Karpov - Move by Move

First published in 2015 by Gloucester Publishers Limited, Northburgh House,
10 Northburgh Street, London EC1V 0AT
Copyright © 2015 Sam Collins
The right of Sam Collins

to be identified as the author of this work has been
asserted in accordance with the Copyrights, Designs and Patents Act 1988.
All rights reserved. No part of this publication may be reproduced, stored in a

retrieval system or transmitted in any form or by any means, electronic,
electrostatic, magnetic tape, photocopying, recording or otherwise,
without prior permission of the publisher.
British Library Cataloguing-in-Publication Data

A catalogue record for this book is available from the British Library.
ISBN Kindle: 978-1-78194-230-7

ISBN epub: 978-1-78194-231-4
Distributed in North America by National Book Network,

15200 NBN Way, Blue Ridge Summit, PA 17214. Ph: 717.794.3800.
Distributed in Europe by Central Books Ltd.,

99 Wallis Road, London E9 5LN. Ph 44(0)845 458 9911.
All other sales enquiries should be directed to Everyman Chess,

Northburgh House, 10 Northburgh Street, London EC1V 0AT
email: info@everymanchess.com; website: www.everymanchess.com
Everyman is the registered trade mark of Random House Inc. and is used in this work under licence from Random House Inc.
Everyman Chess Series

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

Cover design by Horatio Monteverde.
Printed by TJ International Limited, Padstow, Cornwall.
About the Author
Sam Collins is an International Master with two Grandmaster norms, and a former Irish and Japanese
Champion. He has represented Ireland at seven Olympiads, winning an individual gold medal at Bled
2002. He has a wealth of teaching and writing experience, and has produced many books, DVDs and
magazine articles on chess.
Also by the Author

The French Advance
Gambit Busters
The Greatest Ever Chess Strategies
The Tarrasch Defence: Move by Move
Contents
About the author
Bibliography
Introduction
1 Middlegame Themes
2 Key Structures
3 Openings
4 Linares 1994
5 Recent Battles
Index of Complete Games

Bibliography
Books
The Art of Chess Analysis, J.Timman (Cadogan Chess 1997)
Chess Explained: The Meran Semi-Slav, R.Vera (Gambit Publications 2007)
Declining the Queen’s Gambit, J.Cox (Everyman Chess 2011)
The English Opening, Volume Three, M.Marin (Quality Chess 2010)
How Karpov Wins (second edition), E.Mednis (Dover Publications 1994)
How Life Imitates Chess, G.Kasparov (William Heinemann 2007)
How to be a Complete Tournament Player, E.Mednis (Maxwell Macmillan Chess 1991)
Kasparov vs. Karpov: 1975-1985, G.Kasparov (Everyman Chess 2008)
Kasparov vs. Karpov: 1986-1987, G.Kasparov (Everyman Chess 2009)
The Meran & Anti-Meran Variations, A.Dreev (Chess Stars 2011)
My Best Games, A.Karpov (Edition Olms 2007)
My Great Predecessors: Part V, G.Kasparov (Everyman Chess 2006)
Opening Repertoire: The Fianchetto System, D.Lemos (Everyman Chess 2014)
The Road to Chess Improvement, A.Yermolinsky (Gambit Publications 1999)
Sicilian Attacks, Y.Yakovich (New in Chess 2010)
The Strategic Nimzo-Indian, I.Sokolov (New in Chess 2012)
Training for the Tournament Player, M.Dvoretsky & A.Yusupov (Batsford 1993)
Winning Chess Middlegames, I.Sokolov (New in Chess 2008)
Winning Pawn Structures, A.Baburin (Batsford 1998)
DVDs
Meeting the Pirc Defence - The Classical Way, D.Bojkov (Chessbase 2012)
Play the King’s Indian with g3, A.Mikhalchishin (Chessbase 2012)
Databases
Chess Informant (Sahovski Informator)
Mega Database 2015 (Chessbase)
Websites
www.bigthink.com
www.chessbase.com
www.chesscafe.com
www.chessclub.com
www.chessgames.com
www.wikipedia.com
Engines
Houdini
Stockfish
Introduction
The 12th World Champion, Anatoly Karpov, needs no introduction. However, a book dedicated to his
games certainly does. As a reader might note from the bibliography, many books have been published
on Karpov, including several by Karpov himself and, more recently, by Kasparov (most of Volume 5
of My Great Predecessors is dedicated to Karpov, in addition to the books devoted to the great world
championship matches between the two Ks). Karpov’s legacy is a rich one, with a huge number of
model games, as befits the most successful tournament player in history.
Kasparov once observed that many great players, while excellent in all areas, were particularly
outstanding in two. Thus, Kasparov categorized himself as an outstanding opening and middlegame
player. Kramnik was categorized as an outstanding opening and endgame player. And Karpov was
categorized as an outstanding middlegame and endgame player.
My approach in this book has been to select a number of aspects of Karpov’s play which could be
helpful to club players. I could easily have doubled the selection.
In the first chapter, we examine a number of techniques which Karpov used consistently, across a
number of structures and openings, namely prophylaxis, masterful handling of opposite-coloured
bishops in the middlegame, and exchange sacrifices.
In the second chapter, we consider two of the formations in which Karpov excelled, namely the
isolated queen’s pawn and the Carlsbad structure.
The third chapter is devoted to a number of openings which Karpov handled with particular
expertise. I have tried to include games from different stages of his rich career, so the reader will see
Karpov’s transition from 1 e4 to 1 d4.
The fourth chapter is devoted to one of the greatest tournament performances of all time, namely
Karpov’s dominating win at Linares 1994 ahead of all the world’s best players. As is clear from the
notes, this achievement was based on a decent slice of luck (including the worst move I’ve ever seen,
played by Evgeny Bareev in an equal position). Nevertheless, in this tournament Karpov produced a
number of classic games in diverse openings – and in some ways, of all his amazing competitive
achievements, this is the most impressive, and a perfect expression of all the elements of his chess
which made him such a great player.
The fifth and final chapter takes a look at a few of Karpov’s recent efforts. While it is clear that
professional play is no longer the centre of Karpov’s activities, he has managed to maintain a
remarkably high level and continues to impress, even when playing relatively few games each year at
a classical time control.
It’s obvious that the above topics are merely a selection. Karpov has far too many strings to his
bow for them all to be described here. Similarly, he brilliantly handled many more openings and
structures than I am able to cover. But I hope that these examples, as well as being instructive, prompt
the reader to investigate Karpov’s games more deeply. As well as being one of the all-time greatest
world champions (and the most successful tournament player of all time), he is also a master of
reinvention. He has changed from 1 e4 to 1 d4 and moved around in different openings as Black,
helping him to remain competitive (albeit primarily in blitz and rapid events) against top players for
decades.
Karpov on Karpov
I have decided to omit any biography of Karpov, partly because the essential aspects of his life story
are widely known, and partly because I have very little of note to add.
However, a description of Karpov as a player strikes me as something that should be brought to the
reader’s attention. In this regard, Karpov recently gave an interview to www.bigthink.com where he
explained a number of aspects of his style, and his approach to chess, which I found deeply
instructive:
“I had an active positional style. I played quite strongly endings, so this was my advantage also.
And then I could defend difficult positions, which is quite seldom in modern chess, and I could resist
in positions where other players probably would resign. And I was finding interesting ideas how to
defend difficult positions, and I could save many games. So I never gave up, I was stubborn as a
chess player and I tried to defend even very bad positions, and in many cases succeeded.”
In response to the question “How do you remain calm after you realize you’ve made a poor
move?”, Karpov said:
“This is a very important and good question, because many people would call back the situation,
that they missed chances, and then of course it will spoil the rest of the game. It is concerning not only
special situation during the game, but also the bad result of a previous game for the next game you
play. In my life I tried, and I succeeded in many cases, to forget everything what was in the past. Of
course you need to make some analysis and not to repeat mistakes, but it’s extremely important to
accept a situation like it is, the real situation, not with the thoughts and regrets of what you missed
and, okay, two moves ago you had winning position and now you have to defend difficult position,
and probably you might lose the game. So these thoughts shouldn’t be when you play chess game.”
A Taste of what is to Come

The following game could have been in the main body of the book – in particular, the sections on
Prophylaxis, the Sicilian and even, perhaps, the Exchange Sacrifice – but I wanted to describe it here.
When thinking about what to include in this book, this game was one of the first that came to mind. It
contains so many characteristic Karpovian themes: prophylactic thinking, simple chess, manoeuvring
and tactical alertness. It takes amazing ability to make Boris Spassky look like an ordinary player, and
this is precisely what Karpov achieves in this game.
Game 1
A.Karpov-B.Spassky
Candidates semi-final (9th matchgame), Leningrad 1974
Sicilian Defence
1 e4 c5 2 Nf3 e6 3 d4 cxd4 4 Nxd4 Nf6 5 Nc3 d6 6 Be2

Throughout his career with 1 e4 Karpov favoured this classical development against the Sicilian,
no doubt influenced by one of his long-time assistants, GM Efim Geller, who was a leading expert on
this system.
Black, by using a pure Scheveningen move order, allowed the Keres Attack with 6 g4, a system
which Karpov also used to devastating effect, as we will see later. Nursing a one point lead in the
match, Karpov chooses the safer continuation.
6 ... Be7 7 0-0 0-0 8 f4 Nc6 9 Be3 Bd7
W
One of the classical methods of development.
The more modern 9 ... a6 takes the game into the modern interpretation of the Classical
Scheveningen (often arising from a Najdorf move order), as was developed in several key encounters
in the series of matches between Karpov and Kasparov.
In the second game of the match, Spassky had played 9 ... e5 10 Nb3 a5 11 a4 Nb4 12 Bf3 Be6 13
Kh1 Qc7 14 Rf2 Rfd8 15 Rd2 Bc4 16 Nb5 Bxb5 17 axb5 a4 18 Nc1 d5, when Black was on top and
won after 63 moves. In a subsequent game the same year, Geller demonstrated the improvement 12
Kh1, beating Spassky in 30 moves.
Exercise: What is Black’s intention and how should White respond?
Answer:
10 Nb3!
Black intended to trade knights on d4 and bring his bishop to c6. While this freeing manoeuvre was
allowed in many games, here Karpov decides to avoid simplification (a normal preference for the
side with more space) and leave Spassky with more work to do in order to bring his pieces into play.
10 ... a5 11 a4 Nb4 12 Bf3 Bc6 13 Nd4
This is quite an unpleasant position for Black, from which White has scored heavily.
13 ... g6 14 Rf2 e5 15 Nxc6 bxc6 16 fxe5 dxe5
Exercise: Where should the white queen move?
Answer:
17 Qf1!
Of all Karpov’s attributes, perhaps the most difficult to imitate is his uncanny ability to co-ordinate
his pieces. There are some positions where good co-ordination can be achieved easily, or according
to a standard pattern. However, Karpov had a remarkable capacity to make apparently “strange”
moves, after which his pieces prove perfectly placed to deal with any transformation of the position.
17 Qf1 is an excellent illustration of this.
Of course White is not especially keen to exchange queens, when (as often in the Sicilian) most of
Black’s problems would be solved. But why not 17 Qe2 - ? The benefit of Karpov’s move is that he
keeps open options of Rd2 (taking the open file), Be2-c4 (attacking the weak f7-pawn and unleashing
pressure on the half-open f-file), while the queen is both well placed (supporting the rook on f2,
looking at the f6-knight and the f7-pawn) and capable of being improved (c4 is an inviting square).
17 ... Qc8 18 h3!
A typical move, taking control of the g4-square (and so preventing ... Ng4), while creating luft for
the king on h2.
In fact White could anticipate ... Ng4 with a more aggressive move, 18 Qc4, since 18 ... Ng4? fails
tactically to 19 Bxg4 Qxg4 20 Bh6. But this runs into 18 ... Qa6! instead, when Black is alright.
18 ... Nd7
Exercise: What is Black’s idea? How can White stop it? (Get used to this
question, since Karpov asks it of himself on pretty much every move.)
Answer:
19 Bg4!
After a natural move like 19 Rd1, Black’s idea is revealed: 19 ... Bc5!, which trades off one half
of White’s bishop pair and reduces the potential for a direct attack on the black king. Of course 20
Rxd7?? fails to 20 ... Bxe3.
19 ... h5
Now 19 ... Bc5 has been prevented directly in view of 20 Bxc5 or 20 Bxd7, winning a piece.
20 Bxd7 Qxd7 21 Qc4 Bh4?
An odd move, forcing the rook to a good square (from, admittedly, another good square).
Spassky avoids the ugly 21 ... Qe6 22 Qxe6 fxe6, when the compromised black structure means
that the players are playing for two results (a white win or a draw). Such a position is pure torture
against a technical master like Karpov, but this might have been a better try than risking a direct attack
in the middlegame.
Simply 21 ... Rad8 also looks more logical than Spassky’s move.
22 Rd2 Qe7
Again 22 ... Qe6 came into consideration, although this is a worse version for Black than on the
previous move since he has misplaced his bishop.
Exercise: How should White continue?
23 Rf1?!
Answer: 23 Bc5! was even stronger, since White wins after 23 ... Qg5 24 Rad1 (here 24 Rd7 Nxc2 25 Rf1 Ne3 26 Bxe3 Qxe3+ 27
Kh1 leads nowhere for White due to the surprising 27 ... Kg7!, when the king can hide on h6, which is stronger than Timman’s 27 ...
Kh8) 24 ... Rad8 25 Rxd8 Rxd8 26 Rf1 Rd7 27 Bxb4 axb4 28 Qxc6.
23 ... Rfd8
Hort suggested 23 ... Rad8, seeing that Black survives after 24 Bc5 Qb7!. But Black’s position
remains unpleasant if White’s makes the same 24th move as in the game.
Exercise: What should White play now?
Answer:
24 Nb1!!
One of the most memorable moves in chess history. Karpov prepares to re-route his knight from c3
(where it is limited by the black pawn on c6 and the white pawns on a4 and e4) to the kingside, while
protecting his rook and preparing to drive away Black’s only good piece with c2-c3. According to
Timman, this move was predicted in the press room by Semyon Furman, Karpov’s long-standing
trainer.
24 ... Qb7 25 Kh2!
More typical Karpovian play, improving his king and giving himself the option of pushing the
bishop away with g2-g3.
25 ... Kg7 26 c3 Na6
Perhaps Black should have taken the opportunity to trade a pair of rooks with 26 ... Rxd2, even
though this brings the white knight where it wants to go.
27 Re2 Rf8
Answer:
28 Nd2!
Bringing the knight to its ideal square.
28 ... Bd8
28 ... Qxb2? drops a piece to 29 Nf3.
29 Nf3 f6
Exercise: How can White further improve his position?
Answer:
30 Rd2!
Seizing the open d-file, at a time when Black can’t contest it since his bishop is awkwardly placed
on d8. It makes sense to use this rook to control the d-file since the f1-rook is usefully placed on the f-
file. All the white pieces are dramatically more active than their opponents.
30 ... Be7 31 Qe6! Rad8 32 Rxd8 Bxd8
Spassky couldn’t play 32 ... Rxd8, as 33 Nxe5! wins on the spot.
Exercise: How should White proceed now?
Answer:
33 Rd1!
Black can’t defend his last two ranks. As noted by Timman, 33 Nxe5 Qc7 34 Bf4 also wins, but it
is more complicated. Karpov’s move is risk free and utterly decisive, while requiring no calculation.
33 ... Nb8 34 Bc5 Rh8
Exercise: Find a tactical blow to end the game.
Answer:
35 Rxd8! 1-0
Black resigned in view of 35 Rxd8 Rxd8 36 Be7, when his position collapses.
I’d like to thank John Emms, Byron Jacobs and Jonathan Tait for their help with this book.
Sam Collins,
Dublin, July 2015
Chapter One
Middlegame Themes
Prophylaxis
Of course it is trite to say that Karpov is one of the world champions most associated with
prophylaxis (the other being Petrosian, who had a more defensive style). There has been much
outstanding writing on prophylaxis recently, in particular by Dvoretsky, Yusupov and Aagaard. I hope
that this section can serve as an introduction to this extraordinarily useful tool for those of you who
don’t yet use it.
While definitions vary, for me prophylaxis is simply the act of asking oneself “What is my
opponent planning? What would he play if he were to move?” While these questions seem basic, they
are in fact quite counter-intuitive. Most people, when they play chess, are concerned primarily with
their own ideas and plans. A good illustration of this (which I have certainly experienced in my own
games) is how wonderfully creative players can be in finding tactical ideas for themselves, but can
miss even basic tactical ideas for the opponent. But, as Tartakower taught us, the opponent also has
the right to exist, and his plans are not necessarily any weaker than our own.
It should be noted at the outset that prophylaxis does not imply passivity, or abandoning one’s
plans in favour of reacting exclusively to the opponent’s ideas. Rather it takes the opponent’s ideas
into account in deciding which move to play. Clearly, in an inferior position (or when facing a direct
attack), prophylaxis is essential since ignoring an opponent’s threats would be suicide. But
prophylaxis can also be used to anticipate defensive resources when we are attacking, or in intricate
strategic middlegames.
One other factor is that, just because an opponent has an idea, doesn’t mean it’s any good!
Prophylactic thinking can also be useful in “falling” into traps, where the opponent is allowed to
execute his idea, but in so doing worsens his position.
All of Karpov’s best games contain a healthy dose of prophylaxis and so, to an extent, all the
games in this book are examples of prophylaxis. However, the followng two games are masterful
illustrations of the concept.
Game 2
A.Karpov-J.Timman
Montreal 1979
Pirc Defence
This game is one of the best examples of prophylactic play I’ve ever seen. Kasparov praised it as “a
eulogy to prophylaxis!”
1 e4 d6 2 d4 Nf6 3 Nc3 g6 4 g3
Opting for a fianchetto construction. As we will see subsequently, Karpov was very successful
with similar set-ups against the King’s Indian and Grünfeld Defences.
Karpov’s normal recipe against the Pirc was the Classical 4 Nf3, followed by Be2 and 0-0, with a
sound position and a small edge, which he used to create serious problems for his opponents. We
cover this in a separate section.
4 ... Bg7 5 Bg2 0-0 6 Nge2 e5 7 0-0 Na6 8 Re1 c6 9 h3
Karpov notes: “A typical prophylactic move in such situations. White restricts the opponent’s
bishop, and at the same time also creates a ‘no-go area’ on the kingside for the remaining minor
pieces.”
9 ... Re8
Here Karpov writes: “All Black’s hopes of obtaining counterplay are associated with pressure on
the e4-pawn.” As Dvoretsky observes, “ ... this is already prophylactic thinking. Karpov immediately
determines the opponent’s main idea, for which he will carefully watch throughout the entire game.”
10 Bg5 h6?!
This is the move Karpov’s last had provoked, but now Qd2 will come with tempo. Analysis by
Dvoretsky and Kasparov suggests that Black has better chances to hold the balance after the more
dynamic 10 ... Qb6 11 Rb1 exd4 12 Nxd4 d5 or 12 ... Ng4!?.
11 Be3 Qc7
Karpov suggests 11 ... Kh7, “since all the same this move will have to be made sooner or later.”
Then after 12 Qd2, Black has a wider choice of responses. This is a small example of how Karpov
used prophylactic thinking to anticipate attacks.
12 Qd2 Kh7 13 Rad1 Bd7
Answer:
14 g4!
A very typical idea in fianchetto structures. This can be seen in variations like 1 e4 d6 2 d4 Nf6 3
Nc3 e5 4 Nf3 Nbd7 5 g4!? h6 (of course taking the pawn is critical) 6 h3, when White aims to get an
“accelerated fianchetto” following Bg2 and 0-0.
Karpov writes of 14 g4 as follows: “After some thought, I came to the conclusion that
straightforward play in the centre would not get me anywhere. Now, with the aim of seizing fresh
territory, the kingside pawns must be advanced. But I did not wish to play 14 f4 immediately. It would
be illogical to increase the tension straight away – all the same White will subsequently have to play
g3-g4, so why not first utilize a resource for strengthening the position, such as g3-g4 and Ng3. At the
same time White also solves an important strategic problem – he reinforces his e4-pawn.”
14 ... Rad8 15 Ng3 Bc8 16 f4
This advance has been prepared in such a way that White’s centre can’t be attacked.
16 ... b5
Answer: As Dvoretsky notes: “Knowing Karpov’s style, it is very easy to guess his next move.”
17 a3!
Restraining ... b6-b4, which Timman goes for anyway.
17 ... b4?!
This basically weakens Black’s structure for no reason, but it is already hard to give him good
advice. If Karpov is right that all Black’s hopes of counterplay are attached to pressurizing the e4-
pawn, we can see that Black has lost the strategic battle, since the e4-pawn is directly defended by
three pieces and indirectly defended by the rook.
18 axb4 Nxb4
As Kasparov observes, “the only result of the desperate pawn advance is that now it is
advantageous for White to go into practically any endgame.”
19 Nce2
Karpov observes: “Black’s idea was somehow to bring his pieces together, by playing ... a7-a5, ...
Ba6, ... exd4 and ... c6-c5. But this is a lengthy process, and White succeeds in hindering his
opponent’s plan.”
19 ... exd4 20 Nxd4 a5 21 c3 Na6
Answer: There are at least two good moves here!

22 Qc2
Dvoretsky notes: “Probably the best move of the game, which by and large is an excellent
illustration of Nimzowitsch’s idea about the overprotection of strategically important points. Here
two forms of prophylaxis mentioned by Nimzowitsch are simultaneously combined – prevention of
the opponent’s plans and overprotection.” The main point is that 22 ... Nc5 can be strongly met by 23
b4!.
Without wishing to detract from a definite strategic masterpiece by Karpov, it should be noted that
22 Nxc6! might have been even stronger. White has a strategic advantage with his extra central space,
but his pieces are also better developed than Black’s (the knight on a6 is particularly badly placed at
the moment), so he benefits from an opening of the position. After 22 ... Qxc6 23 e5 d5 (there is
nothing better) 24 exf6 Qxf6 25 Qf2, White has a clear advantage, with superior pieces (the a6-knight
remains out of play) and a healthier pawn structure.
22 ... Bd7
22 ... Bb7 looks like a better spot for the bishop.
23 Nf3
Continuing the slow, methodical strategic build-up. Houdini likes 23 g5, claiming an almost
winning advantage after 23 ... hxg5 24 fxg5 Ng8 25 Ngf5. For human players, this attack is both far
from clear and very committal, weakening the white structure and giving important to squares to
Black, in particular the e5-outpost. I prefer Karpov’s move which keeps everything protected and
leaves Black worrying about possibilities of e4-e5 or g4-g5.
23 ... Re7
If 23 ... Nc5 24 e5 Nd5 25 Bxc5 dxc5, Kasparov assesses 26 Qe4 as better for White.
24 Bf2
Karpov notes: “One of the last prophylactic moves. Prior to his decisive offensive, White arranges
his pieces in the most harmonious way possible and ... once again reinforces his central e4-forepost!
24 Qd3 was premature in view of 24 ... Bc8.”
24 ... Be8
This drops a pawn, but it is hard to give Black any good advice at this point.
25 Qd3 Qb7 26 Ra1
Winning a pawn and retaining all his positional advantages. The b2-pawn is immune for the next
couple of moves because of Reb1.
26 ... Nc7 27 Rxa5 Rdd7 28 b4 Ne6 29 Be3!
Karpov didn’t want to allow 29 Qd2 d5 with “something resembling counterplay”.
29 ... c5 30 f5 Nd8 31 b5
Black is completely lost.
31 ... Kh8 32 Bf2 Qc7 33 Ra4 Qb8 34 c4 Ra7 35 Rxa7 Rxa7 36 e5!
Winning more material.

36 ... dxe5 37 Nxe5 Ra2 38 Bxc5 1-0
Game 3
A.Karpov-A.Yusupov
USSR Championship, Moscow 1983
Ruy Lopez
This game made a strong impression on Yusupov (who cites it as one of the rare cases when he felt was completely outplayed), and it
remains an absolute classic. Karpov’s subtle moves with his major pieces on the queenside, each configuration preventing Black’s idea in
a slightly different way, is an example of chess strategy of the very highest class.
1 e4 e5 2 Nf3 Nc6 3 Bb5 a6 4 Ba4 Nf6 5 0-0 Nxe4 6 d4 b5 7 Bb3 d5 8 dxe5 Be6 9 c3 Bc5 10
Nbd2 0-0 11 Bc2
This variation has become less popular for White, losing ground to 9 Nbd2 in particular. One
reason is the range of options Black has on the 11th move – in addition to the line chosen by Yusupov,
11 ... f5 or even 11 ... Nxf2!? 12 Rxf2 f6 (the Dilworth Attack, also played by Yusupov with success)
are quite playable.
11 ... Bf5 12 Nb3 Bg6 13 Nfd4 Bxd4 14 cxd4 a5 15 Be3 a4
Karpov had previously faced 15 ... Nb4 16 Bb1 a4 17 Nd2 a3.
Exercise: How should White deal with the advance of the a-pawn?
Answer: 18 Qc1! (clearly the strongest move; 18 bxa3 Rxa3 19 Qc1 Qa8 gives Black active queenside counterplay, while the more
natural 18 Qb3 allows 18 ... Nc6 with a light-square blockade after, for instance, 19 Bxe4 dxe4 20 Qc3 Ne7 21 bxa3 Nd5 22 Qb2 c6) 18
... Ra6? (18 ... Nxd2 19 Qxd2 Nc6 was better, though White retains an enduring strategic plus since Black cannot achieve ... c7-c5) 19
bxa3 leads to a clear advantage for White. Savon’s attempt to shoot his way out of trouble fails tactically: 19 ... Rc6? 20 Qb2 Nc2 21
Rc1 Nxe3 22 Rxc6 Nxf2 23 Nf1! Qd7 24 Nxe3 and Black resigned in A.Karpov-V.Savon, Moscow 1971.
16 Nd2
In the sixth game of his Merano match with Korchnoi in 1981, Karpov had tried the alternative 16
Nc1, but was convincingly outplayed, losing in 41 moves.
16 ... a3 17 Nxe4 axb2 18 Rb1 Bxe4!
Karpov’s notes to the Savon game had only considered 18 ... dxe4.
19 Rxb2
Yusupov analysed 19 Bxe4 dxe4 20 Qg4 Nxd4 21 Rfd1 c5 22 Bh6 g6 23 Bxf8 Kxf8 with a strong
position for Black. He suggested that White has nothing better than to play for perpetual check with 24
Qxe4 (not 24 Rxb2? Nf3+!) 24 ... Rxa2 25 e6 fxe6 26 Qe5, though it’s unclear whether he will
achieve it after 26 ... Qd5, when Black still seems to have an edge.
19 ... Qd7 20 Bd3
Not 20 Rxb5? Bxc2 21 Qxc2 Nxd4 etc.
20 ... Bxd3 21 Qxd3 Rfb8
Answer:
22 Rfb1!
Yusupov notes that the most active plan for White begins 22 f4!?, aiming for a kingside attack. He
observes as follows: “It is interesting that Karpov approaches the question quite differently. He is
primarily a prophylactic chessplayer. Most probably he immediately asked himself: ‘What does my
opponent wish to do, what is his plan?’ And indeed, what do you think Black wishes to play now? Of
course – ... Na5-c4! But now watch how with each move Karpov prevents this knight manoeuvre.”
22 ... b4
22 ... Na5? drops a pawn to 23 Rxb5.
23 h3
Exercise: How should Black continue?
23 ... h6
This move doesn’t spoil anything, as the later course of the game will demonstrate.
Answer: However, in his analysis Yusupov found 23 ... Rb6!, intending ... Na5-c4. After 24 Qc2 (24 Rc1 Na5 25 Qb1 is harmless in
view of 25 ... Rab8 and Black is fine) 24 ... Rab8 25 Rc1 R8b7 26 Qc5, a draw was agreed in Pe.Popovic-A.Yusupov, Sarajevo 1984.
Yusupov’s other suggestion, 23 ... Ra4!?, allows White to retain an edge after 24 a3 h6 25 axb4
Rbxb4 26 Rxb4 Rxb4 27 Rc1, in view of the Black’s weaknesses along the c-file.
24 Rc1 Rb6
Exercise: How can White respond to the immediate 24 ... Na5 - ?

Answer: 25 Qb1! forces the knight back, since 25 ... Nc4 drops a pawn to 26 Rxb4 in view of the tactic 26 ... Rxb4 27 Qxb4 Rxa2??
28 Qb8+ Kh7 29 Qb1+, winning the rook on a2.
25 Qb1
In a game a couple of years later in Reykjavik, Arnason agreed a draw with Yusupov in this
position.
Exercise: How should Black continue?
25 ... Rab8
After 25 ... Ra7 26 Rc5 Na5?! 27 Rxb4 Nc4 28 Rb3, Black had insufficient compensation in
Pe.Popovic-J.Timman, Sarajevo 1984 (1-0 in 44).
Answer: Yusupov gives 25 ... Ra4! as Black’s best; for instance, 26 Qc2 (or similarly 26 Rbc2 Ra7, threatening ... Na5, when White
appears to have nothing better than repeating moves; whereas 26 Rc5?! Na5 gives Black the edge) 26 ... Ra7 27 Qc5 Ra5 28 Qc2 Ra7
and “White either has to consent to a draw or allow the knight to come to c4.”
26 Rc5
Again preventing ... Na5.
26 ... Nd8 27 Rcc2 Nc6 28 Qc1
Enabling White to triple on the c-file, if required. Now ... Na5 is impossible since the c7-pawn
would hang.
28 ... R8b7 29 Rc5
Yusupov’s comment here is a great advertisement for the benefits of prophylactic play: “Again
there is no opportunity to play ... Na5. I had already become totally confused. When you are not
permitted to carry out your main idea it is very difficult to carry on the fight.”
29 ... Ne7 30 Kh2 Nf5?
Yusupov: “Simply a blunder. It was necessary to stick to defensive tactics with 30 ... c6, retaining
a slightly inferior but solid position. By waiting for some active move from White, say g2-g4, it
would have been possible to try to create counterplay with ... b4-b3.”
Exercise: How can White exploit Black’s error?
Answer:
31 Rbc2
Now c7 collapses.
31 ... Rg6 32 Rxc7 Rxc7 33 Rxc7 Qb5 34 g4 Nh4 35 Rc8+ Kh7 36 Qd1 Qa6 37 Rc2 f5
Exercise: White to play and win!
Answer:
38 Kg3!
Winning the knight on h4. While Karpov’s style is careful, as you can see he’s not afraid to
calculate variations at the decisive moment.
38 ... fxg4 39 Kxh4 gxh3
Yusupov’s idea had been 39 ... g3 40 fxg3 Qa3, intending ... b4-b3 and ... Qe7+, but he realized
that 41 Qc1 refutes the idea (as do several other moves).
40 f4!
As Yusupov notes, 40 Kxh3 Qe6+ 41 Kh2 Qf5 allows Black to create threats (though White is still
winning with careful play). Karpov finishes the game more accurately.
40 ... Qe6 41 Qh5 Qe7+ 42 Kxh3 Qf7
Exercise: What is Black’s threat, and how can it be stopped?
Answer: Black threatened 43 ... Rg3+ 44 Kh4 Rh3+ etc. Karpov plays the only winning move, which he must have foreseen several
moves previously.
43 Rh2!
43 Kh4 doesn’t spoil anything, provided that after 43 ... Qe7+ 44 Kh3 Qf7, White finds 45 Rh2!.
43 ... Qd7+ 44 f5 1-0
Opposite-Coloured Bishops in the Middlegame

One of Karpov’s calling cards is his fantastic handling of opposite-coloured bishops. His prowess
at playing with such bishops in the endgame has been well documented (for instance, Mihail Marin
has a wonderful chapter on it).
However, Karpov’s ability to handle them in the middlegame is also worth noting. As a master of
the endgame, Karpov knew which pieces to exchange (and when) to obtain an advantageous ending,
or bring a difficult position to safety. But he also could exploit the bishops without trading down.
Seizing the initiative is crucial in middlegames with opposite-coloured bishops, since the opponent
will be unable to compete for squares which are controlled by the enemy bishop. Karpov was fully
aware of this, sought the initiative with great creativity, and was precise in fully exploiting any
initiative he obtained.
Game 4
A.Karpov-G.Kasparov
World Championship (4th matchgame), Moscow 1985
Queen’s Gambit Declined
1 d4 d5 2 c4 e6 3 Nc3 Be7 4 Nf3 Nf6 5 Bg5 h6 6 Bxf6 Bxf6 7 e3 0-0 8 Qc2
An interesting sideline which Kasparov used to defeat Timman in the 1984 USSR vs. Rest of the
World match. Karpov subsequently took it up against Kasparov (!) – in the 27th game of their 1984/85
match, and again in this one.
8 ... Na6!?
Kasparov gives the following fascinating account of the genesis of this novelty: “Instead of the
usual 8 ... c5, Black provokes complications with this unexpected knight leap, which I devised
literally just before the match. ( ... ) The move looks ridiculous, but in fact it has a perfectly logical
basis: Black wants to play ... c7-c5 and after d4xc5 not to lose time on regaining the pawn, as in the
main variations. The artificial position of the knight on a6 cannot be exploited – 9 cxd5 Nb4 and 10 ...
Nxd5. ( ... ) In our preparation we considered the variations with 9 c5 b6 10 c6 to be the most
dangerous for Black – we spent considerable time analysing them, but until now they have not
occurred in practice. Of course, Karpov did not take such a risk.”
8 ... c5 9 dxc5 is a more established continuation:
a) 9 ... Qa5?! 10 cxd5 exd5 11 0-0-0 Be6?! (Kasparov suggests 11 ... Bxc3 12 Qxc3 Qxc3+ 13
bxc3 Be6 14 Nd4, “although this too is in White’s favour”) 12 Nxd5 Rc8 13 Kb1! Bxd5 (13 ... Rxc5
14 b4!) 14 Rxd5 Nc6 15 Bc4 Nb4 16 Qd2 was G.Kasparov-J.Timman, London 1984, where “White
gained a very serious advantage” (1-0 in 34).
b) 9 ... dxc4 (this “relieving, essentially drawing variation” was the reason Kasparov gave up the
8 Qc2 system with White) 10 Bxc4 Qa5 11 0-0 Bxc3 12 Qxc3 Qxc3 13 bxc3 Nd7 14 c6 bxc6 15
Rab1 Nb6 16 Be2 c5 17 Rfc1 was A.Karpov-G.Kasparov, World Championship (27th matchgame),
Moscow 1984. This endgame is thoroughly playable, if slightly worse, for Black and strong GMs
have confidently continued to defend it, but Karpov’s masterful performance was deeply unpleasant
for Kasparov (and 1-0 in 59), so it was natural to seek an improvement.
9 Rd1 c5!
This pawn sacrifice is one of the main ideas behind 8 ... Na6.
10 dxc5 Qa5 11 cxd5 Nxc5
Answer:
12 Qd2!?
A deeply practical move, very much in Karpov’s style.
As noted by Kasparov, both 12 Be2 Na4! and 12 Bc4 b5! give Black strong counterplay; but it’s
not immediately apparent whether Black has sufficient compensation after 12 dxe6!? Bxe6 13 Nd4.
Kasparov gives the following variations: 13 ... Rac8 (or 13 ... Bd5 14 Rd2!) 14 Be2 Bd5 (not 14 ...
Bxa2? 15 Ra1) 15 0-0 Ne4 16 Rd3 Nxc3 17 Rxc3 Rxc3 18 bxc3 b6 19 Rd1 “and although, because
of the weakness of his queenside, White has few hopes of success, Black is not guaranteed a draw – a
pawn is a pawn!”
12 ... Rd8 13 Nd4 exd5 14 Be2
Black has almost equalized but still needs to find precise moves in order to justify his isolated d5-
pawn.
14 ... Qb6 15 0-0 Ne4 16 Qc2 Nxc3 17 Qxc3 Be6 18 Qc2 Rac8 19 Qb1 Rc7 20 Rd2
Exercise: What is the cleanest way for Black to equalize?
20 ... Rdc8
Answer: Kasparov recommends the surprising 20 ... Bxd4! 21 Rxd4 Rdc8 22 Bd3 Rc5, followed by ... Qc7 and ... Rc1 “with a quick
draw”.
Answer:
21 Nxe6!
Changing the character of the game. Black now has an unpleasant choice between 21 ... Qxe6,
when the d5-pawn would be horribly weak, and what occurs in the game. Note that 20 ... Qxe6 would
have been the logical response to 20 Nxe6, since in that case Black’s rook is well placed on d8,
defending the pawn and preparing ... d5-d4.
21 ... fxe6
Exercise: What is your assessment of this position?
Answer: White has a small but very long lasting advantage (one of the characteristics of opposite-coloured bishop positions is often
that advantages last, since the bishops can’t be exchanged). His structure is superior, since Black has three pawn islands and White only
has two. The e6-pawn is a target and, if it moves to ... e5, the d5-pawn will become weak. The structure also gives White a natural
pawn break (e3-e4!), while Black’s only pawn break ( ... e6-e5 and ... d5-d4) is both difficult to achieve and would severely weaken the
a2-g8 diagonal.
Linked to his structural problem, Black will have difficulties with king safety for the rest of the
game. His king only has two pawns for defence, and these are both placed on dark squares, leaving
his light squares horribly weak.
Kasparov’s assessment is again deeply instructive: “White obtains a safe position with
possibilities (albeit only slight) of improving it, whereas Black, with no active counterplay, is forced
merely to passively keep an eye on the opponent’s actions. The defence of the d5-pawn has been
temporarily achieved, but this cannot be called a particular achievement: the weakness of the e6-
pawn and the resulting weakening of a complex of light squares on the kingside give White a slight
but persistent positional advantage. This factor cannot be immediately exploited: 22 Qg6? Qa5!,
winning a pawn. The possible future invasion of the white queen on h7 will also not in itself solve
anything, since after moving to e7 the black king will be safe enough. What is required of White is
systematic play, the essence of which can be described as follows: the consolidation of his position
on the queenside, the switching of his queen to the kingside, the opening of the position by e3-e4, and
only then the mounting of an attack on the light squares, making use of the now open e-file. In the game
Karpov skilfully put all these ideas into practice, but, of course, not without substantial ‘help’ on my
part.”
Kasparov notes that, even after the removal of the rook from the d-file, 21 ... Qxe6! was better,
when Black has good chances of achieving ... d5-d4.
22 Bg4 Rc4 23 h3!
Creating luft for the king and securing the bishop on an excellent square.
23 ... Qc6 24 Qd3 Kh8
This move was criticized by Kasparov, since the king is more vulnerable to ideas of Qh7 when in
the corner.
25 Rfd1 a5 26 b3 Rc3 27 Qe2 Rf8 28 Bh5
28 ... b5 29 Bg6 Bd8 30 Bd3 b4 31 Qg4 Qe8 32 e4 Bg5 33 Rc2 Rxc2?

