1 4. . ..i.e7 1 5 l:tadl tiJxb s 16 axb s
position . was played in A.Graf-S. H alkias, Fuerth
7 .Jtxb S 'iVa s 8 .Jtd2 'ilVb6 2002, and here 16 ... 'ilVC7 is about equal .
White has a small lead in development,
but Black h as the bishop pair and a su
perior pawn structure.
9 ... e6
9 .JtC4
Alternatives include:
a) 9 'iVb3 was Karpov-Miles (Game
19).
b) 9 tiJge2 tiJxds 10 0-0 tiJf6 11 tiJf4 10 'ilVb3
e6 12 'iie 2 .Jte7 with an unclear g ame. Going for the endgame - or perh aps
c) 9 'ili'f3 (Graf's speciality) m ore accurately, queen less middle
g ame - but Black shoul d h ave n o prob
lem s in that case. Sharper possibilities
are:
a) 10 e4 tiJxe4 1 1 tiJf3 ! (after 1 1
tiJxe4? exds 12 .JtxdS .JtxdS 1 3 'iVe2
'ii'e 6, as in K.Zietak-K.Miton, Warsaw
rapid 2005, or 11 dxe6? ! fxe6 Black i s
clearly better; while 1 1 tiJ g e 2 tiJf6 12
tiJf4 .Jtd6 is quite okay for Black too)
11 ... 'iVb4! ? (the sharpest m ove, leading
to an unclear position; otherwise the
9 ... tiJa6 10 tiJge2 tiJb4 11 0-0 tiJbxdS quieter 1 1 ... tiJf6 12 0-0 .i.e7 1 3 .Jtf4 0-0
12 e4 tiJC7 13 a4 e6 14 .Jtg s (after 14 looks sufficient for equality) 12 tiJes
.Jte3 tiJxb S 15 axb s 'ii'd 8 16 :xa8 .Jtxa8 'ii'xb2 ! (since 1 3 tiJxe4?? n ow loses to
77
1 3 . . .'ii'x e 5 ; instead 12 . . .tiJxd2 'ili'xd2 1 3 tran sposes accordin g t o where White

'ii'x d2 �d6 1 4 tiJxf7 ! rj;xf7 1 5 dxe6+ develops his king's knight.
g ave White a strong attack in C.Csiszar 12 tiJf3
M.Morvay, Hun g arian Team Champi Instead:
onship 1995) 13 J:.bl 'iix d2+ 14 'ii'x d2 a) 12 tiJge2 exd5 return s to 10 tiJge2
tiJxd2 1 5 11xb7 (15 'ii;J x d2 ? ! is dubious in above, or else Black can just play
view of simply 1 5 ... �c8, when Black is 12 ... tiJb4 1 3 0-0 tiJbxd5 with equality.
slightly better) 1 5 ...�d6 16 tiJxf7 tiJxc4 b) 12 f3 ? ! is too slow, due to
17 tiJxh 8 �e5 18 tiJb5 (not 18 tiJe4? 12 ... tiJb4 13 dxe6 fxe6 14 tiJb5 (or 14
exd5 19 tiJxc5 g6 and Black picks up the tiJge2 tiJd3+ 15 rj;fl �a6) 14 ... c4 15
knight on h 8 with a clear advantage) �xc4 tiJc2+ 16 rj;dl tiJxal 17 tiJC7+ rj;f7
18 ... tiJa5 (18 ... exd5 is m et by 19 f4! ) 19 18 tiJxa8 �xa8 and Black was slightly
lIb6 tiJc4 20 l:tb7 with a draw. better in Y.Zimmerm an-A.Chernin,
b) 10 tiJge2 exd5 11 'ii'b 3 'ii'xb3 12 Hun g arian Team Championship 1999.
�xb3 tiJa6 1 3 0-0 tiJC7 14 .l:1fdl �d6 1 5 12 .. exd 5 13 0-0 �e7 14 :fdl
.
� e l 0-0 16 f 3 lUe8 17 �f2 �e5

(17 ... �f8 and 17 .. Jlab8 also come into
con sideration) 18 lIacl �c6 was played
in A.Beliavsky-D.5ermek, Terme Zrece
2003. White is tryin g to put some pres
sure on the centre, but it is quite well
defended, so Black h as more space and
can look for counterplay on the queen
side. Bl ack h as n o real problems.
10 ...'ii'x b3 11 �xb3 tiJa6
The position is roughly equal . Black

has some nice central pawns and the
possibility of play on the queen side;
while White can try to put some pres
sure on the centre, especially d5.
14 ... 0-0 15 �el tiJC7 16 tiJh4?1 g6 17 g3
l:tfb8 18 tiJg2 �c6 19 l:tabl C4 20 .ic2
tiJe6 21 tiJf4
The point of the White's long
winded knight manoeuvre. Unfortu
The immediate 1 1 ... exd5 followed n ately, Black next m o ve renders it
by 12 ... tiJa6 is also fine, when play worthless.
78
Th e 'Q u i e t ' 5 e 3
H ere the deflection 42 ... g4! 43 i.xg4

l1c6 ! was decisive, when White l oses his
passed pawn s.
43 bS g4 44 b61
A great try.
44 ... gxf3+ 4S �xf3
21 ... d41 22 exd4 i.f3 2 3 tiJfe2

23 Ud2 tiJxf4 24 gxf4 i.b4 is al so
very good for Bl ack.
23 ... i.b4 24 dS tiJgs 2S h4 i.xe2 26
hxgs i.xd1 27 :xd1 tiJe8
4S ...:b2?1
The computer finds a win with
45 ... l:txa4! 46 b7 tiJg4, intending 47
b8'iW l:txf2 m ate, or 47 'itg 2 lIb4! 48 litb5
l:1xf2+ 49 �g l l::lb b2 ! 50 b8'iW ':'g 2+ and
m ates again; but of course that would
be very difficult to see over the board,
especially after the shock of 44 b6.
46 b7 .ll b 3+ 47 �g2 lIxb7 48 tiJxb7 l:ixa4
49 tiJd6 1ld4 SO tiJbS l:tdS S1 l:txdS tiJxds
Bl ack is the exchange up, stands
clearly better and eventually won . I will
go quickly through the rest as it is n ot
relevant for an opening book; though,
as you will see, White puts up a good
fight with his queen side pawn s.
28 a4 i.e7 29 tiJbs i.xgs 30 i.C3 i.f6
31 i.xf6 tiJxf6 32 .l:td4 Ue8 3 3 l:[xC4
tiJxd s 34 ':'d4 l:e1+ 3 S �h2 tiJf6 36 i.d1
�g7 37 b4 ':c8 38 tiJd6 l:tc1 39 i.f3 :C2
40 �g2 lla1 41 :f4 gs 42 l:tfS h6?1
79
Gurevich h as done well t o get this often plays 7 �d2 here.

far, but Benjamin won the knight end
g am e anyway.
52 'iii>f 3 'iii>f6 53 'iii> e 4 liJe7 54 liJC7 h5 5 5
liJe8+ 'iii> g 5 56 f4+ 'iii> g 6 57 'iii> e 5 liJf5 5 8
liJC7 liJxg3 59 liJd 5 liJf5 6 0 'iii> e 4 d6 61
'iW3 liJh6 62 'iii> e 4 liJg8 63 liJe3 liJf6+ 64
�3 liJg4 65 liJC4 d 5 66 liJb6 liJf6 67
liJc8 �5 68 liJe7+ 'iii> e 6 69 liJc6 liJe4 70
liJd4+ 'iii> d 6 71 liJf5+ 'iii> C 5 72 liJg7 h4 73
liJf5 h3 74 liJh6 f6 7 5 liJf5 'iii> C4 76 liJe3+
'iii> d 4 77 liJf1 'iii> d 3 78 liJh2 f5 0-1
7 liJxa6
...
Game summary Not 7 ... �xd5 ? ! 8 �d2 �c6 9 liJf3

In the g am e White gets the queens off liJxa6 10 liJe 5 'iWC7 11 liJxc6 'ii'x c6 12
early, hoping to target the black central 'ii'f3, when White has some advantage
pawn s, but is outplayed in the due to the bishop pair.
queenless middlegame. Although 8 �d 2 1
Gurevich fights hard in the endg ame, The strongest continuation . 8 .ll C4
eventually he is unable to hold on . liJb4 is comfortable for Black; for ex
If White wants to play this way, ample, 9 liJge2 (now if 9 .ll d 2 ? ! , then
then 10 liJge2 m ay be a better try, even 9 ... .ll a 6 ! and Black is already slightly
if objectively Black has no problems better) and th en :
there either; n or after Graf's more
complicated 9 ..wf3 . Both these lines are
seen in the n otes.
Game 21
R�Mohr-J�Polgar
Amsterd a m 1989
1 d4 liJf6 2 C4 C5 3 d 5 b5 4 cxb5 a6 5 e3
�b7 6 liJc3 'iWa 5 7 bxa6
Of course this is n ot possible after a) 9 . liJfxd5 10 0-0 e6 11 e4 liJb6 12
..
5 ... axb 5, so it must be reg arded as an �b 5 �c6 1 3 �xc6 liJxc6 14 �f4 �e7 15
important test of Bl ack's move order b3 ( 1 5 'iWd3 0-0 16 .l:tad1 f5 gives Bl ack
with 5 ... �b7, even if White far more the edge) 1 5 ... 0-0 16 .ll d 6 l:tfd8 and th e
80
Th e 'Q u ie t ' 5 e 3
position is about equal . Black h as fairly J . Polgar, European Team Champion

good central control though, and if ship, H aifa 1989 .
... d7-ds-d4 can be achieved Black will In stead, Black should prefer
have decent ch ances of obtaining an 10 ... .i.e7 ! 11 dxe6 (11 d6? ! fails to
advantage. 11 ... 'ii'x d6 ! 12 es .i.xf3 ! , while 11 .i.C4?!
b) 9 ...liJbxds is also possible; e.g. 10 run s into 11...liJxe4! 12 liJxe4 exds 1 3
0-0 e6 11 .i.d2 'ifb6 with an unclear .i.xa6 .i.xa6 1 4 liJ C 3 'iWe6+ and Bl ack i s
position; or if 10 .i.d2, then 10 ... 'iWC7 clearly better after either 1 5 .i.e3 d 4 or
and Black is again comfortable due a l s liJe2 d4) 11...fxe6 and then :
greater influence in the centre.
S 'ii'b 6 9 e4 e6
...
a) 12 .i.d3 ! ? liJb4 1 3 .i.bl 0-0, when

the open f-file and active pieces gives
Assaulting the white centre in order Black sufficient compensation for the
to open the long diagon al for the light pawn ; e.g. 14 0-0 liJc6 1 5 .i.d3 liJg4 16
squared bishop. At the moment, Black .i.e2 liJf6 17 .i.d3 liJg4 with a repeti
has sacrificed a pawn, but she h as a tion.
good lead in development and is about b) 12 e s liJe4 1 3 .i. e 2 0-0 14 0-0
to win the pawn back. liJxd2 15 'iWxd2 ( 1 5 liJxd2 ? ! allows
10 .i.C4 ls ... 'iWxb 2 ! ) l s ... .i.c6 and in my opinion
10 liJf3 ! is a critical test of s ... .i.b7. Black h as enough compen sation. For
Now it is dangerous to take the pawn in stance, 16 a4 liJb4 17 b3 (or 17 l::tfdl
with 10 ... exds ? ! because of 11 exds 'iVb8 18 lIa3 'iWe8 19 liJbs liJds with
liJxds 12 .i.C4 'iVe6+ 1 3 liJe2, when compensation) 17 ... 'iVd8 18 lIfdl 'ii'e 8
White has a lead in development and a 19 as 'ii'g 6 20 �fl 'ii'h s 2 1 �g l 'ii'g 6 is
dangerous initiative; e.g. 1 3 ... 'ii'g 4? another repetition.
(13 ... liJac7 i s more solid) 14 'iVb3 liJab4 10 liJb41
•..
15 0-0 liJb6 16 .i.xf7+ �d8 17 liJg 3 and Putting further pressure on the
White was already winning in J . Lautier- lig ht squares.
81
0-0 and White has to start thinking

about equalizing .
11 liJxe4! 12 dxe6 fxe6 13 liJxe4 JLxe4
• . .
14 0-0 d 5
1 4. . .JLe7 ! promises Black at least a
slight edge after l s liJg 3 (15 JLc3 ? ! ds is
much better for Black better due to the
strong centre) ls ... JLg6 16 'iWf3 'iWc6.
15 JLxb4 'ii'x b4 16 JLd3 JLxd 3 17 'iWxd3
Now the position is unclear. White
has a lead in development, whereas
This time the tempting 1 0... liJxe4? ! is Black h as a potentially strong centre.
not as good, because White still h as the
option of liJge2 to en able castling; i .e.
1 1 liJxe4 exds 12 JLxa6 JLxa6 1 3 liJg 3
i.d6 14 liJle2 ! 0-0 1 5 0-0 JLes 16 I1el
'iix b2 17 J:tcl and White h ad a definite
advantage in l Fedorowicz-L.Alburt, U S
Champion ship, Cambridge Springs
19BB.
ll liJge2?!
17 JLe7 18 'iWe3 'iWg4?!

•••
lB .. .'it>f7 ! was a better way to defend

the pawn, and if 19 liJf4, then 19 .. :i'e4
20 'iix e4 dxe4 2 1 l:tfel l:thbB, wh en
Bl ack will win a pawn on the queen side
(22 ':e2 JLf6) with a good game.
19 liJf4 .l:i.a4!?
An interesting attempt to play for
the win . The straightforward 19 ... d4 20
But if 11 dxe6 fxe6 12 a3 liJxe4! 1 3 'ii e 4 (or 20 'ii'e s 'iifs 21 'iVxg 7 'iit d 7 22
liJf3 liJd6 1 4 i. e 2 i.xf3 1 5 JLxf3 (or 1 5 'ifh 6 i.fB 23 'ifh4 i.d6 24 liJe2 hs is
g xf3 liJc6) l s ... liJd3+ 16 'ittfl :cB and more complicated) 20 ... 0-0 2 1 'ilxe6+
Bl ack is clearly better; or 1 1 a3 exds 12 'iWxe6 22 liJxe6 lU6 would give Black
exds liJbxds 1 3 liJf3 'ike6+! 14 'iVe2 JLe7 enough compensation for the pawn,
1 5 0-0 liJxc3 16 i.xc3 'iWxe2 17 JLxe2 but few winning chances.
82
Th e 'Q u i e t ' 5 e3
20 ltJxe6 l:te4 21 ltJC7+ �d7 2 2 'ii'C 3 ':'c4 ing 3 s ... d2 3 6 ltJg 2 . After the text the d
23 'ife1 ..td6 24 h3 'ii'f4 25 g3 'iVg5?1 pawn becomes too strong and Black i s
Still playing for the win, although winning .
this is no longer objectively correct. 3 5 d 2 36 l:tbb1 lIe1+ 3 7 � h 2
. • .
Instead, 2s ... 'ii'fs 26 ltJbs 'ii'x h 3 2 7

ltJxd6 llh4! 28 gxh4 'iVg4+ would force
a draw.
26 ltJb5 ..tb8 27 l:td1 l:te8
3 7 . . J:txf4!
A nice finishing touch ; White can do
n othing while Black brings up her kin g .
3 8 gxf4 �d 5 39 �g2 �e4 4 0 b 4 .l:txd1
28 1i'd2? 41 l:txd1 ..txb4 0-1
As we have seen m any times, enter
ing the endg ame does not reduce Game summary
Black's initiative in the Benko. Instead, With s .....tb7 6 ltJC3 1i'as Black contin
the best defence was to counterattack ues to offer the g ambit pawn, and in
with 28 'iVas ! ..txg 3 29 'iVa7+ ..tC7+ (or this g am e White decided finally to ac
29 ... �d8 30 'ii'a 8+ ..tb8+ 3 1 �h l �e7 cept it with 7 bxa6, after which Black
3 2 l:tfe1+ l1e4 33 .l:txe4+ win s a pawn) obtained compen sation by attacking
30 �h l 'iVe s 3 1 ltJxC7 'iVxC7 3 2 l:txds+ on the light squares. The m oves ... 'ifb6,
..t>c6 3 3 .l:txcs+! .l:txcs 34 'ii'a4+ and 3 5 ... e7-e6, ...ltJb4 and ...ltJxe4 are quite
'ii'x e8, when White clearly has all the thematic for this line. In a sharp mid
chances. dlegam e battle, Judit Pol g ar took a few
28 .. Ji'xd2 29 l:txd2 �c6 30 ltJc3?! risks in the quest for a win , but she was
30 b 3 ! l:tce4 3 1 a4 would still have ultimately rewarded as Mohr incor
kept the balance, but perh aps White rectly went into an endgame and then
was in time trouble. defended it poorly. In the opening, 10
30 ... d4 31 ltJe2 ..tC7 32 b3?! ..ta5 3 3 l:tb2 ltJf3 would have been more testing,
d3 34 ltJf4 l:td4 3 5 l:td1? though after 10 ... ..te7 ! Black looks to
The last hope was 35 .l:tbbl, intend- h ave sufficient play all the same.
83
Summary broke through in the centre t o reach a

The 5 e3 line is quite tricky and can be winning position .
dangerous against an unprepared op Finally, should White belatedly ac
ponent. With S ... ..tb7, 6 ... 'i¥as, and then cept the gambit pawn with 7 bxa6 (as
7 ... axbs, Black responds in a sharp way in Game 2 1), Bl ack can follow the same
by targetin g the central light squares. strategy, attackin g on the light squares
As we saw in Gam e 19, White cannot with the thematic moves ... 'i1ias,
simply defend by 'ilfb3 and e 3-e4, since ...liJxa6, ... 'i¥b6, ... liJb4 and ... liJxe4.
Black h as the combination 10 . . .i"t:'lxe4! , Again Bl ack achieves a fine game.
destroyin g t h e centre after all with at Summing up, 5 e3 leads to double
least equality. edged play where a mistake by either
In Game 20, White opted to give up side can easily mean a rapid change in
the centre, hopin g to apply pressure in the evaluation . Although the line is not
turn in a queen less middlegame. H ow played so often these days, since the
ever, Black's pieces are well placed to theoretical chances of an advantage
defend it, and he h as good prospects are n ot that serious, at a practical level
for counterplay on the queen side. After Black has to know what to do if he is to
White tried a faulty plan, Benjamin avoid coming out worse.
84
C h a pte r Four
The 5 f3 Va riation
1 d4 liJf6 2 c4 cS 3 d s bs 4 cxbs a6 S f3 m anoeuvre one of the knights to c 3 .

The second option is 5 ... axb 5, but after
6 e4 'iVa5+ 7 ..td2 b4 S liJa3 ! and 9 liJC4,
I prefer White here too.
Therefore I prefer the n atural 5 ... e6,
attackin g the centre, wh en th e critical
line run s 6 e4 exd5 7 e 5 'iVe7 S 'ii'e 2
liJg S.
With 5 f3 White begins an ambi

tious attempt to build a strong centre.
This was a popular variation in the old
Soviet Union, but as we shall see, Bl ack
has since found some nice ideas
ag ainst it. Nevertheless, the f2-f3 varia
tion can be a dang erous White
weapon, so good preparation is rec In this position Black h as two nice
ommended. central pawn s, while White intends to
After 5 f3 Bl ack has three main develop rapidly and target them with
choices. The standard Benko set-up moves like liJc3, liJh 3 -f4 and often
with 5 ... g6, in my opinion, allows White 0-0-0. Sometimes White can even sacri
to gain an edge after 6 e4, as he has fice a knight on d5, as Nakamura did in
good ch ances to create a blockade on a recent game (see the n otes to Game
the queen side, later play a2-a4, and 24). In this variation both sides need to
85
know wh at they are doing, as a single variation and, as I mentioned above, it

mistake can quickly lead to a loss. is crucial to have concrete knowledge
Practice has mostly shown that in this line, since the sharp character of
Black is okay, with ideas like ... c5-c4, the play often makes even a single,
... Wb4 and, in reply to llxd5, sacrificing slight mistake fatal .
the queen on c3 for long-term compen 6 e4 exd s
sation (as in the first g am e below). In 6 ... C4? ! is a sharp sacrifice of an
response N akamura's new plan (or el se other pawn, based on rapid develop
his computer's) h as been to play liJh 3 - m ent and counterplay on the dark
f4xd5 without even waiting for the squares; e.g. after 7 i.xc4 axb5 8 i.xb5
rook to come to dl, after which White i.C5 . However, I won 't be covering this
gets a lead in development and a dan line as I don 't believe 6 ... C4 is fully
gerous initiative for the sacrificed sound. If Black wants to try it anyway, I
piece. Although the g am e is very un would suggest restricting it to games
clear and further tests are required, I at fast time limits or else as a surprise,
think that Bl ack should be fine. when it m ay not be so easy for White to
prove an advantage.
7 eS
Game 22
B.Lalic-A.Khalifman
Lin a res Ope n 1997
1 d4 liJf6 2 c4 cS 3 d s bs 4 cxbs a6 S f3
e6!
Ag ain, the most principled and

sh arpest reply. In the following varia
tion Black will obtain some nice central
pawn s, but falls behind in develop
ment and so has to pl ay precisely.
After the less popul ar alternative, 7
The principled reply, immediately exd5, Black can equalize more easily
attackin g the white centre. Both sides with 7 ... i.d6 (a logical developing
are fighting for the initiative in this move, which reminds a bit of the so-
86
Th e 5 /3 Va ria t i o n
called Snake Benoni; here White h as .i. e s 1 3 ttJh 3 .i.xbs 1 4 .i.xb s 'iVd6 1 5
the weakening move f2-f3 included ttJg s ttJa6 16 ttJe4 ttJxe4 17 fxe4 f6 with
but, on the other han d, is now a pawn compensation) 12 ... .i.es 13 d6 'fie6 14
up) ttJC7 'fifs 15 �el .i.xfl 16 ttJxa8 .i.d3 17
ttJh 3 ttJc6 18 ttJf2 .i.C4 and Black h as
excellent pl ay for the exchang e with
two active bishops and the exposed
king on el.
a2) 1 1 ttJh 3 ttJc6 ! 12 'fid2 (after 12
dxc6 dxc6 1 3 .i.f4! .i.xbs 14 'ii'd 2 .i.xf4
1 5 ttJxf4 g s ! 16 'iVxd8+ lIxd8 17 .i.xbs
cxb s 18 ttJe2 �g7, White h as to think
about how to equalize) 1 2 ... .i.xbs 13
.i.xbs ttJd4 14 .i.C4 'iliC7 and Bl ack h as
more th an enough for the pawn, when
8 'iVe2+ �8 (Black has to forgo cas you take into account the strong knight
tling since the exch ange of queen s on d4, play on the dark squares, the
would only benefit White, due to the f3 rook on its way to e8 and the weak
weakness and lack of development) 9 white knight on h 3 .
ttJC3 axb s ! (9 ... .i.b7 10 'ii'd 2 ? ! .i.es 1 1 b ) 1 0 .i. e 3 c4! 1 1 �f2 .l:[as 12 g 3
bxa6 ttJxa6 12 .i.xa6? ! .i.xa6 turned out
well for Black in S.Volkov-A. l1jushin,
internet 2004; but White can improve
with 10 ttJh 3 ! , when I think he is
slightly better) and now:
1 2 ... ttJa6 ! (Bareev also suggested

12 ... .i.b7 as good for Bl ack, but after 1 3
a4! the g am e is more unclear, so I
would prefer the text) 1 3 a4 (E.Bareev
A.Khalifm an , Russian Team Champion
a) 10 ttJxbs .i.a6 with a further split: ship 199 5) and n ow 1 3 ... ttJC7 ! (Bareev)
al) 11 'iVd2 .e7+ 12 �f2 (or 12 �dl 14 .i.b6 (14 ttJe4? ! .i.es is slightly better
87
for Black) 1 4.. .l:1a6 1 5 i.xC7 i.xC7 1 6 Thi s has been the main move here
l:tel and although thi s position is also since Hertneck introduced it in 1991.
unclear, it seems to m e that Black (Until then the passive 10 ... 'iIi'd8 was
would h ave m ore winning chances in a thought to be best.) After the text both
practical g ame. (But n ot 16 lDxb S ? .. :ifcs and ... 'ifb4 are made possible,
i.b6+, when Black is clearly better with while Black is pl anning ... a6xbs in some
... lDxdS and ... lDe3 coming.) variations. For instance, with Black to
7 .. :ile7 S 'ike2 move in this position, 11 ... axbs 12
lDxb S ? ? would run in to 12 ...'ifb4+ 13
lDc3 d4, winning a piece. Another main
idea is the thematic queen sacrifice
.. :ilb4xC3+ (in response to White play
ing 0-0-0 and l:txdS); we will see this
both in the current game and in the
n otes below.
11 i. e 3 1
The strongest move in this position.
White develops, enabling queen side
castling, while preventing ... 'iVcs or
S lDgS
••• ... dS-d4. We will examine other moves
There is no other choice, as in Gam e 24.
8 ... lDh S ? ? loses a piece to 9 g4. 11 axbs 12 0-0-0
•••
9 lDc3 i. b7 10 lDh3 The alternative 12 lDxb S is seen in

The knight heads for f4 and then on the next g ame.
to dS, while lDh S can also be a danger 12 lDf4? is mistimed because it al
ous idea in some lines. lows Bl ack to take the nice central eS
10 c41
••• pawn : 12 .. :ifxe s ! 1 3 0-0-0 (13 lDxb S
lDa6 ! is also slightly better for Bl ack,
who has ideas of ... i.c6, or ... i.b4+ fol
lowed by ... lDe7) 13 ... lDf6 ! (instead,
13 ... lDe7 14 lDxb S d4! 15 lDxd4 l1xa2 or
1 5 'iix c4 'iVxe3+ is good for Black; but
14 'iVd2 b4 1 5 lDb S l:txa2 16 i.xc4!
(V.lvanov) or 14 ... i.c6 15 i.cS ! ? is very
unclear) 14 i.b6 (or 14 i.d4 'iVxe2 15
i.xe2 i.d6 ! with a clear advantage)
14 ... i.d6 15 lDfxdS lDxdS 16 lDxdS 0-0
and White was struggling to show any
88
Th e 5 13 Va ria t i o n
compensation for the pawn in .1i.d2) 1 4... 'i!las 1 5 .1i.b6 (not 1 5 ttJfxdS
E.Dalsrud-J .Skjoldborg, Prague 200 5 . ttJxdS 16 ttJxd S ? .1i.xdS 17 :xdS "xa2
12 JlVb4
•• and Black is clearly better, while 16
.li1xdS can now be m et by 16 ....1i.xds 17
iLb6 'ii'x a2 ! 18 ttJxa2 :xa2, again with
good compen sation) ls .. :ika6 16 l:txdS
(or 16 a4! ? iLc6 17 as 'ii'b 7 gives Black
the edge, whereas 16 .1i.cS .1i.c6 or
16 ... 'ikh 6 ! ? is unclear) 16 ... .1i.c6 ! (prepar
ing ... b S -b4 by takin g the b s -square
away from the white knight) and now:
A cool way of clearing the ranks;

whereas taking the e-pawn now would
give White a very dangerous attack
after 12 .. :ii'x e s ? ! 13 f4 "e7 (or 13 .....e6
14 ttJg S) 14 ':xd S ! , as in M.Dlugy
L.Alburt, US Championship, Los Angeles
1991.
13 ttJf4
Now 13 l:txdS can be met by the a) 17 a4 b4 (17 ... ttJc8 18 .l:.xb S ! J.. e 7
thematic 13 .. :ii'x C3+! 14 bxc3 .1i.xds 1 5 19 ttJfdS .1i.xdS 20 ttJxdS "xa2 is al so
ttJf4 .1i.a3+ 16 �bl ttJ e 7 when Black has very unclear) 18 .l:.as 'Wb7 19 l:txa8
excellent compen sation because of the 'i!lxa8 20 ttJbl ttJdS 21 'ikd4 ttJxb6 2 2
exposed king on bl. E . Klein-J .Waitzkin, 'ii'x b6 .1i.e7 2 3 .1i.xC4 .1i.g s 24 g 3 0-0,
USA 1993, continued 17 'ii'd 2 .1i.e6 18 when the king h as fin ally escaped the
ttJh S l:tg8 19 ttJf4 ttJbc6 20 ttJxe6 fxe6 2 1 centre and Black can look to regain his
f4 cJ;;f7 and Black went o n t o win ; while pawn n ext m ove with a good position.
21 ... ttJdS ! ?, planning a well-timed ...b S b) 17 :xb S ! ? (it seem s like the
b4, also came into consideration . lower-rated player prepared a draw
13 ttJe7
... with the computer) 17 ... .1i.xbs 18 ttJe4
Covering the ds-pawn . (now White h as sacrificed a rook, but
14 .1i.b6 he h as the threat of ttJd6 m ate as well
14 l:txdS could still be answered by as a weakness on C7 to aim for)
14 ... 'i!lxc3+!. However, a noteworthy 18 ... ttJc8 (18 ... ttJfS ! ? might be a better
alternative is 14 "f2 ! ? (threatening 1 5 try) 19 ttJdS ttJxb 6 ? ! (now White gets
89
the advantage; but 19 ... l:ta7 l eads to a loses to 20 ... :cl mate !
draw after 20 ttJef6+ g xf6 2 1 ttJxf6+ b) i s llxdS Ji.. c 6 (covering the bS
�e7 2 2 ttJdS+ etc, or 20 Ji.. x a7 'ii'x a7 2 1 pawn ; of course l S ... 'iix C3+ is still an
'iix a7 ttJxa7 2 2 ttJC7+ �d8 2 3 ttJxbS option) 16 g 3 (after 16 �bl Ji..h 6 17
ttJxbS 24 Ji.. x C4 ttJd4 2 S ttJg s �C7 2 6 lld4 Ji.. xf4 18 l:txf4 ttJdS 19 ttJxdS Ji.. x ds
ttJxf7 ':' g 8 2 7 ttJg s with a repetition) 2 0 Black is slightly better) 16 ... Ji..h 6 and
ttJC7+ �d8 2 1 ttJxa6 l:.xa6 (Bl ack doesn 't the position is unclear; for instance,
really h ave enough for the queen here, after 17 Ji.. c s 'iVas 18 Ji.. x e7 Ji.. xf4+ 19
though he can play on) 22 ttJc3 Ji.. c 6 2 3 gxf4 �xe7 Black will either take on dS
Ji.. x C4 ttJxC4 24 'ii'h 4+ Ji.. e 7 2 S 'iVxc4 .l:.f8 next m ove, or if 20 l:td4 then 20 ...b4
2 6 b4 .l:tb6 2 7 a3 ttJa6 2 8 ttJdS Ji.. x ds 29 with the initiative.
'ii'x ds ttJC7 30 'ii'd 3 �c8 3 1 �b2 f6 3 2 c) is �bl Ji..h 6 16 l:txdS Ji.. c 6! (not
'iVxh 7 fxe s 3 3 'ii'x g 7 l:t e 6 34 �b3 ttJb S n ow 16 ... 'ivxc 3 ? ! 17 bxc3 Ji.. x ds 18 Ji.. e 3 !
3 S :el ttJd4+ 3 6 �a2 llff6 3 7 'ii'h 8+ llf8 and White is better) 17 lld4 Ji.. xf4 1 8
3 8 'ii'g 7 l:.ff6 3 9 �8+ �-� l:txf4 0-0 19 'iie 3 l:.e8 with an unclear
A. Bital zadeh-M.Vachier lagrave, E U g ame.
Championship, Liverpool 2008. lS llxd S
14 h S ! ?
...
Preparin g t o swing the king's TOok

into the attack on the queen side via h 6 .
A g ood alternative is 14 ... g 6 , intend
ing ... Ji.. h 6, which gives Black a satisfac
tory position in all lines:
The less direct i s 'iVf2 ! ? lith 6 16 �bl

was tried in M.Lindinger-M. Kopylov,
German league 200S, when Black
should have played 16 ... g S ! 17 ttJfxdS
(or 17 a3 l:txa3 ! 18 Ji.. c s l:tal+ ! 19 �xal
'ii a s+ 20 �bl gxf4 with compensation)
a) i s ttJfxdS ? ! Ji.. x ds 16 l:txdS Ji..h 6+ 17 ... Ji.. x ds 18 Ji.. c s 'iVas 19 Ji.. x e7 Ji.. e 6 20
puts White in difficulties; e.g. 17 �bl Ji.. xf8 �xf8 and the game is still very
liIa3 ! 18 Ji.. d 4 ttJbc6 19 l:lxbS I:[xc3 ! and unclear.
Black win s m aterial, since 20 l:txb4 lS 'ii'XC3+!
•••
90
Again the thematic queen sacrifice. 'irxb6 tDc6, intendin g . . ..:fa8, . . .l:tb2 etc,
16 bXc3 tDxd 5 when the vuln erability of the white
king appears to provide sufficient
counterplay. For example, 24 'irC7
(other lines include 24 'iVxb5 l::tb 8, or 24
'it>cl Ilfa8 2 5 'ii xb5 .llf2 2 6 'iWbl tDxe5 ,
o r 24 f4 l:.fa8 2 5 'iWb7 %:t8a3 ! 2 6 'iVxd7
:b2) 24 ... :fa8 25 'it>el :c2 ! 26 .i.e2
lilaa2 27 .i.dl .ll x c3 2 8 'irxd7 tDb4 2 9
'it>fl : d 3 30 'iie 8+ 'it>h 7 3 1 .i.e2 l:Idd2 3 2
'ij'xf7 ':xe2 and White h as t o take the
draw.
17 ... .i.xd s
17 tDxd S?1
White has done badly after this
natural move. In stead, 17 'ire4! is a key
improvement, which forces Bl ack to be
very accurate: 17 ... .i.a3+ 18 'it>d2 tDxb6
19 'iix b7 .i.C5 20 tDd5 ':'xa2+ 2 1 'it>dl !
0-0 22 tDxb6
Bl ack only has rook and knight for

the queen, but he h as serious counter
play on the queen side, often simply by
... tDc6 and ... b 5-b4, attacking the white
king . In practice, the position is quite
difficult to play for White and hen ce, in
my opinion, slightly better for Black.
and now the game A.Allemann 18 'ii'd 2 .i.e6 19 .i.e2
R.Costantini, Biel 2001, concluded in It is indicative of White's problems
White's favour after 2 2 .. J�al+? 2 3 'it>c2 in this line th at h e doesn 't really h ave
.i.a3 24 f4 b4 25 tDxC4 :cl+ 26 'it>d2 time to develop his pieces. Stohl sug
bxc3+ 27 'it>d3 .i.C5 28 g3 tDc6 29 .i.g2 gests 19 .i.C7 ! ? tDc6 20 .i.d6 .i.xd6 2 1
tDb4+ 30 'ii'xb4! 1-0. exd6 0-0 a s critical, but i t is h ard t o see
Black had to fin d 2 2 ... .i.xb 6 ! 2 3 how White shoul d continue; for exam -
91
ple, 2 2 g 4 ! ? can still b e m et by 2 2 . . .b4! In my opinion, the 5 f3 e6 line is one

23 cxb4 :fb8 with a very strong attack. of the most dangerous for Black in the
19 lDc6 20 f4 b4 21 fS bxc3 22 'ii'x c3
.•. Benko, and it is characteri stic th at both
l:ta3 23 'iWb2? sides h ave to pl ay with precision. My
This loses quickly. White h ad to play advice, therefore, is to prepare better
the unappealing 2 3 Wal, though after th an your opponent !
23 ... .i.xf5 24 .i.xC4 .i.b4 Black is still
clearly better.
23 ... c3 24 'ifbs .i.xfs 2 S l:[fl Game 23
Se.lva nov-A.Kha lifma n
St Petersburg 1997
1 d4 lDf6 2 C4 CS 3 ds bs 4 cxbs a6 5 f3
e6 6 e4 exd s 7 eS 'ife7 8 'ife2 lDg8 9
lDC3 .i.b7 10 lDh3 c4 11 .i.e3 axbs 12
lDxbS
This is the main altern ative to 12
0-0-0 as seen in the previous game.
12 ...'ii' b4+
One of the points of ... c5-c4 and
2 S .l:txa2 1 26 .l:txfs lDb41
.•. ... a7xb 5 .
Threatening ... l:[al m ate. 13 lDc3
2 7 'iVa s g61 0-1
Black doesn 't even need the queen
back.
Game summary
A good win by Kh alifm an, utilizing the
thematic queen sacrifice on C 3 . In the
g am e Lalic could h ave improved with
17 'ilke4, after which Bl ack's attack is
less straightforward, even if still suffi
cient. Therefore Black might prefer
14 ... g6 as given in the n otes. It is logical 13 lDe71?
••.
to develop the bishop to h6, especially The sharpest move in this position .
as the white king is sittin g on Cl, and Kh alifman gives it an " ! " in Informant
Black can reach an unclear g am e in a #69, cl aiming th at White is slightly bet
nice way as shown in the an alysis. ter after 13 .. :iVas or 13 ... .i.c5. However,
92
Th e 5 13 Va ria t i o n
I don 't agree with the latter evaluation, tiJf4 iLc6 20 iL e 2 tiJa6 2 1 l:ld2 0-0 2 2
as I think Bl ack is at least equal after J:.hdl tiJC7 2 3 h4 (takin g some space on
13 ... iLcS, so that the choice between the kin g side in the absence of better
th at move and the text is largely a m at pl an s) 23 .. J�b8 24 g4 g6 with an equal
ter of taste. g ame.
After 13 ... iLcs 14 a3l1
14 O-O-O! i s a m ore logical m ove,
fighting for dS, without wasting tim e
o n luxuries like a2-a3.
14 a3 'iWas 15 iLxcs 'iWxcs 16 'iWf2 !