Strongly criticized by Kasparov – his rook was active on c3 and, had White exchanged it, Black
would have obtained a passed pawn.
34 Bxc2
Now Black has no counterplay.
34 ... Qc6 35 Qe2 Qc5 36 Rf1 Qc3 37 exd5 exd5
Answer:
38 Bb1!
Wonderful technique, avoiding the threatened exchange of queens by 28 ... Qd2 (which would have
significantly eased Black’s defence), while putting the bishop on a protected square (thus, this move
is more accurate than 38 Bg6, which also retains some advantage). As a bonus, White’s only
queenside weakness is now covered (even if going after a2 was not a very realistic possibility).
38 ... Qd2 39 Qe5 Rd8
Granting White access to the f5-square doesn’t help.
40 Qf5 Kg8 41 Qe6+
Karpov sealed this move and the game was resumed later.
41 ... Kh8 42 Qg6 Kg8 43 Qe6+ Kh8 44 Bf5!
Bringing the bishop to a more active position, from which it can land on e6.
44 ... Qc3 45 Qg6 Kg8 46 Be6+ Kh8 47 Bf5 Kg8
Question: What’s the point of all these repetitions (moves 41-43 and 45-47)?
Answer: Repeating moves in a superior position is one of the hallmarks of good technique and is used by virtually all leading players.
The benefits of repeating have been variously described as follows:
1. Gaining clock time (getting closer to a time control, gaining an increment, or both).
2. Emphasizing who’s in charge! It is very hard, when faced with a repetition, not to hope that an
opponent will repeat three times, and so this method can be quite discouraging to face.
3. Giving the opponent the chance to go wrong (by avoiding the repetition himself).
Exercise: How can White improve his position?
Answer:
48 g3!
An excellent plan, both improving the king (Kg2) and giving White the option of dislodging the g5-
bishop with h3-h4.
48 ... Kf8 49 Kg2 Qf6 50 Qh7
Of course White avoids the exchange of queens.
50 ... Qf7 51 h4 Bd2
Black was faced with a tough decision concerning where to put his bishop. Kasparov suggests that
51 ... Bf6 was more tenacious but notes that, even there, White can continue to build up his attack.
Exercise: How can White bring his rook into the game?
Answer:
52 Rd1!
Preparing a rook lift, via d3, to f3. Note how strong White’s attack is, despite the simplified
position, because of the opposite-coloured bishops. Alternatively, 52 f4, aiming to push the f-pawn to
open up the black king, is also liked by the engines but is far more committal and exposes the white
king. I much prefer Karpov’s move.
52 ... Bc3 53 Rd3 Rd6?
Kasparov notes that 53 ... Bf6 was “the only possibility of prolonging resistance”, though after 54 Rf3, White retains a very strong
attack.
54 Rf3 Ke7
If Black sacrifices his queen with 54 ... Rf6 55 Bg6 Rxf3 56 Bxf7 Rxf7, his chances of building a
fortress are slim; while 54 ... Bf6 is met by 55 g4!, preparing 56 g5, winning the bishop.
55 Qh8 d4 56 Qc8 Rf6 57 Qc5+ Ke8 58 Rf4 Qb7+ 59 Re4+
59 ... Kf7
Kasparov notes that 59 ... Re6 fails to 60 Qc4!, since 60 ... Rxe4 61 Qg8+ Ke7 52 Qxg7+ is
curtains.
60 Qc4+ Kf8 61 Bh7 Rf7 62 Qe6 Qd7 63 Qe5 1-0
It’s mate in seven, apparently!
Game 5
A.Karpov-A.Khalifman
Linares 1995
King’s Indian Defence
1 d4 Nf6 2 c4 g6 3 Nf3 Bg7 4 g3

An example of Karpov’s prowess in fianchetto systems against the King’s Indian/Grünfeld, which
could have been included in that section (see Chapter Three) but is placed here due to the wonderful
play with opposite-coloured bishops.
4 ... c5
Khalifman invites Karpov into a Benoni structure after 5 d5, which Karpov declines.
5 Bg2 Qa5+ 6 Nc3 Ne4
7 Bd2
7 Qd3 is the main move, and perhaps more critical. Anyone interested in this line can check a
miniature rout of a King’s Indian expert in M.Vachier Lagrave-M.Hebden, Douglas 2014 (1-0 in 21).
7 ... Nxd2 8 Qxd2 0-0
After 8 ... cxd4 9 Nxd4 Nc6 10 e3 0-0 11 0-0, Black can’t play ... d7-d6 without losing a pawn.
Accordingly he followed with 11 ... Rb8 in A.Karpov-J.Timman, Amsterdam 1988, and had to
continue with some more unusual moves to get his pieces out: 12 Rfd1 e6 13 a3 a6 14 Rac1 Ne5 15
Ne4 Qxd2 16 Rxd2 b5 17 cxb5 axb5 18 b3, when White’s control of the c-file promised him an
enduring plus (1-0 in 73).
Instead, 8 ... d6 9 0-0 Nc6 10 e3 Bg4 11 d5 Bxf3 12 Bxf3 Ne5 13 Be2 Qb4 14 Ne4 Qxd2 15 Nxd2
was A.Karpov-J.Polgar, Budapest (rapid; 7th matchgame) 1998, and it is hard to believe that White
had any advantage (½-½ in 71).
9 e3 d6 10 0-0 Nc6 11 h3 cxd4 12 Nxd4 Bd7 13 Rfd1
As Ftacnik correctly notes, White’s hopes for advantage are connected with the pawn on e3, which
enables him to retain the central tension. If this pawn were on e4 White’s dark squares would be
horribly weak.
13 ... Nxd4
Black goes for simplification but improves White’s structure and central control. In particular, the
g2-bishop is now strengthened and the e7-pawn is potentially weak.
14 exd4 Bc6 15 Nd5 Qd8
Not 15 ... Qxd2?? 16 Nxe7+ Kh8 17 Rxd2 with a clear extra pawn.
16 Rac1 Rc8 17 h4 Qd7 18 b3 Rfd8 19 a4
19 ... Bxd5
Going for opposite-coloured bishops. Otherwise, he might have preferred 19 ... e6 20 Ne3 Bxg2
21 Kxg2 h5 and Black is extremely solid.
20 Bxd5 b6 21 a5 e6 22 Bf3 b5 23 d5 bxc4
Karpov assessed 23 ... e5 24 cxb5 Qxb5 25 b4 as clearly better for White. For the moment the
queenside pawns are blockaded by the queen, but this can be disturbed by Be2. Also, Rc6 is an
important resource, when taking would give White a dangerous passed pawn on c6.
Answer:
24 dxe6!
An essential intermezzo.
Compared with the last variation, after 24 bxc4 e5 if anyone is in danger, it’s White. His control of
the open c-file has been transformed into a weak, backward pawn on c4, while Black can expand on
the kingside (if he wants) at his leisure.
24 ... Qxe6
Ftacnik gives 24 ... fxe6 25 bxc4 as good for White too, but I much prefer this option to the game
continuation since Black at least controls the d5-square.
25 bxc4
This position is a torture to play with Black. The white bishop will go to d5 and exert unopposed
dominance on both sides of the board. White has a queenside advantage (secured by his far advanced
a-pawn, which means that the fall of the a7-pawn would be the end of the game). On the kingside,
White also has something of a free hand, since the bishop on d5 will target f7. Black will probably
need ... h7-h5 to prevent White playing h4-h5 himself, but then his pawns will become strategically
and tactically weak (for instance, White will have the g3-g4 break in some positions; and with the
black king on g8, the f7-pawn is pinned so the g6-pawn is a weakness).
25 ... Qf5 26 Kg2 Bf6 27 a6 h5 28 Re1 Rc5 29 Bd5 Rb8 30 Rc2 Kg7 31 Be4 Qd7 32 Bb7 Qc7
33 Qd3 Rd8 34 Bd5 Qa5 35 Rb1
35 ... Qxa6
Here 35 ... Rc7 would be more tenacious; for instance, 36 Rb7 Rdd7 37 Qe2! Qxa6 38 Ra2
(trapping the queen) 38 ... Qxa2 39 Qxa2 Rxb7 40 Bxb7 Rxb7 and Black can try to establish a
fortress.
36 Rb7 Rf8
Now the black rook is entombed on f8.
37 Re2 Ra5
Karpov suggested 37 ... Qa5 in Informant 63, but with modern engines it can quickly be seen that
after his line 38 Re6 Rc7 39 Rxf6 Kxf6 40 Qd4+ Ke7, White wins with 41 Rb2, since the black king
can’t survive the attack.
38 Re3 Ra1 39 Rf3 Re1
The computer suggests the rather despondent 39 ... Qxb7.
40 Rxf6 Kxf6 41 Qf3+ 1-0
My thanks to John Emms for drawing my attention to the next game.

As already mentioned, I don’t want to discuss Karpov’s amazing technique with opposite-coloured
bishops in the endgame (which has already been excellently covered by a number of authors), but this
game just strikes me as magical so I had to include it. For Karpov to absolutely destroy a strong GM
with Black from a completely equal position is stunning. The slow accumulation of tiny advantages,
obtaining some concession from every exchange, and the seemingly inevitable demise of White’s
position reminds me of the final game of the Capablanca-Marshall match in New York 1909. There,
Black also triumphed in a level rook and bishop endgame with asymmetrical pawns, though in that
game Black had a queenside majority and the bishops were both light-squared bishops.
Game 6
B.Alterman-A.Karpov
European Cup, Tyniste 1995
Queen’s Indian Defence
1 d4 Nf6 2 c4 e6 3 Nf3 b6 4 g3 Ba6 5 b3 b5
An interesting sideline which has been played by several strong players (including English GMs
Michael Adams and Gawain Jones). This was a favourite of Karpov’s, especially in rapid games,
though it’s probably not a wonderful opening compared with the main lines.
6 cxb5 Bxb5 7 Bg2 Bc6 8 0-0 Be7 9 Bg5 0-0 10 Bxf6 Bxf6 11 Nc3 d5 12 Qc2 a5 13 e4 dxe4 14
Nxe4 Bxe4 15 Qxe4 Ra6!
Bringing the rook to d6 is the only way to make sense of Black’s position, even if White retains an
edge.
16 Rac1 Rd6 17 Rc4 c6!
Starting an elegant manoeuvre to co-ordinate all his pieces.
18 Ne5 Qb6 19 Rfc1 Rfd8
Karpov has solidly equalized and one would expect the game to end in a draw in a few moves. But
this isn’t what happens.
20 Nxc6 Nxc6 21 Rxc6 Qxd4 22 Qxd4
22 Rxd6 is also dead equal. However, I don’t want to comment on White’s moves in great detail,
for two reasons:
First, the position remains very close to equal for a long time, and simply suggesting alternatives
which maintain the character of the position doesn’t achieve much.
Secondly, and more importantly, analysing such games with an engine creates a false sense that
such positions are trivial to defend. As demonstrated by Karpov and, in the modern era, emphatically
proven by Carlsen, the mere fact that a position (even a heavily simplified position with strong
drawing tendencies) is equal, does not mean that problems cannot be created by a determined
opponent.
22 ... Rxd4 23 R1c2 Rd1+ 24 Bf1 g6 25 Kg2 Be7 26 Rc7 Bb4 27 Rc8 Rxc8 28 Rxc8+ Kg7 29
Rc2
Question: What possible reason does Black have to play on in this position?
Answer: Firstly, Black isn’t forced to offer or accept a draw here, or at any time, and doesn’t need any special justification for
continuing.
Secondly, looking more closely discloses an important difference between the two sides’
positions. Black has a pawn majority on the kingside and White has a pawn majority on the
queenside, but there is a major qualitative difference between these majorities. Black can slowly
advance his kingside, supported by all of his pieces. In the process, his threats include not only
creating a passed pawn, but also (and perhaps more importantly) creating threats to the white king. A
king, rook, bishop, and one or two attacking pawns is more than enough material to create serious
threats to an enemy king, especially when White can’t compete for control of the dark squares.
What about White’s majority? For the moment, it can’t move. It isn’t even completely clear how
White could give up his a- and b-pawns for the a5-pawn. If White somehow manages to get an outside
passed pawn, it will be easily blockaded and, while it might provide sufficient counterplay, is highly
unlikely to give White serious winning chances.
There is an additional practical difficulty to White’s defence here, namely that it is extremely
passive. The bishop on f1 struggles to find an active role, while if the rook leaves the second rank,
Black might be able to create an attack against the f2-pawn.
So all in all, White’s task is surprisingly unpleasant.
29 ... Kf6 30 Be2 Rd7!
Carefully defending the only potential target in Black’s camp, the f7-pawn.
31 h4 e5 32 h5
Although it looks very natural to trade pawns, in fact this helps Black mobilize his kingside
majority. After 32 Kf3, it is hard to suggest a way to progress for Black.
32 ... Kg5! 33 hxg6 hxg6 34 Kf1 Rd6!
Outstanding technique. After ... f7-f5, the only potential target in Black’s camp is the g6-pawn, so
this is covered in advance. In the meantime, the black king gains access to the f6-square without fear
of harassment by the white rook.
35 Kg2 f5 36 Kf1
Alterman decides not to play f2-f3. I find it hard to evaluate this decision – had he played f2-f3,
I’m sure Karpov would have found a way to continue generating problems.
36 ... Kf6 37 Kg2 e4 38 Kf1?
Maybe miscalculating something, or assuming the position was an easy draw regardless of what
White does.
Now that Black has pushed ... e5-e4, the time was ripe for 38 f3!, which plays to exchange a pawn
and fights for some essential kingside space. The difference from the game, as we’ll see, is that 38 ...
e3 is not a problem, since the white king will be very comfortable on f3 after 39 f4.
38 ... Ke5
38 ... g5 was perhaps a more precise move order, since after 39 f3 e3, playing f3-f4 would create
a weakness for White.
39 Kg2
Here 39 f3 might again be considered since 39 ... e3 is equal after 40 f4+, but Karpov could
consider other moves, such as 39 ... Kd4.
39 ... g5 40 Kf1
Now 40 f3 e3 is similar to the game.
40 ... Rh6 41 Kg2 Rd6 42 Kf1
Black repeats the position, a technique with a host of well-documented psychological benefits
(which were briefly described in the notes to to Black’s 47th move in Game 4). Here we can mention
that, aside from gaining time on the clock, this manoeuvre might have elicited the slight hope that
Black was happy with a draw, a hope which is extinguished in the next couple of moves.
42 ... Rd8 43 Kg2 f4!
Leaving White with the tough choice between allowing ... f4-f3 (for instance, by playing 44 Ba6)
or playing f2-f3 himself. It seems he made the wrong call, although given the trend of the play over the
last dozen moves, it is very hard to concentrate on this position and find the best defence without
cursing oneself about the easier drawing chances from a few moves ago. This is something Karpov is
quite familiar with (see Karpov on Karpov in the Introduction).
44 f3? e3
According to the engines, the position is now winning for Black ...
45 g4 Rd2!
... and this is why. Karpov accurately assesses the arising opposite-coloured bishop endgame as
dead lost for White, despite the equal material – the d2-pawn which is about to arrive will cost White
his bishop.
46 Rxd2 exd2 47 Bd1 Kd4 48 Kf2 Kc3 49 Ke2 Kb2 50 Kd3
Exercise: Black to play and win.
Answer: Karpov finds a study-like solution which he must have already foreseen at move 45.
50 ... Kb1!!
Of course not 50 ... Kxa2?? 51 Kc2 with a dead draw.
51 a3
Instead, 51 Be2 Kc1 or 51 a4 Kb2 52 Be2 Kc1 wins the bishop immediately, while 51 Ke2 Kxa2
is similar to the game.
Exercise: And now what?
Answer:
51 ... Kc1!!
Even 51 ... Kb2 52 axb4 Kc1 53 Be2 axb4 54 Kc4 d1Q 55 Bxd1 Kxd1 56 Kxb4 Ke2 wins, but I
wouldn’t recommend it!
52 Ke2 Bxa3
Winning a pawn, with zugzwang.
53 b4 axb4 54 Ba4 Kb2!
A crisp finish. Karpov doesn’t bother defending his d2-pawn, especially as there are no guarantees
of winning a pure opposite-coloured bishop endgame simply because you are two pawns up. Instead
he calculates and ends the game.
55 Kd1
55 Kxd2 b3 will win the bishop for the b-pawn.
55 ... b3 56 Bc6 Ka1 0-1
A stunning demonstration of the possibilities hidden in simple positions against opponents of any
calibre. As John notes, having seen the game live: “He just made it look so easy!”
Game 7
A.Karpov-E.Miroshnichenko
Russian Team Championship 2005
1 d4 Nf6 2 c4 g6 3 Nc3 Bg7 4 e4 0-0 5 Be3 d6 6 h3 e5 7 d5 a5 8 c5!?

An ambitious and relatively little-played idea. White’s main moves are to continue developing (8
Nf3 or 8 Bd3) or try to restrict Black’s kingside possibilities (8 g4). By contrast, 8 c5 executes a
break which is more associated with the Classical King’s Indian, where White develops his knight on
f3 and his bishop on e2. White opens lines on the queenside and, in particular, aims to exploit Black’s
weaknesses on the c-file, especially the c7-pawn (or c7-square).
8 ... Na6
This is a logical continuation of the plan Black started with 7 ... a5 – he wants to play his knight to
c5 (first taking the b2-b4 break away from White). One drawback to this plan, especially where
Black recaptures on d6 with the queen, is that the knight can be tactically vulnerable on the half-open
c-file.
I think I would prefer the approach of Pavlovic and Fedorov, aiming to exploit White’s slight lag
in development and weakening of the centre with 8 ... c6!? 9 dxc6 Nxc6 10 Nf3 dxc5 11 Bxc5 Re8 12
Bb5 Bd7 13 0-0 Nd4 and Black was okay in Z.Izoria-A.Fedorov, Moscow 2006 (½-½ in 24).
9 cxd6 Qxd6 10 Nf3 c6?
With the knight already committed to a6, this is a poor move which weakens Black’s structure.
Hazai suggests 10 ... Bd7!?.
11 dxc6 Qe7
Exercise: How should White proceed?

Answer:
12 Bc4!
Correctly placing a high value on completing development. Black will be left with a weakened
queenside structure once he recaptures on c6. 12 Qa4 is also suggested as a strong move by the
engines, but I prefer Karpov’s approach which gets his king into safety at the earliest opportunity. In
any event, not 12 cxb7? Bxb7, when the e-pawn is under attack and Black is at least equal in view of
his superior development, which can be further improved by bringing his rooks to the open c- and d-
files.
12 ... bxc6 13 0-0 Nc5 14 Qc2 Nfd7 15 Na4!
Exposing Black’s queenside weaknesses.
15 ... Ne6 16 Rfd1 Bf6 17 Rd2 Re8
Answer:
18 Bxe6!
Karpov is always attentive to his opponent’s plans. Here, Black has been gearing up to play ...
Nd4, so Karpov eliminates this move while cementing his advantage.
Instead, 18 Rad1 would allow Black to execute his only idea, namely 18 ... Nd4. Okay, after 19
Nxd4 exd4 20 Bxd4 Qxe4 21 Qb3, White remains in a dominant position, but Karpov’s solution is
even stronger.
18 ... Qxe6 19 Rad1 Be7 20 Ng5! Bxg5
20 ... Qxa2 21 Qxc6 is no better.
21 Bxg5
Exercise: How would you assess this position?
Answer: Let’s look at the positional features:

1. All of Black’s pieces are considerably less active than their white counterparts.
2. Black’s king is weak and, in general, the dark squares in his position (especially d6, f6 and h6)
are in dire need of protection.
3. Black’s structure is inferior, with both weak pawns (e.g. c6) and weak squares.
Summing up, we can say that Black is lost.
21 ... f6
Desperately trying to gain control of some dark squares, but this pawn soon drops off.
22 Be3 Nf8 23 Rd6 Qxa2
23 ... Qf7 is a slightly tougher defence, albeit still hopeless after 24 Nb6 Rb8 25 Qxc6, winning a
pawn with more to come.
24 Rxf6
24 Qxc6 Be6 25 Bh6 is even stronger according to the computer, but Karpov’s choice is still
winning and is considerably more natural, further weakening Black’s structure and the dark squares
around his king.
24 ... Be6
Bringing the bishop to f7 gives the king some form of protection. Unfortunately, the bishop cannot
compete for the vital dark squares.
25 Nc5 Rab8 26 Rd2 Kg7 27 Rf3 Bf7 28 Nd7
28 ... Rbd8
28 ... Rxb2 gives White a choice between a winning endgame (29 Qxb2) and a direct attack with
29 Qxc6 Qe6 30 Rd6, when the black king will not survive for long.
29 Nf6 1-0
White wins at least an exchange with a continuing attack, so Black threw in the towel.
Exchange Sacrifice
While Karpov was a defensive genius, in my view he was one of the least materialistic of world
champions. If someone made a dubious sacrifice against him, he was fully capable of grabbing
material, calculating the consequences accurately, and bringing the position to safety. But rather than
accept a sacrifice for interesting compensation, he generally preferred to retain control and,
conversely, was comfortable sacrificing material himself. In this chapter we examine one of these
types of sacrifice, namely a rook for a minor piece. Of course this was a trademark of Petrosian, but
Karpov also happily made such trades and expertly handled the resulting positions.
Game 8
A.Karpov-V.Malaniuk
Dutch Defence
1 d4 f5
With his intuitive positional style and keen sense of danger, Karpov has scored heavily against the
Dutch throughout his career. Malaniuk, however, was probably the most dedicated supporter of the
Leningrad Variation at GM level, and so felt honour-bound to test Karpov in this system.
2 g3 Nf6 3 Bg2 g6
Setting up the Leningrad.
Against 3 ... e6, which can lead to the Stonewall (after ... d7-d5) or the Classical (after ... d7-d6),
Karpov favoured formations with Nh3. For example:
a) 4 Nh3 Be7 5 0-0 0-0 6 c4 d6 7 Nc3 (after 7 Qb3 c6 8 Nc3 Na6 9 Ng5 e5!, Black equalized in
A.Karpov-Pr.Nikolic, Reykjavik 1991; ½-½ in 24) 7 ... Qe8 8 Nf4 g5 9 Nd3 Qg6 10 f4 h6 and now,
instead of 11 d5 h6 12 b4 as in A.Karpov-N.Short, Linares 1992 (1-0 in 37), White might gain a
greater advantage by preparing e2-e4; for instance, 11 Nf2 Nbd7 12 e4 with the better chances.
b) 4 c4 d5 5 Nd2 c6 6 Nh3 Be7 (or 6 ... Bd6 7 0-0 0-0 8 Qc2 Bd7 9 Nf3 Qe7 10 Bf4 Be8 11 Bxd6
Qxd6 12 Nf4 with a solid advantage in A.Karpov-B.Bidalis, Rethymnon simul 2001; 1-0 in 64) 7 0-0
0-0 8 Nf4 Qe8 9 Nf3 Ne4 10 Qc2 Bd6 11 Nd3 a5 12 Bf4 Bxf4 13 Nxf4 and the trade of dark-squared
bishops resulted in a classical White advantage in A.Karpov-H.Böhland, Hockenheim (simul) 1994
(1-0 in 44).
An unusual Stonewall formation arose after 3 ... d6 4 Nc3!? d5! 5 Bg5 e6 in A.Karpov-V.Ivanchuk,
Linares 1995 (½-½ in 51).
4 c4 Bg7 5 Nf3
Malaniuk subsequently made a comfortable draw in this line: 5 Nc3 0-0 6 Nh3 Nc6 7 0-0 d6 8 d5
Ne5 9 b3 c5 10 Nf4 Ne8 11 Qc2 Nc7 12 Bd2 a6 13 a4 b6 14 Nd3 Rb8 15 Rae1 b5 with good play
for Black in A.Karpov-V.Malaniuk, Tallinn 2005 (½-½ in 22).
5 ... d6 6 0-0 0-0 7 Nc3 Qe8
8 b3
Karpov has tried a number of moves here.
a) 8 Qb3 c6 9 d5 Na6 10 Be3 Ng4 has been contested twice between Karpov and another
Leningrad Dutch specialist, GM Mikhail Gurevich: 11 Bf4 (an attempted improvement on 11 Bd4 e5
12 dxe6 Ne5 13 Rad1 Qxe6 14 Na4 Qxc4 15 Qxc4+ Nxc4 16 Bxg7 Kxg7 17 b3 Nb6 18 Rxd6 Nxa4
19 bxa4 Nc5 and Black was no worse in A.Karpov-M.Gurevich, Reggio Emilia 1989/90; ½-½ in 53)
11 ... Nc5 12 Qc2 h6 13 h3 e5 14 dxe6 Ne5 15 Rad1 Nxe6 16 Rxd6 Nxc4 17 Rd3 Nxf4 18 gxf4 Be6
and Black had no reason to complain about the outcome of the opening. A.Karpov-M.Gurevich,
Amsterdam 1991 (1-0 in 47).
b) 8 d5 a5 (after 8 ... Na6 9 Rb1 Nc5 10 Nd4 Qf7 11 b4 e5 12 dxe6 Nxe6 13 Nxe6 Bxe6 14 Bxb7
Rab8 15 Bg2 Bxc4 16 Be3 Rbe8, Black was very comfortable in A.Karpov-H.Nakamura, Cap
d’Agde rapid 2008; 0-1 in 52) 9 Nd4 (varying from 9 Rb1 Na6 10 Ne1 Bd7 11 Nd3 h6 12 b3 g5 13
Bb2 Qg6 with balanced play in A.Karpov-Zhang Zhong, Cap d’Agde rapid 2000; 1-0 in 35) 9 ... Na6
10 Rb1 Bd7 11 b3 c6 12 Bb2 Nc7 13 Qd2 Kh8 14 e4 fxe4 15 Nxe4 cxd5 16 Nxf6 exf6 17 cxd5 with
a clear advantage for White in A.Karpov-I.Rotov, Puhajarve (rapid) 2013 (1-0 in 50).
8 ... Na6 9 Ba3 c6 10 Qd3 Bd7 11 Rfe1 Rd8 12 Rad1 Kh8 13 e4!
Karpov’s set-up looks very logical to me. White rapidly develops his pieces to their best squares
and cracks open the centre.
13 ... fxe4 14 Nxe4 Bf5 15 Nxf6 Bxf6 16 Qe3
Generally, if White achieves e2-e4 in the Dutch, he can count on a slight advantage. Here Black
has to guard the e7-pawn and the e6-square, and the knight on a6 is out of play.
16 ... Qf7 17 h3 Nc7 18 Re2!
A perfect multi-purpose move, protecting the weak f2-pawn (and the dark-squared bishop if it
drops back to b2) while preparing to double on the e-file.
18 ... Bc8 19 Ng5!
The knight can only be driven away at the cost of weakening the kingside.
19 ... Qg8 20 Qd2 Ne6 21 Nxe6 Bxe6 22 Rde1 Bd7
22 ... Bc8 is a stronger defence, when 23 Bb2 keeps up the pressure.
Answer:
23 Rxe7! Bxe7 24 Rxe7
For the exchange White gains a pawn and eliminates Black’s best defensive piece. White’s dark-
squared bishop will be dominant on the a1-h8 diagonal. In addition, the d6-pawn is hanging (which
wouldn’t have been the case after 22 ... Bc8).
24 ... Rf6
Trying to block the long diagonal with 24 ... d5 is futile: 25 Bd6! and the bishop gets to e5.
Exercise: Find the best continuation for White.

Answer:
25 d5!
Opening the long diagonal for the bishop.
25 ... Qf8 26 Re3 Kg8 27 Bb2
Black is completely busted.
27 ... Rf5 28 Qd4 Re5
At least this way he eliminates an attacking piece, but White gets two pawns for the exchange and
is still dominating the dark squares.
29 Rxe5 dxe5 30 Qxe5 Kf7 31 d6 Bf5 32 c5 h5 33 g4!
33 Bf1!, aiming for c4, was also strong.
33 ... hxg4 34 hxg4 Bd3
34 ... Bxg4 drops the bishop to 35 Qf6+ Ke8 36 Qxg6+ etc.
Exercise: Find the mate in four.
Answer:
35 Bd5+! 1-0
Since 35 ... cxd5 36 Qxd5+ Ke8 37 Qe6+ Qe7 38 Qxe7 is mate.
Game 9
A.Karpov-B.Spassky
Moscow 1973
Ruy Lopez
1 e4 e5 2 Nf3 Nc6 3 Bb5 a6 4 Ba4 Nf6 5 0-0 Be7 6 Re1 b5 7 Bb3 d6 8 c3 0-0 9 h3 Nb8
The Breyer, an old variation which has become popular again amongst strong GMs who want to
play for a win.
10 d3
Like Spassky, Karpov also favoured the Breyer at various stages in his career. In the 1973
Leningrad Interzonal, he gave a masterclass from both sides of the position after 10 d4 Nbd7:
a) 11 Nbd2 Bb7 12 Bc2 (the main line) 12 ... c5 (Karpov obtained reasonable play with Black
after 12 ... Re8 13 b4 Bf8 14 Bb2 a5 15 Bd3 c6 16 a3 Nb6 in his round 8 encounter with Tal, drawing
in 47 moves) 13 d5 Ne8 14 Nf1 g6 15 Bh6 Ng7 16 Ne3 Nf6 17 a4 Kh8 18 b3 Rb8 19 Qe2 Bc8 20
axb5 axb5 21 Ra7 Ng8 22 Bxg7+ Kxg7 23 Rea1, when Karpov (as White against Gligoric in round
13) dominated the a-file with a clear strategic advantage (1-0 in 63).
b) The unusual 11 c4 c6 12 Bg5 h6 13 Bh4 was tried by Tukmakov in round 3. Karpov quickly
reached a superior position: 13 ... Nh5! (I like this move, trading dark-squared bishops and preparing
to occupy f4 with the knight, as in many lines of the Italian Game) 14 Bxe7 Qxe7 15 cxb5 axb5 16
Nc3 b4 17 Nb1? (this loss of time is more than White’s position can stand; 17 Ne2 was better, though
Black is very comfortable after 17 ... c5) 17 ... Nf4 18 Nbd2 exd4 19 Nxd4 Ne5 20 N2f3 Qf6 21
Nxe5 dxe5
22 Ne2 Rd8 23 Qc2 allows Black to conclude the game with a direct attack: 23 ... Nxh3+!! 24
gxh3 Qf3, when White’s only defence to the threat of ... Bxh3 and mate is the ingenious 25 Qxc6 (25
Kh2 Qxf2+ 26 Kh1 Qf3+ 27 Kh2 Qxh3+ 28 Kg1 Rd3 leads to a decisive attack) 25 ... Bxh3 26 Qg6,
which loses to the equally imaginative 26 ... Qxb3!!.
Thus White opted for 22 Nf5, but after 22 ... Bxf5 23 exf5 Rad8 24 Qf3 Rd2 25 Re3 Rxb2, Black won a pawn and soon the game
(0-1 in 33).
10 ... Bb7 11 Nbd2 Nbd7 12 Nf1 Nc5 13 Bc2 Re8 14 Ng3 Bf8 15 b4 Ncd7 16 d4 h6 17 Bd2 Nb6
18 Bd3 g6
Varying from an earlier Karpov win which had proceeded 18 ... Rc8 19 Qc2 Qd7 20 Rad1 Qc6 21
Be3 Na4 22 Rc1 Nb6 23 Qb1 Qd7 24 Nd2 c5 25 bxc5 dxc5 26 d5 Na4 27 c4 b4 28 Rf1 Qc7 29 f4
Nd7 30 Qc2 Nc3 31 f5, followed by a massive kingside pawn storm in A.Karpov-S.Gligoric, San
Antonio 1972 (1-0 in 52).
19 Qc2 Nfd7 20 Rad1 Bg7 21 dxe5 dxe5 22 c4 bxc4 23 Bxc4 Qe7 24 Bb3 c5 25 a4 c4 26 Ba2
Bc6
Exercise: How should White meet the threat to his a-pawn?
Answer:
27 a5!
A fantastic exchange sacrifice, for which Karpov gains both a pawn and time to prosecute his
kingside attack. 27 Rb1 Bxa4 28 Qc1 Kh7 29 Be3 gives White some compensation, but Karpov’s
move is much stronger.
27 ... Ba4 28 Qc1 Nc8
28 ... Bxd1 29 Rxd1 Nc8 30 Bxh6 transposes.
29 Bxh6 Bxd1 30 Rxd1 Nd6 31 Bxg7 Kxg7
Answer:
32 Qg5!!
Wonderful alertness from Karpov. 32 Qd2? looks strong but runs into 32 ... Nf6!, when Black is on
top since 33 Qxd6? loses to 33 ... Rad8 34 Qxe7 Rxd1+ 35 Kh2 Rxe7.
32 ... f6
There is nothing better. After 32 ... Qxg5 33 Nxg5, Black drops a knight; while 32 ... Qf8 loses
simply to 33 Rxd6, when the rook is immune in view of Nf5+.
33 Qg4
Here 33 Qd2 is possible – and winning, since Qxd6xe7 will come with check (so the tactic
outlined in the note to White’s 32nd no longer works) – but Karpov’s move is even stronger, playing
for the attack.
33 ... Kh7 34 Nh4 1-0
The attack cuts straight through: 34 ... Nf8 (or 34 ... Rg8 35 Rxd6 Qxd6 36 Nhf5 gxf5 37 Qh5+ Kg7
38 Nxf5+) 35 Nxg6! Nxg6 36 Rxd6 Qxd6 37 Nf5! and Black has to shed truckloads of material to
avoid immediate mate.
Game 10
A.Karpov-V.Korchnoi
World Championship (14th matchgame) Baguio 1978
Ruy Lopez
1 e4 e5 2 Nf3 Nc6 3 Bb5 a6 4 Ba4 Nf6 5 0-0 Nxe4 6 d4 b5 7 Bb3 d5 8 dxe5 Be6 9 c3 Bc5 10 Nbd2
0-0 11 Bc2 Bf5 12 Nb3 Bg4
After this defeat Korchnoi switched to 12 ... Bg6, which remains the main theoretical battleground.
13 h3
In games two and four, Karpov had failed to get anywhere after 13 Nxc5 Nxc5 14 Re1.
13 ... Bh5 14 g4 Bg6 15 Bxe4!
Starting a forced sequence resulting in an unpleasant endgame for Black.
15 ... dxe4 16 Nxc5 exf3 17 Bf4
17 ... Qxd1
Perhaps Korchnoi should have played for complications with 17 ... Qe7, though after 18 Qd5 Na5
19 b4 Nc4 20 Qxf3, White is material up – the e5-pawn is immune in view of 20 ... Nxe5 21 Bxe5
Qxe5 22 Nd7, winning the exchange.
18 Raxd1 Nd8!
The best defence, aiming to trade the dominant knight on c5.
19 Rd7 Ne6 20 Nxe6 fxe6 21 Be3 Rac8 22 Rfd1
Black has some hopes based on the opposite-coloured bishops, but with the superior structure and
complete control of the only open file, White is clearly pressing.
22 ... Be4 23 Bc5 Rfe8 24 R7d4 Bd5 25 b3 a5 26 Kh2 Ra8 27 Kg3 Ra6 28 h4 Rc6
Answer:
29 Rxd5!
The points in favour of this exchange sacrifice are as follows:
1. There are few open files in the position. The only open file for the near future will be the d-file,
and Black will struggle to contest it having brought his rook to c6.
2. Black has a number of potential weaknesses on the queenside, as well as a weak f3-pawn.
3. White will dominate the dark squares.
These factors add up to a position which is objectively in White’s favour and, from a practical
point of view, is extremely difficult for Black to defend.
By contrast, 29 c4 Rxc5 30 cxd5 exd5 31 Rxd5 Rxe5 leads to an unclear rook and pawn endgame.
29 ... exd5 30 Rxd5 Rce6 31 Bd4 c6 32 Rc5 Rf8
Black missed a good defensive chance here with 32 ... Rd8! 33 Kxf3 Rd5 34 Rxd5 cxd5. Trading a
pair of rooks is in his favour. White is still favourite with two pawns for the exchange and a mobile
kingside pawn majority, but this seems better for Black than the game.
Exercise: How can White make progress?
Answer:
33 a4!
Winning a pawn on the queenside.
33 ... bxa4 34 bxa4 g6 35 Rxa5
White now has two pawns for the exchange, and Black still has weaknesses on c6 and f3. Note the
tremendous bishop on d4, which blocks all counterplay on the d- and e-files, protects the e5-pawn,
and helps create threats to the black king.
35 ... Ree8 36 Ra7 Rf7 37 Ra6 Rc7 38 Bc5 Rcc8 39 Bd6 Ra8 40 Rxc6 Rxa4 41 Kxf3
With three pawns for the exchange, White is winning.
41 ... h5 42 gxh5 gxh5 43 c4 Ra2 44 Rb6 Kf7 45 c5 Ra4 46 c6 Ke6 47 c7 Kd7 48 Rb8 Rc8 49
Ke3
Simply 49 Rxc8 Kxc8 50 e6 wins. Karpov’s move sets a small trap.
49 ... Rxh4
Falling into it, but there was nothing better.
50 e6+ 1-0
Korchnoi resigned in view of 50 ... Kxe6 51 Bg3, winning a rook.
Chapter Two
Key Structures
All world champions produce model games in the structures they regularly play, and Karpov was no
exception. However, in the two structures I have selected, Karpov’s games are simply required
viewing.
The Isolated Queen’s Pawn

Every strong player has a deep understanding of Isolated Queen’s Pawn (“IQP”) positions. Some
prefer the dynamic chances which come with this pawn, like Short and Spassky. Some, like Kramnik,
are equally comfortable playing with or against an IQP.
Karpov is generally regarded as having a preference for playing against the IQP. The combination
of fantastic defensive skills (which helped him to neutralize his opponent’s initiative) and
unsurpassed technique (enabling him to press indefinitely and without risk, scoring a huge percentage
from endgames which were only slightly in his favour) meant that he was particularly lethal against
the IQP, winning a number of model games from diverse openings. His wins against Garry
Kasparov’s Tarrasch in their first World Championship match are among the most analysed games in
chess history (my two cents on these games are available in Move by Move: The Tarrasch).
Personally, I like playing with an IQP. For this reason, I think it’s useful to begin this section with
Karpov’s education in these structures, since he suffered some crushing defeats.
Game 11
V.Smyslov-A.Karpov
USSR Championship, Leningrad 1971
Semi-Tarrasch Defence
1 c4 c5 2 Nf3 Nf6 3 Nc3 d5 4 cxd5 Nxd5 5 e3 e6 6 d4 cxd4 7 exd4 Be7 8 Bd3 0-0 9 0-0 Nc6 10
Re1
Karpov has extensive experience of this position with both colours. We will focus here on the most
instructive of these encounters, his crushing loss to Smyslov.
10 ... Nf6
Alternatively, 10 ... Bf6 leads to positions with a slightly different flavour from traditional IQP
fare: 11 Be4 Nce7 12 h4 (or 12 Qb3 b6 13 Bf4 Bb7 14 Be5 Bc6 15 Rac1 Nxc3 16 Bxc6 Nxc6 17
Rxc3 Bxe5 18 dxe5 Nd4 19 Nxd4 and a draw was agreed in J.Polgar-A.Karpov, Vienna 1996;
Karpov has also defended the position after 12 Qd3 h6 several times, with good results) 12 ... Nf5 13
Qd3 Nxc3 14 bxc3 h6 (14 ... Bxh4? loses to 15 g4) 15 h5 Nd6 16 Ne5 Nxe4 17 Qxe4 Bxe5 18 dxe5
was reached in V.Anand-A.Karpov, FIDE World Championship (4th matchgame), Lausanne 1998. It is
hard to imagine White losing this position (which arose from a Caro-Kann Defence) but, starting from
here, Karpov – in anticipation of Magnus Carlsen ten years later – gradually outplays his opponent.
Exercise: How can Black activate his game?

Answer: 18 ... f5!? is an excellent move, giving White an immediate choice. Otherwise Anand would have been able to develop his
bishop and queen’s rook in comfort, with good prospects of gaining a strong initiative.
Here 19 exf6 Qxf6 would have been the principled response, when White can continue in resolute
fashion and sacrifice a pawn: 20 f3!? Qxc3 21 Be3 Qc6 22 Qh4 e5 and the position looks quite
unclear. Black has an extra pawn, but in the long run his king is potentially weaker than its
counterpart. The engines even marginally prefer White.
Perhaps informed by the competitive importance of the game, Anand hesitated. Black’s position
has been substantially improved by the last two moves – he has gained kingside space, which reduces
White’s chances of developing an attack there, and has increased his control over the central light
squares. The game continued 19 Qe2 Bd7 20 Rd1 Bb5! (a nice little tactic to improve Black’s co-
ordination) 21 Qf3 Qe8 and now:
a) 22 Qxb7 Bc6 (22 ... Qxh5 23 Bxh6 threatens mate and attacks the bishop on b5, though Black
has some compensation after 23 ... Qxh6 24 Qxb5 f4) 23 Qa6 Qxh5 leads to unclear play. Here 24
Bxh6 isn’t such a great idea, since after 24 ... Rac8, followed by ... Qg6 or ... Qg4, Black is
developing a serious kingside initiative.
b) Anand chose 22 Bf4 Rc8 23 Rd4 Rc4 24 Rad1 Qf7 25 Rxc4 Bxc4 26 a3 Rc8 27 Rd4 Kh7 28
Bd2 Bd5 29 Qh3 b5 30 a4 bxa4 31 Rxa4 Rc4! (I really like this move, trading White’s active rook
and emphasizing the difference in activity between the bishops) 32 Rxc4 Bxc4 33 Qh4 Bb5 34 c4 Be8
35 c5 Qd7 (or 35 ... Qxh5 36 Qxh5 Bxh5 37 c6 Be8 38 c7 Bd7) 36 Bc3 Qd3 37 Qd4 Qxd4 38 Bxd4
a5 39 c6? (Anand cracks under the pressure. Ftacnik analyses two better defensive attempts: 39 f3
Bxh5 40 c6 Be8 41 c7 Bd7 42 Kf2 a4 43 Bc5 or 39 f4 a4 40 Bb2 Bxh5 41 c6 Be8 42 c7 Bd7 43 Kf2
Kg6 44 Ba3 Bc8 45 Ke3 Kf7 46 Kd4) 39 ... Bxc6 40 f3 f4! (fixing the white kingside pawns on light
squares) 41 Bb2 Be8 42 Bc1 a4 43 Bxf4 a3 44 Be3 Bxh5 45 Kf2 Be8 46 Bd4 Bc6 47 Bc3 a2 48 g3
h5 49 g4
Exercise: Black to play and win!
Answer: 49 ... h4! and White resigned, since the creation of a second passed pawn is decisive.
11 a3 b6
Karpov subsequently tried this position as White: 11 ... Qd6 12 Be3 Rd8 13 Qc2 Bd7 14 Rad1
Rac8 15 Bg5 h6 16 Bh4 Qb8 17 Qe2 Be8 18 Bb1 Nd5 19 Qd3 g6 20 Bg3 Bd6 21 Nxd5 exd5 22 Ne5
gave White a clear advantage in A.Karpov-V.Hort, Malta Olympiad 1980 (1-0 in 61).
12 Bc2 Bb7
12 ... Ba6!? is an alternative, as in J.Timman-A.Karpov, Wijk aan Zee 1998 (½-½ in 27), where
Timman avoided the critical 13 b4!.
13 Qd3 Rc8

14 Bg5
Missing a shot.
Answer: 14 d5! is an extremely strong break (thanks to John Emms for drawing this to my attention!). Black has a couple of options,
but nothing quite works:
a) 14 ... exd5 loses to the thematic 15 Bg5! Ne4 (or 15 ... g6 16 Rxe7 Qxe7 17 Nxd5 Nxd5 18
Bxe7 Ndxe7 with insufficient compensation for the queen, M.Novik-A.Zubov, Dos Hermanas blitz
2003; 1-0 in 33) 16 Nxe4 dxe4 17 Qxe4 f5 (or 17 ... g6 18 Qh4 with a decisive initiative) 18 Qh4
Rc7 19 Rad1 with a winning attack at no material cost in D.Adla-D.Garcia Roman, Spanish Team
Championship 2006 (1-0 in 26).
b) 14 ... Na5 15 Bg5! Rxc3 (15 ... g6 leaves White much better after both 16 dxe6 and 16 d6) 16
Qxc3 Qxd5 is probably Black’s best chance, though the compensation for the exchange is clearly
insufficient after 17 Rad1, as in Vl.Petkov-Th.Smith, Erice 2014 (1-0 in 70).
14 ... g6 15 Rad1
Despite White’s missed opportunity on move 14, Black’s position remains extremely dangerous to
defend in practice, even though the engines seem to think it’s equal.
15 ... Nd5
The most popular move, stopping the d5-advance and offering the exchange of dark-squared
bishops. Instead:
a) 15 ... Qc7 16 Bxf6 Bxf6 17 d5 Ne5 18 Nxe5 Bxe5 19 dxe6 was A.Danin-V.Yemelin, Taganrog
2011 (1-0 in 31), when 19 ... fxe6 would have equalized, but 16 Bb3! earlier would have kept Black
under pressure.
b) 15 ... Re8 is also played. The following game is an excellent demonstration of many of White’s
resources in an IQP position: 16 h4! (the h-pawn supports a white piece on g5 and, if the f6-knight
moves, the pawn can advance to h5, softening up the kingside structure) 16 ... a6 (Black tries to
generate some counterplay) 17 Bb3! (another typical idea; once ... g7-g6 has been played the bishop
has done its work on the b1-h7 diagonal – Black’s kingside structure is now less flexible – and is
ideally placed on the a2-g8 diagonal where it supports d5 and tactical shots on e6 and f7) 17 ... Na5
18 Ba2 b5 19 Ne5! (the black knight no longer controls e5, so this knight hurries to occupy the
square) 19 ... Nd5 (19 ... Nc6? would allow a typical shot: 20 Nxf7 Kxf7 21 Bxe6+ Kg7 22 Bxc8,
followed by 23 d5 with a material advantage and a strong initiative; 20 Nxg6 hxg6 21 Bxe6 is good
too) 20 Bxe7 Rxe7 21 Ne4! with a continuing attack in R.Vasquez Schroder-O.Zambrana,
Cochabamba 2013 (1-0 in 34).
16 Bh6 Re8
A recent game continued 16 ... Nxc3 17 bxc3 Re8 18 Ba4 Qd5?! (the engines suggest that Black
can get away with 18 ... Bxa3, when 19 Ra1 Be7 20 Bb5 a5 21 Qe3 gives White compensation, if
perhaps not more) 19 c4 Qh5 (the black queen has manoeuvred to the kingside, but White’s position
in the centre is now extremely strong) 20 Qe3 Red8? (20 ... Bf8 is safer, with an edge for White)
Exercise: How should White play?
Answer: 21 d5! (of course! – this advance is just as effective with hanging pawns as in pure IQP positions) 21 ... exd5 (21 ... Bc5 22
Qf4 Bd6 23 Qg5 is extremely strong; while 22 ... exd5 23 cxd5 Rxd5 loses to 24 g4) 22 cxd5 Na5 23 Qxe7 Qxh6 24 Bd7 Ra8 25 Ne5
and Black was busted in A.Danin-Y.Boidman, Hofheim 2013 (1-0 in 31).
17 Ba4 a6
Here 17 ... Nxc3 18 bxc3 looks safer, transposing to the previous note.
18 Nxd5 Qxd5 19 Qe3 Bf6 20 Bb3 Qh5
Karpov subsequently – 15 years later! – improved with 20 ... Qd7 21 d5 exd5 22 Qxb6 Rxe1+ 23
Rxe1 in A.Beliavsky-A.Karpov, European Cup, Moscow 1986. However, it is debatable whether this
was intended, or Karpov simply fell into the same line twice, since Black’s position is far from
comfortable here, with an inferior structure, less active pieces and a slightly vulnerable king. After 23
... Bxb2 24 Bxd5 Bg7 25 Bxg7 Kxg7, instead of Beliavsky’s 26 h4? (½-½ in 42), White could have
developed decisive activity with 26 Qb2+ Kg8 27 Ba2, threatening Ng5.
Answer:
21 d5!
Utterly thematic and completely decisive. In addition to the normal advantages offered by this
move, Black’s b6-pawn is horribly weak.
21 ... Nd8 22 d6
22 Bg5 Be7 23 Bxe7 Rxe7 24 d6 also wins.
22 ... Rc5 23 d7 Re7 24 Qf4 Bg7 25 Qb8 Qxh6 26 Qxd8+ Bf8
27 Re3
27 Rxe6! is terribly strong too; for example, 27 ... fxe6 28 Qxb6 and Black’s position collapses.
27 ... Bc6 28 Qxf8+ Qxf8 29 d8Q 1-0
I guess even future World Champions have off days.
Game 12
L.Portisch-A.Karpov
Milan (5th matchgame) 1975
Nimzo-Indian Defence
1 c4 Nf6 2 Nc3 e6 3 d4 Bb4 4 e3 c5 5 Bd3 0-0 6 Nf3 d5 7 0-0 cxd4 8 exd4 dxc4 9 Bxc4 b6
This has been dubbed the “Karpov Variation” by several sources. The concept is very much in
keeping with his style, namely quick and healthy development in the opening, and the fact that the
opponent is saddled with an IQP is a bonus.
One of the aspects of Karpov’s play that is not immediately apparent from his best games (which
are the focus of this book) is how resilient he is in terrible or even lost positions. Let’s take a look.
10 Re1
Karpov had put on a positional masterclass in this variation against Taimanov two years before:
10 Qe2 Bb7 11 Rd1 Nbd7 12 Bd2 Rc8 13 Ba6 Bxa6 14 Qxa6 Bxc3 15 bxc3 Rc7 16 Rac1 Qc8 17
Qa4
Exercise: How should Black proceed?
Answer: 17 ... Rc4! (a typical pawn sacrifice in this structure: the pressure down the half-open a- and c-files will restrict the white
pieces; meanwhile, the extra pawn is not felt) 18 Qxa7 Qc6 19 Qa3 Rc8 20 h3 h6 21 Rb1 Ra4 22 Qb3 Nd5 23 Rdc1 Rc4 24 Rb2 f6!?
(an interesting practical decision, though not entirely without risk; Karpov wasn’t satisfied with the immediate draw available after 24 ...
Nxc3 25 Bxc3 Rxc3 26 Rxc3 Qxc3 27 Qxc3 Rxc3 28 Ne5 Nxe5 29 dxe5 Ra3 etc) 25 Re1 Kf7 26 Qd1 Nf8 27 Rb3 Ng6 28 Qb1 Ra8 29
Re4 Rca4 30 Rb2 Nf8 31 Qd3 Rc4 32 Re1 Ra3 33 Qb1 Ng6 34 Rc1? (a significant error; after 34 h4 h5, the position would remain
roughly balanced) 34 ... Nxc3 (here 34 ... Raxc3 is strong, since White can’t hold on to the exchange after 35 Bxc3? Nxc3, threatening
36 ... Ne2+, so he has to play 35 Rxc3 Nxc3 36 Qd3 Nd5 with an edge for Black) 35 Qd3 Ne2+ 36 Qxe2 Rxc1+ 37 Bxc1 Qxc1+ 38
Kh2 Rxf3!? 39 gxf3 Nh4 and White lost on time in M.Taimanov-A.Karpov, Moscow 1973. While the position is unpleasant to play,
especially in time trouble, objectively it’s now equal; for instance, 40 d5 Qf4+ 41 Kh1 Qc1+ 42 Kh2 Qf4+ 43 Kh1 with a draw.
Karpov has a lot of experience with both colours in the main line after 10 Bg5. This variation has
been extensively developed (for instance, game nine in the 2012 Anand-Gelfand World Championship
match) and leads to heavy theory which I’m not going to discuss here. Anyone interested should check
out Ivan Sokolov’s superb book, The Strategic Nimzo-Indian.
10 ... Bb7 11 Bd3 Nc6 12 a3 Be7
Once Black’s knight has gone to c6, it is dubious to play ... Bxc3. In the arising structure the knight
should be placed on d7 where it doesn’t obstruct the black rook on c8.
13 Bc2 Re8
Karpov subsequently improved with 13 ... Rc8 14 Qd3 g6 15 Bh6 Re8 16 Rad1 Bf8 17 Bg5 Be7
18 Bh6 Bf8 and a draw was agreed in F.Olafsson-A.Karpov, Amsterdam 1976.
14 Qd3 Rc8
Exercise: What should White play? And how do you assess the position?
Answer: Karpov has managed to stumble into a position which has been known to be terrible for Black since T.V.Petrosian-
Y.Balashov, USSR 1974 (1-0 in 28).
15 d5!
In fact, attentive readers might notice that this is virtually identical to a position White could have
had at move 14 in the previous game (Smyslov-Karpov). Here Black has the extra move ... Re8;
unfortunately, that doesn’t change the assessment of his position as rubbish.
How does Karpov manage to extricate himself from this situation against a world-class GM? With
large quantities of both luck and skill, as will see.
Exercise: How should Black respond?
Answer:
15 ... exd5
White gets an extremely strong initiative after this move. It’s an interesting defensive choice and,
frankly, I think the selection between 15 ... exd5 and 15 ... Na5 is a matter of taste.
After the text, White needs to play precisely since otherwise Black will escape with equality. As
will become apparent from this game and the notes, several strong GMs were not equal to this
challenge, letting clearly superior positions slip from their grasp.
Instead, 15 ... Na5 at least forces White to think a bit, rather than play on autopilot for the next few
moves: 16 Bg5 (the endgame after 16 dxe6 fxe6 17 Qxd8 Bxd8 isn’t pretty, but is preferable to getting
mated) 16 ... Rxc3 17 Qxc3 Qxd5 was played in R.Skytte-I.Nyzhnyk, Cappelle la Grande 2012 (½-½
in 93). This is yet another example of a strong GM (Nyzhnyk was 2585 at the time) falling into an
opening trap which has been known since 1974. While Nyzhnyk does not deserve praise for reaching
the position after 15 d5, I think his response was excellent. White, it must be acknowledged, still
enjoys a clear advantage in this position, since Black’s compensation for the exchange is not fully
sufficient; but the conversion will require a lot of good moves in a relatively non-standard position,
as indicated by the fact that Skytte, who was 2376, messed up and only drew.
16 Bg5 Ne4
Exercise: What would be White’s response to 16 ... g6 - ?
Answer: 17 Rxe7! (a stock sacrifice in such positions) 17 ... Nxe7 (giving up the queen with 17 ... Qxe7 18 Nxd5 Nxd5 19 Bxe7
Ndxe7 is probably relatively best, when White has some work to do to win, although in A.Lahiri-A.Deshmukh, Commonwealth
Championship, Sangli 2000, Black decided to resign on the next move) 18 Bxf6. Although rook and pawn can sometimes compete against
a bishop and knight, Black’s main problem is that the attack is also extremely strong. White went on to win in B.Lalic-T.Kononenko,
Seville 2007 (1-0 in 28).
If 16 ... Kf8 17 Bxf6 Bxf6 18 Qxh7 Rxe1+ 19 Rxe1 g6, then 20 Bxg6! with an extra pawn and an
attack.
17 Nxe4 dxe4 18 Qxe4 g6
19 Qh4!
Strong GMs have handled this position in different ways. 19 Rad1 is also extremely tempting,
bringing in the last piece with tempo. After 19 ... Qc7 20 Qh4 (20 Bb3? is careless, due to 20 ... Bxg5
21 Qxe8+ Rxe8 22 Rxe8+ Kg7 23 Nxg5 Ne5 and Black was not worse in D.Sermek-V.Bukal,
Croatian Team Championship 2001; 1-0 in 79), Black can transpose to the game with 20 ... h5 though
has some additional options since the e7-bishop isn’t pinned to his queen; for instance, 20 ... Bd6 or
20 ... Bf8 (the latter was played in K.Spraggett-L.Day, Montreal Zonal 1981; 1-0 in 41). While
neither of these moves are great, I would prefer not to give Black additional options.
19 ... h5
The classic Petrosian-Balashov game saw 19 ... Qc7 20 Bb3! (threatening 21 Bxf7+, amongst
other things) 20 ... h5
Answer: 21 Qe4!! (an incredibly powerful move, centralizing the queen and creating the threat of 22 Qxg6+) 21 ... Kg7
Exercise: And what now?
Answer: 22 Bxf7!! (child’s play for Petrosian who, like Karpov, was actually one of the finest tacticians in chess history – a fact
which is sometimes overlooked due to both World Champions’ preference for positional play) 22 ... Kxf7 23 Bh6! (only this move wins;
23 Qe6+? Kg7 24 Nh4 Bxg5 gives White no more than a draw), when Black has a choice of ways to lose:
a) 23 ... Bd6 24 Ng5+ Kf6 25 Nh7+!! and Black resigned in Si.Garcia Martinez-A.Pomar
Salamanca, San Feliu de Guixols 1975, in view of 25 ... Qxh7 (or 25 ... Kf7 26 Qc4+ Re6 27 Qxe6
mate) 26 Qf3+ Bf4 27 Qxf4 mate.
b) Balashov chose 23 ... Qd6 24 Qc4+ Kf6 25 Rad1 (25 Ng5 is even stronger, but Petrosian’s
move is absolutely winning) 25 ... Nd4 26 Qxd4+ Qxd4 27 Rxd4 Rc5 28 h4 1-0. An odd moment for
Black to resign since he now has the best position he’s seen for the last 10 moves! Nevertheless, it’s
still winning for White, and I’m told Petrosian’s technique was normally up to the mark.
R.Hernandez Onna-J.Garcia Padron, Las Palmas 1977, was an attacking masterclass: 19 ... Rc7 20
Bb3! (again, 21 Bxf7+ is the idea) 20 ... Bxg5 21 Nxg5 h5 22 Qf4! Rf8 23 Rad1 Qc8
Exercise: White to play and win.
Answer: 24 Nxf7! Rcxf7 (taking with the other rook meets with the same response) 25 Qf6!! (a wonderful move, having just
sacrificed a piece – the attack is absolutely devastating) 25 ... Qc7 (25 ... Qf5 loses to 26 Bxf7+, when the threatened invasion of white
rooks means that Black has nothing better than 26 ... Kh7 27 Bxg6+ Kh6 28 Qxf5 Rxf5 29 Bxf5 with a depressing material count) 26
Qxg6+ Kh8 27 Qh6+ Kg8 28 Rd6 and Black resigned since his best response is 28 ... Qxd6, and even then it’s mate in twelve.
Returning to Karpov’s 19 ... h5:
Answer:
20 Rad1?!
A very natural move, bringing the rook into the attack with tempo, but this is actually inaccurate. 20
Bb3! is still strongest (even though Bxf7+ isn’t now a threat), when Black has nothing better than 20
... Bxg5 (20 ... Rc7 21 Rad1 Rd7 22 Qe4! Rd6 23 Qf4! was winning for White in M.Kaloskambis-
N.Skalkotas, Greek Championship 1976; 1-0 in 26) 21 Nxg5 Rxe1+ 22 Rxe1 Kg7 23 Qf4 Qc7 24
Qxf7+ Qxf7 25 Nxf7 with a clear extra pawn for White.
20 ... Qc7 21 Bxg6!
Portisch finds a good shot.
21 ... fxg6 22 Qc4+ Kg7 23 Bf4
The black queen is short of squares and it seems like game over.
Exercise: How can Black put up any resistance in this position?
Answer:
23 ... Ba6!!
The only move offering any defensive prospects. It seems that Portisch overlooked this resource.
23 ... b5 appears similar but loses to the computer line 24 Qc1! Qb6 25 Rd7, when Black’s position
collapses.
24 Qc3+ Bf6 25 Bxc7 Bxc3 26 Rxe8 Rxe8 27 bxc3
The opposite-coloured bishops and White’s split queenside pawns give Black excellent drawing
chances, despite his pawn deficit.
27 ... Be2 28 Re1 Rc8 29 Rxe2 Rxc7
Karpov has traded down into a rook and knight endgame, still with the wrecked white queenside.
He didn’t have much difficulty making a draw from here.
30 Re6 Nd8 31 Re3 Kf6 32 Kf1 Ne6 33 g3 g5 34 h3 Nc5 35 Nd2 Rd7 36 Ke2 Rd5 37 c4 Rd4 38 Re8 h4 39 Rf8+ Ke7 40
Rh8 hxg3 41 fxg3 ½-½
Game 13
G.Kasparov-A.Karpov
World Championship (22nd matchgame), London/Leningrad 1986
This was the last decisive game of the 1986 World Championship match. (Kasparov secured the title by drawing the next two games.)
1 d4 Nf6 2 c4 e6 3 Nf3 d5 4 Nc3 Be7 5 Bg5 h6 6 Bxf6 Bxf6 7 e3 0-0 8 Rc1 c6 9 Bd3 Nd7 10 0-
0 dxc4 11 Bxc4 e5 12 h3 exd4 13 exd4 Nb6 14 Bb3
14 ... Bf5
This position has been analysed to, essentially, a forced draw after 14 ... Re8 15 Re1 Bf5 16
Rxe8+ Qxe8 17 Qd2 Qd7 18 Re1 a5 (18 ... Rd8 was played in G.Kasparov-A.Karpov, 23rd
matchgame, Moscow 1985, ½-½ in 41; and 18 ... Nd5 in J.Hjartarson-A.Karpov, Tilburg 1988, ½-½
in 32) 19 a3 Re8 20 Rxe8+ Qxe8 21 Qf4 Be6 22 Bxe6 Qxe6 23 Qc7 (23 Qb8+ Qc8 24 Qa7 Nc4 also
led to a draw, as in the stem game M.Gurevich-P.Van der Sterren, Baku 1986; ½-½ in 33, played a
month after the main game) 23 ... Nc4 24 Qxb7
Exercise: How can Black force a draw?
Answer: 24 ... Bxd4! (24 ... g5 leaves Black with play for the pawn, but the text is a clean equalizer) 25 Nxd4 Qe1+ 26 Kh2 Qe5+
27 g3 Qxd4 28 Qc8+ Kh7 29 Qf5+ Kg8 30 Qc8+ and a draw was agreed in I.Sokolov-P.Van der Sterren, Dutch Championship,
Amsterdam 1994.
15 Re1 a5 16 a3 Re8
Karpov himself played a nice game on the other side of this line after 16 ... Qd7 17 Ne5 Bxe5 18
Rxe5 Rfe8 19 Qe2 Rad8 20 Re1 Rxe5 21 Qxe5 a4 22 Qc5 axb3 23 Re7 Qd6? (23 ... Qxd4? 24 Re8+
Rxe8 25 Qxd4 should also win in the long run; the best option was 23 ... Qc8 24 Qxb6 Rd7 25 Rxd7
Qxd7 26 Qxb3 with some winning chances for White) 24 Re8+ Rxe8 25 Qxd6 and White won in
A.Karpov-A.Beliavsky, European Cup, Moscow 1986 (1-0 in 42).
17 Rxe8+ Qxe8 18 Qd2 Nd7?!
Kasparov’s comments in relation to this move are deeply instructive: “With 18 ... Nd7 Black has
taken control of the e5-square and cleared the way for a possible queen move to b6; true, now his
supervision of the d5-point has been weakened, but d4-d5 leads merely to the simplification of the
position. In what direction then should White operate? In the most critical! He must demonstrate that
control of e5 and the threat of ... Qd8-b6 are mere details and not the main thing in this position, that
the black pieces now suffer from a certain lack of harmony, and that White, by playing energetically
and accurately, can nevertheless force the opponent on to the defensive.”
Black should play 18 ... Qd7, when White does not appear to have anything better than 19 Re1
Re8, transposing into the note to Black’s 14th move.
19 Qf4!
Forcing the f5-bishop to choose between two diagonals.
19 ... Bg6
Kasparov notes that 19 ... Be6 20 Bxe6 Qxe6 21 Qc7 Qb3 22 Ne4! is deeply unpleasant for Black.
20 h4!
More powerful play, undermining the g6-bishop.
20 ... Qd8 21 Na4 h5 22 Re1 b5 23 Nc3 Qb8 24 Qe3 b4 25 Ne4 bxa3 26 Nxf6+ Nxf6 27 bxa3
Nd5 28 Bxd5 cxd5 29 Ne5
Answer: It’s a typical IQP position. The knight on e5 is clearly stronger than the bishop on g6. Objectively White is perhaps only
slightly better, but Black’s defence is very unpleasant.
Ivan Sokolov’s assessment (after Black’s 28th move, in his wonderful book Winning Chess
Middlegames) is instructive: “In itself this looks like a reasonable deal for Black. A lot of pieces
have been exchanged, the pawn structure is totally symmetrical and it seems as if Black is about to
equalize. However, his troubles are far from over – in fact they are just beginning. White gets a
monster knight on e5, which is clearly superior to Black’s virtually useless bishop. It is important to
note that the assessment of this position would have been very much different if Black had had the
time for ... f7-f6, keeping the white knight from its dominant post.”
29 ... Qd8 30 Qf3 Ra6 31 Rc1 Kh7 32 Qh3 Rb6 33 Rc8 Qd6?!
An almost imperceptible inaccuracy. 33 ... Qe7! was better, and if play proceeds as in the game
with 34 Qg3 (34 Ra8 Rb1+ 35 Kh2 Rf1 gives Black counterplay) 34 ... a4, Black has counterplay
against the a3-pawn. White would have nothing better than to force a draw with 35 Qg5 Qxg5 36 hxg5
Rb3 37 Nd7 Bf5 38 Nf8+ Kg8 39 Ne6+ Kh7 40 Nf8+ etc.
34 Qg3 a4 35 Ra8 Qe6
Exercise: Why can’t Black win the a3-pawn with 35 ... Rb3 - ?
Answer: The terrific shot 36 Rh8+! Kxh8 37 Nxf7+ Bxf7 38 Qxd6 wins the queen. I think Black’s chances of setting up a fortress
are minimal here. Note that White’s tactic only works because the black queen is on d6, which is why 33 ... Qe7 was preferable.
36 Rxa4 Qf5 37 Ra7 Rb1+ 38 Kh2 Rc1 39 Rb7 Rc2 40 f3 Rd2
Here the game was adjourned and Kasparov had to seal his move.
Answer: Kasparov decides to finish the game with a direct attack.