(the queen exchange i s the best here;
otherwise the strong central c4- and
ds-pawns will cause White some prob Now Black h as:
lems in the middlegame) 16 ... 'iWxf2+ a) 14 ... tiJfs was given by Kh alifman
(worse is ... d4 17 0-0-0 tiJc6 18 tiJe4 and is quite sharp but, in my opinion,
'ii'x es 19 'ii'g 3 'ii'x g 3 20 hxg 3 - Kh alif leads to a slightly better position for
man, when White is a bit better; e.g. White if he plays precisely: 1 5 iLf2
20 ...iLa6 2 1 tiJd6+ �e7 2 2 tiJxc4) 17 'iWas ! 16 tiJf4 (after 16 g4 d4! 17 iLxd4
'ittxf2 tiJe7 18 l:r.dl Khalifman claim s a tiJxd4 18 litxd4 iLb4 Black h as reason
slight advantage for White, probably able compen sation for the pawn)
because the black pawn s are on cur 16 ... d4! 17 iLxd4 iLb4 ( Kh alifman), and
rently stuck on the light squares. How n ow 18 'iWxC4! (18 'iWf2 tiJxd4 19 'iWxd4
ever, Black can obtain his usual queen 0-0 20 tiJh s iLxC3 21 'ii'x C3 ':'c8 is un
side counterpl ay with ... :as and ... l:1b8, clear) 18 ... tiJc6 (or 18 ... tiJxd4 19 litxd4
and can fight for d4 with the m an oeu iLxC3 20 'iwxc3 ! 'iWxa2 21 tiJh S ! 0-0 2 2
vre ... tiJa6-C7-e6. In my opinion Black : g 4 g 6 2 3 e 6 ! and wins) 19 tiJfdS
has at least equal ch ances, and some tiJcxd4 20 tiJC7+ is g ood for White; e.g.
times even winning chances if his 20 ... 'itt e 7 2 1 'iWxb4+ ! 'iWxb4 2 2 tiJ7dS+
knight can get to e6 without him losing iLxdS 2 3 tiJxdS+ 'itt e 6 24 tiJxb4 with an
control of dS. For example, 18 ... l:tas 19 extra pawn .
93
b) 14 ... ltJa6 ! is stronger, the m ain 15 l:.a61

..•
idea bein g to support the ds-pawn

with ... ltJC7, and then :
A nice and exact move, taking time

to prevent .i.e3-b6 before continuing
bl) 1 5 a3 'ifb3 16 ltJf4 ltJC7 17 'ii'C 2 with ... ltJfS . Now Black is slightly better.
"ii'X C2+ 18 'iti'xc2 g6 (the simple The central pawn s on dS and c4 can be
18 ... .l:la6 ! ? also comes into considera quite dangerous.
tion) 19 .i.cs (or 19 �b6 �h 6 ! ) 19 ... ltJfS ! 16 'ii'd 2?1
20 �b6 d4! 2 1 .i.xd4 .i.h6 2 2 ltJfe2 ltJe6 This is quite slow and allows Black
23 .i.f2 .i.e3 with compen sation for the to increase his advantage. However,
sacrificed pawn . Black would play ... ltJfs after other
b2) 1 5 ltJf4 ltJcS ( l S ... ltJc7 is still pos m oves too; for in stance, 16 ltJf4 ltJfS 17
sible) 16 ltJfxds (or 16 "ii'd 2 nas, indi ltdl (or 17 0-0-0 ltJxe3 18 Wxe3 �cS 19
rectly covering the ds-pawn with an 'iVel .i.xa3 ! ) 17 ... ltJxe3 18 'ii'x e3 :b6 ! 19
unclear g ame) 16 ... ltJxds 17 :xdS (the 'ii'd 2 d4 with a strong initiative; or if 16
only move) 17 ... ltJa4! (undermining the �d2 ltJfs 17 ltJf4, as in S.Uchitel
rook on dS) 18 .i.d4 .i.e7 19 :d6 ! ? .i.c6 V.Marinin, St Petersburg 1997, then
20 ltJe4 c3 ! 2 1 ltJxC3 0-0 and in practice 17 ... .i.cS 18 'ii'e 2 ltJd4 19 'iVdl 'ii'C 7, fol
thi s position would be easier for Black lowed by ... 'iVxes with a big advantage.
to play, even if objectively unclear. Fin ally, the preventative 16 g4 is met
14 JWa 5 15 'iVf2
•. simply by 16 ... ltJg 6 and Black is clearly
After 1 5 O-O-O? ! ltJfS Black is very better ( Khalifman).
comfortable; or 1 5 ltJf4 ltJbc6 16 .l:tdl 16 ...ltJf5 17 .i.f2 .i.C5 18 l1dl
ltJxe s 17 .i.d4 ltJ S g 6 18 ltJxg 6 hxg6 19 Or 18 0-0-0 d4, followed by ... c4-c3,
"ii'f2 l:ta6 ! 20 'iVg 3 l:te6+ 21 .i.e2 (Z.Zhao and White is close to losing.
L.Moylan, Brisbane 2005) and 21 ... ltJc6 ! 18 ... .i.xf2+ 19 ltJxf2 0-0 20 ltJe2 'iVc7
would h ave been very g ood for Black; Avoidin g the queen exchange, since
e.g. 22 �f2 .i.d6 23 f4 ltJe7 24 0-0 g S . White is behind in development.
94
e 6 6 e 4 exd s 7 eS 'ii'e 7 8 'ii'e 2 t'iJg8 9

t'iJC3 .i.b7 10 t'iJh3 c4 11 t'iJf4
Attacking the ds-pawn immediately.
Instead, 11 .i.g S ? ! is a waste of time as
the black queen often goes to cS any
way. M.Notkin-V.L.lvanov, Moscow 1994,
saw 11 .. :iVcS 12 0-0-0 axbs 13 .i.e3 and
now, rather than 13 ... 'iWb4 transposing
to Game 22, Black can play 1 3 .. :iVc6 !
with a slight advantage. For example, 14
t'iJf4 (or 14 :xdS b4 15 t'iJbs l:1xa2)
21 'iWf4 t'iJe7 22 t'iJg4 t'iJg6 2 3 'iVg3 d6 14 ... t'iJe7 15 'itJbl t'iJa6, when 16 t'iJfxdS ?
Khalifman wasn 't sure whether this t'iJxdS 17 t'iJxdS fails t o 17 ...'iVg6+! 1 8
was the most precise move, but it's 'itJal (or 18 'it>cl .i.xds 1 9 l:1xdS t'iJc7)
good enough to win . 18 ... .i.xds 19 1:txdS t'iJb4 and wins.
2 4 exd6 .l:t.xd6 2 S t'iJd4 f S 26 t'iJf2 t'iJc6 2 7 11 ... 'ii'c S
.i.e2 t'iJxd4 28 l1xd4 f4 29 'iVgs .l:t.e8 3 0
t'iJg4 ':de6 3 1 Itd2 c3 3 2 bXc3 'ili'xC3 0-1
Game summary
In gaining a llawn with 12 t'iJxbs White
falls behind in development and Black
reaches an unclear position with good
compensation . After Ivanov's further
inaccuracies (at moves 14 and 16),
Kh alifman obtain s a dangerous initia
tive, with the white king stuck in the
centre. The result is a nice g ame by 12 a4?
Kh alifman, who was one of the leadin g Too slow. N o better is 12 .i.e3 ? ! d4
Benko theoretician s at th at time. 13 .i.f2 axbs, when Black also h as a
clear advantage.
The critical respon se i s the dang er
Game 24 ous sacrifice 12 t'iJfxdS ! , which again
U.Adianto-J.Polga r looks like a computer suggestion. I n
J a ka rta (ra pid m atch, deed, the computer is quite good in the
ga me 2) 1996 5 f3 e6 variation ! Thi s forcin g sequence
continues 12 ... .i.xds 13 .i.e3 'iWb4 14 a3
1 d4 t'iJf6 2 C4 cs 3 d s bs 4 cxbs a6 S f3 'iVas 15 .i.d2 .i.e6 16 t'iJdS and n ow:
95
bl) 18 ... axb 5 ? ! 19 &i:JC7+ llxC7 20

'ili'xC7 &i:Jc6 21 'ili'xd8+ 'itxd8 22 �xb5
g ave White a favourable endgame in
H . N akamura-M.Vachier Lagrave, Cap
d' Agde 2008. White has a rook and two
strong passed pawns for the two
knights, whereas Black is yet to develop
his kingside. The game continued
22 ... &i:Jxe5 ? (a mistake; but even after
2 2 ... &i:Jge7 2 3 f4 g 5 24 fxg 5 �g7 25 0-0
�xe 5 26 �C3 11g 8 27 h4 �g 3 28 �el
a) 1 6 ...'ii'a 4? ! was played i n a recent .te5 29 b4 White has the better
game: 17 &i:Ja+ 'itd8 18 &i:Jxa8 axb5 19 chances with his passed pawns) 23
&i:Jb6 'ii'a 6 20 �e3 �b4+ 21 axb4 'ili'xal+ �a5+ 'ite7 24 ltc8 (now Black is losing
22 'it>f2 and White was clearly better in as he is too far behind in development,
I .Nyzhnyk-S.5iebrecht, Wijk aan Zee though his next move doesn 't help
2011, due to the vulnerability of the m atters) 24 .. .f5 ? ! 25 �b4+ d6 26 ltxf8
black kin g . Indeed, the g ame ended in 'itxf8 2 7 �xd6+ 'itf7 28 �xe5 &i:Jf6 29
an amusing kin g hunt: 22 ...&i:Jc6 23 Wd2 ! 0-0 1:1c8 30 a4 &i:Jd5 3 1 .l:tal .l:tc2 3 2 �d3
&i:Jxe5 24 'ii'd 6 ! 'ii'xb2+ 2 5 'itg l &i:Jc6 26 .l:td2 33 �c4 ltc2 34 b3 &i:Je3 35 .txe6+
'ii'f8+ 'ita 2 7 'iVc8+ 'itd6 28 �C5+ 'ite5 'itxe6 36 �d4 f4 37 �xe3 fxe3 38 as
29 &i:Jxd7+ �xd7 30 'iVxd7 &i:Jf6 31 Wxc6 'itd5 3 9 a6 11c8 40 a7 J:1a8 41 'it>fl l-0.
c3 32 'iWd6+ 'it>f5 33 �d3+ 'itg 5 34 �e3+ Vachier was out-prepared by Na
1-0. It seems that Siebrecht hadn't stud kamura's n ovel knight sacrifice. How
ied the Nakamura game quoted below. ever, if Black reacts precisely he can
b) 16 ... 'ii'd 8 17 'ii'x c4 lla7 18 1:[cl reach an unclear game:
reaches a critical position : b2) 18 ... �xd 5 ! 19 Wxd5 axb5, keep
ing the extra m aterial, is not conspicu
ously bad for Black. For instance, after
20 �xb 5 &i:Je7 2 1 'ii'd 3 &i:Jbc6 22 f4 g6 23
0-0 llb7 24 �e3 'ii'b 8 2 5 a4 �g 7 26 i.C5
0-0 White has definite compensation
for the piece due to the bishop pair,
more space, and two passed pawn s,
but Black can look to unravel with
moves like ... 1:1e8, ... �f8, and ... &i:Jf5 or
... &i:Jc8 and should n ot stand worse.
12 d4!
•••
96
1 9 ttJxg7+ �d8 20 'iie l 'iVb6

Keeping the queen s on . otherwise
20 ... 'iVxel+ 21 'itt x el ttJb3 2 2 .l:tbl 1:[g8
23 ttJh s ttJds was certainly winning for
Bl ack.
2 1 .i.e3
21 c;t> g l :g8 2 2 ttJh s ttJds is n o bet-
ter.
2 1 ... :g8
White should be careful not to allow

this pawn to start advancin g . In this
variation time is crucial . Black is al
ready clearly better.
13 bxa6 ttJxa6 14 ttJe4 'iVb4+
14 .. .'ii' x e s 15 'iVxC4 fs looks even
stronger.
15 �f2 d 3 1 16 'iVe3 i.cs 17 ttJxcS ttJxcS
22 ttJhS?1
White h ad to try 2 2 b4! ? cxb3 23
ttJh s just to break up the black pawn
duo. After the text he is pretty much
lost.
'ii'x b2+ 23 'it>gl 'ili'xes 24 ttJg3 f6 25 f4
'i!kd s 26 'ii' b 4 ttJe4 2 7 i.b6+ c;t>e8 28 lle1
fS 29 as c;t>f7 30 :c1 :ac8 31 .i.e3 hS 3 2
ttJxh S c3 3 3 ttJg3 d 2 34 l:I a 1 d 1'ii' 3 5
l:txd1 'ii'x d1 36 'iix b7 c 2 3 7 'iVb3+ d S 0-1
Now Black is a pawn up and very
comfortable, with good squares for her Game summary
knights on b3 and e6; whereas White This g ame is a good example of what
has an exposed king and is fin ding the can h appen if White plays inaccurately
c4- and d3-pawn s quite annoying. in the opening. Adianto's 12 a4? ! was
18 ttJhS ttJe7 n ot in the spirit of this line: White
18 ... ttJe6 was solid, but Polgar just shoul d develop rapidly. Instead, Polgar
wants to get her pieces out and attack. was able to use her two central pawns
97
and lead i n development t o win con the b s-pawn in stead, but in so doing
vin cingly. falls behind in development, enabling
Far m ore critical is N akamura's 12 Black again to generate strong coun
lLlfxdS ! , which he used to defeat terpl ay.
Vachier Lagrave. H owever, by playing Finally, Game 24 is a good example
to keep the sacrificed m aterial, Black of how things can go wrong for White
can reach an unclear g am e with accu if he plays too slowly and allows the dS
rate defence. N evertheless, there is still pawn to advance. In stead, Nakamura's
plenty of room for practical tests here. dangerous knight sacrifice, seen in the
notes, leads to very interesting play.
Summary With the suggested improvement for
The 5 f3 variation i s both uncompro Black the g ame can be regarded as un
mising and double-edged. With s ... e6 clear, but I think more theoretical dis
Black responds in the m ost principled putes in this line are required before
way, and after 6 e4 exdS 7 eS 'fie7 8 any definitive conclusion can be
We2 lLlg8 9 lLlc3 i.b7 10 lLlh 3 c4! the reached.
battle becomes very complicated, My overall conclusion is that Black is
where a single mistake can easily m ean doing fin e after the sequence 5 f3 e6.
a loss for either side. However, it remains a critical line in
In Game 2 2 it was seen how a direct the Benko, and one th at requires solid
0-0-0 and l1xdS can be m et by the the preparation from both sides. The game
m atic queen sacrifice .. :iix c3 ! , after quickly becomes extremely sharp and a
which Black obtain s excellent compen single mistake can alter the evaluation
sation . In the n ext g ame White takes of the position dramatically.
98
C h a pter Five
White Plays b S -b6
1 d4 lL'lf6 2 c4 cS 3 d s b s 4 cxbs a 6 5 b6 whereas if Bl ack plays ... d7-d6 and

... lL'lbd7xb6, White will often pl ay a4-aS,
followed by lL'ld2-C4 and lL'lc3-a4-b6.
Black should be careful not to allow
such a blockade.
After S b6 Black has several option s.
The straightforward S ... 'i¥xb6 i s the
move I am recommending, though
S ... g6 an d S ... d6 can lead to the same
positions if Bl ack plays ... 'i¥xb6 l ater on .
In stead, s ... e6 i s unclear, but White can
avoid thi s by pl aying lL'lf3 before taking
White return s the pawn in order to on b S (as he in fact did in Games 2 6
interfere with Bl ack's usual queen side and 28 o f this chapter). s ... aS was a
counterpl ay. The plan of b S -b6 used trend for some time, but I reg ard it
not to be regarded as too dangerous move as dubious as it weaken s the b S
for Bl ack. However, in the early nineties square.
Shirov won some impressive attacking After S ... 'i¥xb6 the two sides eng ag e
games with it, and th e line became in a position al battle, where White pro
more popular. GM Vadim Milov al so ceeds with the pl an s outlined above.
won some g ames in a more position al Bl ack, on the other h an d, will aim for
way. play on the b-file and an active g ame,
White's main idea is to seize control with m oves like ... .ltg4 (if allowed),
over the c4-square, and he will usually .. .l::t b 8 and ... lL'les, or sometimes .. ,'¥la7
follow up with lL'lC3, e2-e4 and lL'lf3-d2- and ... lL'lb6 to ch allenge a white knight
c4. Then if Black recaptures on b6 with on the c4-square. In some position s
the queen, lL'lC4 will g ain a tempo; Black can even pl ay ... g 6-g S to secure
99
the e s -square a s a knight outpost, al his dark-squared bishop, plus counter

though this is n aturally somewhat play on the b-file, even though White
risky. Often the queen will be attacked hasn't accepted the offered pawn . In
on b6, either by a knight on c4 or a my opinion, with dynamic and preci se
pawn or as and, dependin g on the cir play, Black can obtain at least an un
cum stances, my recommendation is to clear g ame.
place the queen on a7, c7 or d8! The 7 e4 d6 8 a4 .i.g7 9 a s "ii b 4!
m ain consideration, as in m any Benko
lines, is th at Black shoul d pl ay actively.
I used the ... 'iIi'xb6 line to win a nice
g am e again st Peter Heine Niel sen
(Game 28). He played the opening inac
curately, allowin g a strong and unex
pected knight sacrifice, after which
there was no defence. So what looks
like a positional struggle can quickly
become sharpened, and Black h as to
keep an eye for tactical opportunities.
Another well -known example is Now it is not so easy for White to
that an early a4-aS can be an swered by defend the pawn in a n atural way. Al
... 'iWb4, attackin g the e4-pawn, while if though the position is unclear from an
l:ta4 there is ... JLd7-b S . This is one of the objective point of view, in a practical
reason s, for in stance, why I prefer a g am e I think Black has a good chance
quick .....xb6 to ... d7-d6 and to get the advantage, since White can
... liJbd7xb6. easily end up in a passive position . So
the early a4-aS cannot be recom
mended for White. Indeed, it seem s
Game 25 th at nowadays it mostly occurs wh en
J.Campos Moreno-V.Topalov White is not too well up on the theo
Aviles 1992 retical nuances of the bS-b6 line.
lo l:ta4
1 d4 liJf6 2 c4 cS 3 d s bs 4 cxbs a6 5 b6 Virtually the only move. 10 'iWc2?
'i!Vxb6 and 10 .i.d3 ? both fail to 10 ... liJXe4,
s ... d6 6 liJc3 "xb6 is another move winning at least a pawn ; while 10 f3
order which often reaches the same does not fit in at all with White's in
position ; e.g. after 7 e4 g 6 . tended set-up, as he now has no mean
6 liJC3 g6 ingful way to use his pie c es.
Black n ow h as a good diagon al for lo ... Wb7
1 00
Wh i t e Plays b S - b 6
Planning ... �d7-b 5, taking advan axb5 1 6 'iixb5 "fiC7 1 7 "fia4 l:ta7 ? ! 18
tage of the white rook's unfortun ate �b 5 .l::[b 8 19 0-0 in L.Van Wely
placement, whkh is one of the m ain L. Riemersma, Dutch Championship,
points of 9 .. :iWb4. Am sterdam 199 5 . In this line 17 ... e 6 !
l1 lLlf3 would have offered m ore counterplay,
The main alternative is 11 �C4 �d7 but in that case Black might prefer
12 l:ta3 'ii'b 4 13 'ii'e 2, when Bl ack can 12 ... e6, attackin g the centre at once. For
choose between : example, 13 lLlC4 exd5 14 lLlxd6 (or 14
exd5 .tg4 15 .te2 �xe2 16 'it>xe2 'ike7+
17 <it>fl lLlbd7 18 .tf4 lLle8 19 'i!i'e2 'i!i'f6
20 'i!i'd2 I:r.b8 was unclear in T.Hillarp
Persson-S.Conquest, Torshavn 2000)
14 ... 'iie 7 15 lLlxc8 ':xc8 16 lLlxd5 lLlxd5
17 'iix d5 lLlc6 18 �d3 ? (but if 18 �e2,
then 18 ... 'iVe8 ! intending .. J:td8 or
...lLld4 with the initiative) 18 ...lLlb4 19
'iVC4 l:td8 20 �C2 .l::[ d4 21 'ili'e2 lLld3+
and Black won m aterial in S.Van Gis
berg en - N . Ristic, Groningen 199 5 .
a) 13 ... 0-0 14 lLlf3 �g4 1 5 l::t a4 (or 1 5 12. . .� d 7 13 .l::[ a 3 � b S 1 4 0-0 lLlbd7 15
0-0 lLlfd7 with a n unclear g ame) l::t e 1
15 .. :i+'h7 16 h3 �xf3 17 'ii'xf3 lLlbd7 18
iVe2 lLle5 19 �a2 was S.Sergienko
L.Milov, Nagykanizsa 1993, and here
19 ... l:r.fb8 is okay for Black.
b) 13 ... �b5 ! ? (anyway!) 14 �xb5+
axb 5 15 'ii'xb5+ 'ii'xb 5 16 lLlxb 5 lLla6 17
f3 I:lb8 with compen sation ; e.g. 18 lLla7
.l::[b 7 19 lLlc6 <it>d7 followed by ... e7-e6, or
18 lLlC3 c4 19 lLlge2 lLld7 20 <;t>f2 lLldc5
and Black has sufficient compensation
for the pawn with the bishop on g7, the
rook on the open b-file, and the knight 1s ...lLle8?1
ready to go to d3. A typical m ove in the 5 b6 line and
11 0-0 12 �e2
••• in the Benko in general . The knight is
White has al so tried 12 lLld2, when on its way to c7, and sometimes to b 5
12 ... �d7 13 %:.a3 �b5 turned out well and d4 (although the bishop o n b 5 i s in
for him after 14 'ii'h 3 ! lLlbd7 15 lLlxb5 the way here), or else to support the
1 01
attack o n the centre with . . .e7-e6 (as i n tiJxbs tiJxb s 24 l1xbS 'it'xds 2 S 'it'xds
the current g ame), o r even .. .f7-fS. tiJxds 26 i.d2 would give him a clear
H ere, however, the knight m an oeuvre advantage with the bishop pair and the
is mistimed and premature. dangerous passed a-pawn.
Better was lS ... l:tfcS 16 h3 (prevent- 2 3 ...l:txel+ 24 'ifxel tiJfxd S 2 S i.a4?!
ing ... tiJg4-eS) 16 ... i.xe2 17 'ifxe2 litabS Keeping control of eS was less im
is tiJd2, and n ow lS ... tiJeS 19 tiJC4 tiJC7 portant than the time White wastes
20 i.g s liteS with an unclear g ame. Two with this bishop. Better was 2S i.c4,
ideas are ... tiJbS , and ... 'ila7 followed by when 2 s ... 'it'aS? ! 26 tiJxds tiJxds 27 'iWd2
.. Jlb4. .l:teS 2S l:td3 tiJb4? 29 ':'xd6 'it'xas run s
16 i.gs tiJdf6?! into 30 i.xf7+! 'iii> xf7 3 1 l::t d 7+ 'iii> g S 3 2
Coverin g the e7-pawn and thus in l:txg 7+! 'iii> x g 7 3 3 'it'd7+, while 2 S ... tiJb4
sisting on his erroneous knight m a 2 6 'fie7 'ifc6 27 i.f4 dS 2S 'fixC7 'it'xC7
noeuvre. It was n ot too l ate to go into 29 i.xC7 dXc4 30 l:ta3 is just equal.
reverse and play 16 ... i.xe2 17 'ifxe2 2 s ... 'ifaS 26 tiJxd s tiJxd s 27 i.bs tiJb4
tiJef6, when the game i s still finely bal
anced. Bl ack h as t h e potential for good
play on the queen side with ... .:tfbS,
.. :flIe7 and ... l:tb4, while White h as some
squares to play for too.
17 l:.b3 tiJC7 lS tiJd2 l::t a dS 19 tiJC4 e6?!
19 ... 'ii'a 7 would restrict White to a
slight advantage.
20 tiJa3 exd s 2 1 exd s l:.deS 2 2 tiJaxbs
axbs
2S i.d2?
Allowing the knight to switch to d4
gives Black a definite advantage. White
had to try 2S .te7, and if 2S ... :cS (or
2S ... tiJC2 29 'iWdl) 29 i.xd6 tiJc2, then 30
'it'e7 ! tiJd4 31 i.C4! tiJxb3 32 i.xf7+ 'iii>h S
33 i.es 'ii'x as 34 i.xg7+ leads to a
draw, but 2 9 .. .'iMS ! is still good for
Bl ack.
2S ... tiJC2 29 'iWfl tiJd4
2 3 .txbS? A great square for the knight.
N ow White begin s to go wrong . 2 3 3 0 l:1g3 tiJxbs 3 1 'iWxbs 1:.bS
1 02
Wh i t e Plays b 5 - b 6
In stead, 4 cxbS a 6 S b 6 "ii'x b6 6 liJC3

g 6 7 liJf3 i.g7 8 e4 d6 9 liJd2 0-0 10
i.e2 liJbd7 11 0-0 l'1b8 12 a4 would be a
more standard route to the position at
m ove 1 2 .
White's m ove order with 4 liJf3 is
significant in th at it enables him to
avoid the line 4 cxbS a6 S b6 e6!? True,
he must then be prepared for any of
4 ... b4, 4 ... i.b7, 4 ... bxC4 and 4 ... e6, which
all have their adherents. But none of
After Bl ack win s the b2-pawn the that m atters h ere as I am advising
rest is relatively easy. The c- and d Black to play the con sistent 4 ... g 6 in
pawns are ready to advance, and White any case, aiming for a n orm al Benko
has no real counterplay. after S cxbS a6.
32 'ild3 l:txb2 33 h4 d5 34 i.C3 C4 3 5 4 ... g6 5 cxb5 a6 6 b6 d6 7 liJC3
ii'd1 i.xC3 36 Ibc3 'ii'xa 5 3 7 :e3 'i¥d2
38 'ikf3 'ii'd 4 39 g3 llb1+ 0-1
Game summary
In the opening White plays a4-aS too
early, as 9 ...'iWb4 forces him to cover the
e4-pawn in a rather awkward way with
lIa4, after which Black has pretty much
equalized. Nevertheless, Black's posi
tion doesn 't just play itself, and Topa
lov's inaccurate handling of the early
middlegame allowed White to gain the 7 ...liJbd7
advantage after all . It was only when It doesn't really m atter that Black
Campos also went wrong that Topalov h as delayed takin g on b6 up to n ow,
managed to come out on top. but I think he should definitely play
. .. 'ii'x b6 here. Partly because it is more
flexible and gives Black the option of
Game 2 6 ... i.g4 if White delays liJd2 for too lon g ;
V.Milov-V.Topa lov but m ore because i t avoids the difficul
Prague (ra pid) 2002 ties n oted at move 10 below.
After 7 ... 'ii'xb6 ! play can continue 8
1 d4 liJf6 2 C4 cs 3 d 5 b5 4 liJf3 e4 i.g7 (not yet 8 ... i.g4? ! because of 9
1 03
'ii'a4+ .i.d7 1 0 'illb 3 'ii'C 7 1 1 .i.e2 .i. g 7 1 2 1 1 lIa4 'iVb7, he would be unable to
tiJd2 and White is slightly better) 9 i.e2 continue with the thematic ... .i.d7-b 5
0-0 10 0-0 (here 10 tiJd2 is m ore accu since his knight is already occupying
rate, when 10 ... tiJbd7 11 0-0 l:tb8 12 a4 the d7-square.
tran sposes to the g ame) 10 ... i.g4 11 h 3 10 ... 0-0 11 .i.e2 .l:tb8 12 0-0
(or 1 1 tiJd2 .i.xe2 12 'iVxe2 tiJbd7)
1 1 ... .i.xf3 12 .i.xf3 tiJbd7 and Black h as
equalized.
8 e4 .i. g7 9 a4 'iYxb6
12 Wa7
•..
Black will h ave to retreat his queen

in any case after White plays a4-aS or
tiJC4, and so elects to does so at once, in
10 tiJd 2 order to keep the option s open for his
Milov could h ave exploited Bl ack's other pieces. The a7-square is chosen
m ove order by playin g the sequence 10 so as to leave c7 free for a knight after
as! 'ii'a 7 1 1 tiJd2 0-0 1 2 tiJC4 tiJe8 13 ... tiJe8-c7, or even ... tiJb6-a8-c7 (as in
i.e2 l:tb8 14 0-0, tran sposin g to his the n ote move 14). 12 ... "ii'd 8 and
l ater g ames against Anand and N aka 12 .. .'ii' C 7 are also played, and the for
mura, given in the n ote with 13 as tiJe8 mer is seen (via another move order) in
below. the next game; the latter makes less
In stead, Black might retreat the sense here, however, as after 13 as the
queen to d8, aiming to tran spose to the queen can get in the way and often has
n ext g am e after 10 ... Wd8 1 1 i.e2 0-0 to move agai n .
12 0-0 l::tb 8 1 3 tiJd2, but in that case 13 tiJC4
White h as other options such as 1 1 13 as is a major altern ative here;
h 3 ! ? 0-0 12 i.C4 tiJ e 8 1 3 0-0 tiJ C 7 14 .i.f4 and indeed, as we saw in the note to
with a space advantage. move 10, White may already have
Of course in the previous g am e played this move en rou� e to a tran spo
(Campos-Topal ov) w e saw Black answer sition below. From this exact position
a4-aS by ... 'iVb4, but here after 10 ... 'iVb4 Black h as a few choices:
1 04
ttJxb6 f5 (17 ... i.. d 7 might improve but it

doesn 't equalize; e.g. 18 i.. d 3 e6 19 ,Ubl
lUe8 20 i.. e 3 exd5 21 exd5 ttJd4 2 2 l:tel
and the b6-knight still gives White an
edge) 18 ttJxc8 (after 18 exf5 ? ! i.. xf5 19
i.. g 4 �7 20 l:ta4 ttJd4 Black was fine in
V.Milov-V.Anand, Bastia rapid 2005,
and l ater went on to win) 18 ... l:tbxc8 19
exf5 g xf5 (19 ... l:txf5 20 i.. g 4 skewers the
rooks) 20 i.. C 4 'it>h 8 2 1 lIel i..f6 2 2 'iVh 5
a) 13 .. J:tb4? ! 14 ttJC4! is good for

White. For example, after 14 ... ttJe8 1 5
i.. d 2 ttJe 5 16 ttJb6 White has achieved
one of the main strategical goals in this
line: to make a blockade on b6 with one
of the knights.
b) 1 3 ... ttJe8 14 ttJC4 ttJC7 15 ttJa4
again intends to plant a knight on b6
and block the b-file.
(Milov's improvement h as yielded

him a definite advantage; al so good
was 22 'ikd3 llb8 23 I:tbl - Milov)
22 ... ttJd4 23 lIa3 ttJC2 24 11h 3 e5 2 5 .l:.fl
ttJd4 26 i.. x a6 l:tC7 27 i.. d 3 e4 28 i.. h 6
':' g 8 2 9 ':g 3 lIcc8 30 i.. C 4 i.. e 5 3 1 a 6 f4
3 2 lIxg 8+ ':'xg 8 3 3 �h l 'ilkd7 34 g 3 ttJf3
3 5 i.. e 2 ttJd4 3 6 g xf4 i.. g 7 3 7 i.. g 4 'fIe7
38 i.. x g7+ 'iWxg 7 39 h 3 ttJe2 40 'iWf5 e 3
41 'iWe4 h 5 42 'ilkxe3 h x g 4 43 'ili'xe2 'iVh 6
Anand managed to win from here 44 f3 gxh 3 45 ':'g l ':'xg l+ 46 'it>xg l
as Black with the plan of ... ttJb5-d4 and 'ii'xf4 47 a7 'ii'd4+ 48 �h l 'iWa4 49 'ilke7
.. .f7-f5. However, Milov l ater improved Wdl+ 50 'it>h 2 WC2+ 5 1 �xh 3 'ii'f 5+ 5 2
on that game to beat Nakamura, and 'it>g 3 'iWg 6+ 5 3 'it>h4 1-0 V.Milov
the current status is that 1 3 ... ttJe8 al H . N akamura, Philadelphia 2006. This
lows White a sm all advantage. For in g am e is a good illustration of what
stance, 15 ... ttJb5 16 ttJab6 ttJxb6 17 Black should avoid in thi s line.
1 05
c) 1 3 ... liJes ! i s my recommendation 24 1:[c2 dS 2 5 liJb6 'ii'd 6 and in this un

here. By moving his knight to eS Black clear position Black should not be
pre-empts his opponent's knight ma worse.
noeuvre to c4 and b6, which would c3) 14 <;i(h l g S ! ? (risky but necessary;
n ow simply be swapped off. The obvi Black m aintains the knight on the eS at
ous problem is that the knight might the cost of weakening his king side) 15
be chased away with f2-f4, but at the liJc4 (after 1 5 'ii'c 2 'ii'b 7 16 b3 e6! 17
m om ent it i s safeguarded by tactics on dxe6 fxe6 18 .l:.bl "fie7 19 liJC4 liJxC4 20
a7-g 1 diagonal (one of the reasons be bxc4 l:txbl 21 'ii'x bl liJd7 22 liJdl liJe s
hind 12 ... 'ii'a 7), and otherwise Black Bl ack is doing fine with another strong
g ains the tim e to attack the centre with knight on es) l S ... liJxc4 16 i.xC4 l:tb4
... e7-e6. With 5 b6 White tries to play in 17 'ii'e 2 h 6 18 liJa4 i.d7 19 liJb6 i.bS
a positional way, but tactics still play with an unclear position; e.g. 20 f4
an essential role. 'ikxb6 ! ? 2 1 axb6 i.xc4 2 2 'ii'e l i.xfl 2 3
'ii'xfl l:txb6 24 fxg s hxg s and Black has
enough compen sation for the queen .
So 13 ... liJes is a sharp, uncompro
mising move which leads to an unclear
game.
13 liJb6
...
Cl) 14 f4 liJd3 ! 1 5 �xd3 c4+ 16 <;i(h l

cxd3, and even if the d3-pawn drops off
Bl ack will obtain excellent compen sa
tion with the two bishops and rooks on
the open b- and c-files.
(2) 14 h3 e6 1 5 f4 (or l S liJC4? ! liJxC4
16 i.xc4 exds 17 exds J:.b4 18 b3 liJd7 This move is a useful option wh en
and Black is doing fin e with his active White hasn't played M-aS .
pieces) lS ... liJd3 ! (a nice move to play 14 liJa 5
and n ot really something White should Alternatively:
allow) 16 dxe6 i.xe6 17 fS liJxcl 18 a) 14 liJe3 lle8 15 a � liJa8 (heading
lIxcl gxfs 19 exfs i.ds 20 liJxds liJxds for C7-b S-d4) 16 liJC4 liJd7 17 f4 liJC7 18
2 1 liJC4 i.d4+ 22 <;i(h l liJf6 23 i.f3 'ilVC7 <;i(h l liJb s 19 liJa4 liJd4 20 liJab6 liJxb6
1 06
21 liJxb6 liJxe2 2 2 'ii'x e2 V2-V2 A thematic break i n the centre.

V.Dimitrov-M. Perez Fungueiro, Ponte Unlike in m any queen's pawn open
vedra 2003, though I think Bl ack is ings, the ds-pawn is not supported by a
slightly better after 22 ... e6, with an ac colleague on c4, while if White takes on
tive position and the bishop pair e6 the black pieces become very active,
(unless White wishes to give up his especially after ... d6-ds opens up the
blockading knight). position further. This is what Black is
b) 14 liJa3 liJa8 1 5 liJc4 .:tb4 16 liJa2 aiming for, whereas White h as n ot
':xC4 ! ? (otherwi se 16 .. J::tb 8 is fine) 17 m an aged to m ake a blockade on b6
i.xc4 liJxe4 with compensation in and his knight on as is now m isplaced.
N.V. Pedersen-L.Van Wely, Am sterdam 17 dxe6 i.xe6 18 b3
2002, as it hard for White to do very Not 18 i.xa6 1i'a7 19 1i'd3 c4 and
much ; indeed, Bl ack l ater won after I White loses a piece.
overplayed my hand. At that time I was 18 ... dS 19 exd s liJfxd s 20 liJxd s liJxd s
in spired by the games of Gleizerov and Black is slightly better, due to his
Dreev, who are still playing 5 b6 suc more active pieces and the offside
cessfully as White, but I went back to white knight.
the Fianchetto Variation . 2 1 l::t a cl i.d4 2 2 liJC4 liJf4 2 3 i.xf4 'iWxf4
14 ... i.d7 lS 'ii'c 2
15 liJc6 i.xc6 16 dxc6 :fc8 17 as
liJa8 18 i.f3 liJc7 19 liJa4 liJe6 20 i.e3
liJe8 is also unclear; e.g. 2 1 es i.xes 22
liJb6 lIC7 (or just 2 2 .. Jbb6 ! ? with com
pensation) 2 3 %:tbl liJd4 24 i.xd4 cxd4
with ideas of ... liJf6, ... e7-e6 and ... d6-
ds.
lS ... 'ii'C 7 16 i.d2 e6
Now Black has the bishop pair and

an active position, and Topalov goes on
to convert his advantage into a win .
The b3-pawn and the white king are
obvious targets here.
24 i.d3 �g7 2S :cel 'iWf6 26 i.e4 h S
Aiming t o create weaknesses o n the
white king side.
27 �hl .l:r.b4 2 8 f4 i.xc4 29 bXc4 J:te8 30
107
f S gs 3 1 � d S ':'xe1 3 2 ':'xe1 lib2 3 3 ing the a-pawn is recommended by GM

'iWe4 :f2 34 'iVd3 :xfS 3 S 'iWd1 g4 36 g3 Kiril Georgiev in Squeezing the Ga mbits.
liIf2 3 7 �e4 �c3 38 l1g1 �d4 39 �e1 Note that Topalov's ... 'iVa7 is less
'iig s 40 :f11 appropriate against a quick ttJd2 and
Losing. 40 'iVcl was n ecessary. a4-as, as after 1 1 .. .'ii' a 7 12 ttJC4 White
40 .. :ii'd 2 1 41 'iWxd2 l:txf1+ 0-1 has already achieved base camp to
wards his desired blockade; whereas
Game summary Black is unable to respond with either
With 12 ...'iWa7 and 1 3 ... ttJb6 Topalov ... ttJb6 or ... ttJe s, and so has nothing
obtained good play and won a nice better than 12 ....l:tb8 1 3 0-0 ttJe8 14 0-0,
g ame. White could h ave prevented reaching the slightly unfavourable po
... ttJb6 with 1 3 as, but then 1 3 ... ttJes ! is sition s from Milov-Nakamura in the
al so unclear, as seen to the notes n otes to the previous game.
above. It is characteristic of thi s line 8 e4 g6 9 SLe2
that Black can achieve at least an un Or 9 ttJd2 �g7 10 a4 0-0 11 as 'iVd8
clear g ame with active play. as in the previous note. Note that
In the opening, rather than commit 1 1 ... 'ilVb4 12 1:ta4 'iVb7 is again ineffec
the knight to d7 so early, I would prefer tive, because the bl ack knight is ob
to play ... 'ii'x b6. Then Black keeps the structing the d7-square, thus prevent
option of answering a quick a4-aS with ing ... �d7-bS .
... 'iVb4, ... 'iVb7 and ... .i.d7 (as in Game 9. . .�g7 1 0 ttJd2 'ii'd 8!
2 5), while in some cases the bishop can
al so go to g4.
Game 2 7
J.Hellsten-V.Baklan
German League 1997
1 d4 ttJf6 2 c4 cS 3 d s bs 4 cxbs a6 S b6
d6
Or S ... 'ilVxb6 followed by ... g 7-g6 and
... d7-d6. Another prophylactic retreat, pre
6 ttJC3 'ii'x b6 7 ttJf3 ttJbd7 empting ttJC4 or a4-aS. The natural
H ere 7 ... g6 8 e4 �g 7 9 ttJd2 0-0 10 10 ... 0-0 would give White the option of
a4 ttJbd7 11 as 'iVd8 would be my move 11 ttJC4 followed by 12 �f4, whereas
order, when 12 �e2 l:.b8 13 0-0 trans now 11 ttJC4 might be m et by 11 ... ttJb6.
poses to the g ame. The plan of advanc- 11 a4 0-0 12 0-0 lIb8 13 as
1 08
Now there is no ... liJb6 in an swer to Not n ow 1 6 ... liJd4? 1 7 liJxd6 ! . But
liJC4, so Black must pl ay for ideas like 16 .. .fs ! works especially well with the
... liJe8-C7-bS-d4 - and sometimes ... e7-e6 bishop on g s as .. .l::t xfS g ains a tempo.
or .. .f7-fS . For in stance, 17 exfs (17 f3 h6 18 i.. e 3
13 ... liJe8 14 liJc4 liJC7 15 liJ a4 liJ b 5 fxe4 19 fxe4 :xfl+ 20 'ili'xfl liJd4 i s also
unclear) 17 .. .l::t xfS 18 h4 liJes 19 liJab6
liJxC4 20 liJxc4 liJC7 2 1 liJb6 i.. x b2 22
]::t a 2 .i.d4 2 3 .i.g4 l:tf7 24 liJxc8 .l:lxc8 2S
.i.xc8 .xc8 and Black h as g ood com
pensation for the exchang e with his
active pieces and the weak white
pawn s at as, dS and h4.
17 .l::t b 1
The immediate 17 liJab6 seem s
more accurate.
White is now ready to pl ant a

knight on b6, while Black is ready to
put his on d4, after which the advance
.. .f7-fS is more effective as the ds -pawn
undermined. Obviously havin g an op
posing knight on such a great square
as d4 can quite dangerous for White.
Again active pl ay is the way for Black.
16 i.. g 5
16 liJab6 should be met by 16 ...liJd4!
(Georgiev only covers 16 ... liJxb6?!, which 17 .. :f1C7
may well offer White a slight advan H ere Black might try 17 ... h6 18 .i.f4
tage) 17 .i.d3 liJxb6 18 liJxb6 e6 19 .i.e3 g s 19 .i.e3 e6 20 dxe6 ':'xe6 2 1 f3 liJd4
(19 .i.C4 'it>h 8 20 'ili'd3 exds 2 1 i.. x ds fS ! 22 liJab6 .i.b7 with an unclear g ame.
gave Black good play in I.Golichenko Bl ack h as g ood activity and can even
V.Baklan, Ukrainian Championship, Kiev an swer i.. d 3 with ... g S-g4.
2008) 19 ... exds 20 exds l:te8 21 :bl 18 liJab6 liJf6 19 liJxc8 :bxc8 20 i.. d 3 ? !
'ifh4 22 liJxc8 lIbxc8 2 3 .i.xa6 :a8 24 Correct was 20 .i.xf6, intending liJb6
'it'd3 l1e4 with an unclear game; one and a l ater b2-b4 with a slight advan
idea is 2 S ... liJf3+ 26 gxf3 :g4+ 27 fxg4 tage. The a6-pawn is a long-term
iixg4+ with a draw by perpetual check. weakness and it is n ot easy for Black to
16 ... l:te8 obtain counterplay h ere.
1 09
20 ... ltJd7 Losin g at once, though White was

Now the position is unclear again . much worse in any case.
2 1 ..td2 e6 2S ... ltJec31 29 bxc3 ltJxC3 30 ..tf3
Faced with threats at d1 and e2
White elects to give up his queen, but
his slight initiative is easily neutrali zed.
30 ... ltJxd1 31 ..txd S+ �hS 32 nb7 'ii'xa s
3 3 llxd1 ':'C7 34 ..tf7 �e1+ 3 5 �h2
:xb7 36 :dS+ 'ii'x dS 3 7 ..txdS ':'xf7 0-1
Game summary
This g ame shows some nice ideas
again st an early a4-aS, the primary one
being ... ltJc7-b S-d4, exploiting the fact
The thematic break. As the g am e th at the white infantry can no longer
continuation shows, the pawn roller control b S . Other themes include ... e7-
starting with ... e7-e6 ancl ... d6-dS can e6, or .. .f7-fS (as seen in the note to
be quite dangerous for White. m ove 16), while ... ltJd7-es is som etimes
22 dxe6 fxe6 23 h4?! possible too.
White shoul d keep an eye on the The position from the opening
centre; e.g. 2 3 i.f4 'ii'c 6 24 ltJe3 ..td4 phase was unclear. In the early middle
and the g am e is still in the bal ance. g ame Hellsten m ade a few inaccura
23 dS
••. cies, allowing Black to obtain a strong
Now Black's advancin g pawn s give initiative with ... e7-e6 and ... d6-dS, af
him a definite advantage. ter which it was all one-way traffic.
24 exd s exd s 2 s ltJe3 c4 26 ..te2 ltJf6 27
ltJg4 ltJe4 2S i.gS?
Game 28
P .H.Nlelsen-N.V.Pedersen
Da n ish Tea m
C h a m pions h i p 2006
1 d4 ltJf6 2 ltJf3 cS 3 ds bs 4 c4
Note the rather unusual route to
the Benko. White's other fourth moves,
of which 4 ..tg s is the most significant,
will be examined in Chapter Eight.
4 ... g6 5 cxbS a6 6 b6
110
I used to play the bs-b6 line myself queen then covers d 6 and e s , as well as
as White (see the notes to Game 26 for as. Importantly, Black cannot be pre
example). In this game I ventured the vented h ere from playing ... lDb6 to
Benko as a surprise weapon and it ch allenge the knight on c4, so White
turned out very well. Since then I have has more difficulty in blockading the b
been playing the Benko more regul arly, file.
generally al so with good results. 12 �f4 :b8
6 .. :ifxb6 7 lDC3 d6 8 e4 � g7 9 lDd2
A thematic m ove, bringing the rook

If White is intending to play thi s to the h alf-open b-file. At this point I
way he might a s well d o so at once, saw that the forthcoming sacrifice was
ruling out any ideas of ... �g4. probably problematic for White.
9 0-0 10 �e2 lDbd7 l1 lDc4
... 13 'ili'd2?1
1 3 0-0 i s sounder, when 1 3 ... lDb6
can be m et safely by 14 lDe3 or even 14
e s ! ? In stead, I like 1 3 ... lDe s ! ?, which
m ay not seem that logical, but it's
based on concrete tactics (the b2-pawn
is h anging) and leads to an unclear
position . For example: 14 lDxes (or 14
�xes dxe s intending ... lDe8-d6)
14 ... dxe s 15 .te3 (or 15 �Cl e6 with
counterpl ay again st the centre)
ls ... .:xb2 16 lDa4 J:[xe2 ! ? 17 'i¥xe2
11 'i¥C7
•.. lDxe4 18 lDxcs lDC3 19 'i¥d2 e4 20 lilacl
Both 11...'i¥a7 and 11...'i¥d8 are pos 'ifes 2 1 lDb3 lDxds 22 �d4 was seen in
sible (in the style of the previous two L.Seres-L. D . N isipeanu, Eger Agria 1995,
games), but I prefer 11...'i¥C7 as the and n ow 2 2 ... 'i¥fs 2 3 �xg 7 Wxg 7 24
111
'iWd4+ (or 2 4 11C5 �b7) 2 4... tiJf6 2 5 tiJC5 18 ...'iWxa4 19 'iWxe4 �f5
.i.e6 26 a4 :a8 27 tiJxe6+ 'ilixe6 looks to
give Black reasonable compensation
with two pawns for the exch ange.
13 ...tiJb6 14 tiJe3 ?
Now White can 't cover the knight

on c4 in a good way.
2o 'iWxe7
If 20 'iWe3 1Ixc4 21 b3, then 21 .. :iVd7 !
Other m oves allow Black t o equal 2 2 �xc4 �xal and Black is better.
ize; e.g. 14 0-0 tiJxC4 1 5 �xC4 tiJg4 and 20 .:te8 21 'iWxd6 �xb2! 22 tiJxb2
••.
Bl ack is doing fine. 22 .l:tel would be met by 22 ... �c3

14 ...tiJa4! ! with a decisive advantage.
The idea is to take the pawn on M, 2 2 l:txb2 2 3 �e3
.••
check on as, and play ... llb4. Then This fail s spectacularly to defend
White h as weaknesses at a4, b2, f2 and the e-file.
f4, while his kin g is prevented from 23 .. Jlxe2 !
castling. Back in 2006, the engines I
tested this position on after the g ame
were unable to see th at 14 ... tiJa4 is
g ood for Black.
15 tiJxa4 tiJxe4 16 'iWC2 'iVa 5+ 17 �f1
l:tb4
Attacking the knight on a4, and
with a veiled threat against the bishop
on f4.
18 tiJC4
Not 18 tiJC3 ? tiJxC3 19 bxc3 lIxf4,
while after 18 �C4 'ii'x a4 19 'ili'xe4 �f5 24 'it>xe2
20 tiJxf5 g xf5 Black regains the piece Or 24 'iWc6 �d3 ! 2 5 'ifxa4 !;t2xe3+ 26
with a clear advantage. 'it>g l l:tel+ and mates.
112
24 ... 'i¥C4+ 2 5 �e1 'ifc3+ 26 �e2 iLg4+ reasonable defence i s the artificial
27 f3 'ii'xe3+ 28 �f1 iLf5 0-1 lookin g :a4, but then, after retreatin g
There is no serious defence to the queen, Bl ack can play ... iLd7-b S and
29 ... iLd3+ and mates. equalizes, as seen in G am e 2 5 . Conse
quently, this variation is n o longer so
Game summary popular for White.
In the opening White played an early In the m ain lines White mostly plays
tLld2-C4, to which I replied ... 'ifC7. Then tLld2-C4 and/or a2-a4-as, so that Black
13 0-0 would have been interestin g h as to remove the queen from b6. Just
from a theoretical perspective, when where the queen should go depends on
the unorthodox-looking 1 3 ... tLle s ! ? the precise circumstances.
leads to an unclear game. In stead, Niel In Game 2 6 White played a2-a4,
sen mistakenly delayed castling, allow tLld2, iLe2 and 0-0. H ere ... '¥Ia7 is a good
ing the dangerous, if unexpected, move, intending either 1 3 tLlC4 tLlb6 or
knight sacrifice with 14 ... tLla4 ! ! , after 13 as tLles ! . The l atter is supported by
which I don 't think there is really a the tactical idea of an swering f2-f4
good defence for White. The game was with ... tLld3 ! , as the knight i s indirectly
a good illustration of the activity Black defended by the ... c S-c4 check !
is aiming for in this line, as well as In Game 2 7 White plays a quick
showing how a couple of in accuracies tLld2 and a2-a4-aS (the most accurate
by White can sometimes be disastrous. move order is given in the n otes). In
thi s case the queen does better to drop
Summary back to d8, when Black can generate
In the 5 b6 line White i s tryin g to pl ay counterplay by ... tLle8-c7-b S-d4, exploit
in a more positional way again st the ing the hole on bS, with further ideas
Benko Gambit, using the b6- and c4- of ... e7-e6 or .. .f7-fS .
squares in particular to try and restrict Finally, in G ame 28, White pl ayed a
Black's counterplay and prove an ad quick tLld2-c4 without a2-a4. H ere it
vantage; the advance e4-e S is some m akes sense to play ... 'ifC7, from where
times a theme too. However, after the queen control s more squares, while
... 'iVxb6 Black can still aim for active the white knight on C4 can be chal
play on the b-file, with a rook on b8 lenged b y ...tLl b 6 o r even ... tLle s ! ?
supported by the bishop on g 7 . Often Overall, m y evaluation is that the
Black is able to obtain at least equality b S -b6 line offers ch ances for both sides,
by tactical means. and as long as Black plays with due
White should take care not to play care and precision, I don 't think that
a4-aS too soon, as ... 'ii'b 4 gives him White should really be able to prove a
problems with the e-pawn . The only theoretical advantage here.
113
Cha pter Six
Oth er Options for White
1 d4 lbf6 2 c4 cS 3 d S bS 6 e3 and 6 lbc3 (see Game 3 1).

The rest of the ch apter covers
White's various declinations on move
4: the Trompowsky-like 4 .i.g 5 (Game
3 2), the unthreatening 4 lbd2 bXc4 5 e4
(Games 33 and 34), the un ambitious 4
a4 (Games 3 5 and 3 6), the unch alleng
ing 4 f3 (Game 3 7), and th e rather du
bious 4 e4 (Game 3 8). As the reader will
gather from my choice of descriptors,
none of these lines offers White a seri
ous chance of a th eoretical advantage.
Here we round up White's remain
ing ways of playin g again st the Benko.
First is the Zaitsev Variation, 4 cxb 5 Game 29
a6 5 lbc3 axb 5 6 e4 (Game 29), which is M.Hochstrasser-D.Pikula
sh arp but n ot theoretically dangerous Pontresi n a 2000
for Black. Indeed, it is White who needs
to think about equalizing here. 1 d4 lbf6 2 c4 cS 3 ds bs 4 cxbs a6 5
The next two g ames feature 4 lbf3, lbC3
which is a significant altern ative to 4 The so-called Zaitsev Variation, with
cxb 5 - for White, th at is. For Bl ack I which White aim s for rapid develop
suggest the consistent 4 ... g 6 5 cxb 5 a6, m ent and tactical play from the open
offerin g to tran spose to the m ain lines ing. It is not so popular these days an d
after 6 bxa6 or 6 b6. Other ideas for should perhaps be con �ined to rapid or
White include an early 'ilVc2 (examined blitz games as, in many cases, Bl ack
in Game 30) and the tricky move orders takes the initiative an d the question is
114
O t h e r Op t i o n s fo r Wh i t e
whether White can equalize. a ) 7 dxe6 (giving u p the centre like

this is not usually advisable for White)
7 .. .fxe6 8 e4! ? (the sh arp option; in
stead, 8 e3 ds is just equal : Bl ack has a
nice centre with ideas of ... lLlc6, ... i.e7,
and possibly ... e6-es , ...lLle4, or some-
times ... l:tb8) 8 ...lLlxe4 ( Bl ack takes up
the challenge; otherwise 8 ... lLlc6 is
more quiet, though still unclear) 9
'ilVh s+ (critical ; after 9 lLle2 i.e7 10
lLlec3 lLlxc3 1 1 bxc3 0-0 12 i.d3 .l:tf7
and ... d7-ds Black is doing fine, with his
S axbs 6 e4
... control of the centre) 9 ... g 6 10 'ifes
This line is not critical . Instead, 6 'ii' a s+ 11 'it>dl (or l 1 lLlc3 lLlxC3 1 2 i.d2
lLlxbs i.a6 7 lLlC3 g 6 would transpose :g8) 11 ... lLlxf2+ 12 'it>e2 lLlxh l 1 3 lLlC7+
to the main lines if Black is willing to 'it>f7 14 i.d2 "iVa4 15 'ii'x h 8 with an un
do th at. However, there is an additional clear g am e in O.Uwira-L.Milov, Bad
option in 6 ... e 6 !, which I would prefer Homburg 2009.
as it equalizes more simply; indeed, b) 7 lLlc3 exds 8 lLlxds i.b7
White must be careful n ot to come out
worse from the opening.
After 6 ... e6 Black might be regarded
as having an improved version of the 5
b6 e6 line. True, the white knight has
some active possibilities on bs, but
Black has more control over ds. There
can follow:
9 e4? ! (but after 9 lLlxf6+ 'iVxf6 10

'ii'b 3 i.c6 1 1 lLlf3 ds 1 2 g3 i.d6 Black
h as good compensation with his strong
centre and chances n the queen side)
9 ... lLlxe4 10 i.C4 'ifas+ 11 'it>fl lLlc6 1 2
i.f4 0-0-0 1 3 lLlf3 i.d6 14 "iVcl? ! (14
i.e3 i s still unclear) 14 ... i.xf4 1 5 'iVxf4
'uhe8 and Black was slightly better in
115
G .5osonko-H .5chussler, German League position has been played. However,

1982. Black h as an interesting novelty which
6 b4 7 1ZJbs d6
... my friend I M Rasmus Skytte has found
for this book: 11 ... lZJxe4 ! ? 12 i.xe4 i.g7,
playing again st White's weak points at
al, a2 and b S .
Takin g the e-pawn this time would

be a terrible mistake, due to 7 ... lZJxe4??
8 'ii'e 2 1ZJf6 9 1ZJd6 m ate !
8 i.f4?1 For example, 13 lIbl I:l.xa2 1 4 0-0
The m ain line. White h as ideas of i.a6 15 i.d3 0-0 16 i.C4 'ilfd7 17 'ii'd 3
e4-eS, while still hopin g for 8 ... lZJxe4 9 l:ta8 18 i.f4 l:tas 19 lZJxd6 exd6 with a
'ii'e 2 lZJf6 ? ? 10 lZJxd6+ and wins, al clear advantage to Black, or 13 i.e3
though 9 ... g S ! h ere i s n ow unclear - 'ii'd 7 ! (the knight on bS is toast) 14
and Black does even better to play th at lZJbd4 (14 a4? ! bxa3 15 lZJxa3 'i!i'g4 16
m ove straight away. Therefore, in stead i.d3 l:txa3 ! 17 l:txa3 'ii'b4+ is even
of the g am e continuation, White worse) 14 ... 'i!i'g4 15 lZJg s 'iWxdl+ 16
shoul d probably try: l:txdl cxd4 17 i.xd4 i.xd4 18 l:txd4
a) 8 i.d3 (or 8 lZJf3 lZJbd7 9 i.d3) ':'xa2 19 0-0 h6 20 l:txb4 1ZJxds 21 i.xds
8 ...lZJbd7 9 1ZJf3 g6 10 0-0 i.g7 11 a4 0-0 hxg s and Black is a pawn up. White's
12 ':'el lZJg4 1 3 h3 lZJge s 14 1ZJxes lZJxes best reply m ay be 13 i.d2, intending
1 5 i.fl i.b7 with a level position . 1 3 ...'iWd7 14 a4 bxa3 15 i.c3 ! , but after
b) 8 i.C4 lZJbd7 (avoidin g the sh arp 13 ... i.xal 14 'ili'xal :g 8 15 i.d3 'iWd7 16
lines after 8 ... g 6 9 eS dxe s 10 d6 etc is a4 bxa3 17 lZJxa3 i.a6 18 i.xa6 lha6
clearly Black's strongest option, espe Bl ack looks slightly better even then,
cially as these kin ds of tactics are what due to the exposed knight on a3 and
White is aiming for with the Zaitsev weak pawn on dS.
Variation) 9 1ZJf3 lZJb6 10 i.d3 g6 11 b 3 ; 8 gsl
...
this position h as occurred m any times Although there are- other option s,
and so far only 11 ... i.g7 with an equal this is the strongest move here. Now
116
Bl ack is slightly better, due to the vul 1:txa6 20 Wxg s l:tb6 2 1 l:txb6 Wxb6 with
nerable white knight on b S and pawns g ood winning ch ances for Black due to
on e4 and dS. - the weaknesses at a3 and dS; e.g. 2 2
liJ e 2 "iWbl+ 2 3 liJ c l liJd7 24 0-0 liJf6 2 5
l:tel "iVb2 2 6 liJe2 :g 8 2 7 WfS 'itf8.
9 liJxe4 10 i..f4 liJf61
...
9 iLxg5
White has al so tried:
a) 9 e s ? ! gxf4 10 exf6 liJd7 ! 11 Wf3
(no better is 1 1 fxe7 Wxe7+ followed by The simplest continuation. There is
12 ... liJe s, or ll liJh 3 liJxf6 12 liJxf4 iLh 6) n o n eed at all to enter the complica
11 ... liJxf6 12 iLc4 iLg 7 1 3 liJe2 (V.Milov- tion s of 10 ... i.. g 7 11 'ii'e 2 liJf6 12 liJxd6+
A.Tzermiadianos, Heraklio 1993) 'itf8 13 liJxc8 Wxc8, even if Shirov and
13 ... liJd7 ! with a clear advantage to Khalifman have both played this way.
Bl ack. After the text Black h as a slight ad
b) 9 iLe3 liJxe4 10 iLd3 "iWb6 (saving vantage: the f8-bishop can safely take
the piece tactically) 11 'ii'e 2 l:.as 12 a4 up residence on the long diagonal,
(after 12 i.. x e4 'ii'x bs 13 WxbS+ :xbS whereas White h as problems with both
14 i.. x g s :as 1 5 liJe2 i.. g 7 Black is the knight on b S and the weak dS
clearly better due to his activity and pawn.
queenside pressure) 12 ... bxa3 1 3 bxa3 ll .i.C4
i.. g 7 14 lilbl i.. a 6 1 5 f3 (15 liJf3 allows Now 11 'ili'e2 lla6 g ets White no
Bl ack to swap down to a favourable where, while after l l liJf3 i.. g 7 1 2 i.. d 3
endgame with l S ... liJC3 16 liJxC3 i.. x C3+ l:tas 1 3 0-0 liJxds 14 i.. g 3 0-0 1 5 'ili'C2 h6
17 i.. d 2 'ii'x bl+ 18 i.. x bl i.. x e2 19 i.. x C3 16 l:tfdl i.. e 6 Black is clearly better with
i..xf3 20 gxf3 l:tbS 2 1 i.. c 2 0-0 2 2 h4 g4 his active pieces, greater central con
23 :g l f6, when White's bishops are trol and extra pawn .
in sufficient compensation for the ex 11 i.. g 7 12 liJe2
...
change) l S ... liJC3 16 liJxC3 i.. x C3+ 17 Or 12 liJf3 liJbd7 and Black is doing
i..d 2 i.. x d2+ 18 'ii'x d2 'ii'C 7 19 i.. x a6 fine, since the attackin g try 1 3 liJxd6+! ?
117
exd6 1 4 �xd6 should ultimately prove against the black kin g .

un successful after 14 ... tDe4 15 'ii'e 2 f5 19...tDxC4 20 bXc4 tDe5 21 f4?
16 tDd4 (or 16 tDd2 'ii'f6 17 tDxe4 fxe4 2 1 'ii'e 2 was called for; it was no good
18 'iVxe4+ �d8) 16 ... 'ii'f6 17 tDb5 �f7 18 l ettin g the c4-pawn go.
�C7 �g6, as in B.5alvermoser 2 1 ...tDxC4 2 2 f5 tDxd2 2 3 'ii'xd2 �e5
S.Camilleri, Bad Worishofen 200 5 .
1 2 tDbd7
•..
The b6-square i s great for a black

knight here - attackin g the bishop on
c4 and the pawn on d5.
13 tDg3?!
It i s n atural that White shoul d want to
place his knight on f5 or h 5, but this
isn't really an option unless Black is
especially careless.
13 ...tDb6
White's try at an attack has been
easily parried and Black now has a de
cisive advantage.
24 :f3 f6 2 5 tDh5 l::tf7 26 g4 �h8 27
tDf4 l::t g 7 28 �h1 :ta8 29 .:th3 �xf4 30
'ii'xf4 'ii' b 6 3 1 tDe4 a2 0-1
Game summary
The Zaitsev Variation with 5 tDC3 axb 5
6 e4 can be dangerous if Black pl ays
incorrectly, but if he responds in the
Bl ack is clearly better, as White can right way the risks are reversed! In that
h ardly defend his existing weaknesses, case White's best is probably to defend
n or safeguard his knight without the the e-pawn with 8 �d3 which leads to
further weakening a2-a4. an unclear game. The old main line
14 b3 0-0 15 0-0 �d7 16 �g5 l1a 5 with 8 i.f4 n ow seems to be too slow
The knight on b 5 is the obvious tar because of 8 ... g 5 ! , when Black takes
g et and h as no escape unless White over the initiative with the better posi
pl ays a2-a4. tion . The white knight is quite vulner
17 a4 bxa3 18 tDc3 tDg4 19 �d2?1 able on b5, while the trade of the e4 for
19 'iVc2 followed by tDh 5 might be a g 5 h as left the d5-pawn as a serious
better chance of getting some pl ay weakness. Hochstrasser was unable to
118
O t h e r Op t io n s fo r Wh i t e
find any real counterplay and lost a ttJf3 0-0 9 0-0 and White is slightly bet-
rather one-sided game. ter. After 4 ... 'ii'a s+ we h ave:
Game 30
E.Ba reev-V.Topalov
Sa rajevo 2000
1 d4 ttJf6 2 C4 cS 3 ds bs 4 ttJf3
a) 5 ttJC3 ttJa6 ! 6 cxb S (now after 6

a3 b4 or 6 e4? ! b4 Black is slightly bet
ter) 6 ... ttJb4 7 'iVd2 jLb7 8 e4 (8 a3
ttJbxdS 9 ttJxdS 'ii'x d2+ 10 jLxd2 jLxds
produces a bal an ced endg ame) 8 ... e6
and Bl ack is doing fine. H ere 8 .. ttJxe4? !
.
i s tempting, but after 9 ttJxe4 ttJC2+ 10

This is White's second-most popular c;to>dl 'ii'x d2+ 11 jLxd2 ttJxal 12 ttJC3 the
move (after 4 cxbs). It's prime aim is to knight on al is lost and White's knights
avoid certain lines by Black, such as an must be stronger th an the rook,
early ... e7-e6 in the b S-b6 variation. regardless of his wrecked pawn
However, White does have some inde structure after ... ttJb3 .
pendent option s as well, in particular b ) 5 .1i.d2 (probably the only serious
the idea of an early 'l'c2 . attempt to get an edge)
In fact White can play 4 'iVc2 right
here. This looks a bit odd, but Black
must still play precisely in order to
equalize. In my opinion, the rarely
played 4 .. :iVas+! is the best reply, as it
forces White to place his pieces in a
slightly awkward way. Again st other
moves I think White has good chances
of an edge; for instance, 4 ... ttJa6 5 a3 !
bXc4 6 e4 e6 7 jLxC4 exds 8 exds or
4... bxc4 5 e4 d6 5 jLxC4 g 6 7 f4 jLg 7 8
119
S . . .b 4 6 e 4 d 6 7 .td3 tDbd7 8 f4 (8 .txb2 19 d6 ! "iIIx a3 ? 20 'iWe4! and

tDf3 g 6 9 0-0 .tg7 10 h3 0-0 1 1 a3 .tb7 m ates) 18 l:lael .td7 19 .te4 �f8 20 d6
is TOughly equal) 8 ... g6! (8 ... g S ? ! was 0-0-0 21 dxe7 lIxfl+ 22 ':'xfl .l:%.e8 with
tried in S. Koutsin-D.Trifonov, Kiev 1999, an unclear position.
but then 9 fxg s tDg4 10 tDf3 .tg7 11 0-0 4 g6
...
tDge s 12 .te2 was good for White, and

9 e S dxe s 10 fxg s might be even
stronger) 9 tDf3 .tg7 and then :
This is Black's most con sistent and

popular continuation, still intending to
offer the g ambit pawn after 5 cxbS a6.
bl) 10 0-0 0-0 1 1 h 3 (if 1 1 I:Iel "fIC7 From the other side, White must al so
12 e S , then 12 ... dxe s 13 fxe s tD94) be prepared for 4 ... b4, 4 ... .tb7, 4 ... bxc4,
1 1 ... 'ii'C 7! (keeping an eye on e s ) 12 lIel 4 ... d6 and 4 ... e6, but we don't have to
as 1 3 e S (or 1 3 a3 ':b8 14 .te3 l:te8 1 5 worry about those. The only drawback
tDbd2 e S 16 fxe s tDxe s 17 tDxe s I1xes to the text is that, should White decide
18 axb4 axb4 19 .tf4 �e8, followed by to play an early e2-e3 or tDc3, Black is
... tDd7 with an equal g ame) 1 3 ... dxes 14 n ow committed to a fianchetto. We
fxe s tDh S ! is okay for Black, since 15 g4 will have a quick look at these lines in
is m et by ls ... tDxe s ! (lS ... .txe s 16 .tfl ! the notes to White's 6th move below.
is less clear) 16 tDxes .txe s 17 gxh S s cxbS
.txb 2 ! 18 "fIxb2 "fIg 3+ and draws. 5 'iWc2 is a positional way of playing
b2) 10 e S dxe s 1 1 fxe s tDg4 12 e6! is and leads to a g ame with equal
the direct try and gives White some ch ances. One example: S ... bxC4
initiative, but it is h ard to foresee any (s ... "fIas+ is possible here too) 6 e4 ..tg7
advantage without the queen side 7 .txC4 d6 8 tDC3 0-0 9 0-0 .ta6
pieces in play. For example, 12 ... tDde s (9 ... tDbd7 is a sound altern ative) 10
1 3 tDxe S tDxeS 14 exf7+ tDxf7 1 5 0-0 tDd2 (or 10 tDbs "fId7 11 a4 ..tb7, in
..td4+ 16 �h l tDeS 17 tDa3 (or 17 .te4 tending ... tDa6-b4) 10 .. :ltJfd7 11 .txa6
"fIa6 ! ) 17 ... a6 (not 17 ... tDxd3 ? ! 18 Vxd3 tDxa6 12 tDC4 tDb6 13 'i!Ve2 tDxC4 14
120
'iVxC4 'iVb6 1 5 l:tbl l:tab8 16 J.g s lUe8 8 g 3 d 6 9 .i.g2 and 8 e 4 .i.xfl 9 'ltixfl d6
17 :fel ttJb4 18 'ii'e 2 ? ! (18 .i.f4, plan reach the main lines of Chapters One
ning e4-eS, is' still okay for White) and Two respectively; 6 b6 'ii'x b6 7 ttJc3
18 ... 'fWa6 (now Bl ack's pieces are well i s the m ain line of Chapter Five; while 6
placed and he is slightly better) 19 'ii'd 2 e 3 and 6 ttJC3 are examined in the n ext
'iVd3 20 l:tedl 'ii'x d2 21 l:txd2 .i.xc3 ! ? g ame.
(beginning a correct combination 6 axb5
•••
based on White's weak back rank, Previously Topalov h ad pl ayed

though the preliminary 21 ... h 6 ! would 6 ... .i.g7 7 e4 axbs 8 .i.xbs 'ii' a s+ 9 ttJc3
have made this even stronger) 2 2 bXc3 .i.a6 10 .i.xa6 ttJxa6 11 0-0, when Black
ttJxdS 2 3 l:txb8 ':xb8 24 f3 f6 25 l:txdS did not h ave full compen sation for the
fxg s 26 eS l:tb6 27 exd6 exd6 28 ':xg s pawn in V.Korchn oi-V.Topalov, Ant
c;t>f7 with a difficult rook endgame for werp 1997. This time he was better
White, which he was unable to hold in prepared and showed that Black is fin e
E.Vorobiov-V.Zvjagin sev, Serpukhov after his strong 8 .. :ii'a 4! . Incidentally, I
2008. searched for other ideas for Bl ack, but I
5 a6 6 'fWc2
... think it is difficult to equalize without
this m ove.
This is one of the addition al option s

provided by White's 4 ttJf3 move order, 7 e4
and has been tried by such uncom N aturally 7 'ii'x cs is critical , since
promising players as Korchnoi, Yer this was one of the ideas behind "C2,
molin sky and Ivan Sokolov. It is a but after 7 ... e6 8 'ii'xbs (or 8 d6 ttJc6 9
"greedy" continuation where White is ttJbd2 "as 10 e3 .i.a6 and White can
often out for the c-pawn, but Black can h ardly hope to h an g on to the d-pawn)
expect good compen sation if he plays 8 ... ttJxdS Black h as good compensation
accurately. thanks to his active piece pl ay, aided by
Otherwise 6 bxa6 .i.xa6 7 ttJC3 .i.g7 the exposed queen on b S .
121
7 Ji'a 5+ 8 �d2
. • 11 ... liJg4 12 �f4
Or 8 liJbd2 liJa6 9 �e2 liJb4 10 'ifbl 12 liJb s 0-0 13 �C3 f6 14 e6 dxe6 15
�g7 1 1 0-0 0-0 and Black is fine; e.g. 12 �C4 �a6 is al so unclear.
a3 liJa6 1 3 b4 'iVb6 14 �b2 d6 with an
unclear game.
8 ...'ii'a 4!
12 ... f6
12 ... d6 ! ? is an interesting altern a
tive, when 13 exd6 (or 13 liJb s 0-0 14
This was a novelty at the time. Now exd6 exd6 and Black is doing fine)
on 9 'ifxcS, Bl ack can g ain time on the 1 3 ... a3 ! (the point) 14 'iii> d 2 (other
queen after 9 ... 'ii'x e4+ 10 �e3 (or 10 moves give Black the edge) 14 ... axb2 15
�e2 liJa6 11 'ii'xbs 'ili'xds 12 'ii'x ds .l:i.el e6 16 dxe6 �xe6 17 �bS+ liJd7 18
liJxds 1 3 liJC3 liJac7 14 0-0 �g7 - liJg s 0-0 19 liJxe6 fxe6 20 �xd7 .l:i.xf4 21
Bareev) 10 ... liJa6 1 1 'ii'x bs liJxdS, ag ain �xe6+ 'iii>f8 is again unclear.
with an unclear g ame. 13 exf6 liJxf6 14 a 3
9 'ii'x a4 bxa4 The safest move, preventing ... a4-a3
Although this pawn is potentially and thus fixing the pawn on M.
vulnerable and might be rounded up 14 0-0 15 0-0-0 d6
...
by liJC3 and �bS , as we will see it is n ot Black can even improve here with
so easy for White to arrange that. The l s ... liJe4! 16 liJxe4 .l:i.xf4 17 liJfd2 (or 17
pawn can even be an asset if Black gets liJxcS d6 18 liJd3 .l:i.c4+ 19 'iii>b l �fS with
to play ... M-a3 at some point, sup compen sation - Rogozenko) 17 ... d6, as
ported by the bishop on g 7 . in 5. Ern st- D. Rogozenko, Dieren 2001,
1 0 liJ C 3 � g 7 11 e5 when Bl ack h ad the initiative and a
The immediate 1 1 �b S ? ! would be slight advantage.
met simply by 11...a3, while 1 1 a3 d6 12 16 l:tel lita7 17 �b5 �a6 18 �xa4?!
�d3 liJbd7 1 3 �bS ! ? 0-0 14 �xa4 �a6, 50 White fin ally win s the a-pawn,
followed by ... J:tfb8, gives Black excel but to no avail. He should limit him self
l ent compen sation for the pawn. to 18 �xa6 liJxa6 19 lle2 with equality.
122
O t h e r Op t io n s fo r Wh i t e
24. . .tLla s 2 S lib6 tLl b 3 + 26 <ito>C2

26 <ito>b1, keepin g away from checks,
was a better try.
26 ... tLld4+ 27 tLlxd4 l:[xf2+ 28 :!d2
l:Ixd2+ 29 <ito>xd2 .i.xd4 30 .l:.b8+ <ito>f7 31
l:i.b7+ <ito>g8 32 .l:tb8+ <ito>f7 33 l:tb7+ <ito>g8
34 tLlf6+ .i.xf6 35 .i.xf6 h S 36 :g7+ <ito>f8
37 l:lxg6 cJA7 3 8 l:.h6 d S 39 .i.gs .l:txh6
40 .i.xh6 .i.f1 41 g3 d4 42 a4 <ito>e6 43
'it>e1 .i.d3 44 <itf2 <ito>ds 45 <ito>f3 c4 46
.i.d2 c3 Yz-Yz
18 ... .i.c4
Black could win the pawn back at Game summary
once by 1B ... tLlg4! 19 .i.g 3 .i.xc3 20 bXc3 With the strong novelty B ... 'iVa4! Topa
.i.C4 21 .i.c2 .i.xds with the better game. lov equalized quite easily and, as the
But even after the text Black has suffi game went, could even have obtained a
cient play with his active pieces and the slight edge with either l s ... tLle4 or
exposed ds-pawn and f4-bishop. 1B ... tLlg4. In stead, Bareev reached a
19 .i.gs tLlxd s 20 .i.bs .i.xbs 21 tLlxd s technically favourable ending, though
.i.C4?! was un able to do anything with it and
21 ... tLlc6 ! 22 tLlxe7+ <ito>h B was still the g am e ended in a draw. My impres
fine for Black. sion is th at B ... Wa4! more or less n eu
22 ':'xe7 lIa6 23 litd1 tLlc6 24 l:tb7 tralizes the 6 'ii'C 2 line, and I don't think
we will see too much of it in the future.
The lines with an earlier 'iVc2 are
perhaps more interesting, but after
either 4 'iVC2 'ii'a s+! or 4 tLlf3 g6 5 'ii'C 2
bXc4 (or again s ... 'iVas+) Black can ob
tain an unclear g am e with bal an ced
chances.
Came 31
A.Sorin-D.Serrnek
Preventing ... .l:tbB. White might be World Student Team
said to have some advantage n ow, Cham pionship, Pa ra n a n a 1993
though it is hard to prove it, given
Black's still active pieces. 1 d4 tLlf6 2 c4 cS 3 d s bs 4 tLlf3 g6 5
1 23
cxbS a 6 6 e3 White h as one more attempt at a

tricky move order with 6 4:Jc3 axb s 7 d6
(otherwise 7 4:JxbS ..ia6 8 4:JC3 return s
to the m ain lines), but this is not a
theoretical threat at all for Black. In fact
after 7 ... ..ib7 it is White who has to
think about equalizing :
By holding back e2-e3 until this

mom ent White is tryin g to tran spose to
the 4 cxbS a6 5 e 3 variation with Black
committed to playin g ... g 7-g6 (whereas
S ... ..ib7 was my recommendation in
Chapter Three). Although the n atural
6 ... ..ig7 7 4:JC3 is probably okay for a ) 8 e4? ! 4:Jxe4 9 4:Jxb S (or i f 9 4:Jxe4
Black, there seems to be no reason to ..ixe4 10 ..id3, as in Y.Gozzoli-A.wirig,
acquiesce to the opponent's wishes in Hyeres 2002, then simply 10 .....ic6 11
this way. I n stead, Black can try to ex 0-0 e6 and Black is slightly better)
ploit the omission of 4:Jc3 by playing 9 .. :iWas+ 10 ..id2 4:Jxd2 11 4:Jxd2 ? ! (but
6 ... axb s 7 ..ixbs 'ili'as+ 8 4:JC3 ..ia6, if 11 4:JC7+ 'it>d8 12 4:Jxd2, then
which, 12 ... exd6 ! 13 4:Jxa8 ..ih 6 14 ..ie2 ..ixa8
with strong play for the exch ange)
11...4:Ja6 12 'ifb3 ..ih 6 1 3 4:JC3 ..ixd2+ 14
'it>xd2 ..ic6 was good for Black in
G . Hertneck-M.5choeneberg, Germ an
League 1990.
b) 8 4:Jxb S 'ii a s+ 9 4:Jc3 exd6 10 ..id2
was seen in T. Radjabov-S.Halkias, Greek
Team Championship 2002, when
10 ... 'ifb6 11 e4 ..ig7 12 ..id3 0-0 fol
lowed by ... 4:Jc6 or ... d6-dS would have
been fine for Black.
as we will see in the game, offers c) 8 e3 exd6 9 4:JxbS ( after 9 ..ixb S ? !
him good compensation . dS 1 0 0-0 ..ig 7 11 ..i e 2 0-0 Black was
1 24
O t h e r O p t i o n s fo r Wh i t e
slightly better in Se.lvanov-A.Khalif ently Sorin was worried about his oppo
man, St Petersburg 1997, due to his nent's compensation since he opted for
nice centre, more active pieces and a safe line with little chance of gaining
space advantage) 9 ... d5 10 .i.e2 d6 1 1 an edge: 14 e 5 dxe5 1 5 t"Llxe5 t"Llxe5 16
b3 .tg 7 12 .i.b2 0-0 1 3 0-0 l:. e 8 and l:txe5 'ilVd6 17 .i.f4 t"Llh 5 18 l:txh 5 'ii'xf4 19
Black's centre is hardly weak. ':h 3 lIfd8 and Black soon regained his
D.Blagojevic- I .Marinkovic, Niksic 1997, pawn with no further problems.
continued 14 'ilVC2 t"Llc6 15 .to ? ! (15 a3 9 'ilxa6 10 We2
•..
would be a better try) 1 5 ... d4! (opening In D.Bl agojevic-S. Kasparov, Cutro
up the g ame with a nice initiative) 16 2006, White tried the direct 10 e4 d6 1 1
exd4 t"Lld5 17 .i.b2 ? ! (White had to pl ay e5 dxe 5 12 t"Llxe 5 .i.g7 1 3 'iWd3 0-0 1 4
17 .tC4) 17 ... t"Llf4 and Black was clearly 0-0 'ili'xd3 1 5 t"Llxd3 t"Lla6, when the
better. The game concluded 18 l:tfel ? weak d5-pawn and open a- and b-files
( 1 8 .i.dl was necessary) 18 ... t"Llb4 1 9 g ave Black his usual compen sation . The
'iVdl ':xe2 20 l:[xe2 .i.xf3 2 1 �e8+ 'iVxe8 g ame continued 16 .te3 (if 16 l:tdl,
22 'ii'xf3 ':xa2 0-1. tryin g to protect the d5-pawn, then
6 ... axbs 7 .i.xbs 'ilVa s+ 8 t"Llc3 .i.a6 16 ... .l:tfd8 17 t"Llf4 t"Llb4 18 a3 t"LlC2 19
.l:tbl t"Llg4 with compen sation) 16 ... l:tfb8
17 �adl (after 17 t"Llxc5 ':xb2 Black h as
active play again st the weak a- and d
pawns) 17 ... c4? ! 18 t"Lle5 ':xb2 19 t"LlxC4
l:lc2 20 t"Llb6 and White was slightly
better. Instead, Black should h ave
played 17 ... t"Lld7 ! , preventin g t"Lle5, with
ideas of ... c5-c4 and ... t"Llac5, or ... t"Llb6-
c4, when he still h as good play.
10 .i.g7 11 e4 d6
...
9 .i.xa6
In later games Sorin switched to 9
.te2, but Black has few problems here
either. For instance, 9 ... .i.g 7 10 0-0 d6 11
e4 t"Llbd7 12 l:tel 0-0 13 .i.xa6 'iVxa6 was
A.5orin-R.Felgaer, Rosario 2000, reach
ing a standard-looking Benko position
where Black's typical ideas include
... l:tfb8, ... t"Llb6-c4/a4, ...t"Lle8-c7-b5, or
sometimes ... c5-c4 and ... t"LlC5-d3. Appar-
125
12 lbd2 2 2 0-0 �d4+ 2 3 cith l lbe5 (winning a

12 lbb5 0-0 1 3 a4 lbbd7 14 �d2 piece) 24 lbg 5 llxC4 2 5 litbl lbf6 26 d6
l:.fb8 15 0-0 lbe8 16 �C3 lbC7 17 �xg 7 'ii'x d6 27 �f4 'ii'd 5 28 l:tbdl l:1Ca4 29
was played in Z.Jasnikowski-U.Nielsen, l:!.xd4 lIxd4 30 'ii'x e5 ':dl 0-1 A.5orin
Aarhus 1996, and here 17 .. .'itxg 7 G.Malbran, Buenos Aires 1997.
shoul d be fin e for Black. For instance, 14 ... :fb8 15 l:[Cl lbb6 16 lbxb6 lixb6
18 I1a3 lbxb5 19 axb5 'irb6 20 l:.xa8
':xa8 21 e5 ltb8 ! ? (21 ... dxe 5 22 lbxe5
lbf6 is m ore straightforward, and if 2 3
l:.dl, then 2 3 . . ..:ta4! 24 lbc6 l:te4 2 5 "ii'd 3
c4 2 6 Wd2 "iixb 5 2 7 f3 c3 ! 2 8 'ii'x C3 l1C4
with equality) 22 e6 (or 22 exd6 'iWxd6
23 l:tdl lbf6 and the d5-pawn will soon
drop off) 22 ... lbf8 23 exf7 litb7 24 lbd2
"iix b5 2 5 lbC4 'ii'd 7 (on the way to f5
and possibly d5 or f7) 26 llel 'iWf5 with
an unclear g ame. Play might continue
2 7 lba5 l:tb6 2 8 b 3 'ii'x d5 29 'ili'xe7 .:ta6 Black has typical compensation for
30 lbC4 'ilxf7 31 We2 Wf6 with chances the pawn with his active pieces and
for both sides. Black h as a slightly ex play on the a- and b-files - and, as is
posed kin g , but also nice central pawn s usual for the Benko, this continues into
and good squares for the knight on d4 the endg ame.
or f4. 17 b3 lbd7 18 lba4 l:tb7 19 'ilxa6 ':xa6
12 0-0 13 lbC4 lbbd7 14 �d2
•.. 20 cite2 f5
Another Sorin g am e went 14 f4? ! (as Once more we see this standard
we h ave seen before, this m ove is often thrust in the Benko endgame, seeking
risky for White in the Benko) 14 ... litfb8 to undermine the white centre.
(Bl ack is already fine) 1 5 e 5 dxe 5 16 21 exf5 gxf5 22 l:lhdl citf7 23 .:tc2 lbe5
fxe 5 lbe8 17 �f4 ? ! (but 17 e6 fxe6 18 24 f4 lbg4 25 h 3 lbf6 26 �el h5 27 citf3
'ii'x e6+ 'ii'x e6+ 19 dxe6 lbb6 is more :ta8 2 8 litC4 lbd7 29 �h4?!
than comfortable for Bl ack) 17 ... .:tb4! 2 9 litc2 lbf6 would repeat, though
18 b 3 (forced) 18 .. Jbb3 19 l:tcl ? ! (19 Black might al so try 29 ... �d4 with
axb3 'ii'x al+ 20 lbdl would offer more compen sation.
ch ances of hol ding the g ame, though 29 ...lbb6
Black is at least slightly better) Now Bl ack breaks in on the queen
19 .. J:tb4! (back again; now the white side.
centre is collapsin g ) 20 lbe4? (20 lbdl 30 lbxb6 J::[x a2 31 I:r.el Itxb6 32 ':'xe7+
was n ecessary) 20 ... .i.xe5 21 .i.d2 l:ta4 citf8 33 b4?
126
3 3 .l:te3 would still give White g ood ing the 5 e3 .i.b7 l i n e s o f Chapter
hopes of making a draw. Three. In return Black can exploit this
33 ... .i.d4 specific move order by pl ayin g 6 ... cxbS
7 .i.xbs 'iVas+ 8 liJC3 .i.a6. after which
he obtain s the usual active pl ay for the
pawn . As in m any lines of the Benko.
Black's compen sation continued into
the endgame and in the end it was
White who went astray and lost.
Occasionally 6 liJC3 axbs 7 d6 is
tried as well. but if Black plays accu
rately. startin g with 7 ... .i.b7. it seem s to
be White who h as difficulties in equal
izing .
Suddenly mate is a theme and
White is about to lose the b4-pawn . After examining 4 liJf3 in the previ
34 lie6?! ous two g ames we h ave n ow pretty
After this Black is just winning . The much dealt with all the critical varia
last chance was 34 l:lh 7 llba6 3 5 'it?g 3 tion s in the Benko. What remain s are a
.tf2+ 36 'it?h 2 .i.xh4 3 7 lIxh S .i.e7 3 8 few sidelines which some players like
bxcS dxcS 39 l:txfS+ 'it?g7 40 l:te s .i.f6 41 to try from time to time. None of these
.l:te3 .i.d4 42 ':'g 3 + 'it?f6. when Black is are particularly ch allenging from a
much better but there is still some theoretical perspective. but of course
work to do. it's as well to be prepared.
34 ...l:tba6 3S .i.e7+
35 l:txd4 lit6a3+! 3 6 l:te 3 cxd4 3 7
l:txa3 .l:1xa3+ would have been equally Game 32
hopeless in the long run . E.Grivas-A.Khaliflna n
3S ... 'it?f7 36 'it?g3 .i.f2+ 3 7 'it?h2 :6a 3 38 Len i ngrad 1989
.l::[c 1 .i.g3+ 39 'it?g1 l:ia1 40 1:1f1 l:lxf1+ 41
'it?xf1 :a1+ 42 'it?e2 l:te1+ 43 'it?f3 l:txe6 1 d4 liJf6 2 c4 cS 3 ds bs 4 .i.gS
44 dxe6+ 'it?xe7 4S bxcS h4 46 cxd6+ White is tryin g to play in a Trom
'it?xe6 47 d7 'it?xd7 48 'it?e3 'it?c6 49 'it?f3 powsky-in spired way against the
'it?d S SO 'it?e 3 'it?c4 0-1 Benko. which can be dang erous against
an unprepared opponent. Superficially
Game summary it resembles the line 2 liJf3 cS 3 dS bS 4
With 6 e3 White attempts to manoeu .i.g S liJe4 in Chapter Eight. but here the
vre the game into his territory by avoid- weakness of the queenside dark
127
squares (due t o White's C2-C4) enables a) 6 tiJd2 ? ! bXc4 7 "'C2 tiJxg s 8 hxg s
Black to obtain immediate counterplay g 6 9 "C3 ? ! (but 9 e4 .i.g7 10 .i.xC4 d6
with ... tiJe4, .. :ii'a5+ and ... g7-g 5, and in also gives Black a slight advantage)
fact h as good chances to g ain the ad 9 .....xc3 10 bxc3 .i.g7 11 :1cl d6 12 e4
vantage. tiJd7 13 tiJxC4 tiJb6 14 tiJf3 was played
in N .Murshed-D.5ermek, Moscow
Olympiad 1994, and was shortly
drawn ; but after 14 ... h S Black must be
at least slightly better.
b) 6 .i.d2 is hardly very ambitious,
and Black can g et an easy game quite
comfortably, but it is objectively the
best move here. For example, 6 ... tiJxd2
7 "xd2 "xd2+ 8 tiJxd2 bxc4 9 tiJxC4 d6
10 tiJf3 tiJd7 11 e4 g6 12 0-0-0 lLlf6 13
.i.d3 lLlg4 with an equal game.
4 ...tiJe41 5 ...'ii'a 5+ 6 lLld2 g51
The most testing move. White must
n ow be careful n ot to fall behind in
development.
5 .i.f4?1
Surprisingly, after this obvious
m ove Black m ay already h ave the edge.
5 .i.h4? "a5+ 6 tiJd2 g5 i s similar but
even worse for White. Instead, there is
5 h 4 ! ? (a thematic idea from the Trom
powsky) 5 .. :ii'a5+ and then :
7 .i.e3?1
After this Black is already slightly
better, but other tries are fail to give
White anything either:
a) 7 "c2 gxf4 8 "xe4 bxc4 9 lLlgf3
d6 10 'ii'xf4 .i.g7 and Bl ack's position
was preferable in E. Kahn-J . Brustkern,
Budapest 2002.
b) 7 b4! ? 'ii'x b4 8 lIb 1 'ii'C 3 9 llb3 can
be met by 9 ... tiJxf2 ! (9 ... 'ii'f6 10 lLlxe4
128
'iVxf4, as in A.Botsari-So. Polg ar, Thessa 8 ... d 6 (8 ... g4 9 "c2 ! f s 1 0 tLlh4 is al so

loniki Olympiad 1988, is less clear after unclear) 9 b4 cxb4 (or 9 ...'ifxb4 10 J:tbl)
11 'iVbl ! ) 10 ItxC3 (or 10 �xf2 'ii'f6) 10 tLlxe4 dxes 1 1 e 3 tLld7 1 2 cxbs fs 1 3
10 ... tLlxdl 11 �xb8 tLlxC3 12 �es :g 8 tLlg 3 tLlb6 14 tLl d 2 (14 tLlxe s ? ! b 3 + 1 5
(or even 12 ... tLlxa2 ! ? 13 �xh 8 b4) 1 3 "d2 "xd2+ 16 'iii> x d2 � g 7 i s good for
�xC3 b 4 1 4 �b2 � g 7 1 5 �xg 7 ::'xg 7, Black) 14 ... g4 1 5 tLlb3 'ifa3 16 �e2 and
when the queenside pawns give Black there is still everything to play for.
the better ch ances. 7 . bxc4 8 'ifC2 tLlf6
..
c) 7 �es :g8 and now: Al so possible was 8 .. .fs ! ? 9 f3 tLld6