41 Nd7!
Simply 41 Rb4 would hold the d4-pawn, but it would be difficult to make further progress.
Kasparov’s move is much stronger (in fact, it is completely winning).
41 ... Rxd4 42 Nf8+ Kh6
Exercise: What is White’s winning idea?
Answer:
43 Rb4!!
Preventing the trade of queens with ... Qf4. No prizes for other moves, which permit Black to use
this resource to escape.
43 ... Rc4 44 Rxc4 dxc4 45 Qd6!
Kasparov is precise to the end. 45 Qe1 would allow Black to struggle on with 45 ... Bh7.
45 ... c3
Exercise: Find the only move for White.
Answer:
46 Qd4!! 1-0
Anything else and White would be in trouble in view of Black’s passed c-pawn! Whereas now the
threat of 47 Qe3+ and mates allows White to pick it up for nothing. Karpov resigned, as after 46 ...
Bh7 47 Qxc3, the most he can hope for is a lost king and pawn endgame.
Game 14
U.Andersson-A.Karpov
Nyköping (rapid; 1st matchgame) 1995
Queen’s Gambit Accepted
1 Nf3 d5 2 d4 Nf6 3 c4 dxc4 4 e3 e6 5 Bxc4 c5 6 0-0 a6

Karpov only rarely essayed the Queen’s Gambit Accepted, and it was one of his less successful
openings. Personally I think it was a relatively poor choice. In the Queen’s Gambit Accepted White
has a range of highly forcing continuations, Black needs to know theory deeply to survive, and
Karpov never made any secret of the fact that he didn’t like studying opening theory.
7 Qe2 cxd4
Going straight for an IQP, and a rather dangerous one at that. Most modern QGA specialists tend to
prefer 7 ... b5, when Black can delay taking on d4 with ... Bb7 and, perhaps, ... Nbd7.
8 exd4 Be7 9 Nc3 b5 10 Bb3 0-0 11 Bg5 Bb7 12 Rad1 Nc6 13 Rfe1
13 ... Nb4?
A big mistake. Interestingly, this position has been reached in a number of GM encounters, but with
Black to move! The universal nature of IQP positions, and the number of different variations which
result in these structures, mean that it is not uncommon to reach positions one or even two tempi down
on known theoretical lines.
For instance, M.Illescas Cordoba-V.Anand, Leon (4th matchgame) 1997, went 7 Bb3 Nc6 8 Nc3
cxd4 9 exd4 Be7 10 Re1 0-0 11 Bg5 b5 12 Qd3 Bb7 13 Rad1 Nb4 14 Qe2 and, as promised, we
have reached the same position as in the main game, but with Black to move. After 14 ... Nbd5 15
Ne5 Nxc3 16 bxc3 Nd5 17 Bxe7 Qxe7 18 Qd2 Rac8 19 Rc1 Rfd8 20 Nd3 Nf6 21 Qb2 Qd6 22 Ne5,
the players agreed a draw.
Apart from Karpov’s move, Black has also tried:
a) 13 ... Na5 14 d5! Nxb3 15 dxe6 Qb6 16 axb3 fxe6 17 Nd4 Bd6? (17 ... Rad8 gives better
drawing chances) 18 Qxe6+ left White a pawn up for insufficient compensation in the classic game
I.Boleslavsky-A.Kotov, Zürich Candidates 1953 (1-0 in 42).
b) 13 ... Nd5 14 Bxd5! (Bronstein’s recommendation of 14 Nxd5 Bxg5 15 Nb6 was refuted even
before the computer era: 15 ... Nxd4! 16 Nxd4 Qxb6 17 Qg4 Bf6 18 Nxe6 Bc8 and White resigned in
L.Shamkovich-M.Dlugy, New York Open 1986) 14 ... Bxg5 15 Be4 (or 15 Nxg5 Qxg5 16 Bf3 Rad8
17 d5 exd5, as in V.Hamitevici-R.Soltanici, Kiev 2009, which was agreed drawn after 18 Nxd5 Nd4,
but White could have retained an edge with 18 Bxd5!) 15 ... Bh6 16 a4! (16 d5 exd5 17 Nxd5 only
gives White a slight plus, as in F.Izeta Txabarri-J.Magem Badals, Spanish Team Championship 1995;
0-1 in 42) 16 ... bxa4 (Baburin notes that 16 ... b4 17 d5! exd5 18 Nxd5 is unpleasant for Black) 17
Nxa4 Ra7? (17 ... Rc8 is a better defence) 18 Nc5 Ba8 19 Nxe6! fxe6 20 Bxc6 Bxc6 21 Qxe6+ Raf7
22 Qxc6 and White was winning in P.Wells-J.Magem Badals, Linares Zonal 1995.
c) 13 ... Re8 14 Ne5 Qd6? (14 ... Nxe5 15 dxe5 Nd7 16 Bxe7 Qxe7 17 Rd6 is good for White,
while after 14 ... Nxd4 15 Qe3 Bc5 16 Qf4 Rc8 17 Rd3, White has a very strong initiative for the
pawn) 15 Qe3 occurred in A.Mirzoev-R.Rodriguez Lopez, Poio 2002 (1-0 in 27), but White had
something better at this moment.
Exercise: Can you find the blow which White missed?
Answer: 15 Nxf7! Kxf7 16 d5!! and Black’s position collapses.

Returning to Karpov’s 13 ... Nb4? again:
Exercise: What should White play here?
14 d5!
Answer: As noted by Alexander Baburin, this move had been analysed by the great Soviet analyst Vsevolod Rauzer. Of course it is
deeply thematic for IQP positions.
14 ... Nfxd5
Other captures are no better: 14 ... Nbxd5 (or 14 ... Bxd5 15 Nxd5) 15 Nxd5 Bxd5 (15 ... Nxd5 16
Bxd5 is similar) 16 Bxd5 Nxd5 17 Bxe7 Qxe7 18 Rxd5 wins a piece
15 Nxd5
15 ... Bxg5
There is no way out for Black. 15 ... Nxd5 16 Bxd5 Bxd5 (or similarly 15 ... Bxd5 16 Bxd5 Nxd5)
17 Bxe7 Qxe7 18 Rxd5 again wins a piece.
16 Nxb4 Qe7 17 Nd5 Bxd5 18 Bxd5 1-0
Despite the rapid time control and Karpov’s weak play, I find this game rather instructive. It shows
a lot of the tactical resources White enjoys after a d4-d5 break. More than this, it shows how
dangerous the IQP can be, and how easy it is to stumble into a difficult position with apparently
natural moves. Every now and then I see a top GM lose a game in a similar fashion!
Game 15
G.Kamsky-A.Karpov
FIDE World Championship (4th matchgame), Elista 1996
Caro-Kann Defence
1 e4 c6
The Caro-Kann has been one of Karpov’s main weapons throughout his career. In many lines,
Black refuses to create any pawn weaknesses, thus giving himself good prospects in the event an
endgame is reached. The price he pays is often conceding a temporary initiative to White. Karpov’s
skill set is well suited to this opening, since his sense of danger was useful in dealing with any
initiative, while his outstanding technique meant that Black’s excellent pawn structure could be made
to count later in the game.
2 d4 d5 3 exd5 cxd5 4 c4 Nf6 5 Nc3 e6 6 Nf3 Bb4 7 cxd5 Nxd5 8 Bd2 Nc6 9 Bd3 Be7 10 0-0 0-
0 11 Qe2 Nf6 12 Ne4
12 ... Qb6
Karpov’s improvement on game two of the match, where he had essayed 12 ... Bd7 13 Rad1 Rc8
14 Rfe1 Nd5 (here Baburin notes: “Perhaps Black should have preferred 14 ... Nxe4, although even
then White would keep the initiative after 15 Qxe4 g6 16 a3.”) 15 Nc3! (Kamsky challenges the
knight on d5 – his pieces are better placed for the symmetrical structure which would arise after
Nxd5 exd5, and strengthening White’s centre with 15 ... Nxc3 16 bxc3 is unappealing for Karpov) 15
... Nf6 16 a3 Qc7 17 Bg5 Qa5.
Answer: 18 d5! – a well-calculated shot by Kamsky. As noted by Baburin, “This time the main reason behind the success of this
thematic break in the centre is the lack of protection of the d7-bishop.” (Actually, 18 Bb1! is a strong alternative, improving White’s
position with threats of Ne5 and forming a battery with Qc2 or Qd3. The bishop is perfectly placed on b1, and can come to a2 if needed
on the a2-g8 diagonal.) The game continued 18 ... exd5 19 Bxf6! Bxf6 20 Bxh7+! Kxh7 21 Rxd5, when the black queen and the d7-
bishop are hit, and if 21 ... Qc7 then 22 Qd3+ regains the piece and keeps White a pawn up. Instead, Karpov tried to build a fortress with
21 ... Bxc3 22 Rxa5 Bxa5 23 b4 Kg8 24 bxa5, but Kamsky’s material advantage was decisive and he broke through to win in 65 moves.
13 a3 Bd7
In a later game, Karpov played 13 ... a5 14 Rfd1 Nd5 15 Bc4 Rd8 (here 15 ... Nxd4 16 Nxd4
Qxd4 17 Bxa5 Qe5 looks level) 16 Rab1 Bd7 17 Bg5 Bxg5 18 Nexg5 h6 19 Ne4 Be8 20 Rbc1 Rac8
(Black has placed his pieces ideally and now gradually takes over the initiative) 21 Bxd5?! (I don’t
think White should have traded his strong bishop – if he simply holds the position it’s unclear how
Black could make progress; this decision is very similar to Korchnoi’s 27 Bxd5?! in the next game).
21 ... Rxd5 22 Qe3 was E.Ghaem Maghami-A.Karpov, Tehran (blitz; 6th matchgame) 2009 (0-1 in
68) and here 22 ... Rdd8! would have been strong, aiming for further exchanges with ... Ne7.
14 Rfd1
One of the ideas behind 12 ... Qb6 is that the more aggressive 14 Rad1 might be met by 14 ...
Qxb2.
14 ... Rad8
Now 14 ... Qxb2?? drops the queen to 15 Rdb1, or first 15 Nxf6+ Bxf6 16 Qe4 g6 and then 17
Rdb1.
15 Nxf6+
Kamsky concentrates on the kingside, but Black has sufficient defensive resources. Alternatively,
seizing queenside space with 15 b4 is logical, though after the best response 15 ... Nxe4! 16 Qxe4 (16
Bxe4 f5 is also fine for Black) 16 ... f5 17 Qe3 Bf6 18 Bc3 Ne7 19 Bc4 Kh8 20 Ne5 Ba4 21 Re1
Nd5, Black had a sound position in Hen.Rudolf-R.Zelcic, Nova Gorica 2000 (0-1 in 83).
15 ... Bxf6 16 Qe4 g6 17 Be3
17 Bh6 is well met by the typical 17 ... Nxd4! 18 Bxf8 Kxf8 with good compensation for the
exchange.
17 ... Ne7 18 Ne5 Nf5
As noted by Mikhail Gurevich, this is the modern approach: attacking (rather than blockading) an
IQP.
19 Nc4 Qa6
19 ... Qb3! is even stronger.
20 a4
After 20 Ne5, Black can repeat moves or play for the advantage with 20 ... Bb5!.
20 ... Bc6 21 Qf4 Bd5
Black has an impenetrable blockade on d5 and a solid advantage. The game now enters a lengthy
manoeuvring phase in which Kamsky gets outplayed, but this is not central to our theme and so will
be presented with only one further comment.
22 Ne5 Qb6 23 Bxf5 exf5 24 Rd2 Bg7 25 h4 Rfe8 26 Qg3 Rc8 27 Nd7?! Qc6 28 Nc5 b6 29
Nd3 Qd7 30 a5 Re4! 31 Nf4 b5 32 Rdd1 Bc4 33 Rac1 h6! 34 Rc3 b4 35 Rc2 Rc6 36 Rdc1 Bb5 37
Kh2 Kh7 38 Rxc6 Bxc6 39 Rc4
Exercise: Which of Black’s pieces can be improved?
Answer:
39 ... Bf8!
The bishop was strong on g7 (defending the king and targeting the d4-pawn) but it will be lethal on
d6, creating threats on the h2-b8 diagonal while defending the pawn on b4.
40 Nd3 Qe6 41 d5 Bxd5 42 Rxe4 Bxe4 43 Bxa7 Bd6 44 Nf4 Qe5 45 Nh3 Qe7 0-1
Game 16
V.Korchnoi-A.Karpov
World Championship (9th matchgame), Merano 1981
1 c4 e6 2 Nc3 d5 3 d4 Be7 4 Nf3 Nf6 5 Bg5 h6 6 Bh4 0-0 7 Rc1 dxc4!

A novelty prepared by Karpov for the match.
8 e3
Karpov notes that 8 e4 Nc6! gives Black good play.
8 ... c5 9 Bxc4 cxd4
10 exd4
An ambitious choice. 10 Nxd4 is more conservative and was tried by Korchnoi in game 17 of the
match. After 10 ... Bd7! 11 Be2 Nc6 12 Nb3 Nd5 13 Bxe7 Ncxe7 14 Nxd5 Nxd5 15 Qd4 Bc6 16 Bf3
Ne7 17 Bxc6 Nxc6 18 Qxd8 Rfxd8 19 Ke2 Rac8 20 a3 Kf8 21 Rc2 Ne7 22 Rhc1 Rxc2+ 23 Rxc2
Ke8, a draw was agreed.
Karpov later took on the white side of this variation: 11 0-0 Nc6 12 Nb3 Rc8 13 Be2 Nd5 14
Bxe7 Ncxe7 15 Nxd5 Nxd5 16 Rxc8 Qxc8 17 Qd4 Qb8 18 Bf3 Nf6 19 Nc5 Bb5 20 Rd1 b6 21 Ne4
Nxe4 22 Bxe4 Rc8 was also a draw in A.Karpov-G.Kasparov, World Championship (23rd
matchgame), Moscow 1984. Instead, 11 Bg3 a6 12 e4 Nc6 led to a win for White in A.Karpov-
A.Beliavsky, Dortmund 1995 (1-0 in 60), but the result had nothing to do with the opening.
10 ... Nc6 11 0-0 Nh5!?
A standard idea in IQP positions – the knight is not well placed on the edge, but by anticipating
Bg3 it forces an exchange of minor pieces. All such trades are generally in Black’s favour.
12 Bxe7 Nxe7
13 Bb3
Karpov makes an interesting psychological observation at this point: “Here White could have got
rid of his weak pawn: 13 d5 exd5 14 Nxd5 Nxd5 15 Bxd5 Nf4 16 Be4 Qxd1 17 Rcxd1 Be6 with
equality. But Korchnoi does not want to give up the white pieces so easily, and as a result he ends up
in a difficult position.”
Instead, 13 Re1 Nf6 14 Ne5 Bd7 15 Qb3 Rb8 16 Rcd1 b5 17 Nxd7 (17 Bd3! would have kept
some initiative for White) 17 ... Nxd7 18 Bd3 Nf6 19 Bb1 a6 20 Ne4 was balanced in
L.Christiansen-A.Karpov, London 1982 (½-½ in 45); here 20 ... Nxe4 21 Bxe4 Qa5 was simplest,
when White has nothing better than forcing a draw with 22 d5 Nxd5 23 Bxd5 exd5 24 Qxd5.
13 ... Nf6 14 Ne5 Bd7 15 Qe2 Rc8
16 Ne4?
Not like this! It’s hard to understand why Korchnoi invites piece exchanges which ease Black’s
defence. As Hecht notes: “Now the queen’s pawn is fast becoming a problem child.”
Answer: 16 Rfe1, 16 Rfd1 or even 16 h3 are all normal moves, after which White is at least equal.
16 ... Nxe4 17 Qxe4
Exercise: How should Black meet the threat to his b7-pawn?
Answer:
17 ... Bc6!
A standard idea – as we’ve just said, Black benefits from piece exchanges. Karpov explains his
move as follows: “An important subtlety. Black is not afraid that after the double exchange on c6 he
will also be given an isolated pawn. His knight is capable of both securely defending his own pawn,
and attacking the enemy d4-pawn, whereas the functions of the white bishop are restricted.”
I prefer Karpov’s move order to 17 ... Rxc1 18 Rxc1 Bc6, which gives White control of the c-file.
18 Nxc6 Rxc6 19 Rc3
Exercise: How was Karpov intending to answer 19 Rxc6 - ?
Answer: 19 ... bxc6! is a standard idea in IQP structures. The d4-pawn is fixed (d4-d5 will never happen now) and Black can build
up with ... Qb6 and ... Rd8, with possibilities of playing ... c6-c5. Black is certainly no worse here.
In contrast, 19 ... Nxc6? 20 d5 exd5 21 Bxd5 leaves White with the clearly superior minor piece. Black should be able to hold this
position but there is some suffering ahead.
19 ... Qd6 20 g3 Rd8
The game now enters a lengthy manoeuvring phase. Black’s idea is to triple heavy pieces on the d-
file, after which (should White defend the d-pawn with a rook on d1) the ... e6-e5 break might win the
pawn. But he is also happy with piece exchanges. Over the next few moves Karpov reorganizes his
pieces perfectly. It is hard to suggest how Korchnoi can really improve from here – his position is
strategically difficult.
21 Rd1 Rb6 22 Qe1 Qd7 23 Rcd3 Rd6 24 Qe4 Qc6!
Offering a trade of queens which White can’t afford to accept. This is one of the best things about
nursing a long-term advantage (such as an extra pawn or a superior pawn structure) – by offering
unfavourable exchanges, you can drive your opponent’s pieces from strong squares.
25 Qf4 Nd5 26 Qd2 Qb6 27 Bxd5
Exercise: How would you assess this decision from Korchnoi?
Answer: It has to be classed as a significant mistake. As mentioned before, all trades are in Black’s favour. The pure major piece
position which results is extremely difficult for White. Hecht suggests simply 27 a3, when Black must still demonstrate how he plans to
break through – the knight on d5 looks great but it shields the d4-pawn from direct attack.
Kasparov’s explanation for the decision is probably correct: “Short of time, Korchnoi does not
want to undermine the stability of his bishop by moving the a2-pawn, and he decides to parry the
threat of ... Nd5-b4 in the simplest way. The position after 27 a3 Ne7 28 Qf4 Nc6 29 d5 e5 30 Qe3
Nd4 is also advantageous to Black, but possibly this was the lesser evil.”
27 ... Rxd5 28 Rb3
White gains some time on the black queen, but this is purely temporary and actually removes
defenders from the d4-pawn.
Here Karpov makes another fascinating psychological observation: “The sacrifice of the d4-pawn
would give White practical saving chances in the rook endgame. However, it is psychologically
difficult to decide on such a step of one’s own free will. Meanwhile Black increases the pressure.”
Looking forward, it is certainly the case that Korchnoi, by dealing with tactical threats to his d-
pawn, ends up destroying his own position (in particular, by playing f2-f4 which, as will be seen, is
forced in order to prevent ... e6-e5).
I have probably played through this game a dozen times and, frankly, it had never occurred to me
that White could elect to go into the rook and pawn endgame which, while obviously giving Black
serious winning chances, is not trivial to win.
28 ... Qc6 29 Qc3 Qd7 30 f4
Not a move White wants to make, but Black threatened 30 ... e5, exploiting the pin on the d4-pawn.
30 ... b6!
Really nice play – Karpov wants to force Korchnoi’s rook into a ridiculous position on b4.
31 Rb4 b5!?
The immediate 31 ... e5 was also extremely strong.
32 a4
Otherwise Black could have considered 32 ... a5, driving the b4-rook away from the defence of the
IQP.
32 ... bxa4
White will get this pawn back, but loses time and co-ordination while doing so.
33 Qa3 a5 34 Rxa4 Qb5
Black had several strong moves here – this is one of the most direct, attacking the b2-pawn and
playing with ideas of ... Qe2.
35 Rd2
Karpov expected 35 b3, when he can win the b3-pawn with excellent winning chances in the rook
ending, or (as Kasparov notes) continue as in the game with a strong attack.
Answer:
35 ... e5!
The most incisive continuation.
36 fxe5 Rxe5 37 Qa1
37 Rf2 held on for longer, but White’s position remains horrible.
Answer:
37 ... Qe8!!
Absolutely decisive. Black breaks through to the seventh rank and the white major pieces on the a-
file are completely incapable of defending their king.
38 dxe5 Rxd2 39 Rxa5
Or 39 Qe1 Qd8 and the white king faces far too much firepower.
39 ... Qc6 40 Ra8+ Kh7 41 Qb1+ g6 42 Qf1 Qc5+
Of course not 42 ... Qxa8?? 43 Qxf7+, when Black can only escape perpetual check by hanging his
pawns and rook.
43 Kh1 Qd5+ 0-1
Winning the queen and the game.
Game 17
A.Karpov-W.Uhlmann
Madrid 1973
French Defence
1 e4 e6 2 d4 d5 3 Nd2
Karpov’s normal choice in this opening. The sound positions which arise from this continuation –
in contrast to the relative turbulence of the Winawer Variation after 3 Nc3 Bb4 – made this a perfect
fit for Karpov’s style. Similar considerations underpin the selection of this variation by its top
modern adherent, GM Michael Adams.
Of the move 3 Nd2, Karpov himself commented: “The point is that it leads to a small but lasting
advantage for White – and I stress lasting.”
3 ... c5 4 exd5 exd5 5 Ngf3
5 ... Nc6
Against Vaganian at Skopje 1976, Karpov showcased some of his attacking skills. The game
proceeded 5 ... a6 6 dxc5 Bxc5 7 Nb3 Bb6 (while the bishop is more active on this square than on d6,
the disadvantage is that it becomes easier for White to exchange it with Be3) 8 Bd3 Ne7 9 0-0 Nbc6
10 Re1 Bg4 11 c3 h6 12 h3 Bh5 13 Be3 0-0 14 Bxb6 Qxb6 15 Qe2 Rfd8 16 Rad1 a5! (Black starts
his typical queenside counterplay; see move 18 in the main game for another example) 17 Bb1 (17 g4
Bg6 18 Qe3 is a good alternative) 17 ... Bxf3 18 Qxf3 a4 19 Nd4 Qxb2 20 Nxc6 Nxc6 21 Qf5 (21 h4!
appears stronger, preparing Qd3 and h4-h5 with an attack and good compensation for the pawn) 21 ...
g6 22 Qf6 Rd7 23 Bf5!?
Exercise: How should Black deal with the attack?
Answer: It is vitally important (especially when playing world champions!) not to believe your opponent. Trying to play safe is often
equivalent to allowing your opponent to develop his initiative unchecked. Here 23 ... gxf5! looked dangerous, but by eliminating White’s
last minor piece, Black quells the attack. White is forced to bail out with perpetual after 24 Rd3 (24 Qxh6 Ne7 25 Rxe7 Rxe7 26 Qg5+
Kf8 27 Qh6+ Kg8 28 Qg5+ is also a draw) 24 ... f4! 25 Qxf4 f6! (bringing the rook into the defence and saving the game) 26 Qxf6 (26
Rg3+ Rg7 27 Rxg7+ Kxg7 28 Qg4+ is another draw) 26 ... Rg7 27 Qe6+ Kh7 28 Qf5+ Kh8 29 Qf6 Kh7 30 Qf5+ Kh8;
Vaganian instead elected for 23 ... Re7? 24 Rxe7 Nxe7 25 Bd3 (now White’s initiative is hard to
stop) 25 ... Nf5 (25 ... Re8 26 Re1 wins material since 26 ... Kf8? 27 Rxe7 leads to mate) 26 Bxf5
gxf5 27 Re1 Qxa2 28 Qxh6 (Black is technically a pawn up, but he will shed huge amounts of
material seeking to deal with White’s attack) 28 ... a3 29 Qg5+ Kf8 30 Qf6 Kg8 31 Qxf5 Qd2 32 Re7
Rf8 33 Qg4+ Kh7 34 Re5 Qh6 35 Rh5 Ra8 36 Qf5+ Kg7 37 Rxh6 Kxh6 38 Qf6+ Kh7 39 Qxf7+ Kh8
40 Qxb7 and Black resigned.
6 Bb5 Bd6 7 dxc5 Bxc5 8 0-0 Nge7 9 Nb3 Bd6
The problem with 9 ... Bb6 is that the bishop is easily exchanged by 10 Re1 0-0 11 Be3! (more to
the point than 11 Bg5, as in A.Karpov-R.Vaganian, Budapest 1973; 1-0 in 59); for instance, 11 ... Bg4
12 Bxb6 Qxb6 13 Bxc6 Nxc6 14 Qxd5 Nb4 15 Qe4 Bxf3 16 gxf3 and Black had insufficient
compensation for the pawn in A.Karpov-N.Krogius, Russian Championship, Kuibyshev 1970 (1-0 in
54).
10 Bg5
Michael Adams has mostly preferred 10 Re1 0-0 11 Bd3; for example, 11 ... h6 (11 ... Bg4? is
impossible in view of 12 Bxh7+ Kxh7 13 Ng5+, winning a pawn) 12 h3 Nf5 13 c3 Qf6 14 Bc2 Rd8
15 Qd3 g6 16 Qd2 Bf8 17 Qf4 Bg7 (it seems more precise to repeat with 17 ... Bd6! 18 Qd2 Bf8,
after which Adams varied with 19 Qe2 in M.Adams-I.Zugic, FIDE World Cup, Khanty-Mansiysk
2007, drawing in 51 moves) 18 Bd2 g5 19 Qh2 b6 20 Rad1 Ba6 21 Bxf5 Qxf5 was M.Adams-
A.Yusupov, French Team Championship 2005 (1-0 in 32), and here the undermining 22 h4! would
have left White on top.
10 ... 0-0 11 Bh4
Question: What’s the point behind this bishop retreat?
Answer: As Karpov himself explains: “The idea of this move is simple enough. The advantages of White’s position are associated
with the opponent having an isolated d-pawn. To exploit this weakness, he must try to simplify the position, by exchanging at least some
of the minor pieces, and in particular the opponent’s dark-squared bishop which is ‘holding the zone’ around the d5-pawn.”
11 ... Bg4
Similar play occurred in A.Karpov-G.Kuzmin, Leningrad Interzonal 1973: 11 ... Qc7 12 Bg3!
Bxg3 13 hxg3 Bg4 14 Re1 Rad8 15 c3 Qb6 16 Bd3 Ng6 17 Qc2 Bxf3?! (only White will benefit from
the change of structure on the kingside; Mednis’ suggestion of 17 ... Rfe8 is better) 18 gxf3 Rd6 19 f4!
Rfd8 20 a3 h5 21 Kg2 h4 22 Re2 Nf8 23 Nd2 Rh6 24 Nf3 hxg3 (24 ... h3+ may be more tenacious,
though the pawn is at risk of simply dropping off) 25 fxg3 Nd7 26 Rae1 Kf8 (Karpov now devises a
bold plan) 27 g4! (aiming to restrict the h6-rook and gain space on the kingside – except how will
White defend the f4-pawn?) 27 ... Qc7 28 g5 Rh8 29 Kg3!! (that’s how! Karpov is a master of using
his king actively, as may be seen from Game 3, for example – here the king is completely safe and the
white pieces are ideally placed, while Black struggles to find squares for his knights or to connect his
rooks) 29 ... Nc5 30 Bf5 (30 Rh2! is even stronger, forcing 30 ... Rg8, since 30 ... Rxh2 31 Qxh2
leads to a decisive attack for White) 30 ... g6 31 b4 Ne4+ 32 Bxe4 dxe4 33 Qxe4, when Black had no
compensation for the pawn and lost in a few moves (1-0 in 39).
12 Be2 Bh5 13 Re1 Qb6 14 Nfd4 Bg6
Aiming to keep more pieces on the board and make use of the strong square on e4. After 14 ...
Bxe2 15 Rxe2, White’s task is easier.
15 c3 Rfe8 16 Bf1 Be4 17 Bg3 Bxg3 18 hxg3 a5
This is one of Black’s typical resources in these positions – the knight on b3 is a target for a rapid
advance of the black a-pawn.
19 a4!
Stopping Black’s counterplay. This had to be calculated accurately.
19 ... Nxd4
Karpov suggests that 19 ... Rac8 is more precise.
20 Nxd4 Nc6
Not 20 ... Qxb2? 21 Nb5 and the threats of 22 Re2 and 22 Nc7 win material for White.
21 Bb5 Red8 22 g4!
Taking the f5-square under control in anticipation of the forthcoming endgame.
22 ... Nxd4
Exercise: How should White recapture?

Answer:
23 Qxd4!
23 cxd4? is much weaker, since the queen is perfectly placed on b6, attacking d4 and guarding c7.
Moreover, with queens on the board, White will need to be careful about advancing his kingside
pawns.
23 ... Qxd4 24 cxd4 Rac8
Covering e7 with 24 ... Kf8 looks like a better chance.
25 f3 Bg6 26 Re7 b6 27 Rae1 h6 28 Rb7 Rd6 29 Ree7 h5 30 gxh5 Bxh5
Answer:
31 g4!
The most incisive and clearly the strongest. Karpov has calculated that his king will be quite safe
(it is hard to involve the d6-rook in the attack), while Black’s weaknesses on the seventh rank will
prove fatal. 31 Kh2 is a good move too, but gives Black more chances to survive than the game
continuation.
31 ... Bg6 32 f4!
This is the point – the black bishop is driven away from the defence of the f7-pawn.
32 ... Rc1+ 33 Kf2 Rc2+ 34 Ke3 Be4
Or 34 ... Re6+ 35 Rxe6 fxe6 36 Rxb6 and White wins material.
35 Rxf7 Rg6 36 g5
Here 36 Be2! was even stronger: 36 ... Rxb2 37 f5 Rxe2+ (or 37 ... Rg5 38 Kf4) 38 Kxe2 Rxg4 39
f6 gxf6 40 Rfc7 Kh8 41 Rxb6 with an easy win for White, who trades a pair of rooks with Rb8+ and
then wins the a5-pawn.
36 ... Kh7 37 Rfe7 Rxb2
Exercise: How can White involve the last piece in the attack?
Answer:
38 Be8!
Black’s defences collapse.
38 ... Rb3+ 39 Ke2 Rb2+ 40 Ke1 Rd6 41 Rxg7+ Kh8 42 Rge7 1-0
Vladimir Kramnik’s adoption of the Berlin Wall against Garry Kasparov in their match in 2000 is
one of the most highly regarded decisions in world championship history. Kramnik defended this
variation in games 1, 3, 9 and 13, and Kasparov couldn’t make a dent.
In my view, Korchnoi’s defence of the IQP position after 1 e4 e6 2 d4 d5 3 Nd2 c5 4 exd5 exd5
against Karpov was an even greater achievement. Korchnoi played this variation seven times in the
1974 Candidates Final (which, due to Fischer’s default in 1975, was effectively a title match) and
twice in their 1978 World Championship match, drawing all nine games.
Perhaps Korchnoi’s achievement has been overlooked since he didn’t win either match, or because
his selection didn’t impact top-level chess in the same way as Kramnik’s (every round of every elite
tournament these days seems to feature either a Berlin Defence, or its avoidance via 1 d4 which, as
Shirov quipped, is a good way of avoiding the Marshall and Petroff!). However, I would argue that
Korchnoi (in these games, and his subsequent highly successful patronage of the variation) created the
blueprint which a number of successors (in particular Mikhail Gurevich and Bartosz Socko) have
been using for years to make comfortable draws with Black, and caused a fundamental reassessment
of these positions from “slightly better for White” to “very solid for Black”. As Kasparov notes:
“[Karpov’s] matches with Korchnoi (1974 and 1978) showed that it is not at all easy for White to
win against the position with an isolated pawn, and from 1982 Anatoly also began employing 3 Nc3.”
Game 18
A.Karpov-V.Korchnoi
Candidates final (8th matchgame), Moscow 1974
French Defence
1 e4 e6 2 d4 d5 3 Nd2 c5 4 exd5 exd5 5 Ngf3

In the 1978 match in Baguio, Karpov changed tack with 5 Bb5+ Bd7 6 Qe2+, but Korchnoi
remained resilient, making draws in games 16 and 22.
5 ... Nc6 6 Bb5 Bd6 7 dxc5 Bxc5 8 0-0 Nge7 9 Nb3 Bd6
10 c3
In the 18th game of the Candidates match, Karpov opted for 10 Bg5 0-0 11 Re1 Qc7 12 c3 Bg4 13
h3 Bh5 14 Be2 h6 15 Bxe7 Nxe7 16 Nfd4 Bxe2 17 Qxe2 a6 18 Qf3 Rad8 19 Rad1 Rd7 20 Nf5 Nxf5
21 Qxf5 Rfd8 and Korchnoi comfortably held the draw in 42 moves.
10 ... Bg4 11 Nbd4 0-0 12 Qa4
The 4th game was equal after 12 Be2 Re8 13 Re1 a6 14 Bg5 h6 15 Bh4 Qb6 16 Qb3 Bc5 17 Qxb6
Bxb6 18 Bd3 Kf8 19 a3 Nxd4 20 Nxd4 Bxd4 21 cxd4 Nf5 22 f3 Nxh4 23 fxg4 Ng6, and a draw was
agreed in 45 moves.
12 ... Bh5
The 16th game saw 12 ... Qd7 13 Be3 a6 14 Be2 Nxd4 15 Qxd4 Nc6 16 Qd2 Rfe8 17 Rad1 Rad8
18 Bb6 Bc7 19 Bxc7 Qxc7 20 Rfe1 h6 21 h3 Bf5 22 Bf1 Rxe1 23 Qxe1 Qb6 24 Rd2 Be4, when
White was a long way from demonstrating any weakness of the d5-pawn and a draw was agreed in 67
moves.
13 Re1
In the 10th game Karpov tried 13 Bd3 h6 14 Be3 a6 15 Rfe1 Qc7 16 h3 Na5 17 Nh4 Nc4 18 Qc2
Nxe3 19 Rxe3 Bh2+ 20 Kh1 Bf4 21 Ree1 Bg5 22 Nhf5 Nxf5 23 Nxf5 Bg6 24 Nd4 Bxd3 25 Qxd3
Rfe8 with equality and a draw in 58 moves. Four games later Korchnoi diverged, playing 13 ... Bc5
14 Re1 h6 15 Be3 Bb6 16 h3 Qd6 17 Be2 Rfe8 18 Rad1 Qf6 19 Nh2 Bxe2 20 Rxe2 Nxd4 21 Bxd4
Qc6 22 Qxc6 bxc6 with a solid position and a draw in 30 moves.
The 12th game varied again with 13 Be3 Qc7 14 h3 Na5 15 Bd3 Nc4 16 Nb5 Qd7 17 Bxc4 dxc4
18 Rfd1 Nf5 19 Qxc4 Bxf3 20 gxf3 Nxe3 21 fxe3 Qxh3 22 Nxd6 Qg3+ 23 Kf1 Qxf3+ 24 Ke1 Qg3+
and a draw was agreed.
13 ... Qc7 14 h3 Bg6 15 Bg5 a6 16 Bf1 h6 17 Bxe7 Nxe7 18 Rad1 Nc6 19 Bd3
Exercise: How can Black generate counterplay?
Answer:
19 ... Bh5!
Korchnoi gives Karpov an unpleasant choice. The pin on the h5-d1 diagonal combines with the
central pressure to create a threat of successive captures on d4 and f3, ruining White’s pawn
structure.
20 g4 Bg6
Question: Hasn’t Black just lost two tempi?
Answer: No, since he has extracted a valuable concession in the form of g2-g4. While this seizes kingside space, it weakens the dark
squares around the white king (don’t forget that White has already traded his dark-squared bishop), creating some long-term counterplay.
21 Qc2
Question: Why didn’t Karpov double Black’s pawns with 21 Bxg6 fxg6 - ?
Answer: This decision would have been double-edged even with a white pawn on g2, since the pawn on g6 covers some useful
squares (in particular f5) and gives Black a half-open f-file for his rooks (the ... Rxf3 exchange sacrifice in the French is just as common
as its ... Rxc3 equivalent in the Sicilian). With White having played g2-g4, these concerns become even more compelling. The knight on
f3 has no pawn cover and is the only thing blocking the enemy heavy pieces from the pawn on f2. Note that Black can also use the
structural change to cover his d5-pawn comfortably with ... Qf7, which also increases his pressure on the f-file.
21 ... Bxd3 22 Qxd3
Exercise: How should Black activate his rooks?
Answer:
22 ... Rad8! 23 Re2 Rfe8!
Rooks to the centre. Sometimes chess is an easy game.
24 Nf5 Rxe2 25 Qxe2 Bf4
The movement of the white g-pawn makes itself felt. The bishop now occupies a powerful and
unassailable outpost on f4.
26 Re1 g6 27 Ne7+ Nxe7 28 Qxe7
Exercise: Should Black trade queens?
Answer:
28 ... Qb6!
From a practical point of view, I think this is clearly the best decision. The white king is arguably
slightly weaker than its black colleague, which suggests that a trade of queens would be in White’s
favour in general.
Houdini considers that 28 ... Rc8 29 Qxc7 Rxc7 would also hold comfortably; Black has good pieces and only one weakness, so the
position is definitely tenable, though Karpov would doubtless find a way to play on for a few dozen moves.
Instead, 28 ... Qxe7? would be a horrible move, activating the white rook for no reason.
29 Kg2 Kg7 30 Rd1 Bd6 31 Qe2 Bc7 32 Rd3 Qe6 33 Qd1 Bb6 34 Rd2 Qe4 35 b3 Rd6 36 c4 h5
37 Rxd5 Rxd5 38 Qxd5 Qxd5 39 cxd5 hxg4 40 hxg4 Kf6
Answer: Despite White’s extra pawn, the active black king and limited material ensures an easy draw.
41 Kf1 Ke7 42 Nd2 Bc7 43 Ne4 f5 44 gxf5 gxf5 45 Nc5 Kd6 46 Nxb7+ Kxd5 47 b4 Kc4 48 Nc5 Bb6 49 Nxa6 Kb5 50
Nc5 Kxb4 51 Nb3 Ka3 ½-½
Game 19
A.Karpov-A.Beliavsky
Moscow 1981
1 d4 Nf6 2 c4 e6 3 Nf3 d5 4 Nc3 Be7 5 Bf4 0-0 6 e3

This variation often leads to IQP positions, since Black typically aims for ... c7-c5, either
immediately or after ... Nbd7.
6 ... c5 7 dxc5 Nc6
A straightforward continuation.
8 cxd5
Karpov has also produced a model game after 8 Qc2 Qa5 9 a3 Bxc5 10 Rd1 Be7 11 Nd2 Bd7 12
Be2 Rfc8 13 0-0 Qd8 14 cxd5 exd5 15 Nf3 h6 16 Ne5 Be6 17 Nxc6 (trading a useful black piece) 17
... Rxc6 (Black can’t recapture with 17 ... bxc6 in view of 18 Ba6, winning the exchange) 18 Bf3 Qb6
19 Be5 Ne4 20 Qe2 Nxc3 21 Bxc3 (this position is quite unpleasant for Black – the knights, which
can often create tricks in favour of the IQP-holder, are off the board and White’s bishops are more
active, while the standard plan of tripling on the d-file is very effective) 21 ... Rd8 (21 ... Bxa3? 22
Bxg7! is excellent for White) 22 Rd3 Rcd6 23 Rfd1 R6d7 24 R1d2 Qb5 25 Qd1
(this formation with a queen behind doubled rooks on an open file is known as “Alekhine’s Gun”,
as confirmed by Wikipedia, which informs me that the term derived from a game between Alekhine
and Nimzowitsch at San Remo 1930) 25 ... b6 26 g3 (the straightforward 26 e4? clearly doesn’t work
since the rook on d7 is protected, so Karpov manoeuvres, gradually gaining kingside space) 26 ... Bf8
27 Bg2 Be7 28 Qh5 a6 29 h3 Qc6 30 Kh2 a5 31 f4 f6 32 Qd1 Qb5 33 g4 g5? (Black would be better
off sitting tight with 33 ... Bf7, though his position remains deeply unpleasant; importantly, 34 e4? is
not convincing since it weakens the f4-pawn, as is apparent after 34 ... Qc4 or 34 ... Bd6) 34 Kh1! (so
that there will be no checks on the h2-b8 diagonal) 34 ... Qc6 35 f5 Bf7 36 e4 (the position might
appear to be so blocked that Black can harbour hopes of creating a fortress, even after dropping the
d5-pawn; however, he remains with too many weaknesses to hold, in particular the e6-square and the
f6-pawn) 36 ... Kg7 37 exd5 Qc7
Exercise: How can White make progress?

Answer: 38 Re2! (with two tactical ideas, one of which appears in the game) 38 ... b5? (38 ... Bd6 is more tenacious, but after 39
Re6! Bxe6 40 fxe6, I don’t see how White can be prevented from strolling in on the kingside light squares) 39 Rxe7! Rxe7 40 d6 Qc4 41
b3! and Black resigned in A.Karpov-B.Spassky, Montreal 1979.
8 ... exd5 9 Be2 Bxc5 10 0-0 Be6 11 Rc1 Bb6 12 Qa4 Bd7 13 Rfd1 Nd4
Beliavsky wins the two bishops with a straightforward tactical shot. Nevertheless, simplification
generally favours the side playing against the IQP, so this move does not completely solve Black’s
problems.
14 Qb4 Nxe2+ 15 Nxe2 Bg4 16 Ned4 Ne4 17 h3 Be6 18 Rc2 Rc8 19 Rdc1 Rxc2 20 Rxc2 h6
Answer:
21 a4!
All of White’s pieces are well placed and difficult to improve, so he generates pressure on the
queenside. Black now needs to worry about the possibility of a4-a5, exposing the b7-pawn.
21 ... Re8?!
A careless move by Beliavsky. The conservative 21 ... Bc8 was preferable.
Answer:
22 Nxe6!
The immediate 22 a5 Bxa5 23 Qxb7 allows 23 ... Qb6 with good drawing chances.
22 ... fxe6
22 ... Rxe6 23 a5 is no better.
23 a5! Bxa5 24 Qxb7
White has transformed his advantages. Although the d5-pawn is now secure, White’s pawn
structure is still superior (Black has three “pawn islands” as opposed to White’s two) and his pieces
are all extremely active, while Black’s kingside has been exposed.
24 ... Bb6 25 Ne5
Black is under severe pressure.
25 ... Nd6 26 Qc6 Re7 27 Qa4 Rc7 28 Rxc7 Qxc7
Answer:
29 Nd3!!
A brilliant move. A decent rule of thumb in chess is that queen and knight tend to co-operate better
than queen and bishop, and they form a very dangerous attacking partnership. This applies
particularly in a position like this, where Black’s kingside has been weakened. Karpov aims simply
to trade off the knight on d6. His own knight had to retreat to d3 to prevent ... Qc1+.
29 ... Qe7
Beliavsky could have tried to avoid the dreaded attacking partnership with 29 ... Qc4, but after 30
Qd7 Qxd3 31 Qxe6+ Kh7 32 Bxd6, White is a pawn up with good winning chances.
30 Bxd6!
Simple chess.
30 ... Qxd6 31 Qe8+ Kh7 32 Nf4 e5
Black had more tenacious defences, but everything lost at least a pawn without dampening White’s
initiative.
33 Ne6 Bc5 34 Qf7 1-0
Black resigned, since 34 ... Qe7 35 Qf5+ Kg8 36 Qxe5 would win further material.
Game 20
A.Bisguier-A.Karpov
Skopje Olympiad 1972
Grünfeld Defence
1 c4 c5 2 Nc3 g6
In the closed openings Karpov normally prefers classical set-ups with Black based on ... e7-e6.
However, White’s second move means he can’t now achieve a Maróczy Bind, so Black can play ...
g7-g6 in relatively favourable circumstances.
3 Nf3 Bg7 4 e3
After 4 d4 cxd4 5 Nxd4 Nc6, White doesn’t have time for e2-e4 before playing Be3.
4 ... Nf6 5 d4 0-0
The move order 5 ... cxd4 6 exd4 d5 has been more fashionable recently, possibly because castling
means that Black needs to be prepared for Benoni positions after 6 d5.
6 Be2 cxd4 7 exd4 d5 8 0-0 Nc6
Question: Ever seen this position before?
Answer: We have reached, with reversed colours and an extra tempo for White, a Tarrasch Defence (1 d4 d5 2 c4 e6 3 Nf3 c5 4
cxd5 exd5 5 g3 Nc6 6 Bg2 Nf6 7 0-0 Be7 8 Nc3 0-0). As I noted in my book on the Tarrasch, I don’t believe that White has any
advantage after 8 ... Nc6, although it is playable for both sides. Bisguier rapidly gets outplayed from here.
9 h3 Bf5 10 Be3 dxc4 11 Bxc4 Rc8 12 Be2 Be6
There is nothing wrong with the immediate 12 ... Nd5, but perhaps Karpov wanted to avoid 13
Qb3.
13 Qd2 Qa5 14 Bh6
Question: What is your opinion of this move?
Answer: Trading dark-squared bishops doesn’t weaken the black king as much as it weakens the d4-pawn. However, White was
already running short of constructive options, since Black has a perfect anti-IQP set-up.
14 ... Rfd8!
An argument could also be made for 14 ... Bxh6, which draws the white queen away from the
defence of the d4- and b2-pawns, at the cost of allowing White some extra activity on the kingside.
This is an assessment which must often be made after Bh3 or ... Bh6. In this particular position, while
the engines confirm that White doesn’t have a realistic attack after 14 ... Bxh6, I think Karpov’s
choice is more practical, not allowing White any unnecessary activity.
15 Bxg7 Kxg7 16 Rfd1
Answer:
16 ... Rd6!
Preparing to double on the d-file and lay siege to the d4-pawn. 16 ... a6, simply taking control of
the b5-square, was a good alternative.
17 Qe3 Rcd8 18 a3
Exercise: What is White’s idea and how should Black respond?
Answer: White’s last move took control of the b4-square (thus preventing the standard ... Nb4-d5 manoeuvre) and he might be
planning b2-b4 to gain some queenside space. I am frankly unsure about whether this advance would improve White’s position, but
Karpov puts a stop to it while improving the position of his pieces.
18 ... Bb3!
Now White faces problems defending the d4-pawn, while the black rook will be extremely active
on e6.
19 Rd2
19 Re1 Re6 is also better for Black, who has the more active pieces and the superior structure.
19 ... Re6 20 Qf4
Here 20 Qd3 is more solid.
Exercise: How should Black respond in that case?
Answer: The tactically-astute Karpov would certainly have spotted 20 ... Nb4! 21 Qb5 Qxb5 22 Bxb5 Nbd5 with a solid advantage
for Black.
20 ... Nd5 21 Nxd5 Rxd5!
Now both black rooks are extremely active, both on the central files and controlling key squares on
White’s fifth rank, with an option to swing over to the kingside in some variations. In fact, Black has
an immediate threat of 22 ... Rf5 and 23 ... Rxf3!, exploiting the undefended rook on d2.
22 g4 g5 23 Qg3 Rf6 24 Bd1 Bc4
24 ... Rb5! is even stronger, when White can’t capture on b3 since the rook would come into play
along the sixth rank.
25 b3
Exercise: Black has an excellent tactical opportunity here. See if you can spot it.
25 ... Ba6
Answer: Karpov had the shot 25 ... Nxd4! 26 Nxd4 Rxd4, when taking on d4 leads to disaster after 26 ... Qe1+ and otherwise Black
comes out a lot of material ahead. Not 25 ... Rxf3? 26 Bxf3 Qxd1 27 dxc4 Rxd4, since 28 Bxc6! bxc6 29 Qe5+ is fine for White.
26 b4 Qd8
Black’s pieces have temporarily been driven back, but all the problems in White’s position
remain.
27 Bb3
Exercise: How can Black cash in on his advantage?
Answer: In the same way as before.

27 ... Nxd4! 28 Rxd4
28 Bxd5 Nxf3+ 29 Bxf3 Qxd2 leaves Black a healthy pawn up.
28 ... Rxd4 29 Nxg5?
Desperation. 29 Nxd4 was necessary, even if 29 ... Qxd4 leaves Black a pawn up with excellent
chances.
29 ... Rd3 30 Qh4 h6 31 Nxf7
Exercise: How did Karpov finish the game?
Answer:
31 ... Qd4!
Decisive. Karpov was highly unlikely to fall for 31 ... Rxf7?? 32 Bxf7 Kxf7 33 Qh5+, when White
forces perpetual check or regains his piece with b4-b5. The immediate 33 b5 is also good.
32 Re1 Rxh3! 0-1
A cute deflection of the white queen from its defence of f2.
Game 21
A.Karpov-V.Hort
Budapest 1973
French Defence
1 e4 e6 2 d4 d5 3 Nd2 Nf6 4 e5 Nfd7 5 c3 c5 6 Bd3 Nc6 7 Ne2 Qb6

Karpov also enjoyed success after 7 ... cxd4 8 cxd4 f6 9 exf6 Nxf6 10 Nf3 Bd6 11 0-0 Qc7 (11 ...
0-0 is the main alternative, while 11 ... Qb6 transposes to the main game) 12 Nc3 a6 13 Bg5 0-0 14
Bh4 (aiming for the trade of dark-squared bishops)
a) 14 ... Bf4?! 15 Re1 Kh8 16 Bg3 Nh5 17 Ne2! Nxg3 18 Nxf4 Ne4 19 g3 left White with a clear
strategic advantage, and playing for complications with 19 ... Rxf4?! 20 gxf4 Qxf4 21 Bxe4 dxe4 22
Ne5! only accelerated the end in A.Karpov-De.Chen, Hanover 1983 (1-0 in 27).
b) 14 ... Bd7 15 Re1 Rae8 16 Rc1 Qb8 17 Bb1 Kh8 18 a3 Bf4 19 Rc2 Ng4 20 h3 Nh6 was
A.Karpov-J.Ostos, Malta Olympiad 1980 (1-0 in 46), and here 21 Na4, aiming for c5, would have
been strong.
c) 14 ... Nh5 15 Re1 g6 16 Rc1 Qg7 17 Bf1 Bd7 18 Bg5 h6 19 Be3 Rf7 20 g3 Raf8 21 Bg2 Nf6 22
h3 g5, when Black had played well and enjoyed an active position but was later outplayed in
A.Karpov-A.J.Mestel, London 1984 (1-0 in 61).
8 Nf3 cxd4 9 cxd4 f6 10 exf6 Nxf6 11 0-0 Bd6 12 Nc3 0-0 13 Be3 Qd8 14 Bg5 Bd7 15 Re1
Qb8 16 Bh4!
Trading dark-squared bishops is in White’s favour.
16 ... a6 17 Rc1 b5 18 Bb1 Bf4 19 Bg3 Bxg3 20 hxg3 Qb6 21 Ne2 Rae8 22 Nf4 Nxd4 23 Qxd4
Karpov heads for a clear and strategically superior endgame. As he noted later, 23 Nxd4! was
even stronger: 23 ... e5 24 Nxd5 Nxd5 (24 ... Qxd4 25 Nc7 Rc8 26 Qxd4 exd4 27 Nxa6 wins a pawn)
25 Nf3 and the e-pawn will soon drop off.
23 ... Qxd4 24 Nxd4 e5 25 Nfe6!
25 Nxd5 exd4 was also possible, but I prefer Karpov’s choice which keeps the pawn blockaded
on d5, where it limits Black’s bishop.
25 ... Bxe6 26 Rxe5 Bd7 27 Rxe8 Rxe8 28 f3 Rc8 29 Rxc8+ Bxc8
Exercise: How would you assess this endgame?
Answer: This endgame is pure torture for Black, especially against Karpov. Black has an IQP and has simplified into a very poor
version – he has retained his worst piece (namely the light-squared bishop, which is limited by his pawns on a6, b5 and d5) while trading
all the pieces (major pieces and his dark-squared bishop) that could have created counterplay. The only thing in his favour is the presence
of the knights, since the pure light-squared bishop endgame (or, even worse, a white knight against Black’s light-squared bishop) would
likely be even more difficult.
30 Kf2 Kf7 31 Ke3 Ke7 32 b4
Fixing the pawns on light squares is good technique.
32 ... g6 33 g4 Nd7 34 f4 Nf8 35 g5
The white pawns do an excellent job of restricting Black’s pieces, and now all of his pawns are
stuck on light squares.
35 ... Kd6 36 Kf3 Ne6
Exercise: How would you assess the decision to trade knights?
Answer: This move was given very harsh criticism by Mednis, who writes as follows: “Offering to exchange knights is strategic
suicide and can only be explained by thoughtlessness brought about by extreme time pressure. White’s bishop is much superior to Black’s
so that only the knights are a neutralizing factor. White’s knight looks better at d4 but is of no immediate danger to Black. And Black’s
knight helps keep out White’s king. The endgame should be won for White, but it is up to White to show how after 36 ... Bd7.”
37 Nxe6 Bxe6 38 Ke3 Bg4 39 Bd3 Be6 40 Kd4
Exercise: And how would you assess this position?
Answer: Again, I can’t improve on Mednis’ explanation: “Karpov has achieved a classical won endgame with equal material and
same colour bishops. The winning elements are as follows: White’s king is better placed on the fourth rank than is Black’s on the third. In
fact, Black’s king cannot move because White would infiltrate via either e5 or c5. All of Black’s pawns are fixed on the same colour as
White’s bishop and are thus readily attackable. The primary target is Black’s pawn on d4; secondary targets are a6 and h7. White will
get a passed pawn on the f-file. Note that White’s doubled pawn is no handicap here since the g2- and g5-pawns hold Black’s g6- and
h7-pawns (I would say that simply the g5-pawn holds back both the g6- and h7-pawns – S.C.); thus White’s f-pawn is effectively an
extra pawn. As is evident, Black’s king and pawns cannot afford to move, and White’s winning technique consists of out-tempoing
Black’s bishop.”
40 ... Bg4 41 Bc2 Be6
41 ... Be2, intending 42 Bb3 Bc4, can be comfortably met by 42 f5!, when Black loses a pawn or
falls into zugzwang after 42 ... Bh5 43 a3!; for instance, 43 ... gxf5 44 Bxf5 Bg6 45 g4.
42 Bb3 Bf7 43 Bd1!
An elegant win. The black bishop will be unable to cover the d5-pawn and the h3-c8 diagonal at
the same time.
43 ... Be6
Otherwise 44 Bg4 follows.
44 Bf3 Bf7 45 Bg4 1-0
Immediate resignation is a bit premature, but the position is lost; for instance, 45 ... Bg8 46 Bc8
Kc6 47 a3 (not 47 Bxa6?? Be6, trapping the white bishop) 47 ... Kb6 48 Ke5 Kc6 49 g3 Kb6 50 Bg4
Kc7 51 Bf3 and the d-pawn drops.
Game 22
A.Karpov-A.Yusupov
Candidates semi-final (7th matchgame), London 1989
1 d4 Nf6 2 c4 e6 3 Nf3 d5 4 Nc3 Be7 5 Bg5

The position after 5 cxd5 exd5 6 Bg5 c6 7 Qc2 arose in two games between Karpov and Yusupov.
7 ... g6 (Yusupov later varied with 7 ... Na6 8 e3 Nb4 9 Qd1 Bf5 10 Rc1 a5 11 Be2 0-0 12 0-0
Nd7 13 Bxe7 Qxe7 and Black had a very solid position in A.Karpov-A.Yusupov, Rotterdam 1989;
still 1-0 in 58) 8 e4!? (the position after 8 e3 Bf5 9 Bd3 Bxd3 10 Qxd3 is known to be fine for Black,
while avoiding the exchange gives White nothing either) 8 ... Nxe4?! (8 ... dxe4 is more solid: after 9
Bxf6 Bxf6 10 Qxe4+ Be6 11 Bc4 Qe7 12 Bxe6 Qxe6 13 Qxe6+ fxe6, Karpov had himself held the
endgame quite comfortably against Eingorn in the previous round; ½-½ in 34) 9 Bxe7 Kxe7 (forced,
since 9 ... Qxe7 loses a pawn to the typical 10 Nxd5! shot) 10 Nxe4 dxe4 11 Qxe4+ Be6 12 Bc4
Qa5+ 13 Kf1 Qf5 14 Qe3 Nd7 15 Re1 Rae8 16 d5!
Karpov wonderfully explains the reasoning behind this sacrifice: “White obviously has numerous
tactical possibilities, but I was unable to unite them in a logical chain of moves. It was clear to me
that for the development of the initiative I had to sacrifice my central pawn, but what next? And
suddenly the realization came to me: the weak dark squares in Black’s position are defended by his
knight. And then I immediately found a manoeuvre, the aim of which was to eliminate this knight.”
After 16 ... cxd5 17 Bb5 a6 18 Qa3+ Kd8 19 Qa5+ Ke7 20 Qb4+ Kf6 21 Qd4+ Ke7 22 Bd3,
White had an extremely strong initiative in A.Karpov-A.Yusupov, USSR Championship, Moscow
1988 (1-0 in 43).
5 ... 0-0 6 e3 h6 7 Bh4 Ne4 8 Bxe7 Qxe7
9 Rc1
In game three of the match, Karpov opted for 9 Qc2, but this gave him no advantage after 9 ... Nxc3
10 Qxc3 dxc4 11 Bxc4 b6 12 0-0 Bb7 13 Be2 c5 14 dxc5 Rc8, and he was even worse before making
a draw in 43 moves.
The fifth game continued 9 cxd5 Nxc3 10 bxc3 exd5 11 Qb3 Rd8 12 c4 dxc4 13 Bxc4 Nc6 14 Qc3
Bg4 15 0-0 Bxf3 16 gxf3 Qf6 17 Be2 Rac8! and Black had excellent play. Yusupov was on top and
even winning at one point, but again the game ended in a draw in 46 moves.
9 ... c6 10 Bd3 Nxc3 11 Rxc3 dxc4 12 Bxc4 Nd7 13 0-0 e5
The main alternative is 13 ... b6, as played by Anand in the decisive 12th game of his 2010 World
Championship match with Topalov.
14 Bb3 exd4
Exercise: How should White recapture?

Answer:
15 exd4!
Definitely the strongest move. I can’t improve on Alex Baburin’s excellent explanation: “Here I
should like to take a break and talk about situations when it is objectively necessary to create the
isolated d-pawn in your own camp. I know many club players who would not even consider 15 exd4
here, on the simple grounds that it leads to the isolation of a pawn and therefore it ‘spoils’ the pawn
formation. Such a ‘static’ approach would be quite wrong here, as the dynamic advantages which the
text gives White right now are worth a lot more than some potential weakness of the pawn. Indeed,
the text is much better than the solid but rather drawish 15 Qxd4 – White opens the e-file, gains
control over the c5- and e5-squares and clears the third rank for the c3-rook. The latter as we will
see, is going to play an important part in the game.”
15 ... Nf6
In his excellent repertoire book on the Queen’s Gambit Declined, John Cox notes that Black can do
much better than allowing Ne5, either with his main recommendation of 14 ... Re8!? or, here, by
playing 15 ... Rd8 16 Re1 Qd6, when Black plans ... Nf8 followed by ... Be6 or ... Ne6.
16 Re1 Qd6 17 Ne5 Nd5 18 Rg3 Bf5
18 ... Be6 is more solid, as in L.Portisch-A.Bykhovsky, Suzdal 2011 (½-½ in 26).
19 Qh5 Bh7
Black aims to prevent Bc2 with this unusual bishop manoeuvre. On the other hand, he has
abandoned control of the key g4-square and, as will be seen, the bishop on h7 can become a target in
some variations.
20 Qg4 g5 21 h4 f6 22 hxg5 hxg5 23 f4
23 Rh3! is even stronger, based on the tactical point 23 ... fxe5 (or 23 ... Qe7 24 Ree3! fxe5 25
Rxe5 Qc7 26 Rxg5+ Kh8 27 Rxh7+ Qxh7 28 Rh5, winning the queen and the game) 24 Rxh7! Kxh7
(even the relatively best 24 ... Qf6 loses: 25 Rxe5 Qxf2+ 26 Kh1 Qf1+ 27 Kh2 Qf4+ 28 Qxf4 Rxf4 29
Rxb7 with an extra pawn and a winning position) 25 Bc2+ Kg7 26 Qxg5+ Kf7 27 dxe5 Qe6 28 Bf5
and Black has to give up his queen to avoid mate.
23 ... Rae8 24 fxg5 fxe5
Igor Zaitsev gives the beautiful variation 24 ... Bf5 (trying the keep the g-file closed) 25 gxf6+!!
Bxg4 26 Rxg4+ Kh8 27 Nf7+ Rxf7 28 Rxe8+ Rf8 29 f7 Nf6 30 Rxf8+ Qxf8 31 Rg8+ Nxg8 32
fxg8Q+ Qxg8 33 Bxg8 with a winning king and pawn endgame.
25 g6 Bxg6
Answer:
26 dxe5!
Only this intermezzo gives White an advantage. 26 Qxg6+? Qxg6 27 Rxg6+ Kf7 28 Rg4 exd4 29
Rxe8 Rxe8 30 Rxd4 is just equal.
26 ... Qe6
Exercise: How would you proceed with White?

27 Bxd5?!
Answer: Karpov goes for the rook endgame, which is perhaps a matter of taste. However, an old chess saying teaches us that “all
rook and pawn endgames are drawn” and, while it is fair to say that Karpov’s endgame technique is exceptional, I’m not sure that the
position he goes for would have given him serious winning chances against best defence.
It seems to me that 27 Qxg6+! Qxg6 28 Rxg6+ was a better alternative, when the bishop is at least
as strong as the knight and White has excellent winning chances. It is important to note that White
retains the option of taking on d5 whenever he wants, but might also use his bishop to attack the black
king, or to seek to control key squares such as f7 or h7.
27 ... cxd5 28 Qxg6+ Qxg6 29 Rxg6+
Exercise: Where should Black move his king?
Answer:
29 ... Kh7?
Moving the king away from the centre is rarely good strategy. Perhaps Yusupov was concerned
about checks, but he should have kept his king more active.
Instead, 29 ... Kf7 30 Rd6 Rd8 31 Rf1+ Ke7 leaves all the black pieces ideally placed. Yusupov
assessed 31 Rf1+ as winning for White, presumably because he picks up a second pawn with 32 Rxf8
Rxf8 33 Rxd5. However, after 33 ... Ke6!, White can’t hold on to the e5-pawn without allowing
serious counterplay; for example, 34 Rb5 Rf7 35 g3 a6 36 Ra5 (36 Rc5 Rd7 is similar) 36 ... Rc7
and Black appears to have enough activity to draw. White’s best may be 34 Rd6+ Kxe5 35 Rd7 Rb8,
when the black rook is temporarily passive; but once the queenside pawns are advanced, it may
become active again, while the black king is already extremely active. This looks like good drawing
chances.
30 Rd6 Rc8 31 Re3 Rc2 32 Rd7+ Kg6 33 Rxb7 Re8
Compared with the variation in the previous note, Black has still lost a second pawn but his pieces
are less active. The fact that all rooks remain on the board increases White’s winning chances – the
black king might become an object of attack, for instance.
34 a3 d4 35 Rd3 Rxe5 36 Rxd4 Rg5
Exercise: Work out a way for White to deal with the threat to his g2-pawn.
Answer:
37 Rd6+!
37 g4? would be a horrible mistake in view of 37 ... Re5!, when the threat of mate, and the idea of
doubling rooks on the second rank, gives Black sufficient counterplay to make a comfortable draw.
37 ... Kh5 38 Rh7+ Kg4 39 Rd4+ Kf5
A player of Yusupov’s calibre is unlikely to fall for 39 ... Kg3?? 40 Rh3 mate.
40 Rd5+ Kg6 41 Rg7+ Kxg7 42 Rxg5+ Kf6 43 Rb5
This position is an easy technical win, and Karpov makes light work of the final demonstration.
43 ... a6 44 Rb6+ Ke7 45 Kh2 Kd7 46 Kh3 Kc7 47 Rb3 Kd6 48 g4 Ke5 49 Kh4 Kf6 50 Rb6+
Kg7 51 Kh5 a5 52 Rb7+ Kg8 53 a4 1-0
The Minority Attack
The structure which is examined in the next two games is called the Carlsbad. Nevertheless, I have
dubbed this section “The Minority Attack”, since the plan used is applicable in any position with
asymmetrical pawn majorities (for instance, in the Open Sicilian Black can play a minority attack by
pushing his a- and b-pawns down the board). For those unfamiliar with the concept, it might appear
strange. We know that we shouldn’t play where we are weaker, and surely we are weaker on the side
of the board where we have fewer pawns than the opponent? But in fact the minority attack aims to
gain the initiative on precisely this side of the board and, in particular, to create a pawn weakness
which can be attacked later.
Game 23
A.Karpov-L.Ljubojevic
Linares 1989
1 d4 Nf6 2 c4 e6 3 Nc3 d5
Ljubojevic was best known for his dynamic play, so I have no idea what he was thinking in going
for a solid, passive position against Karpov.
4 cxd5!
In the circumstances, undoubtedly the correct choice. Refraining from the sharp lines with Bf4, or
the main lines such as the Tartakower/Makogonov-Bondarevsky (where Rafael Vaganian and Nigel
Short have scored beautiful attacking wins with Black), Karpov takes some tension out of the position
and steers towards a dour strategic struggle.
4 ... exd5 5 Bg5 c6 6 e3
6 ... Nbd7
Recent super-GM praxis has come back to 6 ... Bf5, a favourite of Nigel Short. (White can avoid
this line with 6 Qc2, but only at the cost of giving Black some additional options, as in the similar
position with 7 Qc2 discussed in the notes to Game 22.) The endgame after 7 Qf3 Bg6 8 Bxf6 Qxf6 9
Qxf6 gxf6 looks depressing for Black in view of his doubled f-pawns and chronically weak f5-
square, but in fact his position has proven quite resilient. He has the bishop pair in case the game
opens up (so executing a minority attack with b4-b5 is quite double-edged for White) and in the
meantime has a few good defensive set-ups (for instance, playing his knight to d6). In fact Karpov
defended this position effortlessly with Black as a young master: 10 Nge2 Bd6 11 g3 Nd7 12 Nf4 Bf5
13 Be2 0-0-0 14 Bh5 Be6 15 0-0-0 Nb6 16 Nce2 Kb8 17 Kb1 Nc8 18 Rc1 Ne7 19 a3 b6 20 Rhd1
Kb7 21 Bf3 Nc8 22 Ka1 and a draw was agreed in A.Shakarov-A.Karpov, USSR U-18
Championship, Moscow 1966.
7 Bd3 Be7 8 Qc2 0-0 9 Nf3
Karpov’s preferred method of handling the Carlsbad structure. By contrast, Kasparov championed
9 Nge2, followed by 0-0 and f2-f3, preparing play in the centre with e3-e4.
9 ... Re8 10 0-0 Nf8
11 Rab1
In a simultaneous game three years later, Karpov excellently demonstrated how effective the
minority attack can be against planless play: 11 Bxf6!? Bxf6 12 b4 (White has given up his dark-
squared bishop in order to accelerate his queenside play) 12 ... h6?! (not really a standard idea –
Black wants to free his knight from f8, but accomplishes no other strategic objectives; Black’s dark-
squared bishop is an asset, which is why I like 12 ... Be7!, intending 13 b5 Bd6, when the bishop is
perfectly placed, eying the white kingside while covering the c5-square – I think chances are
balanced in this position; otherwise, 12 ... g6!? would be a more standard method of resolving the
kingside pressure than that in the game, aiming to play ... Nf8-e6-g7 and ... Bf5, developing Black’s
worst piece – compare this with the fate of the c8-bishop in what follows!) 13 b5 Bd7 14 bxc6 bxc6
15 Na4 Re7 16 Nc5 Be8
(the bishop on e8 covers the c6-pawn, but this is the end of the good news as Black is completely
passive; I should note that, had he traded the a-pawns – for instance, by flicking in 12 ... a6 13 a4 and
then proceeding as in the game – this method of passive defence would have considerably more merit,
since the c6-pawn would be Black’s only weakness; as it is, the a7-pawn is a tempting target for
attack) 17 Rab1 Ne6 18 Nxe6 Rxe6 19 Bh7+! (a standard intermezzo before putting the bishop on f5 –
the black king stands worse on h8 than on g8) 19 ... Kh8 20 Bf5 Re7 21 Ne1! (another instructive
regrouping, bringing the knight to its ideal square on d3 from which it controls, and can sometimes
occupy, c5, as well as looking at other important squares such as e5 and b4) 21 ... Rb8 22 Nd3 g6 23
Bg4 Kg7 24 Nc5 h5 25 Bh3 Rb6 26 Rb3! (breaking Black’s resistance on the b-file: White will
double – or even triple – there, forcing Black to release the tension by taking on b3, so Black decides
to capture straight away) 26 ... Rxb3 27 Qxb3
(opening an avenue of attack to the a-pawn with 27 axb3 was tempting, but I prefer Karpov’s move since he is penetrating down the
b-file) 27 ... Qc7 28 g3 Qd6 29 Rb1 g5 (Black has no kingside attack, but it’s hard to offer him any good advice) 30 Bf5 g4 31 a4!
(bringing up the a-pawn so that the fall of the a7-pawn will be fatal; it can also help control the crucial b7-square) 31 ... Bg5 32 a5 Qf6
33 Qd3 Qd6 34 a6 Rc7 35 Rb8 Qe7 36 Kg2! (typical Karpov; there is no need to rush with Ra8 – first he takes a time out to protect his
king) 36 ... Kh6 37 h3! (further safeguarding the king and leaving Black with additional weaknesses) 37 ... gxh3+ 38 Bxh3 h4 39 g4 Bf6
40 Ra8 Kg7 and Black resigned in A.Karpov-D.Ten Geuzendam, Amsterdam (simul) 1992. White has a number of threats which cannot
be met, including 41 Qf5, followed by 42 g5 and 43 Rxe8.
11 ... Ne4 12 Bxe7 Qxe7 13 b4 a6
14 a4
14 Na4!? is a common idea in such positions, and one often used by Karpov himself. The idea is to
exploit the weakening of the dark squares caused by ... a7-a6 – White aims to play his knight into c5,
and then proceed with his queenside pawn storm. If Black plays ... b7-b6, his c6-pawn will be
desperately weak. If he takes on c5 then, after b4xc5, his b7-pawn will be a target.
14 ... Bf5 15 Ne5 Rad8 16 Rfc1 Ng6?!
This gives Karpov the chance to simplify the position massively. 16 ... f6 17 Nf3 Nxc3 18 Qxc3
Bxd3 19 Qxd3 Ne6 might be preferable. By retaining the knights, Black keeps some (limited)
prospects of counterplay.
17 Bxe4
17 Nxg6 was a more accurate move order.
17 ... Bxe4 18 Nxe4
Exercise: Which knight would you take?
18 ... dxe4?!
Black slightly weakens his pawn structure and allows all the knights to be traded, resulting in quite
an unpleasant major piece endgame.
Answer: I prefer 18 ... Nxe5! 19 Nd2 Nd7 20 b5 axb5 21 axb5 Rc8, when Black retains prospects of a successful defence, though
White still has a solid edge. As I mentioned previously, by keeping the knights Black retains some small prospects of counterplay.
19 Nxg6 hxg6 20 b5!
The minority attack in its pure form. Black can select from a range of structures, but in all of them
he will be left with a queenside weakness.
20 ... cxb5 21 axb5 Rd6
21 ... axb5? 22 Rxb5 is certainly not an improvement, since the b7-pawn is horribly weak and Re5
is an idea in some variations.
22 bxa6 bxa6
Black has opted for a weak a6-pawn, which is probably the best he could get from his position
after 20 b5.
23 Qa4 Qd7 24 Qxd7 Rxd7 25 Rc5!
An excellent move, activating the rook and preparing either Ra5 or Re5 as appropriate.
25 ... Ra7 26 Ra5 Kf8 27 Rb6 Rea8
The black rooks have been forced into total passivity.
28 h4!
Fixing Black’s kingside pawns and opening a route for the white king.
28 ... Ke7 29 Kh2 Kd7 30 Kg3 Kc7 31 Rb2 Rb7 32 Rc5+ Kb8 33 Ra2 Re7 34 Kf4 Kb7 35 Rb2+
Ka7 36 Rc6 Rh8
Answer:
37 Ra2!
As always, Karpov proves willing and able to calculate at the decisive moment. A lazy move like
37 g3? would allow Black to activate his rook with 37 ... Rh5!, when his drawing chances look much
brighter.
37 ... a5
37 ... Rxh4+?? 38 Kg3 wins a rook, since 38 ... Rh5 39 Rcxa6+ Kb8 40 Ra8+ Kc7 41 R2a7+ Kd6
42 Rd8+ Ke6 walks into mate after 43 Ra6+ Kf5 44 Rd5+.
38 Rxa5+ Kb7 39 Rca6 Rxh4+ 40 Kg3 Rh5 41 Ra7+ Kc6 42 R5a6+ Kb5 43 Rxe7 Rg5+ 44 Kh2
Kxa6 45 Rxf7 1-0
Black saved his rook, but not the game.
Game 24
A.Karpov-D.Campora
San Nicolas (2nd matchgame) 1994
1 d4 d5 2 c4 e6 3 Nc3 Nf6 4 cxd5 exd5 5 Bg5 Be7 6 e3 0-0 7 Bd3 Nbd7 8 Nf3 c6 9 Qc2 Re8
10 0-0
Karpov has used the move order 10 h3 too, which retains options of castling long. Nevertheless, as
far as I can see, Karpov almost invariably castled short in the Exchange QGD (unlike Kasparov, who
castled long on several occasions in otherwise quiet QGD positions). A.Karpov-J.Piket, Monte Carlo
(rapid) 2001, continued 10 ... Nf8 11 Bf4 Ng6 12 Bh2 Bd6 13 Bxd6 Qxd6 14 0-0 Be6 15 Rfc1 Re7
16 Rab1 Rae8 17 b4 a6 18 a4 Bd7 19 b5 axb5 20 axb5 Ne4 21 bxc6 bxc6 22 Bxe4 dxe4 23 Nd2 f5,
when Black should be suffering in view of his weak pawns and passive light-squared bishop,
although the game was quickly drawn (perhaps due to the rapid time control).
10 ... Nf8 11 h3 Be6
Other moves:
a) 11 ... g6 (a typical idea, aiming to trade the light-squared bishops after ... Nf8-e6-g7 and ... Bf5)
12 Bxf6 (after 12 Bh6 Nh5 13 Rab1 Ng7 14 b4 a6 15 Bxg7 Kxg7 16 Na4 Bd6 17 Nc5 Qf6 18 Rfc1
Ne6 19 Qd1 h6 20 a4 Re7 21 Bf1 Ng5 22 Nxg5 Qxg5 23 Qf3 h5, Black had at least equalized in
A.Karpov-N.Short, World Team Championship, Lucerne 1989; 1-0 in 66) 12 ... Bxf6 13 b4 was tried
by Karpov in several games.
a1) 13 ... Ne6 14 Rfd1 a6 15 Bf1 Qd6 16 a3 Bd8 17 e4!? dxe4 18 Nxe4 Qf4 19 Re1 Bb6?! 20 Qc3
with a very active position for White in A.Karpov-A.Beliavsky, Linares 1991 (1-0 in 62).
a2) 13 ... Be6 14 a4 Rc8 15 Rfc1 Qd6 16 Qb3 Nd7 17 Ne2 Be7 18 b5 c5! 19 dxc5 Nxc5 20 Qd1
Bf6 gave Black an excellent game in A.Karpov-J.Piket, Monte Carlo (rapid) 1995 (½-½ in 54).
a3) 13 ... a6 14 a4 Be6 15 b5 axb5 16 axb5 Nd7 17 bxc6 bxc6 and White has a typical miniscule
edge in A.Karpov-J.Ehlvest, Vienna 1996 (½-½ in 53).
b) Yermolinsky, a specialist in this line with h2-h3, suggests that 11 ... Nh5 might be Black’s best
defence; for instance, 12 Bxe7 Qxe7 13 Rfe1 Nf6 14 Rab1 a5 15 a3 g6 16 b4 axb4 17 axb4 Ne6 18
b5 Ng7 “with the idea of ... Bf5 – a typical defensive plan. In most cases, Black survives an inferior
ending with a weak c6-pawn, providing that he has managed to exchange light-squared bishops.”
12 Bxf6
Instead, 12 Rfc1 N6d7 13 Bf4 Nb6 14 Rab1 Bd6 15 Ne2 Ng6 16 Bxd6 Qxd6 17 a4 Rac8 18 Qc5
was played in A.Karpov-A.Kharitonov, USSR Championship, Moscow 1988 (1-0 in 42), and now I
like 18 ... Qxc5 19 dxc5 (19 Rxc5 Ne7, aiming to exchange bishops with ... Bf5, looks equal) 19 ...
Nd7 with a solid position.
I lost a game in this line myself: 12 Rab1 Ne4 13 Bxe7 Rxe7 14 b4 f5?! (as noted by Alex
Baburin: “This move may look active, but this is actually a mistake. Black does not really need the
knight on e4. The text weakens the e5-square, while turning the e6-bishop into a bad piece. Prospects
of a pawn storm on the kingside are not very good here”; 14 ... Nxc3 15 Qxc3 Rc7 was better) 15
Rfc1 a6 16 Na4! (another instructive comment from Alex, who is an expert in these structures:
“Normally in positions like this [i.e. where Black has played ... a7-a6 – S.C.] White should get his
knight to c5 first and only then go a2-a4 and b2-b4-b5”) 16 ... Ng6 17 Nc5 Qd6 18 a4 Rae8 19 b5
axb5 20 axb5 Bc8 21 bxc6 bxc6 and I faced a miserable defence in P.Kiriakov-S.Collins, Port Erin
1999 (1-0 in 31).
12 ... Bxf6 13 b4 Rc8
14 Na4
Question: Why doesn’t White simply press