10 �xg s �g 7 11 �f4 tLla6 12 �xd6
exd6 and Black is doing fin e with his
lead in development and the bishop
pair.
Cl) 8 �C7? ! 'ifxd2+ (the simplest

and best solution) 9 "xd2 tLlxd2 10
'iti>xd2 bxc4 11 e3 �a6 12 tLle2 d6 and
Black is doing fine.
C2) 8 b4? ! (fighting for the initiative, 9 �xg5?!
but objectively not the strongest) The fourth indifferent move with
8 .. :ii'xb4 9 tLlgf3 (9 l:.bl ? "xd2+ is this piece fin ally gives Black a clear ad
clearly better for Black) 9 .....as ! (plan vantage. Preferable was 9 'iix C4 �g 7 10
ning ... g s-g4, ... d7-d6 and possibly %1Cl, although 10 ... tLla6 11 �xcs tLlxcs
... bsxC4; the immediate 9 ... g4?! allows 12 'ii'x cs 'ii'x d2+ 13 'iii> x d2 tLle4+, fol
10 llbl ! tLlc3 11 l:txb4 tLlxdl 12 l:lxbs lowed by ... tLlxcs and ... �xb2, would
gxf3 13 exf3 tLlxf2 14 �xf2 d6 1 5 �al still be somewh at favourable for Black
�h 6 16 tLle4 and White had the edge in due to the bishop pair.
N.Murshed-J . H odg son, London Lloyds 9 ... tLlxd 5 10 "XC4 tLl b 6 11 "c3
Bank 1992) 10 'iic 2 fs 11 llbl a6 and Keepin g the queens on with 1 1 "C2
White has very little for the pawn . �g7 12 e 3 tLlc6 is n o improvement for
c3) 8 tLlgf3 ! (this development move, White.
preparing b2-b4, seem s the best here) 11 :iIi'xc3 12 bXc3 � g7
••
1 29
it could hardly be long before Black

won m aterial somewhere. In despera
tion White gives up two knights for a
TOok, but it doesn't ch ange the inevita
ble result.
26 ':'c2 .u.xb1 27 tiJxb1 .i.xb1 28 .l:tc1
.i.d3 29 a4 tiJbC4 30 i.b3 tiJa s 31 i.a2
c4 3 2 i.e3 tiJec6 33 i.b1 i.xb1 34 11xb1
tiJb3 3 S l:tb2 i.xC3 36 ':c2 tiJcd4 0-1
Game summary
Black is clearly better. The bishop on The bishop's excursion to g 5 doesn 't
g7 and knight on b6 are workin g well achieve a great deal in the Benko. In
together, while the b-file is always use fact, as we saw in the game, Bl ack can
ful for the TOok on a8. just use it as a target to generate coun
13 1:[c1 tiJa4 14 tiJb1?! terplay by ... tiJe4 and ... g 7-g 5 . By the
Very passive; it would h ave been time Grivas had m ade four moves with
better just to give up the pawn, with 14 this bishop in the opening, Khalifm an
e4 for in stance. had obtained a clear advantage as
14 ...tiJc6 1 5 e4 l:tb8 16 tiJf3 d6 17 tiJfd2 Bl ack and went on to win comfortably.
i.e6 18 :c2 'iit d 7 19 .i.e2 litb6 Although White still has a few ways
The immediate 19 ... lir.b2 20 ':xb2 (or of achieving equality after 4 i.g 5, I
20 l:tc1 .i.xa2) 20 ... tiJxb2 21 a3 l:tb8 don't think we will see much of this
should al so be decisive. line in the future. If White really wants
20 0-0 lithb8 21 l:lfc1 l:ib2 22 a3 .i.a2 2 3 to pl ay in this style, it is better to pl ay
l:txb2 llxb2 24 i. d 1 tiJb6 2 5 f 3 tiJes i.g 5 on move two.
Game 33
V.Kramnik-P.Leko
Dortm u n d 1998
1 d4 tiJf6 2 c4 cS 3 d s bs 4 tiJd2
Not really a big threat to the Benko
Gambit. White's idea is to neutralize
his any hopes his opponent might have
of an initiative and build- up slowly be-
White is completely tied down and hind the centre;
130
for example, after 4 ... bxC4 5 e4 d6 6 6 bxc3

i.. xC4 g 6 7 b3 i.. g 7 8 i..b 2 0-0 9 lLlgf3 {as The alternative 6 lLlC4 ! ? tries to be
Khalifman had played against Leko the tricky, settin g the trap 6 ... lLlxe4?? 7 'iWe2
previous year}, while if Black delays lLlf6 8 lLld6 m ate. Instead, Black can
taking on c4 White might accept the either play 6 ... d6 with equality {and
gambit pawn after all in more favour then 7 i.. d 3 g 6 8 bxc3 transposes to the
able circum stances. However, as we g ame}; or accept the pawn with
will soon see, Black has a simple anti 6 ... cxb2, when 7 i.. x b2 d6 8 i.. d 3 g6 9
dote which gives him an easy and com i.. xf6 exf6 10 l:.bl i.. e 7 1 1 "ii'a4+ c;i;>f8 is
fortable game. unclear.
4 ... bxC4 5 e4 6 ... g6 7 lLlc4
5 lLlxC4?! i..b 7 is already worse for Other m oves are examined in the
White, who must either play the awk n ext game.
ward 6 lLle3 e6 7 dxe6 fxe6, or else try 6 7 d6
•••
d6 - but then 6 ... exd6 {6 ... lLle4 and Leko is n ot tempted by 7 ... lLlxe4? ! 8
6 . i.. d 5 are al so good} 7 lLlxd6+ i.. x d6 8
.. 'ii'e 2 i.. g 7 9 'iWxe4 i.. x c3+ 10 i.. d 2 i.. x al
"ii'x d6 lLle4! 9 'ii'e 5+ '&W8 10 lLlf3 {not 10 1 1 i.. a 5 "iVxa5+ 1 2 lLlxa5 i.. C 3+ 1 3 c;i;>dl
f3 ? "ii'a 5+ and wins} 10 ... d5 followed by i.. x a5, when Black h as rook, bishop and
... lLlc6 gives Black a very strong initia two pawns for the queen, but they are
tive. mostly sittin g at home. After 14 h4!
5 c31
••• White h as a useful initiative accordin g
Interfering with White's desired set t o Kramnik; for instance, 14. . . i.. C 3
up. 5 ... C 3 ! is Khalifman's move, which {14 ... d6? ? drops the bishop to 1 5 "iVa4+}
has become a standard equalizing idea 15 i..b 5 i..b 7 16 .lir.h 3 and White is bet
against 4 lLld2 . Indeed, from a practical ter.
perspective White must be careful n ot 8 i.. d 3 i.. g 7 9 lLlf3 0-0 10 0-0 lLlbd7 11
to end up in a worse position. i.. d 2 lLlb6
131
1S ... tLlxc3
Stronger was 15 ... iLa6 ! 16 tLlxd6
'ii'x d6 17 exd5 tLlxd5 and Black is clearly
better.
16 �xC3 'ii'xC4 17 �b2 �b7 18 �b3
'ii'a 6 19 eS tLle4?1
Here 19 ... tLlg4 20 l::tb l tLlxe5 21
tLlxe5 dxe5 2 2 �xe 5 �xe5 2 3 ':xe 5 e6
24 �g4 .l:tab8 would also be clearly bet
ter for Black, while Leko's 20 exd6 ! ?
�xb2 2 1 dxe7 lUe8 2 2 �xf7+! c;.t;>xf7 2 3
12 tLla s? 'ii'b 3+ can b e m et by 2 3 ... �d5 ! 2 4 tLlg 5+
Unless tLlc6 is a serious threat the (or 24 'ii'x d5+ c;.t;>g7) 24 ... c;.t;>g7 25 'ii'xb2+
knight is badly placed h ere. White 'ii'f6, when White does not seem to
shoul d h ave preferred 12 tLlxb6 axb6 13 h ave enough for the piece.
a4 (Leko) or 12 tLle3 c4 (or 12 ... e6 1 3 C4) 20 l::t b 1 l::t a c8 21 exd6 'ii'x d6 22 �xf7+?1
13 �c2 with a l evel game. But after 22 �xg 7 c;.t;>xg 7 2 3 'ii'xd6
12 ...c41 exd6 Black has a healthy extra pawn .
N ow Black gets the better chances. 22 ...l:txf7 2 3 �xg7 'ii'x d1 · 24 :exd1
13 �C2 tLlxf2 2 S c;.t;>xf2 �xf3 26 gxf3 ':c2+ 27
13 tLlxC4 tLlxC4 14 �xc4 tLlxe4 is very c;.t;>g3 c;.t;>xg7
comfortable for Black, with a strong
knight on e4 and bishop on g 7 ; e.g. 1 5
tLld4 "ilC7 16 tLl c 6 e6 17 l:tel tLlxd2 18
'ii'x d2 �b7 19 �b5 exd5 with a slight
advantage.
13 ...'ii'c 7 14 l:te1 tLlbxd s i s tLlxC4
This endgame should be lost for

White. I won 't really comment on the
rest as this is not an endgame book. As
it turned out Kramnik managed to
draw after Leko went wrong :
28 Itb3 .l:.xa2 2 9 ':d7 l:tas 3 0 h 4 h 6 3 1
132
lIC3 e6 3 2 l::td 6 'iW6 3 3 lIe3 lle7 34 l:te4 �g 7 9 l:tbl 0-0 10 � d 3 lDbd7 1 1 0-0
g5 35 J:[c6 l:te5 36 .i:ta4 gxh4+ 37 ':'xh4 lDb6, which leads to similar play.
h5 38 .l:!.a6 l:tg7+ 39 �f2 llb7 40 l:tha4
.l:r.b2+ 41 �g3 .l:tg5+ 42 �h3 l:tb3 43
�h4 1lf5 44 f4 l:.b7?
According to Leko 44... .l:te3 ! would
have won ; but not 44 ... l:tf3 ? due to 45
.l:r.xe6+! �xe6 46 l:ta6+ and Black cannot
escape the checks as the white kin g is
stalemated.
45 lIe4 l:te7 46 l:tea4 l:tb5 47 ':xa7 l::t x a7
48 1lxa7 llb4 49 �g3 :b3+ 50 �g2 �f5
51 1Ia 5+ �xf4 52 .l:!.xh5 Yz-Yz
The position is bal an ced and strate
Game summary gically complex. I studied it in my
After 4 lDd2 Bl ack has very few prob preparation for the g am e with Petkov.
lems in equalizing . Black gets a com a) 12 lDb3 ? ! (now Black gets an ini
fortable position with good chances of tiative on the light squares) 12 ... e6 13
an active game, and l ater on Leko even h3? ( 1 3 'ii'c 2 exd5 14 exd5 �g4 15
reached a winning ending . This line lDbd2 l:te8 is equal, but it doesn't seem
can only be recommended as a surpri se like a position worth aiming for with
weapon for White, not as a serious the white pieces) 13 ... exd5 14 exd5
theoretical test of the Benko. �a6 1 5 lDa5
Game 34
V.Petkov-N.V.Pedersen
S ka n d e rborg 2010
1 d4 lDf6 2 c4 cS 3 d 5 b5 4 lDd 2
Petkov used t o decline the Benko
with 5 b6. However, I n oticed he had
been venturin g 4 lDd2 more recently
and, as it is my conclusion that this line
is rather unproblem atic for Black, I was 15 ... lDbxd5 { 1 5 ... 'ii'C 7 ! , threatening
happy to play against it. ... lDbxd5, would be slightly better; e.g.
4 ...bxC4 5 e4 C3! 6 bxc3 g6 7 l:.b1 16 'iWd2 lUe8 17 J:[el llxel+ 18 lDxel
Another option is 7 C4 d6 8 lDgf3 .l:.e8 followed by ... lDe4 and Black is do-
133
i n g fine) 16 liJc6 liJc3 17 'ii'e l 'ii'd 7 18 1 6 g 3 liJd7 1 7 'iiC 2 liJe s 1 8 liJxe s dxe s
liJe7+ c;t>h 8 19 liJxg6+ fxg 6 20 .xC3 19 liJfl 'iid4 20 liJe3 liJf6 2 1 liJds litab8
ii.b7 21 liJel ? ! (quite passive; 21 liJg s , 22 l:.b3 �xds 23 exds e4 24 �xe4 (24
intending 2 1...h 6 ? ! 2 2 liJe6 ! �xg 2 2 3 �e2 l:tb6 25 lIdl 'ii'e s 26 1hb6 axb6 27
�xh 6 ! would still b e unclear) 2 1 . . .l:tae8 c;t>g 2 l:td8 is also equal) 24 ... litxb3 25
22 'iib 3 �c6 23 �e3 liJh s (or 23 ... liJe4! axb3 l:te8 2 6 d6 l:[xe4 27 litdl 'iie s 28
with a clear advantage) 24 �e2 liJf4 2 5 d7 l:tel+ 29 c;t>g 2 ':'xdl 30 .xdl liJxd7
�g4 .f7 2 6 �xf4 .xf4 2 7 �f3 �d7 2 8 3 1 'iix d7 'ii'e 4+ with a level endgame.
ii. d s ii.fs 29 l:ldl? ! (29 liJd3 would b e a 7 ... � g7
better try) 2 9 ... lite2 30 a4 �es 3 1 g 3 Not yet 7 ... d6 in view of 8 �b s+.
'ii'h 6 3 2 liJd3 ii.d4 3 3 liJf4? �C2 34 .f3 8 liJgf3 0-0 9 �d3 d6 10 h3
ii.xdl 0-1 D.5vetushkin -V. Baklan, Ru
manian Team Championship 2009. A
nice g am e by Baklan and a good illus
tration of the dangers awaitin g White
if h e m akes a few inaccuracies.
b) 1 2 l:tel seem s to be the m ost pre
cise m ove, reaching an equivalent posi
tion to the m ain g am e (minus h 2-h 3
and ... 'ii'C 7).
Petkov had pl ayed this before.

White is delaying c3-c4 for a while, so I
hold back on ... e7-es too, since this
would give his knight a nice post on c4.
10 ... liJbd7 11 0-0 liJb6 12 litel 'ii'C 7
12 ... liJh s is a good alternative; e.g.
1 3 c4 'ii'C 7 14 liJfl l:lb8 1 5 �d2 es with
a roughly equal position . White might
try 15 g4!?, which I suppose is one of
V.Akopian- P.Tregubov, Mainz (ra the ideas of h 2-h 3, but it's very com
pid) 2009, continued 12 ... es 13 dxe6 (if mittal and gives Black counter-chances.
White doesn 't take on e6 the position is For example, ls ... liJf6 16 ii.b2 es 17
bal an ced as well; typically Black will dxe6 (17 liJe3 ? ! hs is better for Black)
continue with ... liJh s) 13 ... �xe6 14 17 .. .fxe6 18 es liJe8 19 liJld2 'ii'f7 is un
ii.b2, and n ow 14 ... liJh s offers bal anced clear; both sides have some serious
chances; for example, 15 �xg 7 c;t>xg 7 pawn weaknesses.
134
13 c4 eS 14 lbf1 lbhS is g4 lbf4 16 in trouble .

..Itxf4 exf4 17 'ii'd 2
2 3 lbg4 lbd7 24 .i.C2

17 ... .i.a6 24 gxf6 1:[ae8 25 .i.c2 h S 2 6 lbh 2
More ambitious was 17 ... lbd7 ! (I lIxf6 2 7 lbf3 would be slightly better
al so considered 17 ... lba4, but thought for White, because of the weak pawn s
it was in sufficient) 18 'iVxf4 lbe s, when on e S , c S and g 6 .
Bl ack has decent compen sation in view 24 fxgs 2 5 .i.a4 lbb6 26 .i.c6 lbxC4 27
. . •
of the strong dark-squared bishop and .i.xaS lixaS 2S lled1?!

loose pawn structure in front of the Too straightforward. In stead, 2 8 d6 !
white king . For example, 19 lbxe S .i.xe s 'iVxd6 2 9 1:[edl "fie6 30 1:[dS would give
20 �e3 i.. d4 2 1 1Ve2 "fie7 2 2 lbe3 h S 2 3 White a dangerous initiative.
gxh S .i.xh 3 2 4 'ii'f3 'iih 4 with interest 2S ...�d6
ing play for the pawn .
1S :bc1
The greedy 18 'ii'xf4 .i.xc4 19 lIxb6
.i.xd3 20 :c6 'iVas 21 l:txd6 C4 would
give Bl ack good play with the bishop
pair and active pieces.
1S ... .i.eS 19 lbxes dxes 20 lbh2 lbcS?!
This casual move grants White
some initiative based on the unex
pected weakness of the e s -pawn . I
should have played the solid 20 ...f6,
hindering g4-g S, with equality. N ow the position is unclear.
21 �c3 f6 22 gsl lbb6 29 l:.b1 <iifi> g 7 30 l:dC1 h5 3 1 lbh2 <iifi>f6 3 2
Hurrying back to defend the king 'ii'd 3 lba 5 3 3 'ili'a3 lb b 7 34 'ii'a 4 l:tcS 3 5
side. Not 22 .. .fxg S ? 2 3 lbg4! and Bl ack is 1:[c3 c4
135
H ere Petrov h ad around eight min 44 .li!e3 g3+??

utes to reach the time control at move But this should just have lost. The
40, while I h ad about three. correct course was to retreat with
36 l:ta3 lDC5 44 ... �5, keeping a slight advantage,
The black pieces are n ow very well but then we wouldn 't have seen the
arranged. Petkov shoul d probably tried forthcoming entertaining king march .
to prevent this form ation. 45 fxg3+ �xe3 46 lDg2+ �d3 47 .l:r.d1+
3 7 "a 5 .l:r.C7 3 s lDf3 g4 39 lDd2 �g5?1 �C2 4S lDe3+ �b2
The beginning of a risky winning at 49 'ii'a 4??

tempt, in which the king m akes a solo H ere 49 'iVb5+ �c3 50 lDxC4 wins
trek across White's heartland to the outright; the move played allows Black
distant a3-square. In stead, 39 ... g xh 3 back into the game.
would b e fin e for Black, but I was in an 49 "C5 50 'ii'C 2+ �a 3 5 1 'ii'xe4 �f5 52
•.•
optimistic m ood after the bal ance of 'ii'f 3 c31

the position h ad ch an g ed in my favour,
and was also in some time pressure. Of
course my intuition told me there
should be something for White, but it's
n ot always so easy to fin d it over the
board. Okay, m aybe I wouldn 't play this
way again st Rybka or Fritz.
40 �h2 f3 41 .1:.g1 <oW4?1 42 lDxf3?
The refutation was to play 42 l:te3 !
first, when White win s after lDxf3 or
lDxC4 and 'ilVd2 .
42 lDxe41 43 lDe1?1 �cs
• • • Blocking off the pressure from the
Now I was actually slightly better. queen on f3 .
136
53 d6 .l:[f7 54 l:td 5 'iVb4??

A final exchange of errors. 54 ... .i.e6 Game 35
should draw, whereas the text loses to WI.Schmidt-G.Kasparov
5 5 'iVe2. Instead the mistakes con Dubai Olym piad 1986
tinue ...
5 5 ll:lc2+? .i.xc2 56 'iVxf7 'iVb2 1 d4 ll:lf6 2 C 4 c5 3 d 5 b 5
Kasparov h a s played t h e Benko a
few times, and it's an opening which
suits his dynamic style.
4 a4
57 'iVxa7+??
As Tartakower said: the winner of
the game is the pl ayer who m akes the
next-to-last mistake. In stead, 57 'it>g l
would keep it going. White wants to force his opponent
57 ....i.a4+ 58 'it>gl 'iVb1+ 59 'it>h2 Wxa2+ to m ake a decision about the b 5-pawn .
60 'it>gl 'iVxd 5 0-1 However, Black can reach an unclear
game quite easily and in more than
Game summary one fashion, so 4 a4 is not really the
The game was equal from the opening way to test the Benko G ambit from a
and I had a few ways to get a balanced theoretical viewpoint. Chris Ward is the
position. Instead, I m ade a few inaccu only G M to play this line regularly as
racies and Petrov gained the advan White.
tage, that is until he let it slip. The fin al 4 ... b4
phase of the game was fun if not objec Both this m ove and 4 ... bxC4 (exam
tively correct. Of course the computer ined in the n ext g ame) give Black a
doesn't like it, but Petrov was unable to good g ame. On b4 the pawn controls
find a refutation over the board. So I the c3-square, supporting the fight for
was quite content to play ... 'it>g 5-f4-e3- the long diagonal after ... g 7-g6 and
d3-c2-b2-a3 with that many pieces on ... .i.g7. Then Black can break in the cen
the board and win . tre with ... e7-e6 as in the g ame, or in
13 7
some cases con sider . . . e7-es instead. 7 ... i.. g 7 S g3?!

S ttJ d 2 g6 6 e4 8 ..td3 is a more n atural move, sup
6 b3 ..tg7 7 ..tb2, opposin g the g7- porting the key squares e4 and c4. Then
bishop, is a logical alternative but too play might continue 8 ... 0-0 9 0-0 e6
quiet to offer White any advantage. For (9 ... es is al so possible) 10 h3 ttJbd7 11
example, 7 ... 0-0 8 e4 d6 9 g 3 ? ! (this ttJb3 itC7 12 ..te3 l:.e8 13 'ii'c 2 ..ta6 14
m ove is often dubious in these posi l::tfel l:[e7 with bal anced chances.
tions, as we will see again in the m ain S e6 9 ..th3
•..
g ame; 9 ttJgf3 would be roughly equal) Aiming for the exchange of light
9 ... e 6 ! 10 ..tg2 exds 1 1 cxds lIe8 and squared bishops, which has some logic
Black was already better in A.5urjadnji to it so th at Black can 't play ... ..ta6 hin
R. Ponomariov, Swidnica (rapid) 1998. dering ttJC4.
6 d6
••. The n atural 9 ..tg 2 exds 10 0-0 0-0
(not 10 ... dxe4? ! 11 ttJxe4! ttJxe4 12 'iWds
winning m aterial) 11 cxds ..ta6 12 l:[el
ttJbd7 gives Black a comfortable posi
tion and good chances of dynamic play
in the middlegame, with most of his
pieces active and well placed already.
9 exd s 10 i.. x cs 'iVxcs 11 cxd S
•••
7 ttJ gf3
7 f4 ..tg7 8 ttJgf3 0-0 9 ..td3 is more
ambitious, but Black h as sufficient
counterpl ay h ere too: 9 ... e6 10 0-0 exds
11 cxds l:te8 1 2 l:tel ..ta6 1 3 ji'c2 (or 13
..txa6 ttJxa6 14 e s dxe s 15 fxe s ttJxds
16 ttJC4 ttJac7 17 ..tg s f6 18 ..th4 Y2-Vz
C.Ward-lM. Degraeve, Gent 2005) Now we have a sort of improved
1 3 ... ..txd3 14 ji'xd3 ttJbd7 1 5 as c4! 16 Benoni for Black.
'ii'x C4 l:tc8 17 'ii'x b4 ttJcs 18 l:ta3 was 11 0-0?!
•••
seen in G . Flear-C.Adrian, French Team 11 ... ji'a6 ! 12 'iVe2 0-0 was stronger.
Championship 1997, when 18 ... ttJxd s ! Garry probably didn 't like the exchange
19 exds l:txel+ 20 ttJxel ttJa6 foll owed of queen s, but now White has prob
by ... l:txcl would have equalized. lem s defending the e4-pawn . For ex-
138
O t h e r Op t io n s fo r W h i t e
ample, 13 'ii'x a6 lLlxa6 14 0-0 lUe8 1 5 1 9 llel 'ii'd 5 2 0 'iVdl lLlSd7 2 1 lta2
l:1el l:.ac8 (with the dangerous threat .i.xd4
of ... c4-c3) 16 tt:)c4 lLlxe4 17 l:txe4 l:txe4 In his n otes Kasparov prefers
18 lLlxd6 .l:tee8 19 .l:tbl c4 20 lLlxe8 l:lxe8 2 1 . . .l:Ue8 22 .i.e3 lLlC4 with a decisive
21 .i.e3 .l:.d8 and Black is clearly better advantage, while 2 1...lLlc S ! , threatening
due to the advanced queen side pawns ... .i.xd4 and ... lLlbxa4, l ooks even
and passive rook on bl. stronger.
12 0-0 e4 13 'ii'e 2l! 22 I1d2 lLle5
13 "e2 ! keeps the bal ance: 1 3 ... c3 Declining the second pawn by
14 bxc3 J:te8 (14 ... bxC3 1 5 lLlC4 'ii'a 6 16 2 2 ... .i.xf2+, since White might obtain
lLld4 lLlxdS 17 lLlbS lLlb4 18 lLlC7 'ii'c 6 19 some counterplay on the open files.
lLlxa8 c2 is unclear) 1 5 cxb4 lLlxdS 16 2 3 lLlxd4 lLlbe4 24 :e2 .lilacS
.l:ta3 lLlxb4 17 l:te3 lLl8c6 18 lLlC4 and
White has reached a defendable posi
tion.
13 ... e3 14 bxe3 bxe31
I will leave the rest without n otes.

G arry m an aged to convert the extra
pawn to win this one-sided g ame.
2 5 ':'e3 lLlb6 26 .i.b2 ':'xe3 2 7 .i.xe3 liteS
Now White has a very difficult de 2S .i.al l:te4 29 lLlb5 .l:txa4 30 'iVxd 5
fensive task. The quiet 14 ... 'ivxc3 would lLlxd 5 3 1 .i.xe5 dxe5 3 2 .lilxe5 lLlf6 3 3
lead to a lesser advantage. For instance, ':'e7 a 5 34 lLld6 l1d4 3 5 lLlxf7 'it;JfS 36
15 l:ta2 lLlbd7 16 'ii'b l b3 17 .l:.a3 lLlcs .l:.a7 .l:r.d7 3 7 l:txd7 lLlxd7 3s lLld6 'it;Je7 39
18 lLlxb3 'i'b4 19 lLlfd2 lLlcxe4 20 lLlxe4 lLle4 a4 40 'it;Jfl 'it;Je6 41 'it;Je2 'it;Jd 5 42
lLlxe4 21 lLld2 'ii'x bl 22 lLlxbl l:tfb8 and lLle3+ 'it;Jd4 43 'it;Jd2 lLle5 44 'it;Je2 lLld 3 45
Black is only slightly better. lLldl lLlel+ 46 'it;Jb2 lLlf3 47 h4 lLle5 4S
15 lLlb3 'ilkg4 16 lLlfd4 'ii'x e4 17 'ii'x e3 'it;Ja3 'it;Je4 49 h 5 gxh 5 50 lLle3+ cM3 5 1
lLlxd 5 lS 'iVd2 lLlb6 lLld 5 lLlg4 5 2 lLle7 lLlxf2 5 3 lLlf5 lLle4 5 4
Black is a pawn up and has the 'it;Jxa4 lLlxg3 5 5 lLlh4+ 'it;Je4 56 'it;Jb4 lLlf5
more active pieces as well. 0-1
139
Game summary and 'iie 2 or lDg s-e6 (after .. .fxe6). That

With 4 ...b4, followed by ... g 7-g6, ... .i.g7, Black cannot respond casually to this
and l ater the ... e7-e6 break, it is quite plan can be seen by 7 ... lDbd7? ! 8 lDf3
easy for Black to equalize against 4 a4. g 6 ? ! 9 e s ! dxe s 10 fxes lDg4 11 e6!
The g am e was similar to the Benoni fxe6? (but if 1 1 ... lDde s, then 12 exf7+
Defence, in a somewhat improved ver lDxf7 13 lDg s ! lDxg s 14 'iix g4 lDf7 15
sion for Black, and illustrates well some .i.xc4 and White is clearly better) 12
of the dangers awaitin g White if he lDg s lDgf6 1 3 lDxe6 'i!i'b6 14 lDbs .i.xb s
plays inaccurately or too passively 1 5 axb s :c8 16 .i.xc4 and White had a
again st the Benko. All in all , a convinc huge position in I . Kazantzidi s
ing win by Kasparov. N.Galopoulos, Greek Team Ch ampion
ship 2006.
In stead, Bl ack should play 7 ... g6,
Game 3 6 leaving the d7-square free, and then:
O.Dancevski-J.M.Degraeve
E u ropea n Champions h i p,
Oh rid 2001
1 d4 lDf6 2 c4 cS 3 ds bs 4 a4 bXc4
a) 8 e s (continuing with the aggres

sive approach) 8 ... lDfd7 9 lDf3 .i.g7 10
'iie 2 was pl ayed in C.Ward
A.5ummerscale, British League 2000,
and now 10 ... 'iiC 7, attacking the pawn
A different way of playin g to the on e s , would lead to an unclear game.
4 ... b4 of the previous g ame. For example, 11 e6 (ll lDbs ? ! .i.xbs 12
5 lDC3 d6 6 e4 .ta6 axbs dxe s 1 3 'ii'e 4 'iid 6 is slightly bet
Makin g White work a bit to regain ter for Black) 11 .. .fxe6 12 1r'xe6 .i.xc3+
the pawn . 1 3 bxc3 .i.c8 14 'iie 3 (14 .txc4 ! ? lDe s 15
7 lDf3 .tbs+ 'iit> d 8 16 lDg s is tempting, but
The alternative is 7 f4, intending to after 16 ... 'ii a s ! 17 fxes .i.xe6 18 lDxe6+
attack in the centre with lDf3, e4-es-e6, 'iit> c 8 19 .i.d2 a6 20 .i.d3 c4 21 .i.xC4
140
lLld7 22 exd6 exd6 2 3 0-0 lLle5 White Bl ack h as equalized without too
doesn't have enough for the queen) much effort, just playing n atural active
14 ... lLlb6 15 lLlgS �7 is roughly equal . moves.
b) 8 lLlf3 �g7 9 �e2 i s a quieter way 14 �d2 lLlg4 1 5 .l1bl 'iVc8 16 0-0 lLlges
of playing the position, when 9 ... 0-0 10 17 b3 lLlxf3+ 18 �xf3 �d4 19 �e2 lLles
0-0 lLlbd7 11 lLld2 lLle8 12 �xC4 �xC4 20 �h6 lite8 2 1 lLlxes �xes 22 .tC4
13 lLlxC4 f5 was unclear in A.Yusupov .tbS 23 b4?1
V.Chekhov, Vilnius 1978. From a practi 23 'iVd3 'iVa6 24 �d2 l:rb7 25 Ilfcl
cal point of view this continuation is l:teb8 2 6 b4 keeps the bal ance.
attractive as it attacks the white cen 23 .. :ifa61
tral pawns.
7 ...lLlbd7 8 'iWc2
8 �e2 g6 9 0-0 �g7 10 lLld2 ? ! (10
'iWc2 was stronger) 10 ... lLle 5 ! 1 1 f4 lLld3
12 lLlxc4 �xC4 1 3 �xd3 �xd3 14 'iWxd3
0-0 gave Black a nice g ame in A.Moran
Nuque-J .R.Koch, World Student Team
Ch ampionship, Paran ana 199 3 . The
rook will be a factor on the b-file.
8 ...g6 9 lLlbS lLlb6 10 lLla 3
Going round to regain the pawn .
Another way was 10 �e2 �g 7 11 lLld2 Black i s slightly better: the pawn s
lLlfd7 12 lLlxC4 lLlxC4 13 �xC4 0-0 14 0-0 on b4 and as are weak, and in some
lLlb6 15 �e2 'ii'd 7 16 .:tbl c4 with an lines the rook on bl i s hanging as well.
unclear game. White faces a difficult defensive task.
10...�g7 11 as lLlbd7 12 lLlXC4 0-0 13 24 :fcl?
.te2 l:lb8 After thi s Black win s a pawn for
n othing. White h ad to play 24 �xb5
:xb5 2 5 bxc5 l::t x cS 2 6 'iVa4, when
26 ... :ec8! (26 .. J:td8 ? ! 27 .l1fcl 'iWxa5 28
'ii'x a5 l:.xas 2 9 ':C7 �f6 30 g4 would
give White compen sation for the
pawn) 2 7 'iVd7 'iVxas 2 8 'iVxe7 :5C7 29
'iYh4 'iVa4 gives Black the better
chances, due to his active pieces and
potentially dangerous passed a-pawn.
24 ... �xc4 2 5 'iVxC4 'iVxC4 26 l:[xC4 .l:tbS
27 .td2 l:[eb8 28 a6 fS
141
Game summary
4 a4 is hardly a serious try for an ad
vantage. Both 4 ... b4 and 4 ... bxC4 lead to
unclear positions and it seems a matter
of taste which one Black should play. In
this g ame Degraeve chose the latter
and then defended the pawn by
6 ... i.a6, forcing White to waste time in
winning it back. Later on, standard
counterplay on the b-file enabled him
to win another pawn, after which the
Undermining the centre, a common passed a-pawn m arched through to
theme in Benko endg ames. Whether victory.
White pl ays 29 f3 fxe4 30 fxe4 or takes In the opening White might prefer 7
on fS, the advanced centre becomes f4 as a more aggressive approach, but
isol ated and thus more vulnerable to as long as Black is careful not to play
attack. ... liJbd7 too soon, he shouldn 't have
29 exfs gxfs 30 .l:.b3 l:tSb6 31 g4 ':'xa6 anything to worry about here either.
3 2 gxfS cxb4 3 3 l:tcxb4 l:lxb4 34 l:lxb4
'itt g 7 35 'itt g 2 !!a 3 36 ':g4+ 'iW6 37 .l:.h4
�xfS 38 l:txh7 'iW6 39 i.e3 a s Game 3 7
JJirka-B.Vuckovic
Soz i n a 2004
1 d4 liJf6 2 c4 cS 3 dS bS 4 f3
The a-pawn decides the g ame.

40 �f3 a4 41 l:thS �g6 42 lIh6+ �g7 43
.l:r.h3 .l:r.xe3+! 44 fxe3 a 3 45 :g3+ �f7 46
.l:r.g2 i.b2 47 e4 a 2 48 llxb2 a1"iV 49 l:.f2
'iVd4 50 h4 �g6 51 ':'g2+ �hS 52 �f4 We are n ow really down to the odds
"ilfes+ 0-1 and sods. 4 f3 was once pl ayed by An-
142
and in a rapid game, which is perh aps was just two pawn s up.
the best thing that can be said about it. 7 ... .i g 7 8 0-0 0-0 9 ltJec3
4... bxC4 Intending to deploy both knights on
Here 4 ... g6 5 e4 d6 6 cxbs a6 would the queen side in order to combat
tran spose to the 5 f3 variation, except Black's counterplay there.
th at I am recommending s ... e 6 ! in 9 ltJbd7
••.
Chapter Four (rather th an s ... g6 6 e4

d6, which I think is good for White).
5 e4 d6 6 .ixC4 g6
10 .ie2
Prophyl axis again st ... ltJb6 or ... ltJes,
which might n ow be met by a2-a4 or
7 ltJe2 f2-f4 respectively, but Black can now
7 ltJC3 .ig7 8 .ie3 0-0 9 ltJge2 ltJbd7 refrain from either for the time being.
10 0-0 ltJes is quite comfortable for Instead, Y. Duh ayon- N .V. Pedersen,
Bl ack. In B. Lajth ajm-V.Mal akhatko, Krakow 2008, saw 10 ltJa3 ltJb6 1 1 .ie2
Yugoslav Team Champion ship 2001, e6 12 dxe6 .ixe6 1 3 .ie3 We7 (I h ad
White tried the optimistic 9 h 4 ! ? .ia6 this position in my preparations and
10 .ixa6 ltJxa6 11 Wd2 Was 12 ltJge2 saw that a subsequent ... d6-ds would
l::tfb8 (aiming for counterplay on the b be fine for Black) 14 1:[cl 1:[fd8 1 5 Wc2
file and the dark squares) 13 h s ? ! (a ds 16 ltJxds ? ! (but 16 .ig s h6 is very
thematic pawn sacrifice to open the comfortable for Black) 16 ... ltJfxds 17
file for the rook, but here it fails tacti exds ltJxds and I was clearly better,
cally; instead 13 0-0 ltJC7 would still be since the black pieces are very active,
unclear) 13 ... ltJxh s 14 g4? ltJf6 (the un- while the white knight on a3 is out of
expected 14 ... ltJg 3 ! was more accurate, play. The finish cam e rather quickly: 18
with the same idea of 15 ltJxg 3 ':xb 2 ! .ixcs Wg s 19 .l:.fdl?? ltJC3 ! winning
etc) 15 ltJf4? ltJxg4! (spottin g the tactic m aterial and the g ame.
this time; i.e. 16 fxg4 l:txb2 ! and wins) 10 1:[b 8 11 ltJa3 lle8?1
.••
16 ltJdl Wxd2+ 17 .ixd2 ltJes and Bl ack I would prefer 1 1 ... ltJe8 with an un-
143
clear g ame. Bl ack h as similar ideas as 2 4... tDe5 was still clearly better for
in the b5-b6 line, such as ... tDC7 (an d Black. After the text White might play
possibly ... tDb5-d4), ... e7-e6, o r .. .f7-f5 . 2 5 g 3 with good chances of holding the
12 'iVC2 tDb6 13 i.e3 e6 14 dxe6 i.. x e6 position, but instead he blunders.
15 l1ad1 'ii'e 7 16 tDcbS? 2 S b3??
16 i.g 5 would keep control of the
position with a slight advantage to
White. The text provides Black with a
tactical solution to his problem s.
16 tDfd s l 17 exd s?1
..•
But if 17 i.f2, then 17 ... tDb4 18 'ii'd 2

d5 with the initiative.
17 ... i..f s 18 i.. d 3 'iVxe3+ 19 l:If2 i.xd3
20 I:[xd3 'ii'e 7
2 S tDhs
•..
Now there is no defence. 2 5 .. .l:iel !