on with the minority attack by playing 14 b5 - ?
Answer: As noted by Yermolinsky, Black has the typical response 14 ... c5!, opening lines for his rook on c8 and bishop on f6, after
which he is at least not worse.
14 ... Rc7!
I like this move, preparing to drop the bishop back to c8 to cover the b7-pawn if needed.
15 Rac1 Be7!
Another typical and strong manoeuvre, aiming for d6, which is the best square for the bishop in
this structure. So far, Campora is playing excellently.
16 Qb1 Bd6 17 b5 Qf6 18 bxc6 bxc6 19 Nh2?
19 Nd2 was more circumspect.
Exercise: How should Black meet the slow strategic build-up?
19 ... Qh4?
Answer: In his wonderful book on the Queen’s Gambit Declined, Matthew Sadler analysed 19 ... Bxh3! to a clear advantage for
Black. Nowadays the engines spot it instantly: 20 gxh3 Qh4 21 Ng4 (the best response; defending the h3-pawn runs into 22 ... Rxe3!) 21
... h5 (21 ... Qxh3? is well met by 22 Bf5) 22 Ne5 Bxe5 23 dxe5 and Black cashes in with 23 ... Qxa4 (admittedly a hard move to
calculate when focusing on the kingside).
Exercise: Returning to 19 ... Qh4?, how can White defuse the kingside attack?
Answer:
20 Bf5! Qh5 21 Bxe6 Nxe6
Karpov is out of the woods now. Black is still fully equal but gradually gets outplayed.
22 Nf3 f5?!
While tempting, it is very difficult to make this move work in the Exchange QGD. The dark
squares (in particular e5) are seriously weakened and, as Karpov demonstrates here, ... f5-f4 isn’t a
threat. I made a similar mistake in my game against Kiriakov (see the note to White’s 12th move).
23 Rc3 Nd8 24 Nc5 Bxc5 25 Rxc5
Now Black has too many weak pawns and squares to cover.
25 ... Ne6
Exercise: Where should the rook retreat?
Answer:
26 Rc3!
Child’s play for Karpov. The rook is perfectly placed both for attacking (doubling on the c-file,
attacking the a7-pawn with Ra3) and defensive purposes (over-defending the e3-pawn and looking to
guard the third rank).
26 ... f4 27 e4
A good move, after which White keeps a sizeable edge. 27 Re1 leads to similar play.
27 ... h6
27 ... dxe4 was more tenacious, though after 28 Qxe4 Black has three weak pawns (which is about
three too many when facing Karpov!) and poorly co-ordinated pieces.
28 Re1 Rce7
Campora ditches a pawn to get some threats, but Karpov calculates well and patiently rebuffs
Black’s activity.
29 Rxc6 dxe4 30 Rxe4
Of course Karpov isn’t going to allow nonsense like 30 Qxe4? Nxd4 or the more ambitious 30 ...
Nd8!, when Black is fine.
30 ... Qd5
For the moment it looks like Black might be getting somewhere, but Karpov coolly deals with the
threats.
31 Rc3!
A nice echo of move 26, and again the strongest. It is useful for White to have the f3-knight
defended so that there is no risk of his kingside structure being shattered.
31 ... Qf5 32 Qe1!
An outstanding improvement of White’s co-ordination, getting out of the pin and simultaneously
pinning the e6-knight.
32 ... Qd5 33 Kh1
A safe move (stepping out of any future ... Ne6-g5/d4xf3+ tricks) which preserves White’s winning
advantage.
33 ... Qd6
Exercise: What did Karpov intend after 33 ... Qxa2 - ?
Answer: 34 Re5! and the threat of d4-d5 will win material.

34 Qd2 Ng5
There is nothing better, since Black has nothing to hope for here.
35 Rxe7 Qxe7
Or 35 ... Rxe7 36 Rc8+ Kh7 37 Nxg5+ hxg5 38 Qc2+ g6 39 Rc6 Re2 (the only move) 40 Qc4
Re1+ 41 Kh2 Qf8 42 d5 with an extra pawn and an attack.
36 Qxf4 Qb4 37 Nxg5
37 Rc7 was also strong, but Karpov’s solution is more than sufficient.
37 ... hxg5 38 Qd2 g4 39 hxg4 1-0
Chapter Three
Openings
Karpov was never acknowledged to be particularly dangerous in the openings, and always
preferred the middlegame and endgame. However, he could hardly be said to be a weak opening
player – he was well prepared and had assistance (in tournaments and World Championship matches)
from some of the most creative opening researchers in history, such as Geller. Moreover, with his
smooth positional style, Karpov created a number of model games in the openings he regularly
played.
The following material is not at all a theoretical survey, but rather an introduction to Karpov’s
approach of playing with White, which netted him an enormous winning percentage.
After his defeat to Kasparov, Karpov switched from 1 e4 to closed openings with 1 d4, a move
which probably helped him stay at the top for so many years. Apparently it was his final game in the
second match against Kasparov that was the tipping point – in a sharp Sicilian, Karpov hesitated and
lost, and realized that such positions did not suit his style as well as more closed strategic battles.
Accordingly I have taken a selection of openings and games with both 1 e4 and 1 d4.
Battles in the Pirc

The Pirc has recently undergone something of a revival, in that the world’s elite regularly try it when
playing a must-win game with Black. For a long period Pirc seemed to be regarded as a semi-correct
opening, conceding the centre for unclear compensation. Karpov was lethal when facing it, since he
was granted a healthy position with central space and excellent development as White, from which he
could squeeze his opponents in the middlegame.
Game 25
A.Karpov-Cu.Hansen
Biel 1992
Pirc Defence
1 d4 d6 2 e4 Nf6 3 Nc3 g6 4 Nf3 Bg7 5 Be2 0-0 6 0-0 c6

The most popular continuation, preparing ... b7-b5 and taking control of d5.
7 Bf4
Quite a rare move. In the main White tends to prevent ... b7-b5 by playing 7 a4 immediately, but
it’s worth noting that the immediate ... b7-b5 isn’t so much of a threat. I played this line with White in
my final round game at the 2004 Calvia Olympiad (back when you could obtain GM and IM titles just
by scoring norms in Olympiads!). I think I had already secured my norm and title before this game, but
I was still highly motivated to beat my Belgian opponent, who made things easier for me by
answering 7 h3 with 7 ... b5?!.
Exercise: How should White respond?
Answer: 8 e5! – it turns out that ... b7-b5 has given White queenside targets both on b5 (which can be attacked by a2-a4) and c6
(which now relies on piece support), although my opponent could certainly have defended: 8 ... Ne8 9 a4! b4 10 Ne4 d5 (I don’t like this
plan, after which c5 is horribly weak) 11 Nc5 a5 12 c3 bxc3 13 bxc3 Nc7 14 Ba3 Nca6 15 Re1 Nxc5 16 Bxc5 Nd7 17 Ba3 f6 (this
allows White to seize more space, though it was already difficult to offer Black good advice) 18 e6 Nb6 19 Rb1 Nc4 (this could have
been forced by Bc5 in any event; now the c-pawns will be horribly weak) 20 Bxc4 dxc4 21 Nd2 Ba6 22 Qf3 Bh6 23 Ne4 f5 24 Nc5 Qc7
25 Nxa6 Rxa6 26 Qe2 Bf4 27 Qxc4 with an extra pawn and a dominant position in S.Collins-A.Abolianin, Calvia Olympiad 2004 (1-0 in
50).
7 ... Nbd7
Instead:
a) 7 ... Qc7 8 e5 (8 h3 retains an edge) 8 ... Nh5 9 Bg5 dxe5 10 dxe5 Bg4 11 Re1 was
V.Zvjaginsev-V.Onischuk, St Petersburg 2011, when Black’s offside knight on h5 was a source of
some concern (½-½ in 53).
b) 7 ... Bg4 was tested in a recent game between two strong GMs: after 8 Qd2 Nbd7 9 Rad1 Bxf3
10 Bxf3 e5 11 Be3 Re8 12 Rfe1, White had a stable edge with two bishops and more space in
G.Meier-Z.Efimenko, World Blitz Championship, Dubai 2014 (1-0 in 52).
8 Re1
Other moves:
a) 8 a4 (this seems to mix systems slightly; 7 a4 and 7 Bf4) 8 ... Qc7 9 e5 Nh5 10 exd6 exd6 11
Be3 Nhf6 12 Qd2 Re8 13 Rfe1 Nb6 14 h3 a5 15 b3 Bf5 16 Bd3 Bxd3 and a draw was agreed in
J.Dorfman-V.Tkachiev, French Team Championship 2000.
b) 8 Qd2 Qc7 9 e5 Nh5 10 exd6 exd6 11 Ne4 Ndf6 12 Bxd6 Nxe4 13 Bxc7 Nxd2 14 Nxd2 Bxd4
15 c3 Bb6 16 Bxb6 axb6 17 Bxh5 gxh5 resulted in a roughly level endgame in D.Solak-T.L.Petrosian,
Golden Sands 2014 (½-½ in 39).
8 ... Qb6!?
Hansen’s attempted improvement over his previous play: 8 ... Qa5 9 Nd2 Qc7 10 a4 e5 11 dxe5
dxe5 12 Be3 Rd8 13 f3 b6 14 Qb1 Nf8 15 Nd1 Ne6 16 c3 h5 and a draw was agreed in
U.Andersson-Cu.Hansen, Novi Sad Olympiad 1990. Later, he returned to this line and after 12 Bg3?!
Nc5 13 b4?! Ne6, Black was already better in T.Wedberg-Cu.Hansen, Stockholm 1996 (0-1 in 37).
Exercise: How should White meet the threat to the b2-pawn?
Answer:
9 Qc1!
A multi-purpose move. The queen defends the b2-pawn while stepping off the d-file (which is
likely to open after Black’s ... e7-e5 break, when a rook on d8 would uncomfortably oppose the
queen). White also prepares the standard Bh6 intending a kingside attack.
More ambitious players might prefer 9 Qd2!?, with sufficient compensation after 9 ... Qxb2, if
probably not more. In contrast, 9 Rb1 seems slightly passive, though even here White might be able to
count on an edge. It is unsurprising that Hansen subsequently returned to 8 ... Qa5 given the range of
reasonable options at White’s disposal against 8 ... Qb6.
9 ... e5
Black’s standard break in the centre. White normally takes on e5 and, indeed, in this position there
is little choice, since 10 Be3?! Ng4 is awkward.
10 dxe5 dxe5 11 Be3!
Gaining a tempo on the black queen.
11 ... Qc7
Exercise: What do you think of this position? What are the plans for both sides?
Answer: We have reached a standard pawn formation for the Pirc. Similar structures are sometimes reached in the King’s Indian,
and White’s plan is similar to there, namely to advance gradually on the queenside and, in particular, threaten to invade on d6 with a
knight. Black is more likely to bring a knight to e6, aiming to hop into d4, and might seek to combine this with play on the kingside. The
important difference is that White’s c-pawn is on c2 (rather than c4), so Black doesn’t have a guaranteed outpost on d4.
Over the next few moves the two grandmasters proceed logically with their respective plans.
12 Nd2
Aiming for c4, while preventing an annoying ... Ng4.
12 ... Rd8 13 Rd1
Both rooks occupy the only open file.
13 ... Nf8 14 Nc4 Rxd1+ 15 Qxd1 Ne8
Hansen carefully covers his weak square on d6. The immediate 15 ... Ne6 is also possible, since
after 16 Qd6 Qxd6 17 Nxd6 Nd4, White has nothing.
16 a4 Ne6 17 Nb1!?
A nice echo of Karpov-Spassky (Game 1). The knight on c3 wasn’t well placed – the b5- and d5-
squares are covered, and the e4-pawn doesn’t need protection for the moment – whereas it is useful
to have the pawn on c3, covering the d4-square. Meanwhile, the white knight can re-route to f3,
targeting the e5-pawn.
After 17 a5, Stohl gives 17 ... Nd4 18 Bf1 Be6 with counterplay for Black.
17 ... b6 18 c3 Nf4 19 Bf1 Be6 20 Nbd2 h5
This move is extremely natural. Once the knight goes to f4, Black’s play with ... f7-f5 is more
awkward (since the knight stands in the way of the pawns). Instead, marching the h-pawn up the board
is standard, aiming to force a concession (to stop Black’s pawn landing on h3). And yet 20 ... h5,
almost imperceptibly, weakens the g5-square, which will require constant defence.
It is notable that, throughout the manoeuvring phase of the game, Karpov has been careful to avoid
taking on any weaknesses in the white camp. Hansen is not quite so careful and this proves a
significant aspect in his downfall.
Black would have nothing to fear after 20 ... Rd8.
21 Qc2 Rd8 22 b4
Seizing space on the queenside.
22 ... Nf6
Starting an odd plan of covering the g5-square with the knight, from a very passive location on h7.
22 ... f6 seems more consistent.
23 Nf3 Nh7
Answer:
24 Rd1!
Competing for the only open file. Black has to react in view of the threatened capture on e5 after
trading rooks on d8 (another benefit of 22 ... f6 was that the e5-pawn would be solidly defended).
Suddenly it becomes apparent that Black’s pieces have largely abandoned the queenside, and Karpov
aims to exploit his spatial advantage there.
24 ... Rxd1
24 ... Rc8 is an alternative but also depressing for Black, leaving White in control of the d-file.
25 Qxd1 Bg4

Answer:
26 Qd6!
Trading the only black defender of the queenside. White doesn’t mind taking on some structural
damage on the kingside, since all of his remaining pieces, together with the a- and b-pawns, will
force significant concessions on the other side.
26 ... Qxd6 27 Nxd6 Kf8
Bringing the king into the defence is relatively best.
28 a5 Ke7
As Stohl notes, 28 ... bxa5 29 bxa5 leaves the a7-pawn indefensible.
29 Nc4 Bxf3 30 gxf3 b5 31 Na3 a6
Exercise: Has Black managed to consolidate his defences?
Answer: Absolutely not! Karpov unleashes a stunning sacrifice, destroying the black queenside.
32 Nxb5!!
Interestingly, modern engines (at least the ones I have access to) don’t see this idea until it’s played
on the board, and accordingly misassess the entire sequence from move 24 as equal, whereas in fact
Black’s position is critical.
It is hard to know which aspect of Karpov’s play to praise more, playing into this position or
finding the winning combination. Playing into the position is based on a very subtle intuition, having
regard to the absence of defenders and the far advanced a5-pawn. Importantly, the black king is
outside the “square” of this pawn, and so will need help from the f4-knight – and knights are
notoriously poor at stopping passed rook’s pawns. Probably these factors are more difficult to judge
than the purely tactical nuances. And yet, as will appear, Karpov’s combination is based on some
precise tactical points, so his calculation certainly deserves praise too.
It should be noted that preparing the queenside breakthrough with 32 c4 is good too (albeit not as
strong as Karpov’s move), based on the same motif: 32 ... Kd7 (32 ... Ng5 33 cxb5 Nxf3+ 34 Kh1
cxb5 35 Bxb5! wins for White; 32 ... bxc4 might be best but this position is very unpleasant for
Black) 33 cxb5 cxb5 34 Nxb5! Bf8 (after 34 ... axb5? 35 Bxb5+ and 36 a6, the pawns are too strong)
35 Na7 Bxb4 36 Bb6 Ng5 37 Bxa6 Nxf3+ 38 Kh1, when the a-pawn is hard to stop.
32 ... cxb5
After 32 ... axb5 33 a6! Ne6 34 Bb6!, the a-pawn runs home.
Exercise: How did Karpov continue?
Answer:
33 Bxb5!!
Continuing to play for a passed a-pawn, without counting the material cost.
33 ... Ng5
Other moves also fail to save the game; for instance, 33 ... axb5 34 a6 Ne6 35 Bb6 wins as before.
34 Bxa6
Perhaps the most impressive aspect of Karpov’s combination was that he had to assess this
position. He has three connected passed pawns for the piece which, supported by the bishop pair, are
obviously extremely dangerous. However, the f3-pawn is hanging with check, and it is not obvious (at
least to me) that Black will be completely unable to organize a blockade of the white pawns, or
sacrifice a knight back in favourable circumstances. Again, the placement of the black pieces on the
kingside, far away from the action, is a decisive factor.
34 ... Nge6
Desperately trying to defend some squares on the queenside. The f3-pawn is irrelevant: 34 ...
Nxf3+ 35 Kh1 and White will continue with Bc4 and a5-a6, as in the game.
35 Bc4 Kd8 36 a6
Now the threat is 37 a7 Nc7 38 Bb6, so Black must react.
36 ... Kc8
If 36 ... Nc7, the b-pawn starts rolling and will, at a minimum, win the knight for insufficient
compensation.
37 b5 Bf8 38 b6
The principled move. According to the computer 38 a7 wins as well, but it is much more “human”
to avoid having the pawns blockaded (even temporarily).
38 ... Bc5 39 Bd5!
A lovely winning idea, after which all of Black’s minor pieces are tactically vulnerable.
39 ... Nxd5 40 exd5 Bxe3
40 ... Nd4 41 cxd4 Bxb6 42 dxe5 is equally hopeless.
41 b7+ Kc7 42 dxe6 Bc5 43 exf7 1-0
A nice final position. Black is helpless against the slow march of the white king to e8.
Game 26
A.Karpov-V.Korchnoi
World Championship (32nd matchgame), Baguio 1978
Pirc Defence
1 e4 d6 2 d4 Nf6 3 Nc3 g6 4 Nf3 Bg7 5 Be2 0-0 6 0-0 c5

A relatively rare move.
In the 18th game of the match, Korchnoi had opted for 6 ... Bg4 7 Be3 Nc6 8 Qd3 (again White can
seize space with 8 d5, when Black tends to continue 8 ... Bxf3 9 Bxf3 Ne5 10 Be2 c6, seeking to chip
away at the white centre) 8 ... e5 9 d5 Nb4 10 Qd2 a5 11 h3 Bd7 12 Bg5 Qe8 13 Nh2 Kh8 14 a3 Na6
15 Bh6 Bxh6 16 Qxh6 Ng8 17 Qe3 f5 with reasonable King’s Indianesque counterplay and a draw in
64 moves.
7 d5
Karpov goes for perhaps the most strategically ambitious continuation, aiming for a Benoni-type
structure (albeit with the white pawn on c2). 7 dxc5 dxc5 8 Be3 is also a slightly problematic
continuation for Black, who needs to be precise to equalize.
While 7 Re1 or a similar move is playable, clearly Karpov wouldn’t be tricked into 7 ... cxd4 8
Nxd4, when White has adopted a modest set-up against the Sicilian Dragon.
7 ... Na6 8 Bf4
Supporting the e4-e5 break, which is a key resource for White in such structures.
It’s interesting that Karpov continues to allow the option of ... Bg4, which Korchnoi rejects. Often
this bishop struggles to find work in Benoni structures and so exchanging it off for White’s light-
squared bishop, or the f3-knight (which, after a2-a4, can find a comfortable square on c4) is generally
viewed as easing Black’s defence. However, perhaps both Karpov and Korchnoi had regard to the
fact that the c2- and e7-pawns are still on the board (in an true Benoni, c2-c4 and ... e7-e6 would
have been included, together with ... e6xd5 and c4xd5). Accordingly, the position could open more
rapidly (for instance, if White meets ... e7-e6 with d5xe6), which would favour the bishop pair.
When GM Eugenio Torre essayed the same sideline in the SWIFT World Blitz Championship in
Brussels 1987, Karpov opted for 8 h3 (preventing ... Bg4) 8 ... Nc7 9 a4 b6 10 Re1 Bb7 11 Bc4 a6
12 Qd3, going on to win in 37 moves.
8 ... Nc7 9 a4 b6 10 Re1 Bb7
Question: How do you assess the development of the bishop to b7?
Answer: Perhaps it is a matter of taste – and if you like putting the bishop on b7, you are in good company with Grandmasters
Korchnoi and Torre! The engines don’t see a major problem with it either. However, to my mind the development of the bishop here is
extremely dubious (all the more so in a World Championship match). It’s well known that the light-squared bishop doesn’t belong on b7 in
a Benoni, since it is severely restricted by the d5-pawn and not only fails to support Black’s counterplay with ... b5 (as it would to from
d7), but actually impedes this counterplay (since it blocks a rook on b8).
As it happens, in the present game Korchnoi eventually achieves ... b6-b5, but only after White is
fully developed and starting active play in the centre and on the kingside. The bishop on b7 would be
wonderfully placed if the white centre could be broken down with ... e7-e6 and ... f7-f5, but this is
simply not going to happen: ... e7-e6 would severely weaken the d6-pawn after d5xe6, while ... f7-f5
would weaken the e6-square and the e7-pawn, and invite a very strong e4-e5 break by White.
Chess history celebrates, with good reason, Tony Miles’ famous victory with Black against Karpov
at the European Team Championships in 1980, where the English player used the St. George Defence
(1 e4 a6?!). But frankly it takes more courage to play the line Korchnoi uses here! And for such a
position to be reached in a modern World Championship match is basically unthinkable.
Instead, 10 ... Bg4 was still an option (see the discussion in the notes to White’s 8th move).
11 Bc4!
A strong move, aimed at all three of Black’s potential pawn breaks:
1 ... b6-b5 will be hard to achieve after Qd3.
2 ... e7-e6 will need to be defended after d5xe6 (in addition to the problems discussed above).
3 ... f7-f5 exposes the king after e4-e5 (for instance, ... d6xe5 and d5-d6+ would be a catastrophe)
– again, in addition to the problems discussed above!
11 ... Nh5
This manoeuvre doesn’t achieve much, but already, as you might have guessed, I’m not a big fan of
Black’s position.
12 Bg5 Nf6 13 Qd3 a6 14 Rad1!
Simple, powerful, central play.
14 ... Rb8
Exercise: How would you proceed with White?
15 h3?!
A slight hesitation. While I cannot argue with Karpov’s methodical build-up, which is particularly
effective against such provocative play, here he allows Black to keep the disadvantage within
manageable limits.
Answer: Already 15 e5! would have put Black in a critical position; for example, 15 ... dxe5 16 Nxe5 b5 17 axb5 axb5 18 Bxb5
Nfxd5 (or 18 ... Ncxd5 19 Nxd5 Qxd5 20 Qxd5 Bxd5 21 c4 Bb7 22 Bxf6, followed by Nd7, winning material) 19 Bc4! and White is
much better prepared for the opening of the position.
15 ... Nd7 16 Qe3
Karpov plays simply and well, aiming to trade Black’s best piece (the bishop on g7). 16 Bf4 was
also an option, aiming to support e4-e5 in some lines.
16 ... Ba8
Another very ambitious move by Korchnoi, but this wouldn’t be my preference.
Answer: Black has less space and wants to avoid the trade of dark-squared bishops, so 16 ... Ne5! looks very logical. After 17 Nxe5
Bxe5 18 Bh6 Re8, White remains better, but I think Black’s position has improved.
17 Bh6 b5
This was the idea, finally gaining space on the queenside, but it’s hardly convincing.
18 Bxg7 Kxg7 19 Bf1 Nf6 20 axb5 axb5
The position has stabilized somewhat. Black has managed to push ... b6-b5 and, for the moment,
restrains the e4-e5 break. White has exchanged dark-squared bishops, which means that the black
king’s safety will always need to be taken into account. We now enter a manoeuvring phase with both
sides trying to improve their pieces.
21 Ne2!?
Borrowing a concept from the Ruy Lopez, Karpov aims to put his knight (which could have been
hit by ... b5-b4) on g3, from where it solidly defends e4, supports h3-h4-h5, and can land on f5 on a
good day.
21 ... Bb7
The downside of Karpov’s manoeuvre was that his central control was loosened for a moment,
meaning that, for the first time, 21 ... e6!? was worth considering. After 22 dxe6 Nxe6, Korchnoi’s
faith in the bishop on b7 has been rewarded, as this is now a good piece. All the same, I still prefer
White after 23 e5 Nd5 24 Qd2, whose central pressure is impressive.
22 Ng3 Ra8 23 c3 Ra4 24 Bd3 Qa8?
Korchnoi continues in aggressive style, strengthening his counterplay on the queenside while
putting more pressure on the long diagonal. Unfortunately, the removal of the queen from the centre
has its own problems.
Exercise: What’s the refutation?
Answer:
25 e5!
Deeply thematic and extremely effective.
25 ... dxe5
Relatively best, though Black’s position is collapsing.
Exercise: Why can’t Black take with 25 ... Nfxd5 - ?

Answer: White crashes through with 26 Nh5+!! (Filip’s 26 Nf5+! wins as well) 26 ... gxh5 (or 26 ... Kh8 27 Qh6 Ne6 28 Ng5) 27
Qg5+ Kh8 28 Qh6 f5 29 Ng5 with mate to follow.
26 Qxe5 Ncxd5
Korchnoi finds the best defence. If 26 ... Qd8 27 Bc2 Ra6 28 Ne4, the c5-pawn is falling.
27 Bxb5 Ra7
27 ... Ra5 at least attacks the bishop so that, after 28 c4, Black has time for 28 ... Qb8.
Nevertheless, White is still dominant after 29 Qg5 (threatening Nf5+) 29 ... e6 30 Ne5 and Black has
no defence to the multiple threats.
28 Nh4?!
It was more precise to play 28 c4 Nc7 and only then 29 Nh4, preventing the defence in the next
note.
28 ... Bc8?!
28 ... Qb8! is the most tenacious, keeping the bishop with some prospects on b7.
29 Be2 Be6
30 c4
Karpov cashes in and wins a pawn, leaving himself with connected queenside passers.
Alternatively, he could have played for a direct attack: 30 Bf3 Qb8 31 Qg5 with a very strong
initiative; for instance, 31 ... Rd8 32 Ngf5+ Kh8 33 Rxe6! gxf5 (33 ... fxe6 34 Nxg6+! mates in a few
moves) 34 Rc6 Rg8 35 Qh6 and Black has too many weaknesses to survive.
30 ... Nb4 31 Qxc5 Qb8 32 Bf1
Black is a pawn down for very uncertain compensation, even if it is not easy to push the passed
pawns at the moment.
32 ... Rc8?!
32 ... h6 would have been preferable.
Answer:
33 Qg5!
Just because White has won a pawn doesn’t mean he has to abandon the kingside attack.
Combining play on both sides of the board is a hallmark of many World Champions, and Karpov is no
exception.
33 ... Kh8
It is too late for 33 ... h6? in view of 34 Ngf5+.
34 Rd2
A solid consolidating move, protecting b2 while preparing to double on the d-file.
34 ... Nc6 35 Qh6 Rg8
This apparently solid defensive move (maybe even with ideas of supporting ... g6-g5) actually
makes things easier for White. 35 ... Ne5 would have retained some small defensive chances.
36 Nf3!
Stopping any activity with ... g6-g5 while aiming for that square with the knight.
36 ... Qf8
Black has gone over to kingside defence at the price of abandoning the queenside blockade.
Exercise: Should White trade queens?
Answer: No! Karpov refuses to be blinded by stereotypical thinking. The queen on f8 is obviously worse than the white queen (and,
for that matter, the black king remains a long-term liability).
37 Qe3!
Of course White is much better after 37 Qxf8 Rxf8, but then considerable work remains to be done
to push the queenside pawns. By keeping the queens on, White also creates the tactical threat of b2-
b4!, overstretching the c6-knight.
37 ... Kg7
Going back with 37 ... Qb8 allows White to develop his initiative with 38 Ne4.
38 Ng5 Bd7 39 b4!
Once the pawns start rolling Black’s position is hopeless.
39 ... Qa8 40 b5 Na5 41 b6 Rb7 1-0
Korchnoi resigned during the adjournment, rather than see which of the winning continuations (for
instance, 42 Nxf7!) appealed most to Karpov.
Beating the Sicilian

It is probably not the first thing one associates with Karpov, but in the early part of his career he
was extremely effective against the Sicilian. He played principled variations with 2 Nf3 and 3 d4,
after which he was equally comfortable in quiet strategic lines with the classical Be2, or taking on the
sharpest main lines in the Dragon (where he won almost all his games).
World Championship matches clearly involve their own psychological battles, and one of the main
aims of the participants is to shake the opponent’s confidence in one of his main openings.
A classic example comes from Kasparov’s 1995 World Championship match with Anand.
Following eight draws and a win for Anand in the ninth game, Kasparov (having drawn in the Open
Spanish in game six), Kasparov unleashed Tal’s 14 Bc2!, followed by the spectacular 15 Nb3!!,
winning a wonderful game. Having varied with White in earlier games, Kasparov played 1 e4
exclusively for the rest of the match, with Anand casting around for a reliable defence (including the
Spanish Möller, Sicilian Najdorf and even the Scandinavian). Here is Kasparov’s assessment from
his fascinating book, How Life Imitates Chess:
“In game 10, I played the first part of the new idea on move 14, actually following a suggestion
made long ago by Mikhail Tal. Anand was clearly ready for this approach and replied after just 4
minutes’ thought. After my next move, however, he spent a solid 45 minutes thinking, perhaps a record
for the famously speedy Indian Grandmaster. The trap had been sprung and there was no way out. I
was still playing all my moves almost instantaneously, happy to at last have them out of my head and
on to the board. To his credit, the ‘Tiger of Madras’ played like a true championship contender and
survived the first wave of the attack after first falling into my snare. Only when the smoke had cleared
and my advantage was evident did I slow down in order to make sure to bring home the full point. It
would have been excruciating to waste such a marvellous new idea. With precise play I notched the
win to level the match, which ... I eventually went on to win. The victory was worth one point, but the
psychological effect was devastating. Thanks to this one novelty Anand was forced to put his primary
defence on the shelf for the rest of the match. With the comfort of hindsight, some suggested that he
should have avoided repeating his use of the Ruy Lopez again in the tenth game despite his earlier
success. But he had just taken the lead in the match and he wanted to emphasize his ascendancy by
failing to blink in the psychological opening battle.”
Similarly, Karpov’s success in the following game resulted in Korchnoi abandoning the Dragon for
the rest of the match, instead relying on the Open Spanish and the French.
Game 27
A.Karpov-V.Korchnoi
Candidates final (2nd matchgame), Moscow 1974
Sicilian Defence
1 e4 c5 2 Nf3 d6 3 d4 cxd4 4 Nxd4 Nf6 5 Nc3 g6