2 6 l:. d l tDh 5 ! was even stronger.
26 tDe4 fS 27 g3 'iVh3 28 'iVg2 'i\Vxg2+ 29
<Ji>xg2 fxe4 30 fxe4 ];txe4 31 tDxd6 liIe2+
32 <Ji>h3 :d8 33 tDf7 l:.xd S 34 g4 l:.e7 3 5
':'df3 tDf6 0-1
Game summary
Black h as a strategically won posi Again st 4 f3 Black can equalize easily
tion, due to his strong bishop and con with 4 ... bxC4, followed by ... g7-g6,
trol of the e-file, versus the weak white ... i.g7, and usually a rook on the b-file,
pawns and passive rooks and knights. with further ideas of ... tDbd7-e5, or
2 1 l:tf1 tDd7?1 ... tDb6 and ... e7-e6, or sometimes
The knight was better where it was. ... tDe8-C7.
Black could have forced his way into In the g ame White played 10 i.e2 in
enemy camp by 21 ... a6 22 tDC3 i.d4+ time - after 10 tDa3 there is 10 ... tDb6
23 <Ji>h l i.xC3 ! 24 bxc3 .e2, and 2 5 with the idea of ... e7-e6 and ... d6-d5
'iVxe2 ]::t x e2 2 6 I1bl tDc8 ! , o r 2 5 <Ji>g l and an easy game for Black as in Du
'it'xc2 2 6 tDxc2 tDC4! followed by ... l:.b2 hayon -Pedersen - and as an improve
or ... l:[e2 and win s . ment I suggest 11 ... tDe8 with an un
2 2 tDC3 i.d4+ 2 3 <Ji>h1 'iVh4 24 tDC4 tDf6? clear position . However, it is ch aracter-
144
istic for this line that, even from a fa While 4 e 4 might be said to be too
vourable position, White can quickly active, White's fin al option, 4 b 3 ? ! , suf
land in trouble with inaccurate play, as fers from being too passive. Black can
occurred after 16 ltJcb S ? ltJfdS ! . Vukovic get a comfortable g am e after either
went on to win, though not without 4 ... bxC4 (or 4 ... e6 5 dxe6 fxe6 6 e3, as in
further mistakes from both sides. L.M. Rodrigues-S. Kasparov, Figueira da
Earlier on, the attacking idea of ltJc3 Foz 2009, when 6 ... bxC4 7 bXc4 ltJc6 is
and h2-h4, as in Lajthajm -Ma1 akhatko, fin e for Black with ideas of ... iLe7, ... 0-0,
is not dangerous for Black. In fact it is ... !Ib8 and ... d7-dS) 5 bXc4 d6 6 ltJc3 g 6
White who is more likely to get into 7 iLb2 ? ! (the bishop is vulnerable here
difficulties from the opening, given with the b-fi1e open ; 7 e4 iLg 7 8 llbl
Black's activity and counterp1ay on the would m ake m ore sense, though Black
queen side. is certainly n ot worse) 7 ... iLg 7 8 f3 ? !
ltJbd7 9 e 4 llb8 (White is already i n
trouble) 1 0 'it'C2 'iWas 1 1 'ifilf2 ? (allowing
Game 38 a combination; White h ad to try 1 1 1:[cl
D.Shchukin-A.Khalifma n and 12 iLal, albeit with an unenviable
St Pete rs b u rg 1998 position) 1 1 ... ltJxe4+ ! 12 fxe4 ':'xb 2 ! 1 3
'ii'x b2 iLxC3 1 4 'ii'c l ltJf6 1 5 ltJf3 ltJxe4+
1 d4 ltJf6 2 c4 cS 3 d 5 b5 4 e4?! 16 'ifile3 iLxal 17 'iYxal ltJf6 18 iLd3 0-0
19 h3 iLa6 20 I:[bl e5 21 dxe6 fxe6 2 2
1:tb3 "fiC7 2 3 1:t a 3 iLc8 24 'iWb2 d5 2 5
'it'e5 ? ltJg4+ ! 0-1 U .Masera-P.Benko,
Reggio Emilia 1970/71. An easy and
instructive g am e from Pal Benko.
4 ...ltJxe4 5 'iYf3
The point of the previous m ove,
hoping for 5 ... ltJf6? 6 d6 ! ltJc6 7 cxb5
ltJd4 8 'it'xa8 ltJc2+ 9 'ifildl ltJxal 10
'iVxa7 with a winning position (Mirk
ovic). 5 cxb 5 would h ardly be con sis
Here White tries to seize the initia tent, and indeed s ... 'iVa5+ 6 ltJd2 (6
tive by offering the e-pawn, but it's n ot iLd2 ? ! ltJxd2 7 ltJxd2 g 6 is worse for
really a very good strategy to give away White) 6 ... iLb7 gives Bl ack a nice g ame,
a central pawn like this. Especially since as the trade of the e4- and b 5 -pawn s is
the loss of the e4-pawn often leads to n aturally favourable for him .
the ds-pawn becoming a weakness, as 5 ... 'iYa 5+1
we have seen before in this book. A familiar idea, combining with the
145
knight o n e 4 t o disrupt White's play, as are al so quite unclear) 10 .1i.d3 .i.g7 11

we have seen before in G am e 3 2 . 0-0 b2 12 .l:tael 0-0 1 3 l:lxe7 d6 14 lLlg s
lLld7 l s lLlxh 7 ! "d8 ( l s ... 'iit> xh 7? leads to
m ate: 16 "iVh S+ 'iit> g 8 17 .i.xg 6 et c) 16
.i.g s f6 17 lLlxf8 'ii'xf8 18 'iWe2 lLle s 19
.l:[xe s dxe s 20 .i.e3 and White went on
to win in D.5hchukin-K.Andreev, St Pe
tersburg 2010.
Rather than go into these complica
tions it is simpler to play 7 ... 'iib 6! 8
bxc3 (or 8 .i.xC3 b4 9 .i.d2 'ii'f6 and
Black is just a pawn up) 8 ...bxC4 9 .1i.xc4
(or 9 lLlh 3 'iWf6) 9 ... .i.a6 10 .i.xa6 'iWxa6
6 lLld2?! (Bl ack has no problem s after 10 ... lLlxa6
White doesn 't want to give up the 11 lLle2 "iVf6 either) 11 lLle2 d6 12 0-0
bishop pair by 6 .i.d2 ? ! lLlxd2 7 lLlxd2 lLld7 13 c4 lLles 14 'iib 3, as in I.Nemet
d6 when Black is better, but after the A. Robert, Bern 1990, when 14 ... lLlxC4!
text the knight is somewh at badly would have been good for Black, since
placed. 15 'ii'd 3 can be met by ls ... 'iWa4! 16 lLlc3
Shchukin l ater improved with 6 lLle s .
lLlC3 ! ? lLlxC3 7 .i.d2 ! (after 7 bXc3 d6 8 6. . .lLld6 7 cxbS lLlxbS
cxbS lLld7 9 .1i.d2 a6 10 c4 'ii'a4 Black
h as a nice position)
Black is already slightly better: his

knight has a great square on d4 and
7 ... b4 8 bxc3 b3 9 lLlh 3 ! , when White the white king is rather exposed.
h as definite compen sation for the White's strategy has failed at quite an
pawn s. For example, 9 ... g 6 ? ! (but 9 ... b2 early stage !
10 .l:[bl 'ii'x a2 11 "iVfS ! and 9 ... h6 10 .1i.d2 8 d6?
146
A blank shot. On the other hand, 8 tDxe2 24 �xe2 �e6 2 5 fxe5 fxe 5 26
.lixb5 'iixb5 9 tDe2 i.a6 10 tDe4 d6 1 1 �3 :'e6 2 7 �e4 i.xd6 28 l:td 5 i.e7 29
tD2c3 'iVd3 is hardly pleasant for White tDxe5 :d6 30 ]:txe5 i.b6 3 1 :b5 i.xe3
either. The endgame after 12 'iVxd3 3 2 �xe3 lld 5 3 3 tDf3 l1xb5 34 tDd4+
.lixd3 is quite favourable for Bl ack, due �d 5 35 tDxb5 J:b8 0-1
to the bishop pair and good, long-term
activity. Game summary
8 ... tDe6 9 i.xb5 'iVxb5 10 tDe2 e6 Khalifman won convincingly against 4
10 ... i.b7 and 10 ... i.a6 look even e4? ! . Indeed, after 4 ... tDxe4 5 'iVf3
stronger. 'iVa5 + ! White m ay already be worse.
11 0-0 i.a6 12 :tel 'iVd 3 13 'ii'x d 3 i.xd 3 The best try is probably 5hchukin's
14 tDf4 tD b4 l ater idea of 6 tDC3 ! ? tDxC3 7 .lid2, in
14 ... i.f5 would allow White to de tending 7 ... b4 8 bxc3 b3 9 tDh 3 ! with
fend the pawn with 15 tDc4, when dangerous play for the pawn, but Black
15 ...tDb4 16 i.d2 ! ? tDc2 17 tDd5 creates can avoid this with the simple 7 ... 'iVb6 ! ,
some mischief. when i t is h arder for White t o prove
15 a3?! sufficient compensation. Conversely,
After 15 tDxd3 tDxd3 16 :dl tDxcl the passive 4 b 3 ? ! , as seen in the n oted
17 J:1axcl i.xd6 18 tDb3 i.e7 19 tDxC5 g am e Masera-Benko, gives Black n o
d5 White is a pawn down for n othing. problems at all .
15 ...tDe2 16 tDxd 3 tDxal
Summary
In this chapter we examined various
sidelines for White on the fourth and
fifth m oves.
The Zaitsev Variation, S tDc3 axb5 6
e4, aim s for a quick attack with moves
like .lif4, i.C4 and e4-e5, but after 6 ... b4
7 tDb5 d6 8 .lif4 g 5 ! Black is already
slightly better. 50 White should try an
other option such as 8 i.d3. White h as
no advantage here either, but at l east
The knight cannot be prevented he is n ot worse.
from emerging at c2 or b3, so Black is Another idea is 4 tDf3, which is of
the exch ange and a pawn up and win s ten just a transpositional tool to other
easily once he gets h i s pieces out. lines, in particular 4 ... g6 5 cxb5 a6 6 b6
17 tDe4 tDb3 18 i.e3 :e8 19 tDf4 f6 20 as we saw in Ch apter Five. In stead, in
.l:tdl �f7 21 tDe2 e5 22 f4 tDd4 23 �f2 Gam e 30 White played 6 'iVc2, a choice
147
of Korchn oi and Ivan Sokolov, based on dlegame. With a creative effort involv
the greedy idea of 'iix C5 ; but this was ing an unusual king march I managed
n eutralized by Topalov's strong n ovelty to win the g ame all the same.
8 ... 'iIi'a4 and it is n ot so popular any 4 a4 gives the Benko pl ayer a com
m ore. Gam e 3 1 considers 6 e3, by fortable choice between taking the
which m ove order White avoids the 5 pawn or pushing past. 4 ... b4 was cho
e 3 iLb7 of Chapter Three. H ere 6 ... cxb5 sen by Garry Kasparov in Game 35 and
7 iLxb5 'ili'a5+ 8 lLlC3 .ia6 looks to he quickly seized the initiative in a sort
guarantee Black sufficient play for the of improved Benoni position . Degraeve
pawn in a typical Benko position . preferred 4 ... bxc4 5 lLlc3 d6 6 e4 iLa6 in
4 iLg 5 is inspired by the Trom Game 36 and al so equalized quite eas
powsky (2 iLg 5), but it doesn 't really ily.
work again st the Benko. After 4 ... lLle4!, Finally, the idea of 4 f3 intending
with ideas of ... 'ii a 5+, ... g 7-g 5 and e2-e4 (as in Game 37) is a lot less chal
... iLg7, it i s White who must take care lenging than the 4 cxb5 a 6 5 f3 of
n ot to come out worse from the open Chapter Four; while the immediate 4
ing. In G am e 3 2 Grivas played weakly e4 is somewh at suspect, and 5hchukin
with 5 iLf4? ! , 7 iLe3 ? ! and 9 .ixg 5 ? ! , soon got the worse position as White in
after which Khalifm an was clearly bet Game 38, even if he later found a bet
ter and won quite easily. ter way of playing (6 lLlc3 ! ?) to reach a
A sounder, if unthreatening, idea is more equal game.
4 lLld2, often leadin g to an unclear po To sum up, none of these fourth
sitional g ame. This was played by and fifth move deviations should trou
Kramnik again st Leko in Gam e 3 3, ble Black at all as long as he knows
where Kramnik was lucky to escape wh at he's doing - in which case some
with a draw after a mistake in the of them are even a bit risky for White.
opening. Vladimir Petkov tried the And with th at we complete our cover
sam e line again st m e in G am e 34, and age of the Benko proper and now turn
even g ained the advantage when I our attention to the anti-Benko varia
played inaccurately in a balanced mid- tion s: 2 C4 cS 3 lLlf3 and 2 lLlf3 cS.
148
C h a pter Seven
The Ka s parov G a m bit
1 d4 li:Jf6 2 c4 cS 3 li:Jf3 pawn on dS.

Perh aps unsure wh eth er, after 2 ... cS To n am e thi s g ambit after Kasparov
3 dS, his opponent intends the Benko is a little unfair, given th at Vaganian
or som e variety of Benoni, White de and others had already played it sev
cides to pl ay it safe. Or so he thinks - eral times in the late 1970'S. However,
since Bl ack has another rather danger it was Kasparov's high-profile win in
ous respon se: The Kasparov Gambit ! the first g am e below which led to it
3 ... cxd4 4 li:Jxd4 eS 5 li:Jbs d S 6 cxd S becoming popular for Bl ack, an d his
i.cs n am e h as stuck.
Game 39
A.Mikhalchishin-G.Kaspa rov
USSR C h a m p i o n s h i p,
Fru nze 1981
1 d4 li:Jf6 2 c4 cS 3 li:Jf3
Thi s position can al so be reached
from flank openings, such as l li:Jf3 li:Jf6
2 c4 cS 3 d4 or 1 c4 cS 2 li:Jf3 li:Jf6 3 d4.
Black can gain good activity with 3 ... cxd4 4 li:Jxd4 eS s li:Jbs
this line, and I think it is sound as well. Oth er m oves for White, including 5
Bl ack already has a lead in develop li:JC2, are examined in the l ast g ame of
ment, while the white knights on b S thi s ch apter.
and bl are a bit clum sy, since both of S ... dS 6 cxd S .lies
them would like to go to the c3-square. Obviously n ot 6 ... li:JxdS ? ? 7 'ii'x dS !
Often a theme is simply to regain the and White win s a piece.
149
7 lLlsC3 involves .. :fle7 and .. ,IUd8, followed

Answering the immediate threat of either by ... lLlbd7-b6 (or ... lLles) or just
7 ... 'iVb6, with a double attack on bS and ... lLlc6 if White allows the ds-pawn to
f2. be captured (as in the game). The im
7 e 3 0-0 8 lLlSC3 comes to the sam e portant thing is for Black to develop
thing. White occasionally plays 8 �e2 rapidly and play aggressively.
lLlxdS 9 0-0 with a reversed Sicili an, but 9 � e2
h as no chance of an advantage here; Alternatively:
e.g. 9 ... lLlc6 10 a3 �e6 11 b4 �b6 12 a) 9 �C4? ! secures the ds-pawn but
lLllc3 lLlxc3 13 lLlxc3 f6 and Black was exposes the bishop to attack; for in
fine in B. Kurajica-B.Vuckovic, Bosnian stance, 9 ... lLlbd7 10 'ilc2 lLles 11 lLlxe4
Team Champion ship 2009. �fS 12 lLlxf6+ 'ilxf6 13 e4, as in
Fin ally, the tempting but incorrect 7 R. Parker-C.Matamoros Franco, London
d6? is refuted in G am e 42. 1987, when 13 .. JUe8! would have been
7 0-0 8 e3
... very strong , as 14 0-0 run s into
The first of three alternatives for 14 ... lLlf3+ ! (another reason the bishop
White; 8 h 3 and 8 g 3 (the m ost critical should be on e2) 15 �h l (or 15 gxf3
line) are seen in the n ext two g ames. �h 3 and m ates) l S .. :irh4 16 gxf3 i.. d 6
8 e4
•.• and wins.
b) 9 a3 'ile7 10 lLld2 J:!d8 11 b4 (here
11 �e2 returns to the main game)
11...�d6 12 lLlc4 �c7 1 3 lIa2 (the idea
is to switching the rook to d2 to defend
the d-pawn, but this is easily neutral
ized) 1 3 ... �g4 14 �e2 �xe2 15 l:[xe2
lLlc6 16 �b2 (or 16 l:.d2 lLles 17 d6? !
�xd6 ! 18 lLlxd6 lLld3+! - Krasenkov)
16 ... lLles 17 lLlxes �xe s 18 ':'d2 .:tac8
19 'iWb3 lLlg4! (seeking more th an
19 ... �xc3 20 �xC3 lLlxdS with equality)
The advance ... e S -e4 is the standard 20 h 3 'iVh4 and Black had a strong ini
m ove h ere and it the best. The pawn tiative in M. Krasenkow-I.Smirin, Vilnius
takes control of d3- and f3-squares and 1988.
rules out any l ater idea of the white e 9 'iWe7
...
pawn supporting its colleague on dS. Further protecting the e4-pawn and
My evaluation of this position is m akin g room for ... lId8, putting pres
that Black has sufficient compen sation sure on the ds-pawn - cl central theme
for the pawn . The usual plan of action in the Kasparov Gambit, since regain-
1 50
Th e K a s p a ro v G a m b i t
ing the pawn will give Bl ack a comfort 1 4 'ii'X e4 'ili'xe4 1 5 lLldxe4 lLla6 was
able g ame. equal in I.Morovic Fernandez-J. Liew
(hee Meng , Dubai Olympiad 1986, as
Black will soon regain his pawn .
11 lLlxd SI
...
10 lLld2
10 a3 l:td8 11 lLld2 tran sposes be
low, while 11 b4 �d6 12 �b2 as
(12 ... �es is also good) 1 3 bxas �es 14 Black can simply take the ds-pawn,
1ia4 lLla6 1 5 lLld2 lLlcS 16 'iWb4 �g4 17 since it is rather risky for White to take
lLlC4 �xc3+ 18 'ivxc3 �xe2 19 �xe2 on e4 (see the n ext n ote).
'iWd7 20 J:thdl .:tac8 g ave Black great 12 lLlxd S
play for the pawns in lTimman After 12 lLlcxe4 Black h as excellent
V.Topalov, Wijk aan Zee 1996. compensation for the pawn, and the
10 l:.d8
... sacrifice on e 3 is a serious threat. For
instance, 12 ... �b6 (even 12 ... lLlxe 3 ! ? 1 3
fxe 3 � x e 3 14 'ili'b3 �xd2+ 1 5 lLlxd2
lLlc6 16 lLlf3 1:te8 g ave Black a strong
initiative in E.Barkovsky-Ma.Tseitlin,
Leningrad 1981) 1 3 lLlc3 (after 1 3 'ifb3
lLlc6 14 0-0 lLlas 1 5 'ili'a4 �d7 16 'ili'dl
�e6 or 13 lLlg 3 lLlc6 14 0-0 �e6 Black is
doing fine with his far more active
pieces) 13 ... lLlxe 3 ! ? (a very tempting
sacrifice, though the simple 1 3 ... lLlxc3
14 bXc3 �fS 15 0-0 lLla6 i s also g ood, as
ll a3 White is quite passive and h as weak
11 'ili'c2 �fS 12 a3 lLlxdS 1 3 lLlxdS squares and weak pawns) 14 fxe 3
l:[xdS is another transposition below, 'ii'x e3 1 5 l:tfl lLlc6.
whereas 12 g4! ? lLlxg4 1 3 �xg4 �xg4 This position is very tough for White
151
t o defend: h i s king is stuck in the cen- concession s at all .

tre and i t is difficult for h i m t o extri
cate his pieces, whereas Black h as a
great m any active possibilities.
16 .. 'ii'g S
.
The most active move in the posi

tion and the most uncompromising;
The g am e E . Postny-I , Smirin, I sraeli Garry is ready to give up the e4-pawn
Team Championship 1999, continued to accelerate his initiative.
16 'ifa4? ! (16 'ifb3 i.. e 6 17 lDC4 was a Otherwise Black has several ways to
better try, though Bl ack still has great m aintain at least equality, such as
play after 17 ... 'ifC5, threatening ... lDd4) 16 ... lDe5 17 i.. x e5 lIxe 5 18 lDC4 :c8 19
16 ... lDe5 17 ltdl i.. e 6 18 Ilel ? (18 ltc2 llfdl i.. C 7 20 lDxe 5 i.. x e5 21 'ifbl i.. x al
was perhaps the last chance, when 22 'ii'x al h6 23 'ii'd4 .l:tc2 or 16 ... i.. C 7 17
18 ... l:tac8 19 lDC4 'ifxe2+! 20 lDxe2 l::tfdl lIc8 18 lDC4 l:txdl+ 19 l::t x dl a6,
i.. x c4! 21 ltbl i.. x e2 22 l:[el i.. d 3+ 2 3 while 16 .. Jk8 17 lDC4! i.. C 7 18 .l:.fdl
lta2 l:td4 24 b 4 .l:.d5 2 5 i..f4 i.. C 4+ 2 6 .l:.cd8 19 l:txd5 .l:lxd5 20 l:.dl ':'xdl+ 21
lt b l i.. b 5 reg ains the queen with a 'ii'x dl a6 2 2 lDd2 lDe 5 2 3 g 3 lDd3 24
clear advantage) 18 ... I:td4 (now the i.. d4 h 6 is unclear; White has a good
white king cannot escape and the at bishop on d4 and Black a good knight
tack is decisive) 19 i..f3 'ii'f4 20 'ii'c 2 on d3.
lDxf3 2 1 l::t e 4 'ii'f6 2 2 :xe6 fxe6 23 g xf3 17 lth1 nd6?!
'ifxf3+ 24 ltel 'iff2+ 0-1. A nice g ame Perhaps a bit too direct. 17 ... 1:.c8
by Smirin, illustrating the dangers for looks stronger, and if 18 :adl then
White if h e underestim ates Black's ac 18 ... .l:tdd8 (Kasparov) or just 18 ... lDxb4,
tivity in the Kasparov Gambit. while 18 lDC4 i.. C 7 19 ltfdl 1:.xdl+ 20
12 ...l:txd S 13 'ii'C 2 i..f s 14 b4 i.. b 6 lS l:txdl b5 21 lDd2 i.. xh 2 ! 22 ltxh 2 lDe5
i.. b 2 lDc6 16 0-0 23 'ii'x c8+ i.. x c8 24 lDxe4 'ii'h 4+ 25 <;itg l
Bl ack h as n o problems here, h avin g lDC4 remain s unclear.
regained his pawn without m aking any lS lDxe4 i.. x e4?1
152
Th e Ka s p a r o v G a m b i t
And here Black might have pre 2 3 .i.a5 24 llcS?

•••
ferred 18 .. :.g 6 19 f3 l:.ad8 20 'ii'C 3 Both the obvious m ove and the cru
i.xe4 2 1 fxe4 ':'d2 2 2 1::tf2 (or 2 2 :ae1 cial mistake. After 24 a4! White might
i.C7 with compensation) ':xb2 2 3 still hope to draw; for instance, 24 ... l:tb4
'Wxb2 i.xe3 2 4 i.C4 (if the rook moves, 2 S ':'c4 l:.xc4 26 'ii'x c4 'ii'e 7 27 'ili'c6 :d8
Bl ack has 24 ... l:I.d2 or 24 ... .i.d4) 28 :c1 i.C7 ! 29 g3 i.b8 and although
24 ... i.xf2 2S 'it'xf2 ttJe s 26 i.dS b6 2 7 Black h as fin ally h as the passed pawn
'Wf4 l:Ie8 28 h 3 h 6 , intendin g ... l:.e7 and under control it is n ot clear th at
... 'ii'g s with an unclear g ame. whether can m ake any progress.
19 1i'xe4 .l:td2 24 ... 'it'b5!
20 b5?! Now Mikha1chishin h as nothing

Now it is White's turn to go wrong . better than to go into an endgame
2 0 i.d3 ? fails t o 2 0 .. :�g 6 ! , s o Mikhal with a queen for two rooks, which is
chishin decides t o sacrifice a piece t o favourable for Black.
obtain a pawn o n the seventh, which 25 1:.fc1 'ii'x b7 26 'ii'e S 'ii'x cs 27 'ili'xcs
may well be sufficient to draw. H ow i.d2 2S h3 h6 29 'ili'C4 i.xc1 30 'ili'xc1
ever, he had a stronger, if unexpected, Iixf2
continuation in 20 i.a6 ! bxa6 21 'ili'xc6 The rest is quite simple.
(Kasparov), when 21 ... l:tad8 22 i.c3 is 3 1 'Wc7 a6 3 2 'ili'a7 :f6 3 3 a4 :dS 34 as
actually good for White. .l:1d1+ 35 <ifi>h2 ,U,d2 36 'ii' b S+ 'itt h 7 37
20 .. J!xe2 21 bxc6 l:txb2 22 cxb7 llfS 2 3 'ili'b4 :ff2 3 S 'We4+ f5 0-1
l:tac1
23 J:[fC1!, keepin g the option of Game summary
.l:tab1, was more accurate, since Black Against the 8 e3 line it is a good plan
could hardly take on f2. Kasparov in for Black to target the ds-pawn, to
tended 2 3 ... i.cS 24 :ab1 i.xa3(??) 2 S which end the idea of ... "fie7 and ... l:td8
.l:txb2 i.xb2, but this run s into 2 6 'it'b4! . is quite effective. Even if White tries to
1 53
keep the pawn, Black obtain s easy be quite dangerous if Bl ack is not pre
compensation with his l ead in devel pared to a certain level, so he has to
opment and active pieces. In the game pl ay with some precision to achieve an
Kasparov regained the pawn but, char unclear g ame. The drawback to White's
acteristically, was n ot satisfied with plan is th at it creates weaknesses in his
equality as Black and sacrificed the e4- kin g side.
pawn in order to sharpen the play. Ul 8 .i.fS !
...
tim ately his uncompromising strategy Getting the bishop out before White
was successful, although Mikha1chishin prevents it with g 2 -g4, while planning
should probably h ave drawn, and to meet th at move with ....i.e4. Then
might even have been better at one Black will continue to develop rapidly,
point. with moves like ... ttJbd7-b6(e5), ... l1ac8,
... ttJc4, and possibly ... .i.d4 (should
White play e 3 -e4). If White pushes on
Game 40 the king side with g 2-g4 and h 3 -h4,
A.Raetsicy-D.Tyomkin Black can consider ... h 7-h 6 or ... h7-h 5 or
Biel 1999 sometimes just ignore it.
The advance ... e5 -e4, on the other
1 d4 ttJf6 2 c4 cS 3 ttJf3 cxd4 4 ttJxd4 eS h an d, is more risky for Black in this
5 ttJbs ds 6 cxd s i.. c S 7 ttJSC3 0-0 8 h 3 variation and, if played at all, needs to
be well timed. For in stance, the imme
diate 8 ... e4? ! would met met by 9 g4,
wh en the disruptive pawn sacrifice
9 ... e3 isn't really very good after 10
.i.xe3 .i.xe3 1 1 fxe 3 'iid 6 12 'iid4! 'iig 3 +
1 3 � d 2 intending 'iif4. In the spirit of
the gambit, Bl ack has to pl ay at a good
pace with con stant active and aggres
sive m oves, but it's possible to take
that too far.
9 a3?!
A more ambitious approach th an This move, preventing ... ttJb4 or
the immediate 8 e3 of the previous ... .i.b4 and preparing b2-b4, is often
g ame. With h 2-h 3 White is planning to useful in the Kasparov Gambit, but just
undermine the black e-pawn, if it ad here it seem s like a luxury. In stead,
vances to e4, with g 2-g4 and possibly White h as mostly preferred:
g4-g 5, and at the sam e time intends to a) 9 g4 is an swered by 9 . . .i.e4! and
.
develop the bishop to g 2 . Thi s line can then :
1 54
Th e Kasp a r o v G a m b i t
bly C S ) 1 3 ttJd2 (preparin g e2-e4)

13 ... ttJd6 14 hs i.fs 15 e4 i.d7 16 ttJC4
fs 17 h6 g6 18 a3 ttJxe4 19 ttJxe s i.d6
20 'iVd4 with a very unclear position .
White's pieces are quite unorthodoxly
placed, but he does h ave a m atin g
theme o n g 7 and a potential check on
c4 or ds (if ds-d6 is all owed). For in
stance, 20 ...ttJxC3 (or 20 .. :iIi'e8 2 1 f4
ttJxg s 2 2 �dl) 2 1 i.e3 ! 'iVe7 ! 2 2 ttJxg 6 !
hxg6 2 3 h 7 + �f7 24 h 8'iV l::t x h 8 2 5 l::t x h 8
al) 10 f3 ? is quite bad for White, be l:1xh 8 2 6 'ifxh8 f4 2 7 'iVh 7+ �f8 with a
cause of 10 ... ttJxds ! 11 ttJxe4?! (11 h4 is draw.
required, even though 11 ... ttJxC3 12 b) 9 e 3 ! is the best continuation
"it'xd8 ':xd8 13 ttJxC3 i.c6 is very com here, and then :
fortable for Black) 1 1 .. JWh4+ 12 �d2
':d8 (Palliser), when 13 ttJxcs ttJe3+ 14
ttJd3 ttJxdl win s the queen and proba
bly the game, since most of White's
pieces haven't moved at all .
a2) 1 0 ttJxe4 ttJxe4 is also fi n e for
Black; for instance, 11 e3 i.b4+ 12 ttJd2
"it'f6 13 'i!Ve2 .l:[c8 14 a3 i.xd2+ 15 i.xd2
ttJa6 16 i.g 2 ttJacs 17 i.xe4 ttJxe4 with
more than enough compensation,
thanks to White's numerous weak
nesses. hi) 9 ... ttJa6 ? ! g ave Black a convinc
a3) 10 l::th 2 seems to be the best ing win in the g am e quoted below. I
move when the position is unclear, al will give a bit of coverage on this, since
though I would prefer Black in practice, it is useful for the understandin g of the
given White's vulnerable kin g on el variation, though I think other m oves
and displaced rooK. One possible con are stronger.
tinuation might be 10 ... i.g 6 1 1 g s (see b11) 10 .1i.xa6 ? ! bxa6 1 1 'iVf3 .1i.g6 12
ing as White can 't castle short he might e4 ttJe8 13 g4? ! (a risky m ove, as it cre
as well use the rook actively and try to ates weaknesses around the white
do something on the kingside) 11 ... ttJe8 kin g ; stronger was 13 i.e3 ! i.d4 14 0-0,
(the knight will be well placed on d6) after which Black has insufficient com
12 h4 ttJa6 ! (on the way to b4 or possi- pensation) 1 3 ... ttJd6 14 0-0 i.d4 1 5 ttJd2
1 55
17 .. .fS 18 h S .i.e8 19 g 6 h 6 (or 19 ... hxg6

20 h 6 ! ) 20 'Wb3 and White still has the
extra pawn .
b2) 9 ... e4! ? is quite challenging,
given that h 2-h 3 specifically prepared
to m eet this move; i.e. 10 g4 .i.g6 11
.i.g2, but then 11 ...4:Ja6 (or 11 ... 4:Jbd7,
as in note 'b3 3 ' below) 12 g S ? ! (12 a3
4:JC7 13 h4 looks a better try) 12 ... tiJb4!
1 3 g xf6 4:Jd3+ 14 �d2 (or 14 �fl 'ikxf6
1 5 'ii'd 2 'ikh4) 14 ... 'ii'xf6 gave Black ex
l S .. .fS ? ! (not really sound but it cellent compen sation for the piece in
worked in practice; lS ... ':c8 was more K.Aseev-I.Smirin, Klaipeda 1988. The
sen sible) 16 exfs 4:JxfS 17 g xfS ':'xfS 18 g ame continued 15 l::tfl (not 15 f4?
'We2 ? (after the awkward-lookin g 18 .i.xe3+! and wins) l s .. .1Ue8 16 a3 (16
� 1 ! Black h as no obvious way forward �C2 might offer more hope) 16 ...bS 17
and White is clearly better) 18 ... 'ili'h4 19 :ta2 hS 18 4:JxbS 'Wg s 19 .i.h l? 4:Jxf2 !
4:Jde4? ? (one mistake foll ows another; 20 'iVe2 4:Jxh l 2 1 l:txh l 'WxdS+ 22 �C2
White h ad to accept the inferior end 'Wxa2 and Black went on to win.
g ame after 19 �h 2 'iVf4+ 20 �h l l:th S b3) 9 ... 4:Jbd7 ! looks like the strong
2 1 'Wf3 'Wxf3+ 2 2 4:Jxf3 .i.xc3 2 3 bxc3 est reply. Black is happy to allow e3-e4
.i.e4 24 �h 2 .i.xf3 25 d6 with reason as his pieces can work around the cen
able drawin g chances) 19 .. :ii'xh 3 (now tre:
there is n o defence) 20 4:Jg s l:[xg S+ 2 1
.i.xg s .i.h s 2 2 'We4 .i.f3 2 3 'Wxf3 'Wxf3
24 .i.e3 l:tf8 0-1 J .Werle-V.Mikhalevski,
Hoogeveen 2000. A nice attack, ag ain
illustratin g the dang ers for White in
the Kasparov G ambit, even if it wasn 't
entirely correct.
b12) 10 a3 ! (preventing ... 4:Jb4)
10 ... 4:JC7 11 g4 (this seem s like the right
time for this m ove; after 11 b4? ! .i.e7
Black h as compensation for the pawn)
1 1 ... .i.g6 12 .i.g 2 'ike7 1 3 g s 4:Jfe8 14 h4 b31) 10 a3 tran sposes to the main
e4 (if 14 .. .1:tc8, then 1 5 e4 and h4-h S is a g ame.
strong threat) 1 5 4:Jd2 4:Jd6 16 b4 .i.b6 b 32) 10 4:Jd2 ? ! .i.b 4 (a positional
17 .i.b2 with a slight advantage; e.g. move, aiming to undermine the dS-
156
Th e Kasp a r o v G a m b it
pawn and/or weaken the white struc vantage for Black) 18 ... :c8 ! (18 ... bxc3
ture) 11 a3 Jtxc3 12 bXc3 4JC5 13 c4 19 bXc3 Jtxf2+ 20 'ii'xf2 4Jfxe4 21 Jtxe4
'ii'a 5 14 Jte2 4Jfe4 15 Jtb2 b 5 ! with a 4Jxe4 22 4Jxe4 ..txe4 2 2 d6 is equal) 19
strong attack and the advantage. 4Jb5 4Jxd5 and Bl ack regain s the pawn
Bl ack's pieces will soon be very active, with an unclear g ame.
while the white king is stuck in the cen 9 .. 4Jbd7 10 e3 e4
.
tre. Other m oves were g ood too:

b33) 10 g4 Jtg6 11 Jtg 2 e4! ? 12 g 5 a) 10 ... 4Jb6 ! ? 1 1 b4 Jtd6 1 2 'i'f3
4Je8 13 4Jxe4 Jtxe4 1 4 Jtxe4 'ii'x g 5 1 5 'ii'c 8 ! 1 3 Jtb2 (not 1 3 e4? Jtxe4 ! ) 1 3 ... e4
'iVg4 'ilie7 1 6 4JC3 4Jef6 1 7 'ii'f 5 g 6 18 14 'ii'd l l:td8 with compensation
'iVg 5 4Jxe4 19 ii'xe7 Jtxe7 20 4Jxe4 f5 (Tyomkin).
21 4Jc3 4Jb6 with g ood compen sation. b) 10 ...':c8 1 1 g4 Jtg6 1 2 Jtg 2 4Jb6
Bl ack might well reg ain the d-pawn or 13 h4 (after 13 b4 Jtd6 14 e4 ':'C4 Black
in some cases play more actively with h as g ood play with his lead in devel
... Jtf6, ... l:tac8 and ... 4JC4. opment and White's weaknesses on e4
b34) 10 'ii'f3 Jtg 6 11 e4 Jtd4 12 Jtd3 and on the king side) 1 3 ... h 6 14 e4 4JC4
ttJC5 15 g 5 hxg 5 16 hxg 5 4Jh 7 17 "g4 Jtd4
18 0-0 4Jd6 with compen sation.
11 b4?!
Thi s creates weaknesses on the dark
squares and l ooks mistimed. A less
committal m ove such as 1 1 4Jd2 might
h ave been better.
11 ... ..t d6 12 Jt b2
13 JtC2 (after 13 O-O?! 4Jxd3 14

'i'xd3 4Jxd5 15 'i'e2 4Jxc3 16 4JxC3 'iVh4
Black had the more active position and
the bishop pair and went on to win
R. Palus-A.Vaulin, Pol anica Zdroj 1998)
13 ... b5 14 a3 as 1 5 0-0 b4 16 axb4 axb4
17 l:r.xa8 'ii'x a8 18 ttJd2 ! (White can
leave the knight on c3 because the 12 .. 4Je5
.
bishop on d4 is trapped after b2xC3; 12 ... 4Jb6, intending to target the d

other moves lead to at least a clear ad- pawn after ... Jte5, was also fine for
157
Black, since 1 3 g4 i. g 6 1 4 h4 is safely protected d-pawn is a constant danger

m et by 14 ...h S . for White.
13 tl)d2 l:tc8 1 4 i.e2 24 rJilf2?!
14 g4 i.g6 1 5 gs is risky here, be 24 rJild2 i.bs 2 5 :cl l:ld8 26 'ifl'el!
cause of lS ... l:txC3 ! (a nice and very tl)C4 2 7 i.d2 was a better defensive
logical sacrifice: Black gets the dS- and configuration .
g s -pawns for the exchange) 16 i.xC3 24 ... i.a4 2 5 l:tC1 i.c2 26 i.d4 tl)C4
(16 g xf6 tl)d3+ 17 i.xd3 l:lxd3 18 fxg 7 N ow all White's pieces look more or
l:[e8, foll owed by . . .i.es, is good for less displaced and he was also in severe
Black) 16 ... tl)xds 17 i.d4 "ii'x g s , when tim e trouble.
White faces long-term problem s re 27 i.C5 b6 28 i.e3 f6
g arding the safety of his kin g .
1 4. . .tl)d3+ 1 5 i.xd3 exd3 16 11c1?!
White might still h ave m aintained
equality with 16 'ii'f3 'iid 7 17 1:cl i.es
18 0-0 l:tfe8 19 l:tfel i.xc3 20 l:txc3
l:[xc3 21 i.xC3 tl)xds 22 i.d4 i.g6,
when a draw is the m ost likely result.
16 i.e5 17 'ii' b 3 i.xC3 18 1ixC3 IixC3 19
•••
i.xC3 'ii'x d 5
29 tl)d2?!
Giving up the a-pawn in the wrong
way. Instead, 29 a4 tl)b2 ? ! 30 rJilel tl)xa4
3 1 rJild2 tl)b2 32 tl)a3 tl)c4+ 33 tl)xC4
liIxC4 34 l:tal (Tyomkin) or 29 ... i.xa4?!
30 tl)d2 i.bs (or 30 ... bS 3 1 :al) 31
tl)xC4 i.xC4 3 2 rJilel would offer White
reasonable drawin g chances due to the
opposite-coloured bi shops, although
The endgame is slightly better for 29 ... rJilf7 30 tl)d2 tl)b2 still leaves Bl ack
Black, who h as a dangerous d-pawn, with a favourable endgame.
while the white pieces are rather pas 29 tl)xa 3 30 lla1 tl)b5
.•.
sive. Black has won a pawn while keep

20 'iix d 5 tl)xd 5 2 1 tl)b1 tl)b6 22 f3 :c8 ing the knights on and stands clearly
23 e4 i.d7 better, especially as White had no time
On its way to c2, after which the on the clock to try and find a defence.
1 58
Th e K a s p a r o v G a m b i t
31 h4 h S 3 2 g4 hxg4 3 3 fxg4 1;f7 34 h S

.l:!.C3 3 S 1;f3 .l:. a 3 36 .l:[h1 .l:.a4 3 7 h6 gxh6 Game 41
38 ':'xh6 lbb4 39 1;f4 ttJC3 40 l:th7+ M.Matlakov-Alexa.Evdokimov
1;g6 41 ':xa7 R u s s i a n Tea m
C h a m pi o n s h i p 2010
1 d4 ttJf6 2 c4 cS 3 ttJf3 cxd4 4 ttJxd4 es

5 ttJbs d s 6 cxd s i.cs 7 ttJsc3 0-0 8 g3
41 ... ttJxe4! 42 ttJxe4 �xe4+ 0-1

A nice concluding tactic. If 43 1;xe4
then 43 ... d2+ and the pawn promotes,
or if 43 1;f3 then simply 43 ... bS is win
ning; e.g. 44 l:1b7 b4 4S ltb6 b3 46 g s S g3 is the theoretically most testin g
l:!.xe3+! 47 '\tlxe3 1;xg s (Tyomkin) and line in t h e Kasparov Gambit. It was the
one of the pawn s will go through . choice of Aronian in a g am e again st
Sutovsky (see below). Given one mo
Game summary m ent and White will h ave played i.g2,
Although the move S h3 can be dan 0-0 and e2-e4 with a clear advantage,
gerous against an unprepared oppo so Black n eeds to disrupt this plan right
nent, after the accurate S ... i.fS ! Bl ack n ow.
shouldn 't have any problems. Even 8 .. .'iWb6!?
again st White's best reply of 9 e3, ei It is n ot so easy get an unclear g am e
ther 9 ... e4! ? or 9 ... ttJbd7 ! offers Black in another way here. I spent some time
excellent counterplay for at least on other continuations, but I won't go
equality. In the game, 9 a 3 ? ! and 1 1 into detail since I could n ot fin d a clear
b4? ! were not the most precise moves, path to equality.
and the further error 16 l:tc1 ? ! allowed g e3
Black to reach a favourable endg ame. Obviously the only move.
Aided by his opponent's subsequent 9 ... i.g4!
time trouble, Tyomkin went on to win In stead, 9 ... l:tdS h as some logic to it
convincingly. - attacking the ds-pawn - but then
1 59
White g ets t o play 1 0 �g 2, supporting Black is about to regain the pawn with
the pawn, and stands slightly better. the superior structure. And blocking
For instance, 10 ... �g4 11 'ii'c 2 tLlbd7 12 the bishop by 10 f3 weakens the white
a3 ':'ac8 1 3 0-0 'ii'a 6 14 tLld2 �e7 1 5 king side, so that after 10 ... �d7 Black
tLlb3 tLlb6 16 e 4 tLla4 17 l:t e l tLlxC3 18 has good compensation for the pawn
bxc3 'ii'a4 19 lla2 when Black h as some with his lead in development. For ex
compensation, but probably not ample:
enough. a) 1 1 'ii'e 2 e4 and then :
10 �e2 al) 12 tLlxe4 tLlxd5 13 tLlxC5 'ii'x C5 14

The best reply, even though it 'ii'C 4 'ii'd 6 15 tLlC3 tLlxC3 16 'ili'xC3 l:tc8 17
means Black h as achieved his first aim 'i!Vd3 'i!Ve7 18 "ilr'a3 'i!Ve8 19 e4 tLlc6 (with
of preventin g the fianchetlo. Moving ideas of ... tLld4 or ... tLle 5) 20 i.e3 as 21
the queen would allow Black to play �f2 tLlb4 (taking another route) 22 :Ldl
... �f3 with an easy g ame; e.g. 10 'ii'h 3 ? ! f5 23 exf5 �xf5 24 �f4 �e6 25 l:td2
'ii'x b3 1 1 axb3 �f3 'ii'f7 and the white king is still exposed
with compen sation, while Black is
ready to reg ain the pawn on a2.
a2) 12 fxe4! ? was tried in
V.Tverdohlebov-I . Foigel, Ukrainian
Team Championship 1980, when
12 ... tLlg4 13 tLla3 (or 13 tLld2 i.xe3 14
tLlC4 �f2+ 15 �d2 'ii'c 5) 13 ... �xa3 14
bxa3 ':'c8 15 �d2 tLle 5 16 'i!Vf2 tLla6
g ave Black good play for the pawn s. He
might al so have considered 12 ... �g4 ! ?
1 3 'ii'f2 (or 1 3 'ii'b 5 �f3 1 4 'ii'x b6 axb6
12 l:tg l l:r.d8 13 �C4 �b4, when 15 ':'g l tLlbd7 16 tLld2 tLle 5 17 i.e2
1 60
Th e K a sp a ro v G a m b it
i.xe 3 18 .l:tfl i.g 2 19 :fS l:tfe8) worth investigatin g . The only drawback
13 ... ltJbd7 14 h3 i.h S 1 5 i.g2 ltJe s 16 is that White doesn't n eed to allow it.
0-0 ltJd3 17 ltJa4 ltJxf2 18 ltJxb6 ltJxh 3+ Therefore I would prefer the popular
19 i.xh 3 i.xb6 again with compen sa 11 ... e4! , after which Black can obtain
tion, since the pawn s in the centre are sufficient compensation for the pawn
weak and Black is still very active. with correct play:
b) 11 'ii'd 2 keeps an eye on the dS
pawn, but then White's pieces are even
more clum sily placed, and Bl ack can
gain active counterplay with moves like
... ltJa6-b4, ...l:.ac8, and ...i.b4. Play
might continue 11 ... ltJa6 (not now
11 ... e4? ! 12 ltJxe4! and White is better)
12 ltJa3 lUe8 13 e4 .:tac8 14 ltJc4 'iVd8
15 a4 h6 16 b3 ltJh 7 17 i.a3 i.d4 18
l:[cl 'ii'f6 19 'ii'd l ltJcS 20 i.g2 a6 21 as
i.bs 22 11c2 i.xC4 23 bXc4 h S 24 h4
ltJb3 25 'iVd3 'iVd8 with ongoing com a) 12 'iVh S ? ! (this unambitious m ove
pen sation . gives Black an easy g ame) 12 .. :iVxbS 1 3
10 ... i.xe2 ltJxbs ltJxds 1 4 ltJ l c 3 ltJxc3 1 5 ltJxC3 fS
10 ... i.h 3 ? ! 11 g4!, threatening l:tg l 16 �e2 ltJc6 17 i.d2 ltJes 18 ltJds llad8
g 3 , was good for White in V. Korchnoi V2-V2 P. H . Niel sen-B.Gelfand, Monte
D.Gurevich, Las Vegas 2007. Carlo (blindfold rapid) 2006, and Black
11 'iVxe2 certainly h ad the edge final position .
b) 12 ltJd2 (not the m ost precise; 12
a3 offers White m ore chances of g ain-
ing an advantage) 1 2 ... i.b4! (now the
ds-pawn is weak; 12 ... l:te8? ! would al -
low White to rectify his mistake with
13 a3 ! , tran sposin g to line 'd' below) 1 3
0-0 :t e 8 1 4 lI d l ltJbd7 1 5 a3 (or if 1 5
l:tbl, a s in C . Lingnau-A.Kadatsky, Par
dubice 2001, then lS ... l:.ac8 16 a3 i.f8
17 ltJC4 'iVh 3 ! ? 18 ltJas 'ilb6 19 ltJC4
with a repetition) lS ... i.xc3 16 bxc3
11 ... ltJbd71? ltJxds 17 ltJxe4 'ii'c 6 18 'ii'd 3 ltJ7b6 (or
This prepares a very interestin g idea 18 ... ltJsb6 ! ? 19 ltJd6 ltJes 20 'iVh s 'ii'x bs
on the next move, which is certainly 21 ltJxbs l:ted8 22 ltJd4 ltJa4 23 i.d2
161
J:tacS 2 4 l:tabl b 6 2 5 l:tb4 1DxC3 2 6 i.xC3 In his annotation s Aronian sug

l1xC3 with a level position} 19 litbl ? (19 gested 12 .. ..l:teS ! as a possible im
c4 is equal) 19 ... l:tadS 20 c4 was provement, and in my opinion Black
F . Kristen sen-Ca. H an sen, correspon h as sufficient compensation after this
dence 1994, and n ow Black could just move. N ow 1 3 0-0 lDbd7 would trans
h ave taken 20 ... lDxC4! with a clear ad pose to line 'c' above, while 13 b4 i.d6 !
vantage. The threat is 2 1...f5 and is also fin e for Black. So that leaves 13
2 2 ... lDdxe3, while 2 1 lDg 5 ? loses to lDd2 and then :
2 1...lDf4 ! .
c } 12 0-0 lDbd7 (12 ... i.b4 ! ? is possi
ble here too, when 13 a3 i.xc3 14 1Dxc3
lDbd7 transposes to 1 3 0-0 in the m ain
g ame) 1 3 a3 lUeS ( 1 3 ...l:tacS 14 lDd2
lUeS 15 b4 is slightly better for White -
see 1 2 O-O! in the m ain g ame) 14 b4 (or
if 14 lDd2, then 14 ... 'iIi'c7 ! ) 14 ... i.d6 !
(possible because White doesn 't h ave
lDC4 yet) 15 lDd2 i.e5 16 lDC4 fidS 17
i.b2 i.xc3 is i.xc3 lDxd5 19 i.b2 was
seen in the freestyle g am e " King dl} 1 3 ... a5 14 'ili'C4?! (better was 14
Crusher" IRybka-"Equidistance", Inter b 3 ! lDa6 1 5 i.b2 l:tadS 16 1Da4 'iVd6 17
n et 2006. Black shoul d be able to with lDc4 'iVxd5 is i.xf6 gxf6 19 0-0 with an
stand White's slight pressure h ere; for edge for White) 14 ... lDbd7 15 lDdxe4
instance, 19 ... fie7 20 %1adl We6 2 1 l:td2 lDxe4 Y2-Y2 A.Maly-O. Berezin, Kharkov
l:tacS 22 ,Ufdl 1D7b6 23 lDxb6 lDxb6 24 2000. Black is okay here and could have
i.al h6 25 'Wh 5 1Dc4 2 6 .l:.d7 b6 2 7 'itg 2 played on. For instance, 16 lDxe4 lDe5
a6 2 S J:.ld5 b 5 2 9 i.d4 l:taS 30 g4 %:tacS 17 'iVa4 lDd3+ is 'itfl i.b4 ! ? 19 axb4
31 h4 l:te7 32 l:txe7 'iVxe7 33 g 5 :c6 34 l:[xe4 20 'itg 2 l:txb4 21 'iVc2 'iVg 6 22 'iVe2
'ili'g4 fie6 35 .l:r.dS+ 'ith 7 remain s very 'iWe4+ 23 Wf3 f5 ! is good for Black, due
unclear. to the strong knight on d3 and White's
d} 12 a3 ! is critical . This was played weaknesses at f2, d5 and b2.
in the aforementioned g ame, d2} 1 3 ...'iVC7 (Aronian's move, which
L.Aronian - E .5utovsky, European Team hinders b2-b4 for the time being) 14
Champion ship, H eraklio 2007, in which g 4 ! ? (a drastic respon se thrown up by
White g ained the advantage after the computer to undermine the e4-
12"'lDbd7 13 lDd2 l:tfeS 14 b4 i.fS 1 5 pawn; otherwise 14 0-0 a6 1 5 ,Udl i.d6
0-0 'fiC7 16 i.b2 lDb6 17 lDb5 ! and offers reason able compensation, with
went on to win convincingly. ideas of ... lDbd7, ... l:tacS and ... lDc 5)
1 62
Th e K a s p a r o v G a m b it
14 .. .'ii' e s 1 5 ttJC4 'ii'g s leads to a very Again 1 3 O-O! would h ave been
unclear position . Play might continue m ore testing, although Black gets
16 d6 (or 16 ' b4 .i.fS) 16 ... .i.b6 17 b4 enough compen sation n ow:
ttJbd7 lS .i.b2 .l:r.acS 19 h 3 l:tc6 20 b S
.l:lxC4 ! ? 2 1 'ii'x c4 ttJe s 22 'ii'e 2 ttJd3+ 2 3
'iitfl lIe6 with compensation due t o the
strong knight on d3 and the vuln erable
king on fl.
12 a3
Seemingly following Aronian's plan,
but the slight difference in the position
gives Black an unexpected resource.
More accurate was 12 O-O! lilacS 1 3
a3 ! e 4 1 4 ttJd2 lUeS 1 5 b 4 i.. x e3 16 ttJC4
and White regains the piece with ad 1 3 ... .i.xc3 ( 1 3 .. .l:ifeS?! is a bit too
vantage: 16 ... 'iYC7 17 .i.xe3 (or 17 ttJxe3 fancy: after 14 ttJa2 ! ':'acS 15 exd4 exd4
'iVxc3 lS i..b 2) 17 ... 'ii'x C4 lS 'ii'x C4 1IxC4 16 'iYdl ttJxdS 17 ttJb4 ttJxb4 lS axb4
19 ttJbS lIaS Vz-Vz S.5wapnil-E.Ghaem ttJes 19 ttJd2 a6 20 ttJe4 'ii'x b4 21 f4
Maghami, Chennai 2010. Presumably White is clearly better) 14 ttJxc3 e4 1 5
Black offered a draw before his disap b 4 ttJe s ! 16 .i.b2 'iVd6 17 lIadl ( 1 7 ttJbS
pointing tournament got any worse. 'iYxd S ! lS ttJC7 ttJf3+ ! 19 'ii'xf3 exf3 20
12 .i.d41?
... ttJxdS ttJxdS is roughly equal) 17 ... a6 lS
'iit g 2 lilfeS 19 l:ld4 b S 20 lilfdl ttJC4 2 1 a4
'iVd7 and n ow White might try 2 2
l:txc4 ! ? (after 2 2 axbs axbs 2 3 Ital h S
2 4 h 3 %itxal 2 5 .i.xal ttJd6 Black h as
sufficient compen sation due to his
control of the light squares) 2 2 ... bxc4
23 'ii'x C4 'ii'f s 24 d6 hS 25 h4 ttJg4 2 6
'iWe2 ttJ e s 2 7 ttJ d S ttJd3 2 S ttJf4 (or 2 S
ttJe7+ lIxe7 2 9 dxe7 ];t e S 30 i.. d4 ':'xe7
3 1 l:tbl l:tb7 32 bS axbs 33 as ':'c7 34 a6
b4 35 a7 ];tcS) 2 S ... ttJxf4+ 2 9 exf4 l:tabS
This is only a temporary piece sacri 30 l:td4 l:tbdS 3 1 .i.a3 (if 3 1 'iVxa6 e 3 ! 3 2
fice. Otherwise 12 ... e4 or 12 ... .l:tfeS 1 3 'iYe2 exf2 3 3 'ii'xf2 lIe6 o r 3 1 as lIe6)
ttJd2 e 4 would return t o Aronian 31 ... 'iVcS with an unclear g ame.
Sutovsky. 13 exd4
..•
13 exd4 The threats of ... 1IfeS and ... dxc3 en-
1 63
sure that Black regains the piece. 2 1 c;t>g 2 tiJxd4 2 2 tiJd6 l:ted8 2 3 tiJxb7
14 0-0 lUe8 ':'xdS) 19 ... tiJxC3 20 l:txb2 tiJxdl 21
.l:1xdl tiJf3+ 22 �g 2 tiJxd4 2 3 .l:txd4 l:te7
24 d6 l:td7 25 .:tc2 c;t>f8 is a drawn end
g ame.
b2) 17 c;t>g 2 %1ad8 18 .i.e3 (or 18 .i.g s
"ii'x b2 19 .i.xf6 g xf6 20 "ii'c l "ii'x cl 21
litaxcl l:td7 with equality) 18 ... 'ilxb2 19
.i.d4 tiJxdS ! 20 tiJxds 'iib s 21 tiJe3 (not
2 1 tiJC3 ? ! "ii'C 4 2 2 tiJe2 tiJc6, when Black
emerges a pawn up) 21 ... tiJc6 22 l:tbl
'ii'g s 2 3 tiJC2 b S ! 24 .i.e3 ! ? (otherwise 24
c;t>g l tiJxd4 2 5 tiJxd4 'iif6 26 l:tb4 as !
15 'iVC2 regains the piece) 24 ...l:lxe3 2 5 tiJxe3
Alternatively: ':'xdl 26 l:.fxdl a6 27 a4 (or 27 lld6
a) 15 'ifbs dxc3 16 tiJxC3 'ilxbs 17 'iie s and the knight is indirectly cov
tiJxbs tiJxds is equal, since 18 tiJd6 ered by ... 'ii e 4+) 27 ...'iie s (Black has to
l:teb8 is only a temporary inconven pl ay actively here) 28 axbs axbs 29
ience. c;t>g l g 6 30 l:tdS "ii'e 4 3 1 litbxb 5 tiJd4 and
b) 1 5 'ii'd l dxC3 16 tiJxC3 is more in Bl ack is doing fine. The knight and
teresting, when the queen defends the queen combine well together against
d-pawn, but after 16 ... tiJes Black has the weak light squares around the
compensation due to the weakness of white king - one of the points behind
the ds-pawn and the light squares ... 'iib 6 and ... .i.g4 back in the opening !
around the white kin g . For in stance: 1 5 :tac8 16 .l:!d1
•.•
bl) 17 .i.e3 'ilxb2 18 .i.d4 tiJe4! 19 16 ... tiJ e5

llbl (or 19 tiJxe4 'ii x d4! 20 'iix d4 tiJf3+ Another option was 16 ... tiJcS ! ? 17
1 64
Th e Kasp a r o v G a m b it
�f4 liJce4 18 d6 l:te6, followed by offer White some chances of g aining

... d4xC3 and liJxd6 with equality. an edge. My conclusion is that Black
17 'ii'fs should be doing okay after 8 g 3 if he
Or 17 �g 2 liJc4 18 a4 dxc3 19 liJxc3 plays precisely.
hS with compen sation .
17 ... g6 18 'ii'f4 dxe3 19 liJxe3 l:.e4
Game 42
A.Shvedchlkov-Ma.Tseitlin
USSR Army Cha m pion s h i p,
lvov 1977
1 e4 eS
The g am e starts as a symm etrical
English , but it soon tran sposes.
2 liJf3 liJf6 3 d4 exd4 4 liJxd4 eS S liJbS
d S 6 exd s .ies 7 d6?
Black is very active here, but White

can keep the balance.
20 �e3 "xb2 21 l:.ab1 'iie 2 22 libe1
l::txf4 23 ':'xe2 :e4 24 .:tee1 .l:.ee8 2 s liJbs
YI-YI
Game summary
8 g 3 is currently a popul ar line in the
Kasparov Gambit and I think it is al so
the most critical . In respon se 8 ... 'ii'b 6 9
e3 �g4! prevents the white bi shop go It is unlikely th at any strong players
ing to g 2 in an efficient way. After the will try this m ove nowadays, since it is
further 10 �e2 .ixe2 1 1 'ii'x e2 my rec well known that after 7 ...0-0 Black is
ommendation is 11 ... e4, and if 12 a3 fine. The greedy 8 liJC7 is m et by
then Aronian's idea of 12 ... .:.e8 1 3 liJd2 8 ... liJe4! with ideas of ... .ixf2+ or
"C7, temporarily preventing b2-b4, ... .ib4+ and ... 'ii'x d6, when White can 't
seem s to provide Black with sufficient take the rook in a good way. The calm
compen sation . In the game Evdokimov computer already prefers Black.
found interestin g counterpl ay with 7 0-01
•••
11 ... liJbd7 12 a3 .id4 ! ? and made a This move can almost be regarded as
draw, though in this line 12 0-0 would a refutation of 7 d6, since it gives Black a
1 65
definite advantage. 7. . .tDe4?! , o n the the knight and, behind it, the b2-pawn)
other hand, only leads to an unclear 12 ... tDe4 and Black h as a strong attack.
game, and so must be regarded as ob d) 8 'it'ds ! ? �xf2+ 9 c;t>dl tDf6 10
jectively worse, which is why I h ave ap 'ii'x es + (worse is 10 tDC7+? ! c;t>f8 11 'iVd3
pended the dubious m ark. Nevertheless, e4! 12 'ilfa3 tDa6 13 tDxa8 �cs and
in the book Dangerous Weapons, Flank Black h as a dangerous initiative)
Openings, Palliser suggests that it might 10 ... c;t>f8 11 e3 (or 11 �g s tDbd7 12 iff4
be ventured by players who favour "a 'iVb6 with compensation - Ca. H ansen)
quick knock-out", and it is certainly in 1 1...a6 12 tDsC3 �h4 1 3 'iVcs tDbd7 with
teresting and dangerous for White: a very unclear position in which Black
should not be worse.
S J.. e 3
If White doesn 't do something dras
tic there is simply ... tDc6 and ... a7-a6
with a fine g ame for Black. However,
the intended 8 tDC7 runs into 8 ... tDe4!
a) 8 tDC7+?? (your opponent must be

very cooperative to pl ay like this)
8 .....xC] ! 9 'iVa4+ (9 dXc7 �xf2 is m ate ! )
9 ... 'iVc6 0 - 1 was the rapid e n d to
R . Hoque-l H odgson, London 1987
(along with two other g ames in the
database). 9 e3 (other moves just lose; e.g. 9
b) 8 e 3 O-O! 9 'ii'd s (9 tDC7 tran sposes �e3 ? �xe3 10 fxe3 'ii'f 6, coming in on
to 8 tDC7 in the m ain g ame) 9 ... �b4+ 10 f2, or 9 'iVds �xf2+ 10 c;t>dl tDxd6 11
�d2 tDxd2 1 1 tDxd2 a6 1 2 tDC3 (or 12 tDxa8 tDc6 and the white king is a sit
tDC7 l:ta7) 1 2 ... tDc6 with excellent com ting duck) 9 ... �b4+ ! (stronger than
pensation for the pawn, supposing 9 ... �xd6? ! 10 tDxa8 J..b 4+ 11 tDd2 1r'h4
White even m an ages to h an g on to it. 12 g 3 'iVf6 1 3 f3 tDxd2 14 �xd2 l':td8 15
c) 8 �e3 �xe3 9 fxe 3 0-0 10 'ili'ds J.. x b4 ':'xd1+ 16 l:rxdl tDc6 17 �cs
tDf6 11 'fixe s ? ! (but after 11 'iVC4 tDc6 'iVxf3 18 :g l �g4 19 l:td2 tDd4 20 �xd4
Black is clearly better in any case) exd4 2 1 l:td3 and Whit e managed to
11...tDc6 12 'iVc3 (or 12 'iVf4 'iVb6 hitting hang on in P.Harikrishna-E.5utovsky,
1 66
Th e Kasp a r o v G a m b i t
Hoogeveen 2005) 10 .i.d2 (there is cult defence; e.g. 1 1 'iid 2 .i.d7 12 a4

nothing better: 10 lLld2 'ii'x d6 1 1 lLlxa8 .i.xb s 13 axbs lLlcs, threatening ... lLlb3
l:td8 is clearly worse; 10 �e2 ? loses af or .. JUd8xd6.
ter 10 ... 'ii'h 4 11 g 3 lLlxg 3+! 12 hxg 3 11 'ii'd 2?!
�xh l; while 1 0 lLlc3 lLlxc3 1 1 bxc3 No better i s 1 1 'iih 3 ? ! lLlcs 12 'iih 4
�xC3+ 12 .i.d2 .i.xal 1 3 'ii'x al 'ii'x d6 14 'ii'b 6 1 3 a4 a6 14 as 'ii'c 6 1 5 lLlC7 'ii'x d6
lLlxa8 lLlc6 followed by ... .i.e6 and (threatening ... lLld3+ winning the
... l::t x a8 leaves Black a pawn up) queen) 16 'iib 6 .i.c6 ! 17 l:tdl (or 17
10 ... lLlxd2 11 lLlxd2 'ii'x d6 ! 12 lLlb S (af lLlxa8 lLlfe4 18 lLlxe4 lLlxe4 19 l:tdl 'ii'f6)
ter 12 lLlxa8 ? l::t d 8 Black h as a close to 17 ... 'ii'e 7 18 lLlxa8 l::t x a8 leaves White in
decisive advantage) 12 ... 'iig 6 and Black serious trouble. The threats are ...':'c8
is very active without even being a and ... lLlfd7, or ... lLlfe4 and ... 'iih 4+; for
pawn down, while White will have instance, 19 e4 lLlfxe4 20 lLlxe4 lLlxe4
great difficulty developing his king side. 21 g3 lLlf6 22 :g l lLldS 23 'iWb3 'ii'c s
8 .i.xe3 9 fxe3 lLla6!
... and Black h as a big advantage.
H ere 1 1 a4! seems to be the best de
fence, though after 11 ... 'iVas White
would still h ave his work cut out to try
and save the g ame.
11 .i.xbsI 12 lLlxbs lLle4
.•.
Again Bl ack has a clear advantage:

the weaknesses on e3, d6 and espe
cially el give him a very nice game.
9 ...lLle8 and 10 ... a6 was al so good, but
Tseitlin's move puts more pressure on
his opponent. And n ow Black is winning, since
10 lLllC3 .. :ii'as+ or ... 'ifh4+ cannot be prevented.
There is no time to fianchetto with 13 'iid s
10 g3 because of 10 ... .i.d7 and 11 ... .i.c6. 1 3 'ii'c 2 'ii'h 4+ 1 4 g 3 lLlxg 3 1 5 h xg 3
10 ....i.d7 'ii'x h l i s also a disaster for White.
Another option was 10 ... 'ii'b 6 ! ?, 13 'ii'a s+ 14 lLlc3 lLlacs lS l::t c l l:tfd8 16
..•
when White would again face a diffi- g3 l:txd6 17 'ii'c 4
167
3 2 'iti>d3 tbf6 3 3 .i.C4 'iti>d6 34 :f2 :g6 35

liIc2 l1g4 36 .i.b5 l:tb4 37 .i.C4 l:tb1 38
'iti>e3 e4 39 .i.e2 'iti>e5 40 'iti>f2 tbd5 41 :c6
l::t b 2 42 h3 g5 43 a6 tbb4 44 :C7 tbd3+
45 'iti>f1 :b1+ 46 'iti>g2 .l:tb2 47 'iti>f1 'iti>d4
48 .i.xd3 'iti>xd3 49 ':xa7 .l::t b 1+ Vz-Vz
Game summary
Although 7 d6, threatening tbc7, might
look tempting for White, it is well
known that Bl ack has no problems
17 :1c811
••• here. In fact after 7 ... 0-0 Black already
17 .. Jlb6 would h ave been decisive. h as the advantage. In the game,
The threats are 18 .. J:tb4 19 'il'dS .l:td8 Tseitlin, one of the originators of the
winning the queen, or just .. Jbb2 . The g ambit, pl ayed strongly to reach a
text gives White an unlikely reprieve winning position, until he blundered
which he is very h appy to take. on move 17 and let his no doubt grate
18 b41 tbd 3+ 19 exd 3 l:ixc4 20 bxa 5 ful opponent get away with a draw.
l:[xc3 2 1 llxc3 tbxC3 2 2 'iti>d2 tbxa2 2 3
.i.g2
Game 43
F.Santos Garcia
LAlva rez Naves
Va l l a d o l i d 1975
1 d4 tbf6 2 c4 cS 3 tbf3 cxd4 4 tbxd4 e5

5 tbc2
Somehow White h as escaped into a

slightly worse endgame, which he
eventually m an aged to draw. I will n ot
comment on the rest.
23 tbb4 24 l:1b1 tbxd3 25 .i.xb7 f5 26
•••
'iti>C3 'iti>f7 27 e4 fxe4 28 .i.xe4 tbC5 29

:f1+ 'iti>e7 30 .txh7 .l:th6 31 .i.g8 tbe4+
1 68
Th e Ka s p a ro v G a m b i t
I wouldn 't recommend this for (rapid) 2000.

White. Avoiding the gambit with S liJc2 S dS 6 cxd S 'ii'x d s 7 'ii'x d S liJxd S
..•
(or 5 liJb3) simply means that Black can

pl ay ... d7-dS without havin g to sacrifice
a pawn, and then it is White who must
react precisely to get an equal g ame.
True, exchanging queens sidesteps
many of the complications of the Kas
parov Gambit, but I don 't think the re
sulting endgames (or queenless mid
dlegames) are especially drawish, if it is
that White is aiming for.
Other knight moves are:
a) 5 liJb3 is certainly inferior, since 8 e4
the knight is badly placed on b3 after After a g 3 ? ! tLlc6 9 �g2 �e6 10 0-0
S ... dS 6 cxdS 'ii'x ds 7 liJc3 ? ! (7 'ii'x ds 0-0-0 Black is far m ore active, and 1 1
liJxds 8 e4 liJb4 9 liJa3 it.e6 is al so in it.d2 c;i;>ba 12 tLlc3 liJdb4 1 3 tLlxb4 �xb4
Bl ack's favour; White should probably 14 it.g s f6 15 it.xc6 fxg s 16 �g2 l:td2
opt for 7 it.d2 and 8 liJc3, when Bl ack g ave him a very nice endgame in
has only a slight edge) 7 ... 'ii'x dl+ 8 U .Andersson-B.Gelfand, Tilburg 1990.
<;i;>xdl liJc6 9 liJbS ? ! (9 e4 it.e6 10 f3 8 ... tLl b4 9 tLl xb4
.l:.d8+ 11 'itt e l a6 12 it.e3 liJb4 13 'ittf2 9 �bS + liJac6 10 liJxb4 it.xb4+ 1 1
would keep Black's advantage within it.d2 comes t o the sam e thing.
limits) 9 ....l:.b8 10 liJc7+ 'itt d 8 1 1 liJb s 9 ...it. xb4+ 10 � d 2 liJc6
tLlg4 12 'itt e l liJb4 1 3 liJa3 it.e6 14 f3
tLlxa2 ! 15 ':xa2 it.xb3 16 l::t. a l it.b4+ 17
it.d2 it.xd2+ 18 'itt x d2 liJf6 and Black
was a clear pawn up in R.Cifuentes
Parada- I.Morovic Fernandez, Santiago
1989, and went on to win .
b ) 5 tLlf3 tLlc6 6 tLl c 3 it.b4 7 it.d2 0-0
8 e3 (here 8 a3 ! it.xc3 9 it.xc3 tLle4 10
'iVc2 tLlxc3 11 'ii'x c3 dS 12 cxdS 'ii'x ds is
roughly equal) 8 ... e4 9 tLld4 (or 9 tLlg s
l:[e8) 9 ... tLlxd4 10 exd4 d S ! 1 1 'ii'b 3 it.xc3
12 bxc3 it.e6 1 3 c S ? ! 'ii'C 7 14 litbl litab8 In symmetrical position s such as
15 it.e2 b6 was good for Black in this, the pl ayer "with the m ove" h as
J .C. Dominguez-V.Topalov, Villarrobledo whatever chances are going, and here
1 69
that player is Black who h as one more 18 �fl White should make a draw. The
piece out. White, on the other h an d, text can be met by 14 �xc6 �xc6 15
after just ten m oves is now playing for lLlc4 f6 16 .l:tfdl �e7 17 f3 with equal
a draw at m ost. ity, but in stead White starts to make
11 �b5 more trouble for himself.
Instead, 11 �C4 �e6 12 �xe6 fxe6 14 �C4 0-0 15 lLlf3?!
1 3 lLlC3 was seen in A. Pomar Sala 1 5 lLlb3 was better.
m anca-Bellon Lopez, Las Palmas 1977, 15 ... �g4 16 lLlh4?! lladS?!
when 1 3 ...0-0 14 ':Cl l:.ac8 would give 16 ... lLld4! would have caused the
Black some pressure in the endgame. white knight some problems in return
The doubled e-pawns provide very ing to the g ame. For in stance, 17 �d5
good central control, the knight (or l::t ab8 18 h3 �e6 19 lLlf3 (or 19 �xe6
sometimes the bishop) h as a nice fxe 6 ! ) 19 ... lLlxf3+ 20 gxf3 �xd5 21 exd5
square on d4, while the rooks can g o to 1:tfd8 22 lUdl J:[bc8 23 Jitacl �8 and
c- and d-files with activity on the Black h as good winning chances due to
queenside. the weak white pawn s, especially the d
11 ... �d7 12 0-0 pawn .
Another of Ulf Andersson's g am e 17 f3?!
went 1 2 �xc6 �xd2+ 1 3 �xd2 �xc6 1 4 17 h3 would have freed the knight
lLl c 3 0-0-0+ 1 5 � e 2 l:.d4 16 f 3 .l:lhd8 17 again with equality, whereas the text
lithdl and this time h e m an aged to takes away its main retreat square.
hold this slightly worse position in 17 ... �e6!
U .Andersson-F.Langheinrich, German
League 200 3 .
12 . . .�xd2 13 lLlxd2 a6
is b3?!
18 J:[acl �xc4 19 ':xC4 ':d2 20 l:tf2
lUd8 2 1 ':c2 .l:tdl+ 2 2 .l:[fl g 6 is still
After 1 3 ... lLld4 14 �xd7+ �xd7 15 slightly better for Black due to his more
J::tfdl �e6 16 lLlb3 ! lLlxb3 17 axb3 11hd8 active pieces.
1 70
Th e K a s p a r o v G a m b i t
18 ...J:td2 ct;xb4 h S 46 a4 h4 47 a s h 3 48 a6 h 2 49
More accurate was 18 ... ttJd4! fol a 7 h1'i1 0-1
lowed by 19 ... 95 with a clear advan
tage. The main idea is 20 .ltxe6 fxe 6 ! , Game summary
threatening ... g 7-g 5, when White The alternatives to 5 ttJb5 are all sig
would have to give up the f-pawn in nificantly inferior. Indeed, White h as to
order to save his knight. work to prove equality in these lines,
19 l:tfd1?! and the exch ang e of queen s surren ders
19 f4 would offer more drawin g the initiative completely to Black, even
ch ances, when at least the knight gets if White might expect to m ake a draw.
back in to the game. In the g ame, Santos played a succes
19...l:tfd8 20 iLd S? l:txd1+ 21 ':'xd1 sion of inferior moves, which inevitably
ttJd4?! led to a lost endi ng . Although Black
Here 2 1 ... g 5 2 2 ttJf5 iLxf5 ! 2 3 exf5 conducted the technical phase uncon
�g 7 threatening ... ttJb4 or ... ttJe7 would vincingly, I think the g am e is quite in
have won much more easily. structive up until the fin al diagram.
22 g3 iLxd S 23 exd s :xd S
Summary
The Kasparov G ambit is a sound pawn
sacrifice for which Black gets active
counterplay. In response White h as
three m ain m oves:
The original m ain line with 8 e3 is
quite logical in that closes off the c 5 -
bishop, and this move continues to b e
popul ar, even though Kasparov himself
demon strated back in 19 8 1 (see G am e
39) that 8 ... e4, foll owed b y ... We7 and
I won 't comment on the rest. Black .. .lId8, gives Black a perfectly good
is a pawn up and eventually won in a g ame. Indeed, White h as yet to show
rather unconvincing way. any substantial improvements here.
24 <iW2 g6 2S ttJg2 l:td6 26 ttJe3 bS 27 Secondly, 8 h3 plans to m eet ... e 5 -e4
ttJc2 �c6 28 ttJb4 l::t c 3 29 l:td3 �xd3 30 with .ltg 2 and g 2-g4-g 5, undermining
ttJxd3 ttJc6 31 <iit e 3 fS 3 2 ttJcS a s 3 3 ttJd3 the e-pawn . H ere Black should reply
ct;f7 34 g4 ttJb4 3 S ttJxes+ <iit e 6 36 ttJd3 with 8 ... .ltf5 ! , n ot fearin g g 2 -g4, which
ttJxd 3 3 7 <iitx d3 fxg4 38 fxg4 <iit e s 39 often just weakens the kin g side. Later
<iite 3 b4 40 h3?? gs 41 <iit d 3 <iW4 42 <iit C 4 Black m ay play ... e 5 -e4 in any case, as in
�g3 43 <iit b s <iit x h3 44 <iit x a s <iit xg4 4S G ame 40, with ideas of ... ttJbd7-e5/c5 or
1 71
possibly ... tba6-b4, again with good 0-0, so I would prefer 11 ... e4! as given
play for the pawn . in the notes. H ere Bl ack might improve
The third line, involving 8 g 3 , is cur on Sutovsky's play by 12 ... :te8 13 tbd2
rently reg arded as the most critical, 'ilC7, deterrin g a quick b2-b4, and
with Aroni an -Sutovsky being a key en reaching an unclear g ame.
counter. The primary idea is to rein Attempts by White to avoid the
force the ds-pawn by �g 2 and, ulti m ain line or the g ambit altogether are
m ately, e2-e4. Obviously Black cannot m arkedly inferior. Game 42's 7 d6? is
allow thi s plan to succeed, and 8 ... 'ifb6 simply a mistake, since 7 ...0-0! already
9 e 3 �g4! is an effective hindrance, so gives Black favourable chances. The
that after 10 �e2 �xe2 11 'iWxe2 White various knight retreats on move 5 (seen
i s l eft with weaknesses on the light in Game 43) allow Black to free his po
squares. N evertheless, he still h as his sition without sacrificing, when it is
extra pawn on dS, so Black must play White who is put on the defensive.
precisely to obtain sufficient compen In conclusion, the Kasparov Gambit
sation . appears to be fully viable. Some new
G ame 41 features an interesting ideas and novelties in this chapter, es
idea with 1 1 ... tbbd7 12 a3 �d4 ! ? with pecially in the line with 8 g3, show that
which Evdokim ov achieved equality. Black can prove definite compen sation
H owever, White does better with 12 for the pawn.
1 72
C h a pte r E ight
The Topa lov Variation
1 d4 ttJf6 2 ttJf3 cS 3 d s bs 4 i.. g s ttJe4 first three games below.

The variation bears close resem
blance to the Trompowsky 1 d4 tDf6 2
i.g 5 ttJe4 with the further 3 ... c5 4 d5.
From Black's point of view the m ove
... 'ilVb6 is no lon g er an effective equal
izer since the b-pawn blocks the
queen 's path to b2. On th e oth er h an d,
White is committed to d4-d5, when he
might h ave preferred to pl ay d4xc5 or
f2-f3, while his own developed knight
prevents the l atter m ove altogether.
This position is the startin g point Other fourth m oves are examined
for th e Topalov Variation, which will be in Game 47; while White's third m ove
our main focus in this final ch apter. As option s, 3 e3 and 3 C3, simply defend
you will quickly notice, the m ain differ ing the d-pawn, are covered in G ames
ence from the variation s in the rest of 48 and 49 respectively.
the book is th at the white c-pawn is
still on c2, and yet Bl ack is attempting
to pl ay in Benko-style reg ardless. Game 44
Of course, if White were to oblige O.Romanishin-R. Wojtaszek
with 4 c4 we would reach the Benko E u ropea n C h a m pion s h i p,
after all, specifically the 4 tDf3 line dis D resd e n 2007
cussed in Chapter Six (Games 30 and
3 1). Here we will look at different ideas 1 d4 ttJf6 2 tDf3 cS
for White, of which 4 i.g 5 ttJe4 is the An ambitious continuation, played
most important and the subject of the by Topalov and Sergey Kasparov among
1 73
others. N ow White h as t o decide what other options, such as 4 ... ii.b7, 4 ... d6 or
to do with the nice central d4-pawn . 4 .. :iVb6, but I think these offer White
3 dS more chance of an advantage. With
One of the two m ain tests of 2 . . .c S . 4 ... tDe4 Bl ack aims to gain the bishop
The other, 3 c4, tran sposes t o the pre pair (by ... tDxg S) without making any
vious ch apter on the (Garry) Kasparov concession s to his king side. On the
G ambit. downside, the advanced knight can
3 ... bS become a target, since it cannot be
supported by ... d7-dS, so Black has to
play with some precision in order to
equalize. But White, too, can easily end
up with the worse position, so the
g ame is nicely in the bal ance.
S ii. h4
With this m ove Black g ains space on

the queen side and aim s to get either a
Benko (after 4 c4) or an improved ver
sion of the Schmid Benoni, 1 d4 cS 2 dS,
where a quick tDb1-C3 usually prevents
...b7-b S . The white ds-pawn is n ow a
target. Then again, so is the b s-pawn, The first of four possibilities for
which can be attacked with m oves like White. The others - 5 .1i.f4, 5 h4 and 5
e2-e3, 'ii'd 3 or a2-a4. 'iVd3 - are examined in the next two
4 .1i. g S g ames. Of these, the two bishop re
The principled and critical move in treats are the most testing.
the position . White threatens to com S ...ii. b7
promi se the black pawn structure. In With pawn already on bS, this is the
stead, 4 c4 transposes to the Benko n atural development, putting pressure
G ambit proper (see G ames 30 and 3 1), on the vuln erable ds -pawn .
while 4 a4 and other m oves are covered 6 e3
in G am e 47. A modest-lookin g move compared
4 ...tDe4 with the alternatives, but it is quite
Championed by Topalov. There are reason able. The other two main op-
1 74
Th e Top a / o v Va ria t i o n
tions are: ( 1 4 tiJd4 0-0 1 5 h4 tiJ c s 16 hxg s hxg s

a) 6 'ii'd 3 and now 6 .. .fs is the typi 17 f3 e S 18 dxe6 tiJxe6 is m ore unclear)
cal reaction and the most precise in my 14 ... tiJcS 15 axbs tiJe4 16 'ii'e 2 'iWxb s 17
opinion . Black simply defends the .tes .txe s 18 tiJxes .txdS and Black is
knight on e4, since if White takes on b S doing fine.
then the ds-pawn is hanging . Black a2) 7 g4! ? (an uncompromising at
would be quite content to exchange his tempt to undermine the knight; it h as
b-pawn for the more valuable one in been a trend for some years to play g 2 -
the centre. g4 in m any different opening varia
tion s)
al) 7 tiJbd2 (a position ally sound

way of challenging the knight on e4) 7 ... c4 8 'ili'd4 'iWas + ! 9 c3 (or 9 tiJbd2
7 ... c4 (a them atic way of eliminating e 6 ! 10 dxe6 dxe6 11 g xfs tiJxd2 12
the threat to bS) 8 'ii'd4 tiJa6 (threaten fixd2 fixd2+ 1 3 'iti>xd2 exfs with equal
ing ... tiJb4) 9 c3 (9 tiJxe4 fxe4 10 'ili'xe4 ity) 9 ... b4! (Bl ack equalizes by precise
tiJb4 is unclear, as 11 l::t d l is m et by play) 10 cxb4 (10 g xfS ? ! .txdS 11 cxb4
11 ... tiJxdS ! 12 .li:txdS 'ii'a s+! and ... 'ii'x a2) 'ii'x b4+ 12 tiJc3 e6 13 fxe6 dxe6 14 0-0-0
9 ...tiJxd2 10 'ii'x d2 (R.Hiibner-V.Topalov, tiJxC3 15 'ii'x C3 tiJc6 is slightly better for
Istanbul Olympiad 2000, saw 10 tiJxd2 Black since the white king is quite vul
'i!i'b6 11 'ii'e s, when 11 ... tiJC7 12 'ii'xfs n erable) 10 ... 'ii'x ds 1 1 'ii'x ds .txdS 1 2
tiJxds would keep the tension in an :g l (not 12 g xfS ? g S ! 1 3 fxg 6 .t g 7 , fol-
unclear position) 10 ... 'ii'b 6 11 l::. d l (in lowed by ... .txb2, and Black is clearly
stead 11 0-0-0 g 6 12 'iVd4? ! 1:[g 8 gave better) 12 ... e6 with an unclear game.
Black an active position in V.Kramnik a3) 7 e3 (this is too quiet to give
V.Topalov, Monte Carlo rapid 2000, a Bl ack any problems) 7 ... c4 8 'ii'd4 'ii'a s+
game which Topalov won) 1 1 ... h 6 12 e3 9 tiJbd2 ? ! (this puts the white queen at
g s 13 .tg 3 .tg7 and Black has equal risk, but after 9 c3 e6 10 dxe6 dxe6 1 1
ized comfortably. For example, 14 a4? ! 'ii'd l tiJd7 Black h as a n active position)
1 75
9 . . .e6 (now Black is slightly better) 10 for White to play as well.