Of course this was resurrected in World Championship play by Kasparov in his 1995 match with
Anand.
6 Be3 Bg7 7 f3 Nc6 8 Qd2 0-0 9 Bc4
Karpov also enjoyed success with 9 g4 (preventing 9 ... d5? in view of 10 g5), but to keep the
notes to this game within manageable limits we will focus on 9 Bc4.
9 ... Bd7
The usual continuation. 9 ... a5 10 a4 Nxd4 11 Bxd4 Be6 12 Bb5 Rc8 13 0-0-0 Nd7 14 Bxg7 Kxg7
15 f4 Nf6 16 Rhe1 gave White a pleasant positional advantage in A.Karpov-D.Byrne, San Antonio
1972 (1-0 in 41).
10 h4
Question: What are the pros and cons of this move vs. 10 g4 - ?
Answer: Once White castles queenside, the kings will be on opposite sides of the board and the question of whose attack lands first
will be critical. White wants to open the h-file to attack Black’s king. By playing h2-h4, he gives himself the option of playing h4-h5
without loss of time (even as a pawn sacrifice). By avoiding g2-g4 White also maintains a tighter kingside structure, since f3 isn’t
weakened – a more dramatic illustration of the same point arises where Black, after ... Rc8 and ... Ne5, manages to land some
combination based on ... Rxc3, ... Bxg4 (or ... Nxc3) and ... Nxe4, destroying White’s centre and opening up the ‘Dragon’ bishop on g7
towards White’s king.
The main downside of h2-h4 is that it allows Black to block the kingside temporarily with ... h7-
h5. Recent GM praxis in this line tends to favour this prophylactic approach, although allowing the
h4-h5 pawn sacrifice (as Korchnoi does in this game) has not been clearly refuted.
Other moves:
a) 10 Bb3 Rc8 11 0-0-0 Ne5 12 Kb1 (an improvement on a training game against Korchnoi(!),
which proceeded 12 Bh6 Nc4 13 Bxc4 Bxh6 14 Qxh6 Rxc4 15 h4 Qa5 16 Nb3 Qe5 17 Nd5 Rfc8
with a comfortable game for Black, who even went on to win in A.Karpov-V.Korchnoi, Leningrad
1971; 0-1 in 43) 12 ... Nc4 13 Bxc4 Rxc4 14 g4 b5 15 Nce2 Qc7 and Black had equalized in
A.Karpov-G.Sosonko, World Blitz Championship, Brussels 1987 (1-0 in 34).
b) 10 0-0-0 Qb8 (10 ... Qa5 11 h4 Ne5 12 Bb3 Rfc8 transposes to the next note) 11 Bb3 a5 (this
irregular continuation didn’t disturb Karpov) 12 Ndb5 a4 13 Bxa4 Rc8 14 Kb1 Na5 15 Bb3 Nc4 16
Qe2 Nxe3 17 Qxe3 Rc5 18 Nd4 (18 a4!? might be stronger) 18 ... Qa7 (18 ... b5!? gives Black some
compensation) 19 a3 Ra5 20 Qd2 Ne8 21 Nd5 e6 22 Nb4 Nc7 23 c3 Na6 24 Ndc2 and White had
solidly consolidated the queenside, going on to win with his extra pawn in A.Karpov-D.Velimirovic,
Portoroz/Ljubljana 1975 (1-0 in 50).
10 ... Rc8
The alternative approach 10 ... Ne5 11 Bb3 Qa5 12 0-0-0 Rfc8 was a long-standing favourite of
British Champion and Dragon enthusiast GM Chris Ward. Karpov didn’t have much trouble dealing
with this variation: 13 h5 Nxh5 14 Bh6 Bxh6 15 Qxh6 Rxc3 16 bxc3 Qxc3?! (too hasty; 16 ... Nf6
was correct)
Exercise: How can Black’s threat of 17 ... Qa1+ 18 Kd2 Qxd4+ be dealt with?
Answer: 17 Ne2! (the only move that leads to an advantage; after 17 Kb1 Nc4 18 Bxc4 Qxc4 and White has nothing better than to
return the exchange for equality with 19 Rxh5 gxh5 20 Qg5+ Kf8 21 Qh6+ etc) 17 ... Qc5 (Karpov gives the variations 17 ... Qa1+ 18
Kd2 Qb2 19 Rb1 Qa3 20 Rxh5 gxh5 21 Rh1 with a winning attack, or 17 ... Nd3+ 18 Rxd3 Qa1+ 19 Kd2 Qxh1 20 g4 Ng3 21 Qxh1
Nxh1 22 Ke3! “and the knight is trapped”) 18 g4 (the immediate 18 Rxh5! gxh5 19 Rh1 is strongest, when Black has to give up material
to buy off the attack; for instance, 19 ... Ng4 20 fxg4 Bxg4 21 Nf4 or 19 ... Bg4 20 fxg4 Nxg4 21 Qxh5 Qxh5 22 Rxh5 with a clear
advantage for White) 18 ... Nf6 19 g5 Nh5 20 Rxh5 gxh5 21 Rh1 Qe3+? (Houdini suggests 21 ... Ng6 as a better defensive try) 22 Kb1
Qxf3 23 Rxh5 e6 (now 23 ... Ng6 is met by the stunning 24 Rh1!, when Houdini assures me that Black is defenceless).
Exercise: Find the best move for White.
Answer: 24 g6!! (a wonderful pawn sacrifice; no other moves are sufficient, and 24 Qxh7+? Kf8 25 Qh8+ Ke7 26 Qxa8 Qxh5
would even lose) 24 ... Nxg6 (now 24 ... fxg6 25 Qxh7+ Kf8 26 Qh8+ Ke7 fails to 27 Rh7+ Nf7 28 Qxa8 Qxe2 29 Qxb7 with a decisive
material advantage) 25 Qxh7+ Kf8
Answer: 26 Rf5!! is the only move. As Karpov notes: “The geometrical idea contained in this manoeuvre is rather elegant. Two lines
– the a2-g8 diagonal and the f-file – intersect at the critical point f7. Mate by Qxf7 is threatened, the rook supporting the queen along the
file, and the bishop – in the event of 26 ... exf5 – along the diagonal. As a result Black is forced to part with his queen”; i.e. 26 ... Qxb3+
27 axb3 exf5 28 Nf4 Rd8 29 Qh6+! Ke8 30 Nxg6 fxg6 31 Qxg6+ Ke7 32 Qg5+ Ke8 33 exf5 Rc8 34 Qg8+ Ke7 35 Qg7+ Kd8 36 f6 and
Black resigned in A.Karpov-E.Gik, Moscow 1968.
In a later game Karpov varied with 13 Kb1 Nc4 14 Bxc4 Rxc4 15 Nb3 Qd8 16 Bh6 Qf8? (here
Mednis wrote: “A move recommended by Levy in his generally excellent book The Sicilian Dragon.
The supposition that Black’s queen will be well placed on g7, however, is mistaken. Considerably
more logical is 16 ... Bxh6!? 17 Qxh6 Qf8, and if White retreats the queen, then Black is practically
speaking a tempo ahead of the game’s continuation. Black would like to keep his dark-squared bishop
with 16 ... Bh8, but V.Ciocaltea-D.Drimer, Rumania 1968, demonstrated that White’s attack is too
strong after 17 h5! Rac8 18 hxg6 fxg6 19 e5!.”) 17 Bxg7 Qxg7 18 g4 and White had a strong attack in
A.Karpov-A.Whiteley, Bath 1973 (1-0 in 32).
11 Bb3
Exercise: Why not 11 0-0-0 - ?
Answer: 11 ... Nxd4 12 Qxd4 Ng4 is a typical shot, winning immediately since both white bishops are hanging.
11 ... Ne5 12 0-0-0 Nc4
As discussed in the notes to White’s 10th move, the alternative is 12 ... h5. For example, 13 Bg5
Rc5 14 Kb1 (alternatively, 14 Rhe1 b5 15 f4 Nc4 16 Bxc4 bxc4 17 Bxf6 Bxf6 18 e5 Bg7 19 e6 Bc8
20 exf7+ Rxf7 21 Ne6 Bxe6 22 Rxe6 Qa5 23 Qe3 Bxc3 24 bxc3! left White on top in A.Karpov-
G.Sosonko, Tilburg 1979; Karpov won quickly after 24 ... Qxa2? 25 Rxg6+ Kf8 26 Qe4 Qa6 27 Rd5
Rf6 28 Rxc5 Rxg6 29 Rxh5 d5 30 Rxd5 and Black resigned) 14 ... b5 15 g4 hxg4 16 h5 Nxh5 17 Nd5
Re8 18 Rxh5 gxh5 19 Qh2 is an extremely sharp position which I’m not even going to attempt to
evaluate.
Dozens of games have been played here with various continuations, including the typical Dragon
exchange sacrifice 19 ... Rxd5!?. Karpov won twice against inferior continuations (and inferior
opposition) in 1986.
One game proceeded 19 ... Ng6? 20 Qxh5 Qa5 21 Rh1! (even 21 Qxg6 is possible, since taking the
queen runs into a cute mate after 21 ... fxg6 22 Nf6+ Kf8 23 Nxd7 or 22 ... Kh8 23 Rh1+ Bh6 24
Rxh6+ Kg7 25 Rh7+ Kf8 and White has a choice of three mates in one!) 21 ... Rxd5 22 Qh7+ Kf8 23
Bh6 Bxh6 24 Qxh6+ Kg8 25 Bxd5 e6 26 Nf5 and Black resigned in A.Karpov-P.Petri, Groningen
(clock simul) 1986.
The other game went 19 ... Rc4? 20 Bxc4 bxc4 21 Qxh5 f6 (Black is a pawn up but the strength of the opposing attack means that
he is on the ropes) 22 f4! Nf7 (22 ... fxg5 23 fxe5 dxe5 24 Rh1 gives White a winning attack) 23 Bh4 Qb8 24 Rh1 c3 25 b3 Qb7 26 f5
Ne5 27 Ne6 and Black resigned in A.Karpov-A.Sznapik, Dubai Olympiad 1986. It is unclear why Black didn’t follow Kiril Georgiev’s
example from three rounds earlier, where he managed a draw with Karpov in 54 moves after 15 ... a5!?.
13 Bxc4 Rxc4 14 h5!
Opening the h-file at the cost of a pawn is probably the most dangerous approach.
14 ... Nxh5 15 g4 Nf6
This position has been extensively tested and many moves have been tried for White. In principle,
White would like to exchange the dark-squared bishop (which is the black king’s main defender), but
the immediate 16 Bh6 leaves the d4-knight vulnerable to the tactical shot 16 ... Nxe4! 17 Qe3 Rxc3
18 bxc3 Nf6, as in E.Geller-V.Korchnoi, Moscow (4th matchgame) 1971 (1-0 in 43).
16 Nde2
Karpov’s choice withdraws this piece from its vulnerable position, while overprotecting the knight
on c3 against Black’s standard ... Rxc3 exchange sacrifice.
Here’s his own description: “A logical explanation for this early knight retreat from the centre runs
roughly as follows. The square c3 is the most important point in the white piece set-up, on which
Black is concentrating his blows. Here one often sees the exchange sacrifice on c3, when the
opponent obtains a strong attack (it is notable, too, that after doubling the white pawns on the c-file,
Black’s position is so rich in possibilities that, even if he has no attack, he can hold the balance; this
was magnificently demonstrated in some of his games by the great grandmaster Leonid Stein). Thus,
the basic idea of 16 Nde2 is to strengthen the square c3. In addition, the knight from e2 can easily
transfer to a direct attack on the enemy.”
Exercise: What is White’s intention, and how should Black respond?
16 ... Qa5?!
Allowing the trade of dark-squared bishops, which gives White a dangerous attack.
Answer: As noted by Yakovich in his excellent Sicilian Attacks: “Later it was established that the typical Dragon prophylaxis 16 ...
Re8! promises Black sufficient counterchances. In order to justify the retreat of the knight to e2, White should make the breakthrough
e4-e5. But both after the immediate 17 e5 Nxg4! 18 fxg4 Bxg4 19 exd6 Qxd6! and after the preliminary 17 Bh6 Bh8 18 e5 Nxg4 19 fxg4
Bxe5! (since the white bishop is no longer ‘hanging’ on e3 after the opening of the e-file, Black does better to take this pawn) 20 Bf4
Qa5!? or 20 ... Bg7!? Black obtains sufficient compensation for the piece.”
17 Bh6 Bxh6
Black’s other typical exchange sacrifice, 17 ... Bh8, looks insufficient here. In any event, the
players were still playing known moves and Karpov’s novelty only lands on move 19.
18 Qxh6 Rfc8 19 Rd3!?
A typical Karpov move, overprotecting the knight on c3. Korchnoi might even have been willing to
sacrifice two exchanges if it had broken the pawn cover around White’s king.
19 ... R4c5
Black can try to hang on with 19 ... Qd8.
20 g5 Rxg5
20 ... Nh5 21 Nf4 Rxc3 (21 ... Rxg5 22 Rd5! wins) 22 bxc3 Qe5 23 Nxh5 gxh5 24 Rxh5 Qg7 25 f4
is also horrible for Black.
21 Rd5! Rxd5 22 Nxd5 Re8
Exercise: How can White continue the attack?
Answer:
23 Nef4!
Bringing in the last piece, with the threat of 24 Nxf6+ exf6 25 Nd5, is the only way to win. A
straightforward continuation like 23 Nxf6+ exf6 24 Qxh7+ Kf8 leads nowhere; for instance, 25 Nf4
Bc6 and Black covers the d5-square, with equality.
23 ... Bc6
23 ... Be6 is no better, due to 24 Nxe6 fxe6 25 Nxf6+ exf6 26 Qxh7+ Kf8 27 Qd7 and wins
(Botvinnik) or indeed 27 Qxb7 as in Schleupner-Beuchler, German League 1999 (1-0 in 30).
Exercise: Find the tactical blow to finish the game.

Answer:
24 e5!
Instead, 24 Nxf6+ exf6 25 Nh5 looks equally conclusive, but Black has a defence: 25 ... Qg5+! 26
Qxg5 fxg5 27 Nf6+ Kg7 28 Nxe8+ Bxe8 and the endgame after 29 Rd1 Kf6 30 Rxd6+ Ke5 is quite
unclear – frankly, I would prefer Black, who is about to create a dangerous passed pawn on the
kingside.
Actually, 24 e5! works against 23 ... Be6 as well, since 24 ... Bxd5 transposes to the game, while
24 ... dxe5 25 Nxe6 fxe6 26 Nxf6+ exf6 27 Qxh7+ Kf8 28 Qxg6 leads to a decisive attack.
24 ... Bxd5
24 ... dxe5 25 Nxf6+ exf6 26 Nh5! wins on the spot. Black resigned here in C.Metz-I.Kocsis,
Austrian Junior Championship 1993 – a striking example of what happens when you don’t know your
chess history!
25 exf6 exf6 26 Qxh7+
Now 26 Nh5?? would be a disaster: 26 ... Re1+ 27 Rxe1 Qxe1 mate.
26 ... Kf8 27 Qh8+ 1-0
It’s never too late to blunder: 27 Nxd5?? Re1+ mates again.
Korchnoi resigned in view of 27 ... Ke7 28 Nxd5+ Qxd5 29 Re1+, winning either the queen or the
rook.
Game 28
A.Karpov-J.Dorfman
Sicilian Defence
1 e4 c5 2 Nf3 d6 3 d4 cxd4 4 Nxd4 Nf6 5 Nc3 e6 6 g4

The Keres Attack is probably the main reason that Scheveningen players nowadays prefer to reach
their favoured structure via a Najdorf move order with 5 ... a6 (although GMs such as Movsesian,
Nisipeanu and Hou Yifan have demonstrated that, even in the Keres Attack, Black has resources).
Karpov was equally lethal against the Sicilian with sharp approaches like this one, as well as quieter,
more classical play.
6 ... Be7
In the next game we will examine what happens after the more principled 6 ... h6.
One of the main justifications of 6 g4 is that the “thematic” strike in the centre with 6 ... d5 doesn’t
equalize: 7 exd5 Nxd5 8 Bb5+ Bd7 9 Nxd5 exd5 10 Qe2+ has been considered good for White since
the classic R.J.Fischer-S.Reshevsky, US Championship, New York 1966 (1-0 in 43).
However, 6 ... e5!? is considerably more interesting, as recently essayed by Nisipeanu and
Kurnosov, amongst others.
7 g5 Nfd7 8 h4
To modern eyes Black has allowed White to gain kingside space a bit too easily.
8 ... Nc6 9 Be3 a6 10 Qe2
The queen is usually developed to d2 in such structures but, as this game demonstrates, this piece
can also be very effective on e2, where it looks down the e-file, prevents an eventual ... Nc4 and
retains options of going to g4 or h5.
10 ... Qc7 11 0-0-0
Interestingly, the excellently-prepared Hou Yifan went for this position in 2012. Nevertheless,
after 11 Bh3 Nc5 12 0-0-0 Nxd4 13 Bxd4 0-0, as in A.Hunt-Hou Yifan, Gibraltar 2012 (0-1 in 28), I
like 14 f4 b5 15 a3 for White, since Black will struggle to progress with his queenside counterplay.
11 ... b5 12 Nxc6 Qxc6 13 Bd4 b4
Exercise: Where should White retreat the knight?
Answer:
14 Nd5!
Okay, trick question. But this sacrifice is absolutely standard in Open Sicilians. Here it is
relatively easy to find, not least because 14 Nb1? is depressing.
What does White get for the knight? A few things:
1. The main advantage is that the e-file is opened, with the black king sitting uncomfortably at the
end of it. Clearly the Nd5 sacrifice is much harder to justify after Black has castled.
2. White gains a certain amount of time. In Open Sicilians generally, White enjoys a development
advantage. Black has spent additional time on ... b5-b4 to hit the knight. By advancing rather than
retreating, White can argue that this time was wasted (and, in this particular position, he gains a
tempo on the queen).
3. White gains a pawn (here, in fact he will gain a second pawn after taking on g7). Moreover, the
structural change which occurs is heavily in his favour. As I mentioned above, the e-file is opened,
which is bad news for the black king. But the white e-pawn is also brought to d5, where it controls
key squares on c6 (an outpost for a white piece) and e6 (thus preventing Black from blocking the file
with ... Be6, for instance). The loss of Black’s e-pawn weakens the f5-square as well – while that
isn’t so significant in this position, when a white knight is on d4, Nf5 can be a very useful resource.
14 ... exd5
Black has to accept the offer. 14 ... Bb7 leaves White with a superior position for free after 15
Nxe7 Kxe7.
15 Bxg7!
The correct move order. Not 15 exd5? Qxd5, when Black attacks the rook on h1 and is ready to
block the e-file with ... Ne5.
15 ... Rg8 16 exd5
Black is faced with a difficult choice of where to move the queen.
16 ... Qc7?
Houdini suggests that 16 ... Qc5 would have held the balance. After 17 Re1 Ne5 18 Bxe5 dxe5 19 Qxe5 Kf8, White has three
pawns for the piece, but the black king has found some measure of safety and the bishop pair could count for something. The position
seems completely unclear.
17 Bf6 Ne5 18 Bxe5 dxe5
At first sight, this looks better than the position after 16 ... Qc5 – after all, Black holds the e5-
pawn.
Answer:
19 f4! Bf5
There is nothing better.
a) 19 ... exf4?? loses to 20 d6 with a fatal pin on the e-file.
b) Trying to grab the e-pawn with 19 ... Bd6 20 fxe5 Bxe5 loses to a number of moves, most
simply 21 d6 Qa5 (21 ... Qc5 22 Bg2 threatens both the a8-rook and 23 Rd5) 22 Re1 and White
regains the piece with interest.
c) Trying to keep the e-file closed fails tactically: 19 ... e4 20 d6! and 20 ... Bxd6 drops a rook to
21 Qxe4+.
20 Bh3!?
As always, Karpov’s primary concern is the opponent’s counterplay. He doesn’t want to allow 20
fxe5 Rc8, although White could defend comfortably with 21 Rh2 and is clearly better.
20 ... Bxh3 21 Rxh3 Rc8 22 fxe5
White now has three pawns for the piece and a tremendous central pawn majority. He threatens to
win straightforwardly with d5-d6.
22 ... Qc4 23 Rdd3?
Karpov stumbles, selecting a solid move which gives up White’s advantage. The endgame after 23
Qxc4? Rxc4 is quite playable for Black too – for one thing, White must deal with the immediate threat
of 24 ... Rxh4 25 Rxh4 Bxg5+.
Computer analysis shows that 23 Qf2! was the strongest; for instance, 23 ... Qxa2 24 d6 Qa1+ 25
Kd2 Qxb2 (or 25 ... Qa4 26 Rb3) 26 Ke1 Rc6 (not 26 ... Rxc2? 27 d7+ Kd8 28 Qb6+ and mates) 27
Qf4! Qxc2 28 dxe7 Kxe7 29 Rhd3 and White’s initiative is too strong. 23 Rhd3!? is also a good
move, as after 23 ... Qxh4 24 Kb1, both e5-e6 and d5-d6 are coming.
23 ... Qf4+ 24 Kb1 Rc4!
Dorfman defends tenaciously, preparing to activate his rook via e4. White’s weak back rank is
significant in some variations.
25 d6 Re4 26 Rhe3 Rxe3 27 Rxe3 Qxh4!
Now 28 dxe7? fails to 28 ... Qh1+ 29 Qe1 Qxe1+ 30 Rxe1 Rxg5.
28 Qf3!
White defends against the threatened ... Qh1+ and sets up ideas of Qc6+ and Qa8+.
28 ... Qxg5?
Now Black goes wrong. After simply 28 ... Bd8, the engines can’t see more than a draw for White.
29 Re1?!
Another inaccuracy. Although White retains some advantage after this move, there was a stronger
continuation available.
Answer: This was the time to cash in with 29 Qc6+! Kf8 (29 ... Kd8 30 Rg3! wins) 30 dxe7+ Qxe7 (after 30 ... Kxe7 31 Qd6+ Ke8
32 Rd3 and White has a decisive attack, while easily covering his back rank; and 30 ... Kg7 loses to the fantastic 31 Rg3!! Qxg3 32 Qf6
mate) 31 Qh6+ Rg7 32 Re1 and, despite the equal and limited material, Black faces a difficult defence. Once White has made some
provision for his king’s safety, he can set about exploiting the weak enemy pawns, and Black’s king is uncomfortable too.
29 ... Qg2
The relatively best 29 ... Qg4 (here 29 ... Bd8 30 Qc6+ Kf8 31 Qd7 Kg7 32 e6 is strong for White)
30 Qc6+ Qd7 31 Qe4 Bd8 32 Qxh7 leaves White with two pawns and a continuing initiative for the
piece. Houdini’s top line runs 32 ... Qg4 33 e6 Rg7 34 d7+ Ke7 35 exf7+ Kxf7 36 Qd3, when Black
still faces severe practical problems defending his king and stopping the d7-pawn.
30 Qf5 Rg6 31 Rf1 Qd5 32 dxe7 Kxe7
The position has finally clarified. Material is equal, but that’s where the good news ends for
Black. With four weak pawns and an unsafe king, he is in for a very hard time.
33 Qf4 a5 34 Qh4+ Ke8 35 Qxh7
Now White has an extra pawn to add to his other advantages.
35 ... Qf3 36 Qh8+ Ke7 37 Qh4+ Ke8 38 Qc4 Qb7 39 b3
Finally Karpov finds the time to create some luft for his king.
39 ... Re6 40 Rg1 Rxe5 41 Rg8+ Ke7 42 Qh4+ Kd7 43 Qf6!
This quiet move is the strongest, hitting the rook and the f7-pawn while retaining all White’s
threats.
43 ... Re7 44 Qf5+
44 Rg4! was even stronger, with the fatal threat of Rd4+.
44 ... Kd6 45 Qxa5 Re5 46 Qd8+ Ke6 47 Kb2!
Good technique. Karpov takes time out to safeguard his king, while retaining all the threats.
47 ... f6 48 Rf8 Qg7 49 Qc8+ Kd5 50 Qc4+ 1-0
After 50 ... Kd6 51 Rd8+, White wins the queen and the game.
Game 29
A.Karpov-G.Sax
Linares 1983
Sicilian Defence
1 e4 c5 2 Nf3 e6 3 d4 cxd4 4 Nxd4 Nf6 5 Nc3 d6 6 g4 h6

A more combative choice than Dorfman’s 6 ... Be7.
7 Rg1
7 g5 is possible but spoils the white structure. After 7 ... hxg5 8 Bxg5 Nc6 9 h4 a6 10 Be2 Qb6 11
Nb3 Bd7 12 Qd2 Qc7 13 0-0-0 b5 14 a3 Ne5 15 f4 Nc4 16 Qd4 Rc8 17 h5, a draw was agreed in
A.Karpov-A.Yusupov, Vienna 1996.
7 ... Be7 8 Be3 Nc6 9 Qe2
Karpov explains: “The idea behind this somewhat unusual move is to castle queenside, and then to
make the quickest possible use of the white rook and black queen’s opposition along the d-file.”
9 ... Bd7
Dvoretsky assesses the position after 9 ... Nxd4 10 Bxd4 e5 11 Be3 Be6 12 0-0-0 Nd7 13 Kb1 a6
14 f4 exf4 15 Bxf4 Bf6 as “either unclear, or slightly favourable to White”, A.Karpov-U.Andersson,
Turin 1982 (½-½ in 61).
10 h4 Nxd4 11 Bxd4 e5 12 Be3 Bc6
12 ... Rc8 may be well met by 13 g5.
13 Qd3
If 13 f3, Dvoretsky suggests that 13 ... Nh7 is “an unpleasant reply”, although with the pawn
sacrifice 14 g5!? White gets serious compensation; for instance, 14 ... hxg5 15 hxg5 Bxg5 (or 15 ...
Nxg5 16 0-0-0 Ne6 17 Qd2 Qd7 18 Bc4) 16 0-0-0 Bxe3+ 17 Qxe3 Kf8 18 Bc4 Nf6 19 Kb1 with
continuing pressure.
Exercise: Can Black play 13 ... d5 - ?
13 ... Qa5
Answer: After 13 ... d5?!, Dvoretsky gives the variation 14 exd5 Nxd5 15 0-0-0! (15 Nxd5 Qxd5 16 Qxd5 Bxd5 17 Bb5+ Bc6 18
Bxc6+ bxc6 19 h5 gives White the better endgame, but Black might improve with 17 ... Kf8) 15 ... Nxe3 16 Qxe3 Qa5 17 Bb5! 0-0 (or
17 ... Bxb5 18 Qxe5 “with a great advantage”) 18 g5 “when Black’s position grows shaky.”
Dvoretsky’s 13 ... Nh7! again seems best, when 14 h5 gives Black the g5-square for his bishop,
although 14 g5!? still keeps a strong initiative.
14 0-0-0 Nxe4
The Hungarian GM launches a bold combination to free his position. It takes even modern
computers a relatively long time to realize that this combination leaves White with an advantage,
based on the line selected by Karpov.
15 Nxe4 d5
An extraordinarily complex position.
Exercise: How would you play with White?
Answer:
16 Qb3!?
Karpov’s wonderful pawn sacrifice forces Black to open the position, when the position of his
king on e8 will tell. White has a number of tempting alternatives:
a) 16 Qc3 Qxa2 (16 ... Qxc3 17 Nxc3 d4 18 Nb5 is good for White) 17 Bc5 leads to enormous
complications which seem to favour White.
b) Dvoretsky prefers Viktor Bologan’s idea: 16 Nd2!! d4 17 Nb3 Qxa2 18 Qf5 dxe3 19 Bc4 0-0
20 g5 “when White has a decisive attack.” Houdini is initially unimpressed, giving complete equality
(or “0 00” in modern parlance), but soon changes its mind and agrees!
16 ... dxe4 17 Bc4 Rf8
Of course not 17 ... 0-0? 18 g5 with a completely winning attack.
18 Rd5!?
Karpov sacrifices an exchange to dominate the light squares. The simple 18 g5 would have been
very strong too.
18 ... Bxd5 19 Bxd5 Rd8!
20 Bc4!?
White is now the exchange and a pawn down, but the threat of Bb5+ is hard to meet. The obvious
20 Bxb7 allows 20 ... Bc5 with a possible draw after, for instance, 21 Bc6+ Ke7 22 Qb7+ Ke6 23
Qb3+ Ke7.
20 ... Bb4!
Sax finds the best defence. As Dvoretsky notes, 20 ... Bc5 is now impossible due to 21 Bb5+,
while 20 ... Rd6 21 Qxb7 Kd8 22 g5 gives White a strong initiative.
21 c3 b5! 22 Be2 Bd6 23 Qd5!?
Again spurning material in order to keep the initiative. After 23 Bxb5+ Ke7, the black king is
reasonably comfortable in the centre, while the half-open b-file could prove problematic for the
white king – the position is quite unclear. Instead, Dvoretsky suggests continuing the attack with 23
g5.
23 ... Ke7?
In the face of such unusual and powerful play, Sax stumbles.
Karpov himself recommended 23 ... Qc7 with unclear play after, for instance, 24 Bxb5+ Ke7 25
Qxe4.
Dvoretsky also notes: “For some reason, the World Champion did not examine the natural 23 ...
a6!?, protecting the pawn and limiting the mobility of White’s lightsquare bishop. (Here again, the
inclusion of the moves 18 g5 hxg5 19 hxg5 g6 would have been good for White: his bishop could now
get out via g4.) The continuation 24 Rd1 Qc7 25 Qxe4 leaves White with enough compensation for
the exchange, but apparently no more than that.” However, Houdini finds 24 Qc6+! Ke7 (24 ... Rd7
25 Rd1 Ke7 transposes) 25 Rd1 Rd7 26 g5!, creating the threat of Bg4 after all, when Black is in
serious trouble.
Exercise: How can White tactically exploit Black’s last move?
Answer:
24 Bc5! Bxc5
Note that 24 ... f6? loses at once to the crushing 25 Bc4!, since taking the c4-bishop drops the
queen and otherwise White walks in on the light squares.
25 Qxe5+!
This intermezzo is the whole point. After 25 Qxc5+ Rd6, White has no more than a draw with 26
Qxe5+ Re6 27 Qc5+ Rd6 28 Qe5+.
25 ... Kd7 26 Qxc5
White will now collect one or two pawns for the exchange and still has a strong initiative.
26 ... Qc7
Answer:
27 Qf5+!
Cutting off the black king’s escape to the queenside. After 27 Bxb5+ Kc8 28 Qf5+ Kb8 29 Qxe4,
White has two pawns for the exchange but has relinquished the initiative.
27 ... Ke7
The black king faced a difficult choice. Houdini suggests 27 ... Kc6 as better, but White is still on
top after 28 Qxb5+ Kd6 29 Qb4+ Qc5 30 Qxe4. Compared with the previous note, the black king is
on d6 rather than b8, which leaves White with a strong attack in a position of approximate material
equality.
28 Qxe4+ Kd7 29 Qf5+ Ke7
Exercise: What should White play now?
Answer:
30 Re1!
Setting up a horrible discovered check. This is more logical than 30 Bxb5 since the bishop might
move to a different square. 30 g5 is a decent alternative, but gives Black some breathing space.
30 ... Rd6
The black king can seek safety on the kingside with 30 ... Rh8 31 Bxb5+ Kf8, but after the simple
32 Kb1, White is clearly better. He has two pawns for the exchange, more active pieces (in
particular, the rook on h8 struggles to get back into play), and a safer king.
Exercise: Find the best continuation of the attack.
Answer:
31 Bc4+!
The materialistic 31 Bxb5+? would allow 31 ... Re6 32 Rd1 Rd8, when Black is not worse.
31 ... Kd8 32 Bxb5!
By “losing” a tempo, Karpov has prevented Black from using his rooks effectively in defence.
32 ... a6 33 Ba4 g6 34 Qf3 Kc8
The poor black king seeks refuge again on the queenside.
Exercise: Find the decisive combination.
Answer:
35 Re7!!
A superb shot.
35 ... Rd1+
As good or bad as anything else. After 35 ... Qxe7 36 Qa8+ Kc7 37 Qa7+, Black has to give up the
queen since 37 ... Kd8 38 Qb8 is mate.
36 Kxd1 Qxe7?
Sax has had enough. 36 ... Rd8+ forces White into an easily winning queen endgame after 37
Bd7+! Rxd7+ 38 Rxd7 Qxd7+ 39 Kc2.
37 Qa8+ Kc7 38 Qa7+ Kd6 39 Qb6+ 1-0
It’s mate in two: 39 ... Ke5 40 Qd4+ Ke6 41 Bb3 mate.
A really tremendous attacking game and a forgotten classic in my view.
Game 30
A.Karpov-S.Dolmatov
Amsterdam IBM 1980
Sicilian Defence
1 e4 c5 2 Nf3 Nc6 3 d4 cxd4 4 Nxd4 Nf6 5 Nc3 e5 6 Ndb5 d6 7 Bg5 a6 8 Na3 b5 9 Nd5
Karpov preferred this primarily positional continuation to the more chaotic variations that arise
after 9 Bxf6.
9 ... Be7 10 Bxf6 Bxf6 11 c3 0-0 12 Nc2 Rb8
A prophylactic move, aiming to discourage White’s a2-a4 break.
13 Be2 Bg5 14 0-0 Be6 15 Qd3 Qd7 16 Qg3 f6 17 Rfd1 a5 18 Na3 Na7
Answer:
19 h3!
An outstanding positional move, preparing Bg4, trading Black’s important light-squared bishop
which covers d5 and helps generate counterplay.
19 ... Kh8 20 Bg4 Rfc8?
The Sveshnikov can’t be handled in a passive manner. After the typical 20 ... f5 21 exf5 Bxf5,
Black would have had reasonable counterplay.
21 Bxe6 Qxe6
Exercise: Black intends 22 ... b4 with queenside

counterplay. How might White prepare for this?
Answer:
22 Qd3!
A perfect prophylactic move, combining White’s aims of pressure on the d6-pawn with
anticipation of Black’s only active plan.
22 ... Rc5
Now 22 ... b4 is met by 23 Nc4 bxc3 (23 ... Rc5 24 b3 leaves White completely stable on the
queenside) 24 bxc3 with a clear advantage for White, who will occupy the b-file, while the a5- and
d6-pawns need constant protection. Black isn’t in a position for an open battle on the queenside,
especially since the g5-bishop doesn’t actively participate.
23 Nc2 g6 24 b4!?
This move, coupled with the following sequence to move 27, is a fascinating decision. Karpov
eliminates Black’s queenside pawns, leaving a white pawn on c3 which, while technically a
weakness, will always be protected by a knight on d5. This transformation of the position has two
purposes:
First, the game will now be played effectively for two results, since Black’s winning chances will
be minimal even if things go wrong for White.
Second, Black will need to make sure that the white rooks can’t penetrate down the open a- and b-
files, where they would combine very unpleasantly with the knight on d5.
Karpov’s choice was not forced by any means; for instance, 24 a3 would have kept a solid
advantage.
24 ... axb4 25 Ncxb4
Playing for a passed b-pawn with 25 cxb4!? is interesting too.
25 ... Rb7 26 a4 bxa4 27 Rxa4 f5!
A correct decision, trying to undermine White’s centre.
28 Ra6
Aiming at d6.
28 ... Nc8 29 Ra8
A white rook has finally landed on the eighth rank. Although its activity is contained for the
moment, this is certainly not a good sign for Black.
29 ... Kg7 30 Qe2 fxe4 31 Qxe4 Qf5 32 Qe2 Rf7 33 c4 Bh4 34 Rf1
Karpov easily deals with Black’s counterplay against the f2-pawn.
34 ... Ne7
Characterized as a positional mistake by Dvoretsky. At the moment White has two knights
competing for one square (d5) and Black’s last move resolves this problem for him.
Dvoretsky states that “ ... by continuing 34 ... e4! Black would have got a good game. Indeed,
White has to reckon with the possibility that after ... Qf5-e5-d4 there will be threats to his only real
weakness – the pawn on c4.” However, after 35 Ra3!, intending to target the e4-pawn with Re3,
Black’s position is still unenviable, and the knight on c8 hardly impresses.
35 Ra6 Qd7 36 Qe4 Nxd5 37 Nxd5
White has achieved the “dream” scenario of a good knight on d5 vs. a passive dark-squared
bishop. Defending such a position in practice is deeply unpleasant since the d5-knight can never be
exchanged. Nevertheless, considerable work remains to be done.
37 ... Be7 38 Rfa1 Bf8
Dolmatov decides to shore up his weak pawn on d6 (and, perhaps, defend some dark squares
around his king); the price of this decision is that his game becomes increasingly passive.
39 Qe2 Rc6
Answer:
40 R6a3!
Based on trading the f7-rook, which will have the effect of disastrously weakening Black’s second
rank. This wasn’t the only move to achieve this idea – 40 R6a5! and 40 R6a4 are also good, if
followed by 41 R1a3! and 42 Rf3. But it’s an idea which, if obvious in retrospect, is hard to see
when focused on the queenside.
Kasparov’s admiration for the move is palpable: “Karpov’s ability to see the entire board is
staggering. After doubling rooks on the a-file, he suddenly sharply changes the direction of the attack
and makes an unexpected manoeuvre, which decisively strengthens his position. Since Black does not
have a more useful move than the return of his rook to c5, White removes the rook on f7 from the
game, thereby weakening the defence of the king. His other rook can now threaten not only to invade
on the a-file, but also to switch to the kingside via a3 and f3. What becomes noticeable here is the
dominating role of the knight at d5, which prevents the rook at c5 from returning in time to the aid of
its king, leaving it performing the role of an observer.”
40 ... Rc5 41 Rf3
41 h4!? was another option, trying to force a kingside concession, but I prefer Karpov’s
methodical play.
41 ... Rxf3 42 Qxf3 Qf7 43 Qg4
Here 43 Qb3! was extremely strong, defending c4, while preparing to bring the rook to a7 or a8.
43 ... h5 44 Qe4 Rc8
Answer:
45 Ra3!
The same rook lift as at move 40 – only this time, rather than exchanging a defender, the rook will
be a powerful attacker. Something like 45 g4, as well as being objectively weaker than Karpov’s
solution, gives Black more practical chances since the white king is exposed.
45 ... Qf5 46 Ra7+ Kh6
Of course not 46 ... Kg8 47 Qxf5 gxf5 48 Nf6+ Kh8 49 Rh7 mate; while 46 ... Kh8 allows an idea
based on the same motif: 47 Qxf5 (Houdini prefers 47 Nf6!? Qxf6 48 Qb7 Qg7 49 Qxc8 Qxa7 50
Qxf8+ Kh7 51 Qxd6, which has to be winning with the extra passed pawn, the weak e5-pawn, and a
safer white king; all the same, queen endgames are notoriously tricky, so this might not be the best
practical decision) 47 ... gxf5 48 Nf6 Bg7 49 Nxh5, winning a pawn.
47 Qe3+
Forcing a weakening of the black king’s pawn cover, since 47 ... Qg5 drops the queen to 48 Rh7+.
47 ... g5 48 Qe2
Unusually, Karpov allows his opponent to activate some pieces. 48 Qb3 prevented ... Rb8 and
retained total control, though White is winning either way.
48 ... Rb8 49 g4 hxg4 50 hxg4 Qb1+ 51 Kg2 Rb7 52 Rxb7 Qxb7
Answer: Black might have harboured hopes of drawing chances in view of the reduced material, but his position is dead lost. The
knight is overwhelmingly stronger than the bishop, which can’t compete for the light squares. In addition, it is well established that queen
and knight can create threats more efficiently than queen and bishop.
53 Qf3 Qc8
53 ... Kg7 was more tenacious, though Black loses the g5-pawn for starters after 54 Qf6+.
54 Qf6+ Kh7 55 Qf7+ 1-0
56 Nf6 will lead to mate after any Black move, so Dolmatov threw in the towel.
King’s Fianchetto
In designing his opening repertoire after 1 d4, Karpov moved between several systems. However,
one set-up which always seemed particularly to suit his strengths is that of the King’s Fianchetto
against ... g7-g6 defences, where White safely plays for a small edge and looks to neutralize Black’s
counterplay.
Game 31
A.Karpov-J.Polgar
Las Palmas 1994
1 d4 Nf6 2 Nf3 g6 3 c4 Bg7 4 g3 0-0 5 Bg2 d6 6 0-0 Nc6 7 Nc3 e5