C 3 ? ? (giving up a pawn and the centre; b} 6 a4 is one of the critical tests of
White h ad to try 10 dxe6 dxe6 11 b4! , Black's set up. This was Burmakin's im
although 1 1 ... 'iib 6 12 lbxe4 fxe4 1 3 provement after his loss in the above
lb d 2 lb c 6 14 'it'c3 as is still good for g ame.
Black) 10 ... .ixdS 1 1 lbxe4 fxe4 12 lbg s
lbc6 1 3 'it'd2 .ie7
bl} 6 ... a6 ? ! 7 'it'd3 (the insertion of 6

a4 a6 to line 'a' improves White's
and Bl ack is winning: the knight on chances substantially) 7 .. .fS 8 lbbd2
g s and bishop on h4 are awkwardly 'it'as+ (unfortunately ... lba6 is impossi
placed, and Black plan s ... h 7-h 6, ... g 7-g S ble) 9 c3 e6 (9 ... c4 10 'ii'd4 doesn 't help
or ...b S -b4. here) 10 lbxe4 fxe4 11 'it'xe4 .ixdS 12
V.Burm akin -A.5h ariyazdanov, Biel 'it'f4 lbc6 1 3 e4 .ib3 14 .ie2 bxa4 15 0-0
2000, continued 14 f3 h 6 ? ! (14 ... exf3 1 5 and White was clearly better in
gxf3 0-0, intending 16 e4 lbes ! , looks a V.Burm akin - B.Avrukh, Ubeda 2001.
simpler way to win) 1 5 fxe4 .ixg s 16 b2} 6 ... b4? ! 7 'ili'd3 fS 8 lbbd2 is simi
.ixg s .ixe4 17 .if4 e S 18 .ig3 dS 19 l arly good for White. Notice that Black
.ie2 0-0 20 .if3 ? ! (20 ltf1 ! might offer is un able to pl ay ... cS-c4 and ... e7-e6
more chances to h old) 20 ... .ixf3 2 1 due to his b-pawn advance, while
g xf3 d 4 2 2 0-0 l:tad8 2 3 :adl 'it'b6 2 4 8 ... .ia6 9 'ili'e3 lbxd2 10 lbxd2 d6 11 g4!
l:.del dxc3 2 5 'it'xc3 l:td3 2 6 'it'cl 'it'cs 27 fxg4 12 h3 g3 1 3 .ixg 3 .ib7 14 lbe4 led
l:te2 'it'ds 2 8 l:tef2 'iVcs 2 9 l:te2 lbb4 30 to a crushing win for White in
a3 lbdS 31 l:tdl 'ii'b 6 3 2 .ixes 'ii'g 6+ 33 F.5chmitt-L.Guidarelli, French Team
.tg 3 l:txf3 34 l:tdel 'it'e4 35 .ib8 I1fxe3 Ch ampionship 2007.
3 6 l:txe3 lbxe3 0-1. Quite a convincing b3} 6 .. :iVas+?! has been the usual
win by Shariyazdanov again st a solid choice for Black. Then the obvious 7
grandmaster, which ag ain illustrates c 3 ? ! is met by 7 ... b4! 8 1rd3 bXc3 9 bxc3
that the position is sharp and difficult (not 9 lbxC 3 ? ! lbxC3 10 bxc3 .txdS! win -
1 76
winning a pawn) 9 .. .f5 10 g4 (or 10 e3 numerous times since G . Kasparov

.ta6 11 'iVc2 �xfl 12 l:txfl c4 13 tLJbd2 A.Miles, Basel (3rd m atch g ame) 1986.
'ilVxd5 14 tLJxe4- 'iVxe4 15 'iVxe4 fxe4 16 Although Kasparov won that g ame, its
tLJd2 tLJa6 17 tLJxe4 d5 18 tLJd2 e6 and theoretical status is less clear. I would
Black has the better chances in the prefer the simple 9 e3, when White is
endgame) 10 ... c4! 11 'ii'x c4 was pl ayed again at least slightly better.
in J.Maherramzade-S.Citak, European b4) 6 .. :�b 6 ! (according to my data
Championship, Batumi 2002, when base this is a n ovelty, and I think it's
11 ... g5 12 �xg 5 �g 7 would h ave given quite an interestin g one given that
Black a nice initiative. other moves fail to equalize):
However, 7 tLJbd2 ! is a lot stronger,
tran sposing to the old line 5 .. :�a5+?! 6
tLJbd2 �b7 7 a4! , long known to be
good for White. Then 7 ... �xd5 (or
7 ... tLJxd2 8 tLJxd2 ! �xd5 9 axb 5) 8 axb5
b 3 1) 7 tLJbd2 (following line 'b3')

7 ...i.xd5 8 axb5 'iVb7 ! (the difference) 9
tLJxe4 �xe4 10 e 3 a6 1 1 .:ta5 (or 1 1 c3
g 6 12 'iVa4 �g7) 1 1 .. .f6 (this might look
a bit artificial , but it shuts out the
8 ... 'ifb4? 9 c4 �b7 10 'ilc2 a6 1 1 %:ta4 bishop on h4 and prepares to play ... e7-
won quickly in J.Nogueiras Santiago e 5 ) 12 �d3 �xd3 1 3 'iVxd3 e 5 14 'iWb3
R.Vag anian, Mexico 1980. Of course the (not 14 O-O?? e4, winning a piece)
queen has other squares than b4, but 14 ... �d6 1 5 0-0 i.C7 16 l:ta4 axb5 17
White gains the advantage in any case. l1xa8 'it'xa8 18 'iVxb5 'iVc6 and Black is
For in stance, 8 ... 'iVb6 9 e3 tLJxd2 10 fine.
'iVxd2 'iVb7 11 c4 �e4 12 �d3 d6 1 3 0-0 b 32) 7 'iVd3 bxa4! 8 tLJbd2 (not 8
tLJd7 14 �xe4 'iVxe4 was played in 'iVxe4?? 'iVxb2) 8 .. .fS 9 tLJC4 'iVh 6 10 g4! ?
E.Cosma- B.Chatalbashev, Metz 2002, tLJa6 1 1 c3 (11 g xfS tLJb4 12 'ilxe4??
and now 15 tLJg 5 'iVb7 16 f4 would have loses to 1 2 ... �xdS) 1 1 .. .fxg4 12 'iVxe4
been good for White. Finally, 8 ... 'iVC7 9 gxf3 with an unclear position.
l:ta4 ! ? 'ii'b 7 10 c4 tLJxd2 has been seen b33) 7 tLJfd2 'iVh 6 ! 8 tLJxe4 'ilxh4 9
1 77
Play th e B e n k o G a m b i t
'iWd3 lDa6 1 0 'i!Vxbs �xds l 1 lDec3 lDb4 clear game with ch ances for both sides.
12 lDxds lDxds 13 lDd2 e6 is also okay 7 ... Wa s?!
for Black, since 14 'iWb7 is safely an In my opinion 7 ... 'iWb6 ! is more pre
swered by 14 ... 'i!Vd8 ! . cise, when Bl ack should equalize:
b 34) 7 axbs a 6 8 lDc3 (otherwise 8
bxa6? ! Wxb2 is g ood for Black; 8 Wd3
Wxbs 9 lDa3 'iWb4+! 10 c3 'iWxb2 l1 l:tbl
'i!Vxa3 1 2 'ii'x e4 lIa7 is unclear; while 8
lDfd2 .l1.xds 9 lDxe4 �xe4 10 lDc3 �b7
11 e4 axb s 12 l:1xa8 �xa8 13 �xbs
.l1.xe4 14 0-0 �c6 1 5 Wd3 gives White
just enough play for the pawn) 8 ... lDxc3
9 bXc3 g6 10 c4 (or 10 bxa6 �g 7 ! )
1 0 ... �g 7 1 1 :bl axb s 12 :xb s WC7
with compensation.
6 ... g6 7 c3 a) 8 'ii'd 3 fs 9 lDbd2 c4 10 "ii d4 'ii'xd4
11 lDxd4 lDxd2 12 'itt x d2 �xds 13 lDxb s
lDa6 14 b3 �g 7 1 5 bxc4 .l1.c6 was fine
for Black in O.Romanishin-D.5ermek,
Solin (rapid) 2002.
b) 8 a4 b4 9 as (or 9 C4 .l1.g7 with an
unclear game) 9 ... Wd6 10 c4 e6 11 a6
�xa6 12 lDbd2 lDxd2 1 3 lDxd2 �g 7 14
lDe4 'ii'e s 1 5 �d3 was E.Gausel
P.5vedenborg, Norwegian Team Cham
pionship 2000, and now Bl ack should
have played l s ... .l1.xc4! 16 �g 3 'iWxb2
Blockin g the long diagonal on which 17 lIbl i.c3+! 18 'ittfl �xd3+ 19 Wxd3
Bl ack's dark-squared bishop i s intend 'ilfa2 20 �xb8 exds with four pawn s for
ing to take up residence. Timman won the piece.
a game against Topalov with this c) 8 lDbd2 (a thematic move in these
move, though I don 't think Bl ack is positions) 8 ... �xds 9 lDxe4 �xe4 10 a4
really worse. b4 1 1 lDg s bXc 3 ! 12 bXc3 (not 12
Altern atively, 7 a4 �g7 (the most lDxe4? ! cxb2 1 3 l:tbl "iVb4+ and Black
n atural m ove; 7 ... b4 would cede the regains a piece on the fourth rank with
initiative) 8 axbs �xb2 9 lla4 lDc3 10 a clear advantage) 12 ... �c6 is unclear
lDxc3 �xc3+ l 1 lDd2 a6 12 l:ta3 �b4 13 once more.
d6 f6 14 dxe7 'ii'x e7 reaches a very un- S 'ii'd 3
1 78
In this position Timman played 8 bxc3 I:tf7 16 'iVd2 would h ave given
tLlbd2 ! ? tLlxd2 (after 8 ... �xd5 ? ! 9 tLlb3 White a slight but safe advantage. In
�xb3 10 axb 3 - 'iVb6 11 1i'd3 tLld6 12 stead, the players n ow trade mistakes.
�e2 �g 7 1 3 0-0 White has more than
enough compensation for the pawn,
with the bishop pair and lead in devel
opment) 9 'it'xd2 (9 tLlxd2 ! ? is still pos
sible; e.g. 9 ... �xd5 10 tLlb3 �xb3 1 1
axb3 'iWb 6 12 1i'd3 a 6 1 3 �e2 ttJc6 14
0-0 �g 7 15 l:tfdl l:ta7 16 f4 0-0 17 �f2
d6 18 e4 with compensation) 9 ... �g 7
10 e4 d6 11 �d3 tLld7 12 0-0 and White
was slightly better in J .Timman
V.Topalov, Wijk aan Zee 2001, due to
his useful space advantage. 13 e6??
•..
8 ...fS 9 tLlbd2 �g71 After thi s Rom anishin coul d h ave won
with 14 dxe 6 ! d5 (or 14 ... dxe6 15 �xe6+
..ti>h 8 16 0-0) 1 5 e7 J:.e8 16 1i'e4! dxc4 (or
16 ... ..ti>h 8 17 �xd5) 17 'ii'e 6+ ..ti>h 8 18
ttJe 5 etc. Conversely, Wojtaszek should
h ave played 1 3 ...bxC 3 ! 14 b 3 e6, when
15 dxe6 d5 16 e7 1:[e8 17 'iVe4? ! is
thwarted by 17 ... c2+ 18 ..ti>e2 1i'a6 ! 19
�xa6 �xa6+ 20 'iVd3 �xd3+ 21 ..ti>xd3
�xal 22 ]::t x al ttJc6 and Black comes
out on top.
14 .l:tdl? bxC3 lS 0-0 exd s
10 tLlxe4
10 1i'xb 5 'it'xb5 11 �xb5 tLlxd2 12
tLlxd2 �xd5 is nothing to worry about.
The endgame is bal an ced.
10...fxe4 ll 1i'xe4 0-0 12 'iVC2
Not 12 �xe7? J:te8 13 ttJd2 �f8 and
Bl ack win s material , while 12 tLld2 e6
13 �e7 �xd5 14 1i'h4 .uf7 1 5 �xc5 tLla6
16 �d6 b4! is unclear.
12 ... b4 13 �C4?!
Here 13 ':cl �xd5 14 �xe7 bXc3 1 5
1 79
16 .i.xd S+?? White's own sore spots at dS and on

A further error. 16 Wh3 ! dXc4 17 the queen side. The trade of the bS
'iWxb7 'ili'a6 18 'ifdS+ ii'e6 ! 19 ii'xe6+ pawn and ds -pawn s is usually in
(after 19 'ili'xa8? ! cxb2 20 'ifxa7 c3 21 Black's favour. Sometimes the b-pawn
ii'a4 dS the black pawn s are very dan can be advanced further (as in the
gerous) 19 ... dxe6 20 bxc3 would h ave g ame) to soften up the white queen
kept a slight advantage for White. side.
16 ... .i.xd S 17 :xd S 'ii'x a2 The conclusion is th at Black is okay
Hitting the rook on dS as well as the after 5 .i.h4 as long he plays with preci
pawn on b2. White can h ardly let the sion . For the three main lines above I
rook go, but after 18 ... cxb2 the passed am recommending 6 'ii'd 3 fS ! , 6 a4
queen side pawn s will be decisive. 'ifb6 ! and 6 e3 g 6 7 c3 ii'b6 ! as the best
18 'ii'e 4 cxb2 19 ttJd2 a s l ways for Bl ack.
Game 45
E.Relange-Ch.Bauer
F re n c h C h a m pion s h i p.
Ma rse i l les 2001
1 d4 ttJf6 2 ttJf3 cS 3 d s bs 4 .i.gs ttJe4 s

.i.f4 .i.b7
20 .i. g 3 .t:t a 6 2 1 'uxcS .l:tc6 2 2 %:txc6 ttJxc6

23 ii'd 3 ii'e6 24 ttJC4 Ilc8 2S ttJd6 ttJb4
0-1
Game summary
This was a very sh arp g am e with both
grandm asters m aking serious mis
takes, which only reflects that the posi
tion s are difficult for both sides. As we
saw, White's play in this line generally Again this n atural developing move
focuses on the vulnerable pawn on b S seem s best. So we have the same posi
and knight o n e4. In turn Black must tion as in the previous game, except
pl ay precisely to minimize these tacti th at the white bishop is on a different
cal weaknesses, while targetin g square. White al so has the same op-
1 80
Th e Top a l o v Va ria t i o n
tion s as before, though there are n atu 0-0-0 (11 c 3 lDxdS o r 1 1 l:tdl lDxd S ! is
rally some differences in the play, the fin e for Black) 11...e6 ! ? (otherwise
main one being that the bl ack knight is 11...lDxa2+ ! 12 �bl 'ii a s 1 3 lDes lDb4
free to return to f6, attackin g dS, since 14 'iifs 0-0-0 g ets at least a draw; e.g.
the bishop is no longer able to take it. 1 5 lDf7 'ii a 2+ 16 �Cl C3! 17 bxc3 e6 18
6 a4 dxe6 'ii a 3+ 19 �bl 'ii a 2+ etc) 12 'iie s
The critical move here. �xdS 1 3 e 4 lD c 6 14 'ifh S+ g 6 1 5 'ifh 3 ? !
a) 6 e3 is now well met by 6 .. :ifas+ 7 ( 1 5 'iVg s would b e unclear) l s ... �xe4
c3 lDf6 ! , when 8 d6 (8 a4? ! lDxdS 9 16 lDg S ? ! 'iif6 ! and Black was clearly
�xb8 ':xb8 10 �xb S 'iiC 7 1 1 0-0 e6 12 better in Butnorius-R.Fel g aer, Gibraltar
lDbd2 �e7 13 'iiC 2 0-0 was good for 2008.
Black in S.Yuferov-D.5ermek, Bled 1994) b2) 9 c3 lDac s ? ! was played in
8 .'iYb6 9 dxe7 �xe7 10 �e2 0-0 11
.. G .Mittel m an - B.Avrukh, Israeli Team
'iVd3 a6 12 lDbd2 dS 1 3 0-0 lDbd7 gave Ch ampionship, and n ow 10 g4 fxg4 1 1
Bl ack a comfortable g ame in E.5uppa lDxe4 lDxe4 12 'iVxe4 gxf3 1 3 'ii'xf3
V. Lazarev, Porto San Giorgio 2000. would be slightly better for White.
b) 6 'iid 3 is more frequently played, Avrukh concluded back that 8 ... lDa6
but after 6 .. .fs Bl ack is doing fine. was dubious, but I don't think so. It was
Again Black would be more th an con his n ext m ove which was n ot the best.
tent to see 7 'iix bs �xdS, while 7 lDC3 In stead, Bl ack should play 9 ... lDc7 !
c4 8 'iid4 'iia s (8 ... lDa6 ! ? 9 lDxe4 fxe4 10 lDxe4 fxe4 11 �xC7 (11 'iix e4 lDxdS
10 'iix e4 lDb4 is al so promising) 9 lDd2 would be unclear) 11...'ivxc7 12 lDg s, as
lDxd2 10 �xd2 b4 11 lDdl 'iix ds was in M.Hebden-l Hodgson, British Ch am
fine for Black in S.Boroday-A.Zubarev, pionship, Millfield 2000, and n ow
Poltava 2009. So play usually continues 1 2 ... 'iYb6 would lead to an unclear
7 lDbd2 c4 8 'iVd4 lDa6 and then : g ame.
bl) 9 lDxe4?! fxe4 10 'iix e4 lDb4 1 1 For in stance, 1 3 'ii x e4 g 6 (Bl ack
181
plans ... �g7, ... 0-0 and ... e7-e6) 14 11dl the m ove 10 .. .fS would lead to bal
(or 14 h4 �g 7 1 5 e 3 0-0 16 h S h 6 17 anced chances.
liJf3 gs with compen sation) 14 ... �g 7
1 5 e 3 0-0 16 'iix e7 (or 16 �e2 e6 17 0-0
l::t a e8 18 'ii'h 4 h6) 16 ... l:tae8 17 'iVxd7
�c8 ! 18 'ii'c 6 l:Ixe 3 + ! 19 fxe 3 'fi'xe3+ 20
�e2 'ii'f2+ 21 'ii;J d 2 'ii'f4+ 22 'ii;J e l 'ii'f2+
with perpetual check.
6 ... b4!
For in stance, 11 lta2 (not 11 �es ? !

�xe s 12 liJxe s 'ii'C 7 and Black has a
slight advantage due to the weak d
pawn ; while after 11 liJbd2 liJxd2 12
liJxd2 �xb2 13 l:tbl �g 7 14 0-0 d6
White l acks sufficient compensation
for the pawn : 15 liJf3 liJd7 16 liJg s 0-0
Tsesarsky recommends the imme 17 liJe6 can be met by 17 ... 'ii'a s 18 liJxf8
diate 6 ... liJf6 in his MegaBase ann ota ':'xf8 and ... liJb6, when Black has a very
tions, intending 7 axbs liJxdS . H ow dynamic position) 11...0-0 12 h4 (or 12
ever, in that case White can retreat the as �a6) 12 ... liJd6 1 3 �xd6 exd6 14 as
bish op, chase the knight from dS and liJa6 1 5 'ii'd 3 'ii'c 8 (or l S ... b 3 ! ? 16 'ii'xb3
g ain a slight advantage. l:tb8 17 c3 liJC7) 16 liJbd2 l::tb 8 with an
7 'ii'd 3 unclear position . Black is planning
Other m oves: ... liJC7 and and possibly ... 'ii'e 8-f7 trying
a) 7 c4 e6 8 'ii'd 3 liJf6 9 e4 tran s to win the ds-pawn, or if 17 �b3 then
poses to the g ame. 17 ... c4! 18 liJxC4 liJcs with compensa
a) 7 e 3 g6! (7 ... e6?! 8 dxe6 fxe6 9 tion .
liJbd2 was slightly better for White in 7 ... liJf6 8 e4
A. Rotstein-1M.Degraeve, German In this position White has more
League 2001, due to his slight lead in space in the centre, but Black can neu
development and the weaknesses on tralize th at by a plan of ... e7-e6, ... e4xds
the light squares) 8 �es f6 9 iLf4 �g 7 and ... �e7, followed possibly by ... liJh s
10 �c4 was played in T.Minogina and .. .f7-fS.
E . Fatalibekova, Moscow 1981, and n ow 8 ... e6
1 82
1 2 . . .0-0 ( 1 2 . . .lLlbd7 ! ? might b e m ore

accurate, when White must al so con
sider the possibility of the black king
going long) 1 3 'iVc2 iLc8 (13 ... b 3 ! ) 14
iLd3 b3 ! 1 5 lLlxb3 and n ow, rather than
ls ... iLd7? ? 16 lLlxc s ! and White won in
D.Komarov-1M. Degraeve, French Team
Championship 2002, simply ls ... lLla6
followed by ... l::tb 8 and ... iLd7 offers
Black good compensation for the pawn .
Of course White could prevent the
9 (4 pawn sacrifice with 1 3 b 3 , but then
9 dxe6 is tame and allows Black to Black h as more tim e to organize his
equalize rather easily: 9 .. .fxe6 10 lLlbd2 forces with, say, ... lLlh s /lLlg4, .. .f7-fS,
iLe7 11 eS lLlh S ! (the knight doesn't ... g 7-g6, ... lLld7 and ... lLles . For example,
block the bishop here and will soon re 1 3 ... lLlbd7 (or immediately 1 3 ... lLlh S 14
turn to the centre; 11 ... lLldS would be iLe3 fS 15 g3 iLf6) 14 'iVc2 lLlh s 1 5 iLe3
less precise due to 12 iLg 3 0-0 13 'iVe4! ) fS 16 g3 g 6 17 iLd3 iLf6 with an un
12 iL e 3 0-0, followed by . . .lLlf4 is unclear. clear g ame; or 14 h 3 lle8 1 5 'ii'C 2 iLf8
13 'ii'C4 fails to prevent this due to 16 iLd3 g 6, and if 17 g4? ! then
13 ... lLlf4 14 iLxf4 iLdS !, while 14 iLxcS 17 ... lLlxd S ! 18 cxdS 'iVf6 19 iLg 3 "ii'a l+
.i.dS ! 15 iLxe7 (or 15 'iVxb4 lLlc6) 20 "i!Vbl (20 lLlbl ? ! iLxdS) 20 ... 'iVC3 + 2 1
ls ...'iVxe7 16 'iVd4 lLlc6 17 'iVe3 'iVf7 gives 'iVc2 ( 2 1 iLc2 ? ! iLxdS) 2 1 . . .'iWal+ with
Black very good play for the pawn. perpetual check.
9 exd s 10 exd s d6 11 iL e2
... 11 iL e7 12 0-0 0-0 13 lLlbd2 Yz-Yz
•••
More challenging is 1 1 lLlbd2 iLe7

12 0-0-0 ! ?
H ere the g am e ended in a draw.

13 ... lLlh S 14 iLe3 fS 15 lLlb3 g6 16 iLh 6
1 83
l:[eS 17 as ..tf6 could b e a possible con but this is not really all that threaten
tinuation with a roughly equal posi ing, as long as Black develops rapidly
tion. and fights for the initiative.
The final option is S 'iVd3, hitting
Game summary the knight on e4 and pawn on bS. The
S .i.f4 is one of the critical responses to drawback to this move is that the ex
Black's ambitious play with 4 ... liJe4. The ch an g e on gs leaves the knight is
one drawback of this bishop move is rather misplaced. White can try to jus
that it allows the black knight to return tify it with ideas of 'iVf3 or liJxh 7, or
to f6 unm olested. After ... liJf6 and ... e7- sometimes 'iVC3 (after ... g 7-g6), but
e6, followed by ... e6xds, ... d7-d6, ... ..te7 Bl ack can equalize without too many
and ... 0-0, with the further idea of problems.
... liJh S and .. .f7-fS, Bauer was able to After S ... liJxg s 6 liJxg s g 6 ! ? (the
equalize the g ame and a draw was m ost ambitious move; 6 ... e6 is a safer
agreed. option, since the tactics don't work for
White; e.g. 7 liJxf7 ? ! 'iVf6 ! S liJxh S 'iVxb2
9 'iWC3 ? ? 'ii'C l m ate, or similarly 7
Game 4 6 liJxh 7 ? ! c4 S 'iVf3 llxh 7 9 dxe6 'ii'f6 ! 10
Z.Kozul-V.Topalov exf7+ 'ifi>dS 1 1 'iVxaS?? 'iVxb2), there are
I sta n b u l Olym p i a d 2000 several interesting lines to con sider:
1 d4 liJf6 2 liJf3 cS 3 ds bs 4 .i.gs liJe4 s

h4
a) 7 'iWxb S ? ! .i.g 7 gives Bl ack more

th an enough compen sation, since it is
quite difficult for White to defend b2.
This continuation h asn 't found For example, S c3 (or S liJd2 liJa6 9. c3
m any followers, since Black can equal :bS 10 'iVd3 l:txb2 11 liJC4 :bS and
ize rather easily. The idea is to open the Black was fin e in J. Bonin: M.Molner, US
h -file the rook after the likely ... liJxg s, Chess League 2007) S ...liJa6 9 'iVa4 l:tbS
1 84
10 'ii'c 2 'ii'b 6 11 b3 ttJb4 12 'iVd2 fS with when Bl ack's superior structure and
excellent play for the pawn . bishop pair provide full compensation
b) 7 'fic3 ! ? - (Vaganian ) 7 .. .f6 8 ttJf3 for the pawn .
.1g 7 9 e4 (or 9 'fixc S ? ! fS 10 c3 ttJa6 1 1 d) 7 e4? ! was played in V.Golod
'ii'e 3 0-0 with compen sation) 9 . . .�6 10 V.Topalov, European Cup, Bugojno
a4 bxa4 with an unclear g ame. 1999, which continued 7 ... .i.g 7 ! 8 c3 (8
c) 7 �f3 ! ? should be met by fS ! (the 'iVxb S ? ! e6 9 ttJf3 exds 10 exds 0-0
idea was 7 .. .f6 8 ttJe6 ! - Avrukh) and would allow Black a very strong initia
then : tive) 9 ... 0-0 (White is already somewhat
worse) 9 .i.e2 e6 10 ttJf3 exds 1 1 'iVxds
(or 11 exds .l:le8 12 0-0 d6 with an edge
for Black) 11 ... 'fib6 12 ttJbd2 (not 1 2
Wxa8? ? .i.b7 and the queen is trapped)
12 ... .i.b7 13 'irb3 c4 14 'iVdl d6 15 a4 a6
16 axbs axbs 17 llxa8 .i.xa8 18 0-0
ttJd7
Cl) 8 'ii'C 3 ':'g8 9 'iVxc s ttJa6 ! (9 ... e6

10 "iie 3 is good for White) 10 �e3 (or
10 "ii'x bs l1b8 11 'iVd3 l:txb2 12 c3 ttJcS
13 "iie 3 'fib6 14 ttJd2 h6 15 ttJgf3 .1g7)
10 ... h6 (or 10 ... 'ifb6 ! ? 11 'iVxb6 axb6
intending ... ttJb4 or ... .i.g7) 1 1 ttJf3 .i.b7
12 d6 e6 13 ttJC3 'iVas and Bl ack h as
excellent counterplay with m oves like and Topalov h ad a clear advantage,
... bS-b4, ... ttJb4, ... .i.g7 and ... ':c8. with the bishop pair, m ore space, and
(2) 8 d6 ! ? ttJc6 9 'iVds e6 10 ttJxe 6 ! targets on e4, b 2 and C3.
dxe6 11 'iVxc6+ .i.d7 12 'iVxcs 'ifb6 The g am e went on : 19 l:.el ttJcS 20
(12 ... .i.g 7 ! ? 1 3 c3 :c8 al so gives Black a .1fl 11e8 2 1 'iVc2 h 6 2 2 g3 .1b7 2 3 'ifbl
surprising amount of play for the h S 24 h4 'iVc6 2 5 .i.g2 ttJd3 2 6 lIfl ? !
pawn s) 1 3 'iVc3 (13 'iVxb6 axb6 hardly (the rook is passive here; but 2 6 l:te3
lessens Black's initiative at all) 13 ... :g8 .i.h 6 and 2 6 l:te2 b4 are al so very good
14 e3 .i.xd6 1 5 ttJd2 :c8 16 'ifb3 �e7 for Black) 2 6 ... dS (26 ... b4! looks even
17 0-0-0 .i.c6 18 �bl .1dS 19 'iVxb s strong er) 27 b 3 ? ! (27 ttJel ttJcS 28 exds
"iixb s 20 .i.xb s .i.xg 2 2 1 .:thg l .i.ds, 'ikd7 2 9 d6 .i.xg 2 30 ttJxg 2 would have
1 85
been m ore tenacious) 2 7 ...WcS 28 exds Alternatively, 9 iLxb S l:tb8 10 a4

iLxdS 29 bxc4 bxc4 30 lLlg s .i.xg 2 3 1 .i.xb2 11 .l:tbl (after 11 lLlC4 .i.xal 12
�xg 2 WdS+ 3 2 lLlgf3 iLxC3 3 3 'iVC2 'ikxal 0-0 1 3 'iVe s d6 14 'iVh 2 h S ! 15
.i.xd2 34 'iWxd2 l:les (the knight on f3 is gxh 6 lLlC7 or 1 3 d6 exd6 14 lLlxd6 .l:.b6
lost) 35 l:.al l:.fS 36 'ili'e3 lLles 0-1. A 15 'ii'e s 'ii'C 7 16 'ii'h 2 h S White lacks
convincing win by Topalov. compensation for the exchange)
s ... g6 1 1 ... .i.g7 12 0-0 (or 12 d6 0-0 and Black
This tim e S ... .i.b7 is n ot quite as ef is fine) 12 ... 0-0 13 l:tel d6 is roughly
fective, as after 6 .. .fS 7 lLlbd2 c4 (or if equal . Bl ack might follow with ... lLlC7 or
7 .. :ifas 8 c3 e6, as in H .Obermeier ... lLlb4.
H .Teske, Bad Wiessee 2009, then 9 9 ... .i.xb2 10 axbs lLlC7 11 .l:.a2 .i. g7 12
lLlxe4 fxe4 10 'iVxe4 .i.xdS 11 'iVf4! with c4 d6 13 'ii' b 3
an edge for White) 8 Wd4 lLla6 9 lLlxe4 After 1 3 .i.d3 e6 14 0-0 Bl ack takes
fxe4 10 'ii'x e4 lLlb4 11 l:tdl, the usual control of the h-file by 14 ... h 6 ! 15 gxh 6
11 ... lLlxd S ? fail s to 12 l:.xdS 'ikas + 1 3 .l:.xh 6 1 5 .l:l e l �f8, a s in the game. In
�dl ! e6 1 4 'iVf4! iLxdS 1 5 lLl e s and stead, Tsesarsky suggests 1 3 e s t ?, when
wins; while 6 ... lLlxg s 7 hxg s c4 8 Wd4 13 ... dxe s (or 13 ... .i.fs 14 'iWa4) 14 lLle4
e6 9 e4 also offers White a slight ad .i.fs 15 lLlxcs 'iid 6 16 lLla6 would still
vantage, as the rook on h8 and bishop be unclear.
on f8 are n ot th at easy to activate. 13 ... e6 14 .i.d3 �f81
6 lLlbd2
Of course the Trompowsky idea of
'ii'd 3-C3 is n ot possible here, since 6
'ii'd 3 lLlxg s 7 'ii'C 3 ? ? just loses a piece to
7 ... lLlxf3+.
6 ... lLlxgs 7 hxgs .i. g7 8 e4 lLla6
Topalov decides not to castle, since

the king on f8 is less badly pl aced than
the king on el; while if Kozul castles
short, then ... h7-h 6 becomes possible
again, with ch ances for an initiative on
the h-file.
lS :a3 exd s 16 exd s a6 17 0-0
186
Rather obliging, but White had 2 2 ... �f6?!

nothing better at this point. 17 b 6 ? ! A mistake in turn, which allows
would b e met by 17 ... 11b8 and 18 ... ttJa8 White a surprising tactic (see the n ext
winning the pawn for nothin g . n ote). In stead, 22 ... �xf3 2 3 ttJxf3 axb5
1 7. . .!:tb8 18 l:tel h6! 1 9 gxh6 :xh6 2 0 24 cxb5 �f6 would h ave been g ood for
ttJe4 11b6 21 'ii'a 2!? � g4 Black.
After 21 ... axb 5 22 lIa7 White gets 23 'iVd2 l:thS
some counterplay; for in stance 22 ... b4?
loses to 23 'ii'a 5 ! lIb7 24 ]::t a 8, threaten
ing ttJxd6.
22 ttJegs?!
The exchange sacrifice 2 2 l:txa6!
ttJxa6 2 3 bxa6 was stronger, when the
passed pawn gives White sufficient
compensation . For example:
24 'iWa s?!
H ere White could have pl ayed 24
ttJe6+! fxe6 25 �xg6 lir.h 8 (or 25 ... 'iti>g 7
2 6 it.xh 5 �xh 5 2 7 bxa6 exd5 2 8 ttJh 2 !
and the a3-rook swings across to the
king side) 26 'ikf4 with an unclear g ame.
The alternative 24 ...ttJxe6 25 dxe6 'iti>g7
2 6 exf7 'ii'h 8 2 7 'iti>f1 is actually better
a) 2 3 .. .lIb8 24 a7 lla8 2 5 'ii'a 6 �xf3 for White, as the black king is more vul
26 'iVxd6+! 'ii'x d6 27 ttJxd6 (threatening n erable; for example, 2 7 ... :h 1+ 2 8 'iti>e2
lIe8+! ) 27 ... g5 28 ttJb 5 �xd5 29 cxd5 l:txe1+ 29 'iti>xe1 axb5 30 cxb5 d5 3 1
J:r.a6 30 ttJC7 l::t 6 xa7 31 ttJxa8 ':'xa8 with f8'ii'+ ! 'ii'xf8 3 2 lta7+ 'iti>g8 3 3 �xg 6.
equality. 24...ttJa8 2 S ttJe4?
b) 23 ... �xf3 24 a7 ! �xe4 (or A third mi stake. The only chance
24 .. :iVa8 25 'ii'a 5 ! l:tb2 26 gxf3 �d4 2 7 was 25 l:.e4! and if 25 ... �f5 then 26 g4!
J:r.e2 ':'xe2 28 �xe2 'iti>g 7 2 9 �f1 l:lh 8 30 �xe4 2 7 g xh 5 �xd3 2 8 ttJh 7+ 'iti>g 7 2 9
'iWC7) 2 5 a8 Q l:tb8 2 6 'ii'x b8 'ii'x b8 2 7 ttJxf6 'ii'xf6 30 l:txd3 (Tsesarsky) re
it.xe4 �d4 28 g 3 'iti> g 7 when Black is m ains unclear.
perhaps slightly better, but a draw is 2S ...�xf3 26 gxf3 � d4
the most likely result. Now Topalov g ets a decisive attack.
187
5 . . .liJxg 5 6 liJxg 5 and then either 6 . . .e6

or, more ambitiously, 6 ... g 6 ! ?, when
Black h as a comfortable game, espe
cially if White doesn't reply with 7 'ii'C 3
or 7 'ii'f3 .
Game 4 7
R.Vaganian-V.Topa lov
I sta n b u l Olym piad 2000
27 'itfl?1 1 d4 liJf6 2 liJf3 cS 3 d s bs 4 a4

Makin g it easy. 2 7 'itg 2 'lid7 2 8 <itfl
f5 29 liJg 3 l:Ih 2 30 l:Ia2 'iVh 7 3 1 .l:.e6
would h ave been more resilient, al
though Black win s in any case after
3 1 ... 'iVh 3 + 32 'it>el i.. xf2+ 3 3 lbf2 'ii'x g3
34 l:tee2 'ifh4 as White can hardly
m ove.
27 ... l:th2 28 'it>e2 fS 29 liJd2 i.. xf2 30
l:f.eal ..te3+1 3 1 <Ji>xe3 "ii'g S+ 0-1
Game summary
With 5 h4 White aim s to get an unbal With this move White seeks to cl ar
anced position, giving up the bishop ify the situation on the queen side be
pair to open the file for the rook on h l fore decidin g what to do next. Apart
and keep its counterpart o n h 8 out of from 4 ..tg 5 and 4 c4, White has a cou
play. Ironically, in the g am e it was ple of further tries, but neither of them
Black who attacked down the h -file af are very ambitious and Bl ack equalizes
ter Kozul l ater castled short, while To rather easily:
palov played ... 'it>f8 and ... h 7-h 6. Per a) 4 g3 i..b 7 5 c4 e6! (5 ... g6 6 i.. g 2
h aps deterred by the result, few other bXc4 7 liJC3 ..tg 7 8 0-0 0-0 9 liJe5 d6 10
players h ave tried 5 h4, but the g ame liJxC4 liJbd7 11 .l:[el ..ta6 12 'ii'a4 'ii'c 8
was actually unclear for a long time 13 liJa5 liJb6 14 'iVh4 lle8 15 ..tg 5 fiC7
and Topalov only won after Kozul went 16 liJc6 i..b 7 17 e4 liJbd7 18 f4! was
wrong in the middlegame. somewhat better for White in
The altern ative variation seen in the S.Johannessen- R.J . Fischer; Havana
n otes, 5 'ii'd 3, can be answered by Olympiad 1966) 6 dxe6 (or 6 cxb5
1 88
tbxds) 6 .. .fxe6 7 cxb s i.e7 (after 7 ... d s ? ! lead in development.

8 ..th 3 ! 'iVb6 9 tbg s is a bit annoying) 8 c) 5 tbbd2 is the sharpest m ove and
.i.g2 0-0 9 O-O 'was played in G.Marras requires a precise response: s ... tbxds 6
R.Bernardi, Palau 2009, and now 9 ... a6 a4 c4! 7 c3 e6 8 e4 tbf6 9 tbd4 ds is
(or 9 .. :ifb6 10 tbc3 a6) 10 bxa6 tbxa6 or comfortable for Black.
10 ... l:Ixa6 would give Black standard d) 5 i.d2 (the most common reply in
and good Benko compensation for the practice) s ... b4 and n ow:
pawn .
b) 4 e3 can al so be met by 4 ... i.b7,
but I like 4 .. :ii'as+ ! ?
dl) 6 c4 e6 (this break is both the

m atic and logical here; 6 ... g6 7 i.d3
i.g7 8 a3 0-0 9 0-0 tba6 10 e4 d6 is al so
This check can often b e ann oying fin e for Black) 7 dxe6? ! (7 'iVc2 would be
for White when the ds-pawn is vulner more consistent) 7 ... dxe6 8 i.d3 i.b7 9
able, since all responses mean placing 0-0 tbbd7 10 llel i.d6 1 1 e4 tbe s 12
the pieces in a less than optimal way. i.g s ':'d8 was good for Black in
a) 5 tbC3 is the n atural square for R.5eifarth-T.Zoltek, Pardubice 1992.
th e knight, but it can be chased away; d2) 6 i.d3 i.a6 ? ! (6 ...tbxd s ! would
i.e. s ... b4 6 tbe2 i.b7 7 tbf4 e6 8 i.C4 give White sufficient compensation, but
fiC7 intending ... i.d6 is unclear. no more than that) 7 'ife2 ? ! (7 c4! would
b) 5 'iVd2 (a rather passive move) be slightly better for White, since
s ... b4 6 c4 (or 6 i.d3 i.b7 7 e4 c4 8 7 ... tbxds is now met by 8 i.e4! tbb6 9
i.xc4 tbxe4 9 'iVd4 tbf6 10 tbbd2 'iVb6 is i.xa8 tbxa8 10 b3 g 6 11 a3 i.g7 12 l:[a2
al so unclear) 6 ... e6 7 b3 i.b7 8 i.b2 i.b7 13 0-0 tba6 14 i.cl and Black
exds 9 i.xf6 gxf6 10 cxds i.g7 11 a3 fs doesn't have enough for the exchange)
12 lla2 i.O 13 tbxC3 bxc3 14 WCl 7 ...tbxds 8 e4 tbC7 9 0-0 g 6 10 tba3 i.g7
i.xds 15 :C2 i.xb3 16 l:txc3 i.e6, when 11 tbC4?! i.xc4 12 i.xC4 i.xb2 13 l:tabl
Black's extra pawn is balan ced by his i.g7 was also good for Black in
slightly worse structure and White's E.5osulin-V.Krasnov, Donskoj 2002.
1 89
4 ... b4! ship, Chartres 2005, saw 6 b3 i.g7 7

�b2 0-0 8 g 3 ? ! (it wasn't yet too late for
8 lDbd2) 8 ... e6!
This is the best m ove in this position

and leads to an unclear game. Second
best, in my opinion, is 4 ... 'ii a 5+ with (now this break is quite effective
chances for both sides, though I 'm n ot since White is unable to support the
sayin g any m ore about that. After 4 ... b4 centre with e2-e4) 9 �g 2 ? ! (here 9 dxe6
Bl ack g ains space, reinforces his control probably offers more ch ances of keep
of the long dark-squared diagonal in ing the balance) 9 ... exd5 10 cxd5 !le8
co-operation with his (soon to be) fi (now Black has a kind of improved
anchettoed bishop, and takes the c 3 - Modern Benoni and a slight advantage)
and a3-squares away from the b l 11 0-0 �a6 (the bishop is well placed
knight. The one drawback is that White on a6, exerting pressure on the light
n ow h as access to c4, which might be a squares) 12 J:1el lDe4! (12 ... d6 13 lDfd2
nice square for a knight. lDg4 14 i.xg 7 c;txg 7 1 5 lIa2 "iff6 16 lDf3
5 �g5 lDd7 17 lDbd2 'ili'e7 18 lDfl f6 19 .l:!.c2
Threatening to compromise Bl ack's lDge 5 20 lDe3 fS 21 h4 was less clear in
pawn structure with �xf6. This follows A.Bagheri-J . Lautier, French Ch ampion
the example of the previous g ames in ship, Chartres 2005) 1 3 �xg7 �xg 7 14
this chapter, and we will continue in lita2 'ii'f6 1 5 l:tc2 d6 16 e3 lDd7 17 lDh4
the sam e fashion with 5 . ..t2Je4. c4! 18 bXC4?! (but if 18 �xe4 .l:!.xe4 19
White's other m ain option is 5 c4, lDd2, then 19 ... lIxh4! 20 gxh4 c3 21
bringing about a familiar Benko struc lDe4 "ifxh4 and Black is clearly better)
ture from Gam e 3 5 . Indeed, 5 ... g6 (but 18 ... b3 19 lIcl 'ii'xf2+ 20 c;th l lDdc5
n ot 5 ... bxC3 ? ! 6 lDxC3 and White is g ave Black a great position in
slightly better) 6 lDbd2 �g7 7 e4 d6 S.suvrajit-J . Deepan Chakl<ravarthy, In
would tran spose to that game. In stead, dian Championship, Che n n ai 2008.
A. Bagh eri-J . Lautier, French Ch ampion- 5 ... lD e4
1 90
6 g3?! C l ) 7 'iVd3 liJf6 8 liJbd2 .i. a 6 9 liJC4

Not really a very good move. In d6 10 e4 .i.g7 11 .i.e2 0-0 12 as liJg4 1 3
stead, White should investigate the 'i¥b3 liJd7 with a n unclear g ame.
various continuation s from Games 44- C2) 7 .i.es f6 (7 ... liJf6, allowing 8
46 and see how the inclusion of the .i.xf6, would be incon sistent) 8 .i.f4
two extra pawn moves affects things. .i.g7 9 liJbd2 fs ! ? (9 ... liJxd2 10 'iix d2 fs
In those games Black did not always is al so fine) 10 liJxe4 fxe4 1 1 ..tes 0-0
an swer a2-a4 with ... b s-b4, whereas 12 .i.xg 7 <i>xg 7 13 liJd2 e 3 ! 14 fxe 3 e6
here he is already committed, but apart and Bl ack h as good play for the pawn .
from that Black has more leeway, since c3) 7 liJbd2 liJxd2 8 'iWxd2 ..tg7 9
the b-pawn is no longer vuln erable. In l1bl (or 9 c3 bxc3 10 bxc3 'iVas) 9 ... 0-0
particular he can forego ... .i.b7 in fa 10 .i.h 6 d6 11 h4 (11 ..txg 7 <i>xg 7 12 e4
vour of ... g7-g6. Then : .i.g4 is equal) 11 ... .i.xh 6 12 'iVxh 6 f6 1 3
a) 6 h4 g 6 7 liJbd2 liJxg s 8 hxg s .i.g 7 e 4 'iVe8 and Black is fine.
9 liJC4 d6 is fine for Black. d) 6 ..th4 g6! (6 ... ..tb7 ? ! is wrong
b) 6 'ii'd 3 liJxg s 7 liJxg s g 6 8 'iVf3 fs ! here - see the n otes to Game 44) and:
(8 .. .f6, as in M.Hell-P.5ahm, Volklingen
2003, could still be met by 9 liJe6 ! ) 9 d6
liJc6 10 'iVds e6 11 liJxe6 dxe6 12
'i¥xc6+ .i.d7 13 'iix cs .i.g7 (or
13 .. :iVb6 ! ?) 14 c3 as and with a lead in
development and ideas of ... .:.c8 and
... 0-0, or 1 5 e4 'iWg s ! , Black h as very
good play for the pawns.
c) 6 .i.f4 g6! (otherwise 6 ... .i.b7
tran sposes to Game 45) and n ow we
have:
191
dl) 7 'iVd3 liJd6 (now the m ain ideas h ave anything on the light squares to
are ... liJfs, ... .lta6 and ... .ltg7; Black is compensate for the bishop pair. For
comfortable) 8 c4? ! (but after 8 liJbd2 example, 8 .ltg 2 .ltg7 (8 ... d6 ? ! allows 9
.ltg 7 9 Itbl .lta6 10 'iVe3 liJfs 11 'ii'x cs liJe6 ! ) 9 d6 (or 9 c3 as 10 d6 lita6)
liJxh4 12 liJxh4 "iVb6 1 3 'iVxb6 axb6 14 9 ... liJc6 and Black is doing fine.
b3 bS 15 axbs .ltxbs Black h as active 7 .ltg2 d6 8 .ltcl?!
pl ay for the pawn) 8 ... .ltg 7 9 .liIa2 b 3 ! 10
'iVxb3 liJa6 11 liJbd2 .l:tb8 was good for
Bl ack in M.Tratar-A.Escobar Foreno,
World Student Team Ch ampionship,
Paranana 1993 .
d2) 7 liJbd2 liJxd2 8 'iVxd2 .ltg 7 !
(8 ... d 6 9 e 4 .ltg 7 10 .ltb S+ liJd7 1 1 c3 a 6
12 .lt e 2 0-0 1 3 0-0 as 1 4 'iVc2 liJb6 1 5
liJd2 bxc3 16 bxc3 'iVC7 17 litfbl was
good for White in A.5ultanov
V.Grechihin, Russian Championship,
Sam ara 2000) 9 c3 (9 d6? ! is n othing to Vaganian decides to keep the
fear: 9 ... .ltxb2 10 .ltxe7 "iVb6 11 'iVf4 bishop after all, but he will have severe
.ltC3+ 12 �dl fS 1 3 litbl liJc6 and Black trouble getting it out again later.
is better) 9 ... bxC3 10 bXc3 'iVas 11 .l:ta3 8 .lt g7 9 liJfd2 liJxd2 10 liJxd2 liJd7 11
.••
.lta6 12 e4 .ltxfl 1 3 �xfl d6 14 �g l 0-0 0-0 12 lit e 1 a s

liJd7 1 5 h 3 c4 16 �h 2 liJcs with a com Con solidatin g the queen side and al
fortable g ames for Black. lowing the knight to go to b6 unmo
lested.
13 e4
13 liJC4 would just be met by
1 3 ... .lta6.
13 ...liJ b6 14 .l:i.a2
Preparing to let the dark-squared
bishop out again, so Bl ack must strike
quickly.
14 e6!
.•.
An often occurring move with the

knight on b6. Here it is even more ef
6 ... g6 fective, since ... .ltxe6 threatens the rook
6 ... liJxg S ! 7 liJxg s g6 seem s more on a2; for instance, 15 dxe-6 .ltxe6 16 b3
precise to me. White doesn 't really .ltC3 17 .ltb2 ? run s into 17 ... liJXa4!.
1 92
A desperate attempt t o free h i s po

sition . White might h ave tried prepar
ing thi s with 26 ttJdl, when 26 .. J:te8 2 7
c 4 .lir.el+ (or 2 7 . . .bxc3 2 8 'iVxc3) 2 8 .tfl
.ta6 29 .:te2 ! l:txdl 30 J:.e6 'ii'a l 3 1 'ii'h 3 !
would b e unclear, but after 2 6 ... .txd S !
27 'iVxdS+ ttJxdS 2 8 .txdS+ 'iti> g 7 29
.txa8 'iVd4! the queen triumph s over
the uncoordinated white forces; e.g. 30
ttJe3 (or 30 .tf3 'iVxf4 3 1 'iti>g 2 'iVcl and
Bl ack win s by advancing the pawns
15 b3?! exd 5 16 exd 5 .tb7 17 ttJfl .:te8 since White has n o m oves) 30 ... 'ii'xf4 3 1
18 Iixe8+ 'iVxe8 19 h4 'iti>fl 'iWd4 3 2 .tf3 f4 3 3 ttJC4 d S 34 ttJxas
Unfortun ately, White can 't play 19 'ii'd 2, when ... 'ifcl-bl win s the rook.
.tb2 because of 19 ....txb2 20 'iVxb2 26 ... bxC3 27 l:le2
'iVe s, winning the ds-pawn . The text 27 1:[c2 lIe8 is little better.
prepares the development .tf4 and 27 ... .ta6 28 ttJC4 ttJXC4 29 bXC4 .l:lb8
ttJe3 by preventing ... g 6-g s, but al so
weaken s the king side.
19 ...f5 20 .tf4 .te5 21 ttJe3 .txf4!
Topalov is only relinquishing control
of the long diagonal temporarily, as his
queen will soon take up residence
there in stead.
22 gxf4 'iVe7 23 h5 'iVh4 24 hxg6 hxg6
25 'iVf3 'iVf6!
Black h as a decisive advantage. The

c-pawn is very dangerous and the
white king is vulnerable. Vag anian's
next move allows Topalov to end things
quickly.
30 lite6? 'iVxe61
Only a temporary queen sacrifice as
Black will soon g et it back.
31 dxe6 c2 32 'ii'e 3 l:tbl+ 33 'iti>h2 Cl'iV
26 c4 34 'iVg3 'iti>g7 35 'iVh4 'iVgl+ 36 'iti>h3 0-1
1 93
Game summary and .i.b2 is a reason able try for an ad

Black got the upper h an d from the vantage.
opening phase after Vaganian wasted 2 e 3 liJf6 3 liJf3 g6
tim e with his dark-squared bishop. To
palov's realization of his advantage in
the middlegame was very in structive
and he won convincingly.
White would do better to retreat
the bishop on m ove 6, either to h4 or
f4, followin g the example of G ames 44
and 4 S . H owever, Black can get a com
fortable g am e by discardin g an early
... .fi.b7 in favour of ... g 7-g6 as examined
in the notes. In short, Black h as n othing
to fear after 4 a4 b4! . 4 .i.e2
White has al so tried:
a) 4 .i.d3 dS S c3 (very passive: the
Game 48 Colle set-up is not very effective
D.Brumen-I.Smiri n ag ainst ... g 7-g6, so Bl ack should be
C roatia n C u p, Ra bac 2004 h appy if anyone pl ays like this)
S ...liJbd7 6 liJbd2 .i.g7 7 0-0 0-0 8 l:tel
l d4 c5 (not S liJe s ? liJxes 9 dxes liJg4 and
Smirin probably played this way to Black won a pawn in F.Wegerer-M. Roiz,
divert Brumen from his usual Trom Oberwart 2004) 8 . b6 (or S ... lIeS 9 e4
..
powsky (1 ... liJf6 2 .fi.g S). The recom cxd4 10 cxd4 dxe4 11 liJxe4 liJxe4 1 2
mended move order in this book is 1 d4 .fi.xe4 liJf6 with an equal game) 9 e4
liJf6 2 liJf3 cS, when 3 e3 tran sposes to cxd4 10 cxd4 (not 10 liJxd4?! liJcs 1 1
the g ame. If somebody plays 3 e3 one .i.c2 ? ! dxe4 12 liJxe4?? liJfXe4 13 .i.xe4
can probably assume they are rather liJxe4 14 l:txe4 eS 0-1 A. lbanez Lozano
unambitious or else want to avoid a R.Cifuentes Parada, Dos Herm anas
debate on opening theory. 2004) 10 ... dxe4 11 liJxe4 .i.b7 12 liJc3
Similarly, 3 b3 is n ot the m ost ag liJdS 13 ii'e2 liJ7f6 14 liJxdS liJxdS and
gressive continuation . Black replies Bl ack was slightly better in P.short
3 ... g 6 and often ... csxd4 and ... d7-dS (cf S.Conquest, Iri sh Championship, Dublin
4 b3 below). The queen side fianchetto 2007, as the isol ated queen 's pawn is a
m akes more sense to me when Black lon g -term weakness.
h as pl ayed ... e7-e6, aiming for a Nimzo b) 4 b3 .i.g 7 S .i.b2 0-0 6 liJbd2 (6
Indian or Queen's Indian , when b2-b3 .i.d3 allows 6 ...cxd4 7 exd4 liJdS ! 8 0-0
1 94
Th e Top a l o v Va ria t i o n
ttJf4 9 l:tel ttJxd3 10 'ilt'xd3 d6 1 1 C4 ttJc6 ble, tryin g t o open u p the g am e with a
12 h3 d5 and Black was better in well-timed ... d6-d5 or ... a7-a6 and ... b6-
R.Wiemer-F.Gheorghiu, Remscheid b5) 17 ttJel a6 18 �f3 ':'c7 19 h3 l:[dc8
1986) 6 ... cxd4 7 exd4 ttJc6 8 a3 d5 9 20 .i.xb7 'ii'x b7 21 1:.c2 ttJe8 22 �xg 7
.i.d3 ttJh 5 10 g 3 was seen in ttJxg 7 2 3 l:tdcl ttJe6 24 'ilt'a2 ttJe5 2 5
V.Vain shtein-Z.Gofshtein, Israeli 'ilt'b3 ttJd7 2 6 'ilt'd3 ttJe5 2 7 1i'b3 b 5 ! 2 8
Championship, Ramat Gan 1992, and cxb5 axb5 (this structure is fine for
now 10 .. :iWb6! 11 c3 l:te8 would have Black) 2 9 l:txC7 l:txC7 30 ':xC7 ttJxC7 3 1
given Black the edge. ttJef3 ttJxf3+ 3 2 ttJxf3 'ilt'c6
c) 4 dxc 5 ! ? (White can't keep this
pawn, but he does gain a bit of time to
seize some space on the queen side af
ter Black recaptures) 4 ... 'iIt'a5+ 5 ttJbd2
'i!i'xC5 6 a3 .i.g7 7 b4 'fIC7 8 �b2 0-0 9 c4
b6 10 �e2 .i.b7 11 ':cl d6 12 0-0 ttJbd7
13 11fh3 11ac8.
(the endgame is favourable for

Black, who h as the potential to create a
passed pawn in the centre, whereas it
is difficult for White to advance on the
queen side) 3 3 ttJd4 'ili'cl+ 34 rJi>h 2 e 5 3 5
'iVc2 ? ! (exch anging queen s only m akes
things worse for White) 35 ... 'iVxc2 3 6
This is actually a standard position ttJxC2 f5 (the threat is n ow clear: . . .<3;f7-
for the 4 dXc5 variation . White has a bit e6-d5-c4-b3) 3 7 rJi>g 3 ttJd5 3 8 f4 r3;f7 3 9
more space, as intended, while Black r3;f3 ttJ c 3 (preventing both e 3-e4 and
has an extra centre pawn and well a3-a4) 40 fxe 5 dxe5 41 ttJal <3;e6 42
placed pieces; chances are TOughly ttJb3 <3;d5 43 ttJC5 ? ! (43 ttJd2 was a bet
equal . We'll have a look at one game to ter try, intending 43 ... ttJe4 44 ttJxe4
see how pl ay might continue: 14 ':'fdl fxe4 45 'it>g4 <3;C4 46 <3;g 5 <3;d3 47 a4! ,
(or 14 ttJd4 a6 15 .i.f3 ttJe 5 16 .i.xb7 though the resultin g queen endgame
'ii'x b7 with an equal g ame) 14 ... 1Ifh8 1 5 would still be lost) 43 ... e4+ 44 <3;f2 <3;C4
'i!i'a2 'fIa8 16 'ilt'al l:tfd8 (Black h as ar 45 g4 ttJbl 46 g xf5 g xf5 47 <3;g 3 ttJxa3
ranged his pieces as actively as possi- 48 <3;f4 <3;xb4 (now the b-pawn decides
195
Play th e B e n k o G a m b i t
the g ame) 49 iDe6 <ii> C 4 50 <ii> xf5 iDc2 5 1 chances against the weak d4- and b2-
<ii> x e4 b 4 5 2 iDf4 iDel ! 5 3 iDd5 b 3 5 4 pawns) 12 ....l:.d8 (or 12 ... :b8) 13 .i.e3
iDb6+ <ii>b 4 5 5 iDd5+ <ii> C 5 5 6 iDC3 <ii> C 4 was played in A.5avvopul 0 - 1 . Laben sky,
57 iDa4 iDd3 58 iDb6+ <ii> c 3 59 iDd5+ Ukrainian Championship, Alushta
<ii> c 2 60 iDb6 iDf2+ 6 1 <ii>f3 iDxh 3 6 2 iDa4 2009, and now 13 ... :b8 would be
iDg 5+ 63 <ii>f4 iDe6+ 64 <ii>f 5 iDC5 6 5 slightly better for Bl ack.
iDxC5 b2 0 - 1 S.5uvrajit-E.Ghaem 8 ... iDc6 g iDc3 iDe4
Maghami, Kolkata 2008. An instructive The standard riposte to the c-pawn
win by the Iranian GM. advance. The knight cannot be taken
4 .i.g7 S 0-0 0-0 6 c4 cxd4 7 exd4 d S
•.. because the d-pawn would drop off.
Black h as equalized. In fact, it is White
who h as to play accurately to maintain
the bal ance.
10 �e3 b6
I like this idea ... b7-b6 and ... iDxC3,
trying to play against the weak c3- and
a2-pawns.
11 cxb6 iDxc3 12 bxc3 axb6 13 'ii'd 2
.i.g4 14 l:.fbl iDa s
N ow White h as a reversed Queen's

Gambit Tarrasch with an extra tempo,
but it is n ot really of much help to him
since the play is primarily strategical in
n ature.
8 cs
Alternatively, 8 iDc3 iDc6 9 iDes ? !
(but 9 h 3 dxc4 1 0 .i.xc4 iDa5 1 1 .i.d3
.i.e6 12 'ii'e 2 :'c8, as in J . Levitt
C.Beaumont, British League 1999, is ls iDes?!
fin e for Black with ideas like ... �c4, It is a strategic mistake to exch ange
... iDd5 and ... iDc6) 9 ... dxC4 (now it is the light-squared bishop, as it only
White who h as to prove that he can m akes the c4-square even more attrac
equalize) 10 iDxc6 bxc6 11 .i.xC4 'ii'd 6 tive for the bl ack knight.
12 .i.e2 ? ! (12 h3 was a preferable way lS ....i.xe2 16 'ii'x e2 f6 17 iDf3 'ii'd 6 18
to prevent ... iDg4, though Black still has "iWbS?!
a comfortable g am e with long-term Ceding the c4-square straightaway.
196
18 tDd2 �fc8 19 'iVd3 :c6 20 %:tb2 would ter o f a reversed Queen's G ambit Tar
have been more tenacious. rasch, where White's extra tempo was
18 ...tDC4 19 �C1 'iWe6 20 �b3 :a s 21 not enough to outweigh the long-term
i.f4 'iWfs strategic weaknesses in his position :
the queen side pawn s. White might
hold the bal ance with active play, but
he cannot really hope for more th an
th at. As it turned out Smirin scored a
convincing win in a rather one-sided
game. The various alternatives for
White given in the n otes are h ardly any
more promising either.
Game 49
22 i.e3?1 P.Velikov-M.Tal
This just creates another target on Berl i n 1986
e3. 22 i.g 3 was correct, followed by
l:[dl and tDd2, when White still h as 1 d4 tDf6 2 tDf3 cS 3 c3
some drawing chances. This rather diffident move is often
22 :a3 2 3 �b4 :fa8 24 1If1
..• played when White is aiming for a solid
Hoping for 24 .. J:lxa2 ? 2 5 ':'xa2 ':'xa2 queen 's pawn opening, such as the
26 'iVxe7, but that is easily dealt with . Torre, Colle, or London system s.
24 .. :iVd71 3 e6
•••
Now White can't help losing a pawn

and the result is virtually assured.
2S tDd2 tDxe3 26 fxe3 .l:r.xa2 27 ];txa2
l:txa2 28 c4 i.h6 29 :f3 'iVe6 30 'i!i'C3
dXC4 31 l::t h 3?
Hastening the end.
31 ..t>g7 3 2 tDxC4 'i!i'g4 0-1
.•.
Game summary
Brumen played un ambitiously in the
opening with e2-e3 and i.e2, when
Black equalized rather easily by fi The m ost popular reply: White is
anchettoing his king's bishop. The buildin g a rock-like foundation on the
game subsequently took on the ch arac- dark squares as a prelude to l ater ac-
197
tion, s o Black adopts a counter-strategy tiJxds was good for Black in G . Kam sky
on the light squares, intending to fol M.Carlsen, Moscow (blitz) 2008) ls ... h6
low with ... b7-b6, .....tb7 and ... ..te7. 16 tiJge4 tiJxe4 17 ..txe4 ..txe4 18 tiJxe4
4 ..tgS "iVc6 19 tiJd2 ds 20 ..tf4 bs with easy
Thi s takes the g am e into Torre terri equality in K.Arke11-Hmms, Southend
tory. Black can play in the sam e way 2009.
agai n st other system s: 4 ... ..t e 7 5 e3
a) 4 e 3 (Co11e) 4 ... b6 5 ..td3 ..tb7 6 Here 5 ..txf6 ..txf6 6 dxcs ! ? tiJa6 (or
tiJbd2 ..te7 7 0-0 0-0 8 'ilke2 (or 8 e4 6 ...b6 ! ? 7 cxb6 axb6 8 e3 0-0 9 ..td3 ds
cxd4 9 cxd4 ..ta6 with equality) 8 ... tiJc6 10 0-0 tiJd7 with compensation arising
(altern atively, 8 ... cxd4 9 exd4 d6 10 .l:[el from Black's bishop pair and more
tiJbd7 and 8 ... ds 9 tiJes tiJbd7 10 f4 space) 7 e4 (7 b4? just a110ws 7 ... tiJxb4!)
tiJe4! are al so perfectly sound) 9 e4 tiJxcs 8 e s ..te7 9 b4 tiJe4 10 ..td3 ds 11
cxd4 10 tiJxd4 (10 cxd4 a110ws 'ilfe2 as 12 bS tiJxc3 ! ? (otherwi se
10 ... tiJb4! ) 10 ... tiJes 11 ..tc2 'ili'c8 12 f4 12 ... tiJcS is equal) 13 tiJxc3 ..tb4 14 l:1cl
..ta6 1 3 'ilfdl tiJc6 14 l:If3 g 6 with an d4 15 0-0 dxc3 16 llfdl ..td7 led to
unclear position in E.Co11e great complications in I .Miladinovic
lR.Capabl anca, Carlsbad 1929. P.Kiriakov, H alkida 1996.
b) 4 ..tf4 (London) 4 ... b6 5 e 3 ..tb7 6 S b6 6 tiJbd2 ..tb7 7 ..td 3 0-0 8 0-0 tiJc6
..•
tiJbd2 ..te7 (threatening ... tiJh s) 7 h 3

0-0 8 ..td3 cxd4 9 exd4 d 6 (blunting the
bishop on f4) 10 0-0 tiJbd7 1 1 l::t e l l:te8
12 ..th 2 a6 13 a4 'ilfC7 14 'ilfb3 ..tf8
A standard and balanced position

for this opening.
g 'ili'e 2
Another Tal game saw 9 l:.el l:tc8 ! ?
(a typical reorganization of Black's ( 9. . .cxd4 1 0 exd4 tiJds is more usual) 10
pieces) 1 5 tiJg s (or 1 5 l:te2 ..tc6 16 c4 a3 tiJds 11 ..txe7 'ilfxe7 12 'ilfc2 cxd4 13
"ilfb7 17 lIae l ? ! ds! 18 cxds ..txds 19 ..txh 7+? (13 exd4 fs 14 c4 tiJf4 15 ..tfl
..tC4 llac8 20 ..txds 'ii'x ds 2 1 "iVxds 'iWf6 or 15 ds tiJxd3 16 'it'xd3 'it'd6 would
198
Th e To p a / o v Va ria t i o n
be unclear) 13 <ith 8 14 exd4 f5 15 c4

'iYi>xh7! (15 ...lDf4 16 d5 lDa5 17 .i.xf5 was
presumably White's idea) 16 cxd5 lDxd4
17 'iVd3 lDc2 18 lDe5 $2 .i.xd5 19 'iVh 3+
'iYi>g8 20 lDg6 'ifg 5 ! and Black won in
L. Lukovski-M.Tal. Porz 1991.
White has al so tried 9 e4 cxd4 10
cxd4 (or 10 lDxd4 d5 1 1 lDxc6 .i.xc6 12
e5 lDe4 1 3 .i.xe7 'ii'x e7 14 lDf3 f6 1 5
'fie2 .i.b7 16 exf6 'ii'xf6 17 l::t ael l:i.ad8 !
and Black equalized in A.Yusupov
S.Hmadi. Tunis Interzonal 1985. as 18 12 .i.e4
i.. x e4 dxe4 19 lDd2 would be met by Similar play can foll ow after 12 i.. c 2
19 ... 'ikg 6 20 lDxe4 .i.xe4 21 'ii'x e4 'ifxe4 (or 12 g3 f5) 1 2 .. .f5 1 3 lIfel lDg 6 14 'ili'fl
22 ':xe4 l::t d 2 and the rook endg am e is <ith 8 1 5 :acl lDh4 16 lDxh4 'iVxh4 17
drawn) 10 h 6 11 .i.xf6 ? ! (11 .i.e3 lDb4 .i.dl 'ii'g 5 18 :c2 :ac8 with a nice
12 .i.bl .i.a6 13 litel lDd3 or 11 .i.h4 game for Black in E.5h aposhnikov
lDh 5 is equal - Adorjan) 11 ... .i.xf6 12 e5 R.Popov, Russian Cup qualifier (blitz)
i.. e 7 13 a3 d6 14 .i.e4 b 5 1 5 'ili'e2 'iVb6 2004.
16 lDb3 dxe 5 17 dxe5 :fd8 was good 12 ...lDf4 13 'iVdl 'ili'C7 14 l:tel fS lS
for Black in D.Barlov-A.Adorjan, New .i.xb7 'iVxb7 16 lDC4 l:tf6
York Open 1985. Attack!
9...cxd4 10 exd4 17 lD ces gs
Or 10 cxd4 lDb4 11 i..b 5 (11 .i.bl ? Since White h as prevented .. J:tg6,
run s into 1 1 ... .i.a6. winning the ex Tal attacks g2 in a different way.
ch ange) 11 ... a6 12 .i.a4 lDe4 13 .i.xe7 ls lDd3 lDeg6
'fixe7 with equality, while after the fur Black can already con sider
ther 14 a3 lDxd2 1 5 lDxd2 lDd5 16 e4 18 ... lDxg 2 19 <itxg 2 g4, but this would
lDf4 17 'ifg4? 'ii'd 6 ! Black was clearly only be equal.
better in N.Vyskocil- I . N aumkin, Livigno 19 lDxf4 lDxf4 20 l:.e3?!
2010. Here 20 d5 ! lDxd5 (or 20 ... g4 21
10 . lDd S!
. . dxe 6 ! g xf3 2 2 exd7 l:.ff8 2 3 'ii'd 6 lDe2+
The thematic equalizing move. 24 lilxe2 fxe 2 25 'iVe6+ ltf7 2 6 'ii'e 8+
11 .i.xe7 lDcxe7 with a draw) 2 1 lDxg 5 :g6 2 2 'it'h 5 lDf6
This set-up leads to a roughly equal 2 3 'iVf3 ! would be a simpler way to
position, though I would prefer to be keep the bal ance.
Black, as he has greater prospects for 20 g4 2 1 lDel :cS 2 2 'ii'd 2 l:.g6
•••
dynamic play. Although the position is objectively
1 99
still about equal, Black h as m ore prac which m akes the position very difficult
tical chances. Indeed, White n ow be for White. Although Tal did not con
gins to go astray, wastin g tim e with his duct the endgame with complete accu
rook. racy, he m an aged to win in the end.
23 :e5?1 d6 24 ':e3 h5 25 :c1 b5?1 38 ...<it>e4 39 <it>e2 f4 40 l:th4 d5 41 b4

Since this doesn 't prevent c3-c4, <it>d4 42 h 3 gxh 3 43 g3 �C7 44 gxf4 e4
Black should h ave pl ayed 2S ... h4! at 45 l:txh 3 ':c2+ 46 <it>f1 ':xa2 47 :h7
once, when 26 c4 h3 open s up the <it>c3?
white kin g position (27 g 3 ? loses to 47 ...<it>C4 was correct, intending 48
27 ... tDg 2 ! ) with a strong initiative. fS 1:ta3 and ... l:!.f3, when Black should
26 b3?1 win .
Equally there was n o need to pre 4 8 f 5 d4 4 9 f6 l:ta1+ 5 0 <it>g2 lla6 5 1
pare this m ove. H ere 26 c4 llxC4 2 7 liC7+?
l:!.xc4 bxc4 2 8 l:[ c 3 'iie 4 2 9 <it>fl would Here S l l:th 6 ! would have drawn .
h ave been unclear again, as 2 9 ... d S ? 51 ... <it>d2 52 f7 .l:lf6 5 3 l:te7 d3 54 b5 e3
loses a piece to 30 f3 ! . 5 5 l:txe3 l:txf7 56 J::t e 6 <it>d1 57 <it>f1 d2 58
26. . .'ilVd 5 2 7 c4 l:tc6 ':f5 59 b6 axb6 60 l:txb6 :C5 61
Now Velikov feels he h as to sacrifice lId6 <it>c2 0-1
a pawn, to try and fight for a draw in
the endgame. Game summary
27 bxC4 2 8 l:lec3 'iix d4 29 'iix d4 tDe2+
..• The set-up with ... e7-e6, ... b7-b6, ... i.b7
30 <it>f1 tDxd4 3 1 l:txc4 :xc4 32 :xc4 e5 and ... i.. e 7 can be used ag ainst any of
3 3 tDc2 tDxc2 34 1Ixc2 l:ig7 3 5 11c8+ <it>f7 White's solid system s arising after 3 c3.
36 l:th8 <it>e6 3 7 :h6+ <it>d 5 3 8 l:txh 5 In the Torre, the scheme followed by Tal
Velikov h as regained his pawn, but with ... csxd4, ... tDds and ... tDcxe7 both
the price was high. Black's kin g and equalizes and can lead to a dynamic
rook n ow have much greater activity, g ame, in contrast to numerous other
200
Th e To p a l o v Va ria t i o n
options which can be rather too i..b 7 6 e3 g 6 7 c3 iVb6 ! and 6 a 4 'iVb6 !

drawish for an ambitious Black player. (seen in the n otes t o G am e 44) and 5
In the middlegame Velikov had diffi .i.f4 .i.b7 6 a4 b4! followed by ... ltJf6
culty finding a constructive plan, m ade and ... e7-e6 (as in Gam e 45). If White
a few inaccuracies and ceded the ini opts for 5 h4 or 5 'iVd3 Black equalizes
tiative, which Tal later converted to a more easily by takin g the bishop and
superior endgame. playing ... g 7-g6.
Other continuations are less threat
Summary ening . 4 a4 is well met by 4 ... b4, as the
In this chapter we have seen th at 2 ... c5 inclusion of these moves in previous
is al so feasible against 1 d4 ltJf6 2 ltJf3 variation s (i.e. after 5 c4 g6 or 5 i.. g 5
without c2-c4. If White advances with 3 ltJe4) does n ot help White at all . De
d5, Black can play 3 ... b 5 in any case, fen ding the d4-pawn with 3 e3 or 3 c3
trying to gain space and undermine i s of course a valid option, but these
the d5-pawn . The most testing reply to moves are h ardly challenging from a
this plan is probably 4 .i.g 5, and here I theoretical perspective. Black obtained
recommend 4 ... ltJe4, as championed by good play in both the reversed Tarrasch
Topalov. of Gam e 48 and the Torre Attack of
Then the critical moves are 5 .i.h4 Gam e 49.
and 5 .i.f4, retreating the bishop. Bl ack To conclude, the Topalov Variation
should be okay if he responds with i s a very sharp and double-edged open
5 ... .i.b7, but he needs to proceed with ing for Black, but at the mom ent it
some accuracy. Key lines are 5 i..h 4 seems to be holding up pretty well.
201
Index of Variations
1 d4 tiJf6 2 c4 cS 3 d 5 b 5
Chapter 1: T h e Fianchetto Variation

4 cxb 5 a6 5 bxa6 g6
6 tiJC3
6 9 3 d6 7 b3 - 24
6 ... �xa6 7 g3 �g7 8 �g2
8 tiJf3 d6 9 � h 3 - 26
8 . . d6 9 tiJf3
.
9 tiJh 3 - 29
9 . . tiJfd7 10 0-0
.
10 'ili'c2 tiJb6 11 h4 - 1 6
1 0 � d2 1 0 �f4 - 20
,
10 ... tiJb6 11 Vc2 tiJ8d7 12 l:tdl 0-0 13 libl

1 3 e4 - 13
1 3 b3 - 1 6
13 ... tiJC4 14 b3
1 4 tiJ d2 - 1 8
14 .. :i!i'a 5 - 20
202
I n dex of Va ria t i o n s
Chapter 2: The Classical Variation

4 cxb5 a6 5 bxa6 g6 6 lDc3 .i.xa6 7 e4
7 f4 - 6 7
7 ... .i.xfl 8 'itxfl d6
9 g3
9 lDf3 .i. g 7 10 h 3 lDbd7 1 1 'itg l 0-0 12 'ith 2 Was 1 3 l:tel l:lfb8
14 l:te2 - 56
14 Wc2 - 6 0
9 lDge2 - 61
9 g4 hS 10 g s lDfd7 - 64
9 ....i.g7 10 �g2 0-0 l1 lDf3 lDbd7 12 h3
12 :el lDg4
13 We2 - 34
13 lDd2 - 40
1 3 others - 34
12 .. :ii'b6
12 ... l:ta6 1 3 l:tel Was
14 WC2 - 50
14 J:.e2 - 53
13 l:tel l:tfb8
14 l:te2 - 43
14 e S - 4 7
Chapter 3: The 'Quiet' 5 e3

4 cxb5 a6 5 e3 .i.b7 6 lDC3
6 bxa6 73 -
6 Wa 5 7 .i.d2
.••
7 bxa6 80 -
2 03
7 ... axb5 8 �xb5 'ifb6
9 'ilb3 - 73
9 �C4 - 76
Chapter 4: The 5 f3 Variation

4 cxb5 a6 5 f3 e6 6 e4 exd 5 7 e5 'iVe7 8 'iVe2 liJg8 9 liJC3 �b7 10 liJh3 c4
11 �e3
ll liJf4 - 95
11 ... axb5
12 0-0-0 86
-
12 liJxbs 92
-
Chapter 5: White plays b5-b6

4 cxb5 a6 5 b6 'iVxb6 6 liJC3 g6 7 liJf3
7 e4 d6 8 a4 � g 7 9 as 'ilb4 - 1 00
7 ... �g7 8 e4 d6 9 liJd2 0-0 10 �e2
10 a4 liJbd7 11 as 'iVd8 12 �e2 .l:tb8 1 3 0-0 - 10 � e2
1o ... liJbd7 11 0-0
2 04
I n dex of Va ria t i o n s
11 4.:JC4 'fiC7 - 1 1 0
1 1 l:tb8 1 2 a4
..•
12 ...'iVa7 - 1 03
12 ... 'iVd8 - 1 08
Chapter 6: Other Options for White

4 4.:Jf3
4 cxb5 a6 5 4.:JC3 - 1 1 4
4 'ili'c2 - 1 1 9
4 �g 5 4.:Je4 - 1 2 7
4 4.:Jd2 bXc4 5 e 4 c3 6 bxc3 g 6
7 4.:JC4 - 130
7 l:.bl - 133
4 a4 b4 - 13 7
4 ... bxC4 - 140
4 f3 - 142
4 e4 - 145
4 b3 - 145
4 g6 5 cxb5 a6
•••
6 'fic2 - 1 1 9
6 e3 - 123
6 4.:JC3 - 1 23
6 b6 - Chapter 5
6 bxa6 �xa6 - Ch apters 1 and 2
Chapter 7: The Kasparov Gambit

1 d4 4.:Jf6 2 c4 cS 3 4.:Jf3 cxd4 4 4.:Jxd4 e 5
5 4.:Jb5
205
5 tDc2 (and other moves) - 1 68

S d S 6 cxd S i.. c S 7 tDSC3
•••
7 d6 - 1 65
7 0-0
•••
8 e 3 - 149
8 h 3 - 1 54
8 g 3 - 159
Chapter 8: The Topaloy Variation
1 d4 tDf6 2 tDf3 CS
3 dS
3 e 3 - 194
3 c3 - 19 7
3 c 4 - Chapter 7
3 bS 4 i.. g s
•••
4 a4 (and other m oves) - 1 88

4 tDe4
•••
5 i..h 4 - 1 73
5 i..f4 - 1 80
5 h4 - 1 84
206
Index of Com p l ete Ci a mes
Aagaard.J-Nisipea nu.L.D, Berlin 1997 . 50

........................................................... ...............
Adianto.U-Polgar.J, Jakarta (rapid match, g ame 2) 1996 95

........................................
Aseev.K-Ponomariov.R, European Champion ship, Ohrid 2001.. 29

...........................
Bareev.E-Topalov.V, Sarajevo 2000 119

.............................................................................
Blees.A-Greenfeld.A, Tel Aviv 1 9 8 8 56

................................................................................
Brumen.D-Smirin.l, Croatian Cup, Rabac 2004 . .

........................... 1 94
..... ......................
Burmakin.V-Kasparov.S, Deizisau 2003 13

.......................................................................
Campos Moreno.J-Topalov.V, Aviles 1992 . 1 00

.............................................. .................
Comas Fabrego.L-Topalov.V, Pamplona 1994 . .

.................................... 53 .......................
Cossin.S-Vachier Lagrave.M, Ch alons (rapid) 2008 34

...................................................
Dancevski.O-Degraeve.J.M, European Ch ampionship, Ohrid 2001 140 ..................
Furman.S-Geller.E, USSR Team Champion ship 1 97 5 26

................................................
Gabriel.C-Sermek.D, Pul a 2000 60

.......................................................................................
Gallardo Garcia.A-Ortega Hermida,D, Mondariz 2007 20

.............................................
Goganov,A-Kasparov.S, Peterhof 2008 16

.........................................................................
Grivas.E-Khalifman.A, Leningrad 1989 127

......................................................................
Gurevich.D-Benjamin.J, Chicago 1986 76

..........................................................................
Hellsten.J-Baklan.V, German League 1997 1 08

...............................................................
Hochstrasser.M-Pikula.D, Pontresina 2000 . .............. 1 14

...............................................
Ivanov.Se-Khalifman.A, St Petersburg 1997 . 92

.............................................. ................
Jirka.J-Vuckovic.B, Sozina 2004 . 142

...................................................................... .............
Kachar.V-Arbakov.V, Russian Cup, Moscow 1999 . ................... 47

.................................
Karpov.A-Miles.A, Tilburg 1986 73

......................................................................................
Kozul.Z-Topalov.V, Istanbul Olympiad 2000 1 84

...........................................................
Kramnik.V-Leko.P, Dortmund 1998 . ..

.................................................. 130
................ .......
Lalic.B-Khalifman.A, Lin ares Open 1997 .

....................................... 86
..............................
Lugovoi.A-Khalifman.A, St Petersburg 1996 ............................................................. .. 40

Mancini.M-Degraeve.J.M, French Team Championship 2004 ............................... 6 7
Matlakov.M-Evdokimov.Alexa, Russian Team Championship 2010 1 59 ................
207
Mikhalchishin.A-Kasparov.G, U S S R Championship, Frunze 1981 ...................... 149

Milov.V-Topalov.V, Prague (rapid) 2002 ................................................................... 103
Mohr.R-Polgar.J, Am sterdam 1989 ................................................................................ 80
Neverov.V-Pedersen.N.V, Dieren 2010 ......................................................................... 18
Nielsen.P.H-Pedersen.N.V, Danish Team Champion ship 2006 .......................... 110
Pavlovic.S-Giorgadze.T, Lugano 198 5 ........................................................................... 24
Petkov.V-Pedersen.N.V, Skanderborg 2010 ............................................................. 133
Piket.J-Topalov.V, Monte Carlo (rapid) 1997 .............................................................. 43
Raetsky.A-Tyomkin.D, Biel 1999 .................................................................................. 154
Relange.E-Bauer.Ch, French Championship, Marseilles 2001 ............................ 1 80
Romanishin.O-Wojtaszek.R, European Championship, Dresden 2007 ............ 1 73
Sakovich.I-Lanka.Z, Riga 1980 ......................................................................................... 64
Santos Garcia.F-Alvarez Naves.L, Valladolid 1975 .................................................. 1 68
Schmidt.WI-Kasparov.G, Dubai Olympiad 1986 ..................................................... 13 7
Shchukin.D-Khalifma n.A, St Petersburg 1998 ......................................................... 145
Shulman.Y-Khalifma n.A, F I D E World Cup (playoff), Kh anty Man siysk 2005 ..... 61
Shvedchikov.A-Tseitlin.Ma, USSR Army Championship, Lvov 1977 .................. 1 65
Sorin.A-Sermek.D, World Student Team Championship, Paran ana 1993 ...... 123
Vaganian.R-Topalov.V, Istanbul Olympiad 2000 .................................................... 188
Velikov.P-Tal.M, Berlin 1986 ......................................................................................... 19 7
208

Play The Benko Gambit (PDFDrive)

Uploaded by

Copyright:

Available Formats

Play The Benko Gambit (PDFDrive)

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Play The Benko Gambit (PDFDrive)

Uploaded by

Copyright:

Available Formats

Nicolai V.

Copyright © 2011 Nicolai V. Pedersen

All rights reserved. No part of this publication m ay be reproduced, stored in a

British Libra ry Cataloguing-in-Publication Data

I S B N : 978 1 8 5 744 634 0

Everyman Chess Series

Typeset and edited by First Rank Publishing, Brighton .

Index of Variation s 202

Websites and periodicals

style. Furthermore, endgames can •.•C5-C4

One example, from Comas-Topalov

Here the white knight on h 3 can't

...tt:Jd7-e5 (or ...tt:Jg4-e5), seen m any times in the g ames and

.. :iVa5, 'iVb6, 'iVC7, or 'iVa7

The move ... 1Was is especially thematic,

A classical formation or Themes for White

M.Gurevich -Z.Azm aiparashvili, Stras­

White h as just an swered Black's

Prophylactic moves (such as l:.e2,

white king, as in Matlakov-Evdokimov ttJf6 2 ttJf3, but these cannot b e re­

1 d4 tLlf6 2 c4 cS 3 d s bs 4 cxbs a6 S strong bishop on g 7, and the other

Yearbook #94. In m y recent g am e Benko Gambit, s o much so th at I think

To reiterate wh at I 've already said, Before castling Bl ack continues to

The queen is usually placed o n c 2 i n still seems to be equalizing . For in­

think this shoul d cause Black any prob­

played in lVeng-R.5kytte, Copenhagen

a) 17 :a2 ? ! (this move is dubious, in

'iVb1 ttJxe 3 20 i.. x e3 litb8 2 1 liIe1 i.. C 4

3 4 'ii'u ?? strong opponent.

H ere the game C.Horvath -M.Pap,

In stead, E.Bacrot-P.Tregubov, Paris position via the 9 . . .ttJbd7 move order

slightly better due to the bishop pair lildcl ttJxd S 23 'iVf3

The position i s unclear. The bishop

bxa6 g6 6 tDc3 .ixa6 7 g3 d6 8 .ig2

Now White h as to pl ay rather artifi ­

chan g e sacrifice 2 2 .i.xg S ! ? i.xal 2 3

14 b3 l:tdcl l:ta3 19 .i.xe7 l:te8 20 .i.xd6 .i.xd5

Black h as regained the pawn with a

White faces a difficult defen sive

White resigned without waiting for

10 ... 0-0) 11 l':.el 0-0 transposes below. 11...�C4

Preventing ... lLlg4-eS which, as we

:a2 ..txfl lS l:txfl lLleS

A string of inaccurate moves have

Game summary White plans t o blockade the queenside,

K.Aseev-R.Ponomariov ll .i.d2 l:ta7 leads to similar play.

12 b3 "a8 13 l:tb1 l:tb8 1 4 .l:.d1!? �xe4 2 3 i.. x e4 i.. xh 3 with equal

m an ages t o free his pieces, so that Po­

28 ltJd1 ltJxd2+ 29 'iWxd2 .ixa2 30 .ig2

1 d4 t2Jf6 2 c4 cS 3 d5 b5 4 cxb5 a6 5 bxa6 This i s one of the sounder ways to

bl) 13 ... 'ii'a 5 14 .i.g 5 (attackin g the

was slightly better in F.Fleury­ a) 14 h 3 (not the most critical move)

Given the unopposed bishop on b2

compensation . The position i s similar indeed sufficient t o draw in the g ame

Cl) 15 liJb3 'ii'a 6 16 bxC3 liJge5 is

Black has no problem s here; indeed

suggested a s a n improvement for Black precision not to get in trouble. White's

a ) 18 liJb3 ? i s a tactical mistake: i.d2 l:ta4 27 'itt e 2 l::t x c4 28 'itt d 3 l:txd2+

.!:[xb2 34 'iti>f4 g5+ 35 'iti>e4 h6 36 g4 'iti>g6

M.Gurevich -Z.Azm aiparashvili, Stras

white king, as in Matlakov-Evdokimov ttJf6 2 ttJf3, but these cannot b e re

The queen is usually placed o n c 2 i n still seems to be equalizing . For in

think this shoul d cause Black any prob

Now White h as to pl ay rather artifi

m an ages t o free his pieces, so that Po

was slightly better in F.Fleury a) 14 h 3 (not the most critical move)

bl) 16 ... C4? ! was played in M.Gure

C32) 1 7 ... lLles 18 lLlxes i.xes 1 9 'iVd2 in B.5passky-1Koch, French Champion

the immediate 16 . . .ttJd7 was m ore ac

possibility, which I suppose is forgiv bxa6 g6 6 ttJc3 .txa6 7 e4 iLxf1 8 �xf1

17 l:thfl followed by 18 ttJel, ch al and . . .c5-c4, and obtained compensa

position is about equal . Black h as fairly J . Polgar, European Team Champion

With 5 f3 White begins an ambi