Against the most popular move 7 ... a6, Karpov has generally steered clear of the critical
variations after 8 d5 and preferred the solid system with 8 Re1 Rb8 9 Rb1; for instance, 9 ... Bd7 10
b4 e5 11 dxe5 Nxe5 12 Nxe5 dxe5 13 b5 Be6?! (13 ... axb5 gives Black good chances to equalize) 14
Ba3 Re8 15 bxa6 bxa6?! (White’s queen will be much more effective than Black’s, so 15 ... Qxd1
was better) 16 Bc6 (here 16 Rxb8, avoiding the queen exchange, was more precise) 16 ... Nd7? 17
Rxb8 Qxb8 18 Nd5 Qd8 19 Qa4 and White had built up a winning advantage through simple means in
A.Karpov-A.Shirov, Eurotel Trophy (rapid; 4th matchgame), Prague 2002 (1-0 in 50).
8 d5 Ne7
Black’s play is consistent with the main lines of the Classical King’s Indian; for example, 1 d4 Nf6
2 c4 g6 3 Nc3 Bg7 4 e4 d6 5 Nf3 0-0 6 Be2 e5 7 0-0 Nc6 8 d5 Ne7.
Question: What is the difference in this position, and whom does it favour?
Answer: Obviously, the difference lies in the placement of White’s light-squared bishop, which is on g2 rather than e2. At the risk of
oversimplifying, the fianchettoed bishop makes it much more difficult for Black to carry out a standard King’s Indian attack with ... Nd7,
... f7-f5-f4, and pushing the g-pawn. We see Polgar trying to carry out a similar plan in this game, and getting nowhere.
9 e4
As noted by Mikhalchishin, after 9 c5 Ne8 10 cxd6, rather than the apparently automatic recapture
10 ... cxd6, Black gets good play with 10 ... Nxd6!. Judit successfully defended this position twice: in
A.Adorjan-J.Polgar, Hungarian Championship, Budapest 1991 (½-½ in 20) and L.Portisch-J.Polgar,
Veterans vs. Women, Monte Carlo 1994 (½-½ in 43).
Mikhalchishin also suggests the interesting 9 a4, aiming to gain space on the queenside with a5.
9 ... Nd7 10 Ne1 f5 11 Nd3 Nf6 12 Bg5 fxe4
White’s idea after 12 ... h6 is to attack the centre with 13 Bxf6 Bxf6 14 f4. Karpov had this in a
game the month before, which proceeded 14 ... exf4 15 Nxf4 Be5 16 exf5 Bxf5 17 Qd2 c6 18 Kh1
Qb6 19 Rae1, when Dautov suggests that White holds a small advantage in view of the weakness of
the e6-square, A.Karpov-B.Gelfand, Dos Hermanas 1994 (½-½ in 59).
13 Nxe4 Nf5 14 Re1
A strong move and a novelty, entirely in Karpov’s style – he continues development, increases his
central control and anticipates any Black counterplay.
As it happens, White has been very successful with the more concrete 14 g4!?, when the rook can
prove effective on f1 after a subsequent f2-f4, trying to exploit the pin on the f6-knight. Karpov would
not play such a move unless he was sure of the consequences (probably through prior preparation).
14 ... h6 15 Nxf6+ Bxf6 16 Bd2
16 Bxf6 Qxf6 seems comfortable for Black.
16 ... Kh7 17 Bc3 Bd7 18 c5!
This is a thematic break for White in many King’s Indian positions, and there is no reason to
refrain from it here.
18 ... b6?!
Question: What’s wrong with this move?

Answer: In general terms, it violates an important strategic rule (which, like all rules in chess, has some exceptions!): namely, not to
make pawn moves on the side where you’re weaker. In the long run, White is winning on the queenside, and no arrangement of the black
pawns can establish parity on this side of the board once White gets c4-c5 in. More concretely, this move allows White to play c5-c6,
when the pawn severely cramps Black’s queenside, while White retains pawn control over the e6-square.
I think Polgar would not have played in this fashion if she had seen a way for White to break
through to e6. However, it must be acknowledged that Black’s position is already deeply unpleasant.
For instance, Karpov gives 18 ... Nd4 19 Bxd4 exd4 20 c6 bxc6 21 dxc6 Bf5 22 Nf4 as clearly better
for White.
19 c6
Very strong, although White had a promising alternative: Mikhalchishin assesses 19 cxd6 cxd6 20
Nb4 as strategically winning for White.
19 ... Be8
Answer:
20 f4!
Breaking down Black’s centre and gaining access to e6 for the white knight. The f2-f4 break is a
very typical resource in the Fianchetto King’s Indian, which allows White to fight for space on the
kingside. The fact that Black cannot obtain a kingside space advantage is one of the main reasons that
the white king tends to be much safer in the Fianchetto lines than in the Classical Variation with the
bishop on e2.
20 ... exf4 21 Nxf4 Bf7
Black hurries to protect the horribly weak e6-square.
22 Qd3 Bxc3 23 bxc3 Qg5
Answer:
24 h4!
White has a range of strong continuations, but this one is clearly the best.
24 ... Qf6
If Black takes on g3, then after 24 ... Qxg3 25 Qxg3 Nxg3 26 Re7, the c7-pawn drops, leaving
White with a monster passed pawn on c6.
25 Ne6 Bxe6 26 Rxe6 Qg7
Exercise: How can White finish the game?
Answer:
27 h5!
Cracking open the light squares and decisively undermining the knight on f5.
27 ... Ne7 28 Rae1 1-0
As with many of Karpov’s strategic masterpieces, Polgar resigns in a position with equal material,
but facing irresistible threats.
Game 32
A.Karpov-B.Gelfand
FIDE Candidates final (4th matchgame), Sanghi Nagar 1995
Grünfeld Defence
1 d4 Nf6 2 c4 g6 3 g3 c6
Gelfand goes for the most solid set-up, namely ... c7-c6 and ... d7-d5. Kasparov successfully
defended this structure against Karpov in their World Championship matches in 1986 and 1987, in
games which substantially mapped out the best plans for both sides.
4 Bg2 d5 5 cxd5 cxd5 6 Nf3 Bg7 7 Ne5 0-0 8 Nc3
8 ... e6
Gelfand easily equalized in a previous encounter after 8 ... Bf5 9 0-0 Ne4 10 Be3 Nxc3 11 bxc3
Nc6 12 Qb3 e6 13 Nxc6 bxc6 in A.Karpov-B.Gelfand, Linares 1991 (1-0 in 57), but as Kasparov
notes, 10 Bf4! is more active.
9 0-0
Karpov twice essayed an alternative idea, 9 Bg5 Qb6 10 Qd2, against Timman: 10 ... Nfd7 11 Be3
(11 Nf3 Nc6 12 Rd1 Nf6 13 0-0 Bd7 14 Bxf6 Bxf6 15 e4 was A.Karpov-J.Timman, Bugojno 1986,
1-0 in 70; here 15 ... dxe4 16 Nxe4 Bg7 would have given Black comfortable play) 11 ... Nc6 12
Nxc6 bxc6 13 Rc1 Qb4 14 0-0 Rb8 15 b3 c5 16 Rfd1 cxd4 17 Bxd4 Bxd4 18 Qxd4 Qxd4 19 Rxd4
with a slightly better endgame for White in A.Karpov-J.Timman, Kuala Lumpur (2nd matchgame)
1990 (½-½ in 42).
9 ... Nfd7
9 ... Nc6 leads to queenside weaknesses after 10 Nxc6 bxc6 11 Na4. The following game is a good
demonstration of White’s ideas: 11 ... Nd7 12 Bf4 Qa5 13 Bd6! (a standard idea, transferring the
bishop to its best diagonal, namely a3-f8; if Black trades with ... Bf8 he will be left with weak dark
squares) 13 ... Re8 14 b3 Ba6 15 Re1 Bb5 16 Nc5 Nxc5 17 Bxc5 Qd8 (perhaps Black should seek
activity with 17 ... e5, though I prefer White in view of his superior structure) 18 a4 Ba6 19 Rc1 Qd7
20 Qd2 Reb8 21 Qc3 Bc8 22 b4 Ba6 23 e4! (a very thematic break in the Symmetrical Grünfeld –
capturing on e4 is basically out of the question, since the c6-pawn would become horribly weak and
White’s light-squared bishop would be extremely strong on the long diagonal; 23 e3 is good too,
preparing to trade light-squared bishops with Bf1, when the c6-pawn will again be vulnerable) 23 ...
Bc4 24 e5 with the better chances for White in A.Karpov-M.Tortarolo, San Giorgio (rapid simul)
1995 (1-0 in 56).
10 f4
Karpov was less successful with 10 Nf3 Nc6 (playing 10 ... Nf6 first avoids the next bracket) 11
Bf4 (Mamedyarov reinvigorated the line with 11 e4!?, when White gets quite an annoying initiative to
compensate for his IQP) 11 ... Nf6 (11 ... Qb6 is also possible, as in A.Karpov-J.Timman, World
Blitz Championship, Brussels 1987; 0-1 in 55) 12 Ne5 Bd7 13 Qd2 Nxe5 14 Bxe5 Bc6 15 Rfd1 Nd7
16 Bxg7 Kxg7 and Black was extremely solid in A.Karpov-G.Kasparov, World Championship (3rd
matchgame), London/Leningrad 1986 (½-½ in 35).
10 ... Nc6 11 Be3
11 ... f6
Other moves:
a) 11 ... Ndxe5 12 fxe5 f6 13 exf6 Rxf6 14 Qd2 Bd7 15 Kh1 Rxf1+ 16 Rxf1 Qe7 17 Rd1 Rc8 18
a3 Bf6 19 Bg1 Bg5 20 Qe1 Nd8 21 e4 was better for White in A.Karpov-J.Timman, Amsterdam 1987
(½-½ in 57).
b) 11 ... Nb6 12 Bf2 Bd7 (12 ... Ne7 13 a4 a5 14 Qb3 Bd7 15 Rfc1 was slightly better for White in
A.Karpov-G.Kasparov, World Championship, 3rd matchgame, Seville 1987; ½-½ in 29) 13 e4 Ne7
14 Nxd7 Qxd7 15 e5 Rac8 (after 15 ... Rfc8 16 Rc1 Bf8 17 Bf3 Rc7 18 b3 Rac8, Black had a solid
position in A.Karpov-G.Kasparov, World Championship, 1st matchgame, Seville 1987, ½-½ in 30;
but as Makarychev and Kasparov note, 17 g4!? was more dangerous, with similar play to the Kamsky
game) 16 Rc1 a6 17 b3 Rc7 18 Qd2 Rfc8 19 g4 Bf8 20 Qe3 Nc6 21 f5! gave White a strong initiative
in A.Karpov-G.Kamsky, Moscow 1992 (1-0 in 54).
12 Nf3
The position after 12 Nd3 Nb6 13 b3 was tested in two more games between Karpov and Timman:
13 ... Bd7 (an improvement over the premature 13 ... Qe7?!, which allowed White’s bishop to be
improved with tempo after 14 a4 Bd7 15 Bc1! Rfd8 16 e3 Be8 17 Ba3 Qf7 18 Rc1 Bf8 19 Bxf8 Qxf8
20 g4!? and White had a pleasant advantage in A.Karpov-J.Timman, 4th matchgame, Kuala Lumpur
1990; 1-0 in 75) 14 Bf2 Qe7 15 Rc1 Rad8 16 Rc2 Be8 17 Rd2 Nc8 18 e3 Nd6 19 g4 Kh8 20 Qe2 f5
21 g5 Ne4 with equality in A.Karpov-J.Timman, FIDE World Championship (12th matchgame),
Netherlands 1993 (½-½ in 51).
12 ... Nb6 13 Bf2
13 ... Bd7
Black has mostly preferred 13 ... f5 14 Ne5 Bd7 as in A.Karpov-G.Kasparov, World
Championship (13th matchgame), London/Leningrad 1986 (½-½ in 40). As Kasparov observes: “The
manoeuvring battle that now commences promises to be protracted. White has a slight initiative
(mainly due to the difference in the strengths of the knights on e5 and b6), but the character of the
position does not promise him much, so long as the opponent is patient and accurate. One would think
that, with his style, Karpov could not have wished for anything more, but in this match he did not play
the set-up with g2-g3 again.”
14 e4
Taking advantage of Black’s omission of 13 ... f5 and seizing space in the centre.
14 ... dxe4 15 Nxe4 Nd5 16 Re1 b6 17 Nc3 Nce7 18 Qb3 a5 19 a3 a4 20 Qd1 Nc7
Gelfand rearranges his knights to afford extra protection to the e6-pawn, but this manoeuvre seems
rather cumbersome. 20 ... Nxc3 21 bxc3 Rc8 22 Rc1 b5 leaves Black with a solid position.
21 Rc1 Ned5
Exercise: How can White improve his pieces?
Answer:
22 Nd2!
Unleashing the g2-bishop on the long diagonal while re-routing the knight to a better square.
22 ... Re8 23 Nc4 Bf8 24 Ne4 Bb5 25 Re2 Be7 26 Rec2 Rb8 27 Qd2 Rf8 28 h4!
Gaining space on the kingside and giving White options of h4-h5.
28 ... Ne8 29 Ne3 Ng7 30 Nc3
Changing the structure was also worthwhile; i.e. 30 Nxd5 exd5 31 Nc3 and Black has trouble
covering his weaknesses on a4 and d5.
30 ... Nxc3 31 Rxc3 g5!
Gelfand finds the best chance to generate counterplay.

32 hxg5 fxg5 33 Ng4
33 fxg5 Bxg5 34 d5 exd5 35 Bxd5+ Kh8 36 Rd1 also offers White some advantage.
33 ... gxf4 34 gxf4 Bd6 35 Rf3 Be8
Here 35 ... Ba6! was a better idea , rerouting the bishop to the long diagonal, when 36 Ne5 Bxe5
37 fxe5 Nf5 leads to a balanced position.
36 Ne5 Nf5 37 d5!
37 ... Bxe5
Now Black will struggle to cover the dark squares. Otherwise 37 ... exd5 38 Qxd5+ Kh8 39 Nc6
Bxc6 40 Rxc6 leaves White with more active pieces and the bishop pair.
38 fxe5 Rb7
Gelfand resists tenaciously, bringing his worst piece into the defence. After 38 ... exd5 39 Rf4!
Qe7 (39 ... d4 40 Rf1 Qg5 41 Bh4 Qg6 42 Qe2 wins material in view of the threat of Rg4) 40 Qxd5+
Bf7 41 Qe4, White has the two bishops and a powerful passed e-pawn.
39 Kh2 Rg7 40 Bh3 Bh5 41 Rf4 Kh8?
Allowing White two connected passed pawns in the centre is hopeless. 41 ... Bg6 was tougher,
White is still dominates after 42 d6.
42 Bxf5 exf5
Exercise: Find the most convincing continuation for White.
Answer:
43 Bh4!
Bringing the bishop to f6 breaks the defence.
43 ... Qe8 44 Bf6 Bg4 45 Bxg7+ Kxg7 46 Rc7+ Kh8 47 e6 Qh5+ 48 Kg1 Qg5 49 Kf2 Qh4+ 50
Ke3 1-0
A Weapon Against the Queen’s Indian Defence

Following on from the previous section, it is unsurprising that Karpov regularly employed a
fianchetto against ... e7-e6 systems as well.
Game 33
A.Karpov-V.Salov
Linares 1993
1 d4 Nf6 2 c4 e6 3 Nf3 b6 4 g3 Bb7

Of course Black has other options here, including 4 ... Ba6, when 5 Qc2 c5 6 d5!? leads to a sharp
gambit which is currently in vogue.
5 Bg2 Be7 6 Nc3 Ne4 7 Bd2 Bf6 8 0-0 0-0 9 Rc1
I was alerted to Karpov’s fantastic results in this line by reading Mark Dvoretsky’s excellent
Chess Cafe Column, The Instructor, where the renowned trainer wrote as follows:
“Most likely, Black will exchange on d2 – then, by refraining from Qc2, White will have saved a
tempo. On the other hand, the tempo saved doesn’t mean too much here – in fact, this whole opening
variation has the reputation of being rather toothless, almost a drawing line. Anatoly Karpov would
no doubt disagree with this evaluation: he has scored many victories with it. He loves to manoeuvre
back and forth in quiet positions with a tiny edge for himself, provoking and then expertly capitalizing
on the tiniest inaccuracy from his opponent.”
9 ... c5
Karpov has played both sides of this variation:
a) 9 ... d6 10 d5 Nxd2 11 Nxd2 Kh8 12 Nde4 Bxc3 13 Rxc3 Nd7 14 f4 Qe7 15 Qc2 Nf6 16 Nxf6
Qxf6 17 e4 a5 gave Black a solid position in L.Ftacnik-A.Karpov, Dubai Olympiad 1986 (½-½ in
34).
b) 9 ... d5 10 cxd5 exd5 11 Be3 Na6 12 Qa4 c5 13 Rfd1 Qe8 14 Qa3 Nxc3 15 Qxc3 Qe7 16 Qd2
Rfe8 17 b3 Rad8 18 Re1 cxd4 19 Bxd4 left White with a slight structural advantage based on the
weakness of the IQP (Black’s activity doesn’t look fully sufficient to compensate). As discussed
elsewhere in this book, Karpov tended to be lethal in such positions, and went on to win in
A.Karpov-J.Van der Wiel, Amsterdam 1988 (1-0 in 39).
c) 9 ... Nxd2 10 Qxd2 d6 11 d5 Qe7 (Karpov’s own choice of 11 ... e5 12 b4 Nd7 13 Ne1 a5 14 a3
axb4 15 axb4 was enough for equality in E.Torre-A.Karpov, Brussels 1987; ½-½ in 31) 12 e4 Nd7
13 Nd4 Bxd4 14 Qxd4 e5 15 Qd2 a5 16 f4 Nc5 17 f5 with an equal position from which Karpov
went on to win a long manoeuvring game in A.Karpov-V.Salov, Reykjavik 1991 (1-0 in 67).
10 d5 exd5 11 cxd5 Nxd2 12 Nxd2 d6
13 Nde4
Karpov chooses an aggressive continuation.
Five years earlier, he put on a positional masterclass against another top GM: 13 Nc4 Ba6 14 Qb3
Bxc4 15 Qxc4 a6 16 a4 Nd7 17 e3 Ne5 18 Qe2 c4 19 Be4! (I really like this move, re-routing the
bishop to the b1-h7 diagonal and covering the d3-square) 19 ... Re8 20 Bc2 Rc8 21 Ne4 Be7 22 b3
cxb3 (Black would like to support his c-pawn with 22 ... b5, but after 23 axb5 axb5 24 f4 Nd7 25
bxc4 bxc4 26 Ba4! or 25 ... Rxc4 26 Bd3 Rxc1 27 Rxc1, the dislocated pawn is more of a weakness
than a strength) 23 Bxb3
(White has now traded off Black’s major asset, and enjoys good prospects of pressurizing the
slightly weak a6- and b6-pawns; in addition, the c6-outpost beckons for a white knight – Karpov’s
manoeuvring during the next phase of the game is extremely instructive) 23 ... Qd7 24 Rxc8 Qxc8 25
Rb1 Qf5? (the conservative 25 ... Qb7 was more appropriate) 26 Nd2 (here 26 Bc2!, intending f2-f4
and Bd3, was even stronger) 26 ... Rb8 27 Bd1! (overprotecting the f3-square and pressurizing the
b6-pawn; Karpov wasn’t likely to fall for 27 Qxa6?? Nf3+ and wins) 27 ... Qc8 28 Nb3 (the effect of
Black’s queen raid has been to drive the white pieces where they wanted to go) 28 ... Bf6 29 Nd4
Qb7 30 Qc2 g6 (30 ... Qxd5? 31 Rxb6 leads to an immediate collapse) 31 Be2 Kg7 (now 31 ... Qxd5
32 Qc7! leaves Black struggling to defend his weak queenside pawns) 32 Qb3 Nd7 (waiting tactics
with 32 ... h5 33 h4 Bd8 allow White to increase the pressure with 34 e4, followed by Qb4 – though
this was perhaps a better try than allowing the knight into c6) 33 Nc6 (the knight finally occupies its
ideal outpost) 33 ... Ra8 34 Qb4 Qc7 35 Kg2! (typical Karpov: when there isn’t an obvious way to
improve his other pieces, he safeguards his king) 35 ... h5 36 h3 Kg8 37 Bd1 Re8 38 g4!? (interesting
timing; Karpov could have continued manoeuvring but judges that the time is ripe for concrete action)
38 ... hxg4 39 Bxg4!
(an unconventional capture, aiming for a good knight vs. bad bishop position) 39 ... Kg7 40 Bxd7
Qxd7 41 Qf4 (a good alternative was 41 e4 Rh8 42 Qa3, when the queen can come to g3 to defend the
kingside and pressurize the d6-pawn; but not 41 Qxb6??, since 41 ... Rh8 42 Rh1 Qf5 gives Black an
extremely strong attack) 41 ... Rh8 42 Qg4 (42 Qf3 was slightly better, not allowing Black to gain a
tempo with ... Rh4) 42 ... Qe8 43 Rxb6 (taking the pawn under relatively favourable circumstances)
43 ... Rh4 44 Qf3 Rxa4 45 Rb8 Qd7 46 Ra8 Bh4 47 e4 Bf6 (Black has defended extremely well and
should be within the drawing margin) 48 Qd3 Qb7 49 Rb8 Qd7 50 Qc2 Ra3 51 Rb3 Ra1? (the
position after 51 ... Rxb3 52 Qxb3 looks quite unpleasant for Black, who has the less effective pairing
of queen and bishop, as well as two weak pawns to tend to – all the same, he had to go for this)
Exercise: Find the best continuation of the attack.
Answer: 52 Rf3! (the only winning move) 52 ... Qb7
Exercise: And what now?
Answer: 53 Rxf6! Qb5 (53 ... Kxf6 gives White the pleasant choice between taking the rook on a1 and delivering mate in four: 54
Qc3+ Kg5 55 f4+ Kxf4 56 Qg3+ Kxe4 57 Qf3 mate or 55 ... Kh6 56 Qh8 mate) 54 Qc3! (White had a number of ways to deal with
Black’s intermezzo; Karpov picks the best one) 54 ... Qf1+ 55 Kg3 Qg1+ 56 Kh4 and Black resigned in A.Karpov-L.Portisch, Tilburg
1988.
13 ... Be7 14 f4 Nd7 15 g4 a6 16 a4
16 ... Re8
It is stronger to start queenside counterplay immediately with 16 ... b5! 17 axb5 axb5 18 Nxb5
Qb6, as in Pa.Kiss-S.Zagrebelny, Gyula 1991 (0-1 in 39).
17 g5 Bf8 18 Kh1 b5 19 axb5 axb5 20 Nxb5 Qb6 21 Nbc3 Qb4 22 Qd3 Nb6 23 Qg3
23 ... Kh8
The d5-pawn was indirectly defended: 23 ... Nxd5?? 24 Nxd5 Bxd5 25 Nf6+ gxf6 26 gxf6+ Kh8
27 Bxd5 and wins.
24 Rcd1 Nc4
24 ... Na4 avoided the idea in the game but allows 25 f5 Nxb2 26 Rb1 Ra3 27 Rf4 with an
overwhelming initiative.
25 b3 Nb6 26 g6!
White had other good moves here; for instance, 26 f5 and 26 Qh4 were also strong.
26 ... fxg6
Black just gets mated after 26 ... hxg6 27 Qh4+ Kg8 28 Ng5.
27 f5! gxf5 28 Rxf5
At the cost of a pawn, White has opened the kingside. The black pieces are marooned on the
queenside and unable to come to the defence of their king.
28 ... Nd7
A logical choice – Black manoeuvres his knight to the e5-square, hoping that it will prove
sufficiently solid to resist White’s attack.
29 Rdf1
The computer initially prefers continuations like 29 Rf4 and 29 Ng5 but eventually agrees with
Karpov’s decision to bring in another piece.
29 ... Ne5 30 R5f4!
Very precise – now unpleasant discoveries against the black queen are threatened.
30 ... Qb6 31 Ng5 Ng6
Falling for a cute mate, but everything loses at this point.
32 Nf7+ Kg8
Answer:
33 Qxg6!! 1-0
Black resigned since 33 ... hxg6 34 Rh4 is mate next move. A sparkling finish.
Game 34
A.Karpov-U.Adianto
Bali 2000
1 c4 Nf6 2 d4 e6 3 Nf3 b6 4 g3 Ba6 5 b3 Bb4+ 6 Bd2 Be7 7 Bg2 c6 8 Bc3 d5 9 Ne5 Nfd7 10 Nxd7
Nxd7 11 Nd2 0-0 12 0-0
One of the absolute main lines of the Queen’s Indian, and a position with which Karpov has
colossal experience (with both colours) in competitions at the very highest level, including World
Championship matches.
12 ... Rc8
Kamsky used 12 ... Rb8 to make two draws in his FIDE World Championship match with Karpov,
Elista 1996.
An entire book could be written about Karpov’s games after 12 ... Nf6 13 e4 (the less usual 13
Re1 has also been tested by Karpov, both as White against Anand, and as Black against Van Wely) 13
... b5 (the immediate 13 ... dxe4 led to an interesting exchange sacrifice in A.Karpov-V.Anand,
Calatrava rapid 2007: 14 Nxe4 b5 15 Nxf6+ Bxf6 16 Bb4 bxc4!? 17 Bxf8 Kxf8 18 bxc4 Bxc4 19
Re1 Bd5 with reasonable compensation; ½-½ in 28) 14 Re1 (the alternative 14 exd5 exd5 15 Re1
Rb8 16 c5 Bc8 17 h3 Qc7 18 b4 Re8 19 a4 gave White an edge in A.Karpov-D.Sadvakasov, Astana
2004; rapid, 8th matchgame, 0-1 in 68) 14 ... dxe4, which he has played as White against Adams,
Anand (twice), Ehlvest, Leko, Polgar, Portisch and Timman; and as Black against Yusupov. In fact the
Ehlvest game came two rounds after the present one, continuing 15 Qc2 Rc8 16 Rad1 Nd5 17 Bxe4
Nxc3 18 Qxc3 Bf6 19 c5!? Bxd4 20 Qc2 f5 21 Bg2 Qf6 22 b4 with promising compensation for the
pawn in A.Karpov-J.Ehlvest, Bali 2000 (1-0 in 83).
13 e4
This straightforward continuation was Karpov’s preference. He successfully defended Black’s
position after 13 Re1 against Fridman, Timman and Vyzmanavin.
13 ... dxc4
Very forcing play results after 13 ... c5 14 exd5 exd5 15 dxc5 dxc4 16 c6 cxb3. While Karpov
initially championed this line with White against Beliavsky, Korchnoi and Polugaevsky, he
subsequently chose to defend the Black side against Kamsky, Leko, Van der Sterren and Van Wely.
Later, Karpov was unsuccessful with Black after 13 ... dxe4 14 Nxe4 b5 15 Re1 bxc4 16 Bf1,
losing to both Bacrot and Kasimdzhanov in 2007 (albeit at blitz and rapidplay).
14 bxc4 b5 15 Re1
15 ... bxc4
Not the only move here. 15 ... Nb6 has featured in two Karpov victories: with Black against
Kasparov (World Championship, 6th match game, Moscow 1984; 0-1 in 70), and then with White
against Van der Wiel (Amsterdam 1987; 1-0 in 39).
16 Qc2
Karpov preferred the nuance 16 Qa4 Bb5 17 Qc2 against Mikhail Gurevich (USSR Championship,
Moscow 1988; 1-0 in 60) and Anand (Las Palmas 1996; ½-½ in 44), but for this encounter returned
to a line he had essayed against Kasparov in 1986.
16 ... c5
An ambitious move. 16 ... Qc7 was Kasparov’s choice in the 21st game of their 1986 World
Championship match (London/Leningrad; ½-½ in 45).

Answer:
17 d5!
White creates a strong passed d-pawn which will be well supported by all his pieces. Black
would be better after all other moves.
17 ... exd5
Black has to make this capture eventually, as otherwise the long-term threats of d5xe6 and (in some
positions) d5-d6 will be too annoying.
18 exd5 Bf6 19 Rad1 Nb6 20 Bxf6 Qxf6 21 a4!
Threatening the stability of the b6-knight.
21 ... Rfd8 22 Ne4 Qg6 23 d6
23 ... Rb8
23 ... Bb7 24 a5 Nd7 25 Qxc4 is much better for White.
24 Qc3 Bb7?
It seems that Black should have taken the pawn: 24 ... Nxa4, when 25 Qa5 Bb5 26 Qxa7 c3 gives
him some counterplay.
25 a5?!
Simply 25 Nxc5! was better, as after 25 ... Bxg2 26 Kxg2, the d-pawn is taboo: 26 ... Rxd6? 27
Qe5! and White wins a rook.
25 ... Na4 26 Qa3
26 Qe5 was another option, when 26 ... Nb2 can be met by 27 Qxb2, transposing to the game, or
else 27 Nxc5!? Bxg2 28 d7! h6 29 Rd6 f6 30 Kxg2 fxe5 31 Rxg6 Nd3 32 Nxd3 cxd3 33 Rd6 Rb7 34
Rxd3 Rbxd7 35 Rxd7 Rxd7 36 Rxe5 with an extra pawn in the rook ending.
26 ... Nb2! 27 Qxb2 Bxe4
Answer:
28 Qxb8!!
An extremely strong queen sacrifice.
It should be noted that White retains an advantage with 28 Qe5 too; for instance 28 ... Bd3 29 Bd5
(after 29 d7 Qe6 30 Qxe6 fxe6 31 Bh3 Bf5!, the draw looks more likely) 29 ... Qf5 30 Qxf5 Bxf5 31
Bxc4 with the better chances. However, Karpov’s continuation is much more incisive.
28 ... Rxb8 29 Rxe4!
Again the strongest continuation. The immediate 29 d7 allows 29 ... f5. White is still better after
30 d8Q+ Rxd8 31 Rxd8+ Kf7 32 Rd7+ Ke8 33 Rd5, but he has given up his main asset, namely the d-
pawn.
29 ... f5
29 ... Rd8 loses to the surprising 30 Rde1!!. It is highly counterintuitive to reduce the defence of
the d-pawn, but the direct attack on Black’s king is decisive: 30 ... f6 (or 30 ... f5 31 Re6 Qh5 32
Bc6, followed by 33 d7 etc) 31 d7 Rxd7 32 Re8+ Qxe8 (32 ... Kf7 33 Bd5+!! Rxd5 34 R1e7 is a
beautiful mate) 33 Rxe8+ Kf7 34 Rc8 and White should win the endgame.
30 Rxc4 Rd8 31 d7
From a material perspective, Black is doing fine with queen for rook and bishop. However, the
advanced d-pawn is enormously strong, and I think White is just winning here.
31 ... Qa6 32 Rxc5 Qe2 33 Rcc1
33 Rdd5!, followed by Bf1, looks even better.
33 ... Kf7
33 ... g6 is more tenacious.
34 Bd5+ Kf6?
This walks into a strong attack; 34 ... Kf8 was necessary.
35 Re1!
Now Re6+ or Rc6+ is a huge threat.
35 ... Qd3
36 Red1
Houdini gives 36 Rc6+! Kg5 37 h4+ Kg4 38 Rd6!! and there is no defence.
36 ... Qe2 37 Bc6
Going round again with 37 Re1! was stronger. Karpov doesn’t spot the mating ideas and simply
holds on to his d-pawn, which soon proves sufficient to win.
37 ... g6?
37 ... Qe5 was a better try.
38 Re1 Qd2 39 Rcd1! Qc2 40 Rd6+ Kf7 41 Re8 1-0
White Against the Semi-Slav

In relation to the Queen’s Gambit, we have already experienced Karpov’s handling of the
Orthodox Defence by exchanging on d5 and executing a minority attack (Games 23 and 24). Against
the extremely theoretical Semi-Slav Karpov chose a similar approach, going for e2-e3, Qc2 and
quietly developing his king’s bishop, rather than becoming embroiled in sharp discussions after 5 Bg5
or in the Meran Variation (though, as we will see in the next chapter, Karpov’s win in the Meran
against Kramnik in Linares 1994 was outstanding).
Game 35
A.Karpov-V.Kramnik
Vienna 1996
Semi-Slav Defence
1 d4 d5 2 c4 c6 3 Nc3 Nf6 4 Nf3 e6 5 e3

Karpov rarely essayed 5 Bg5 and probably didn’t have an intimate knowledge of the nuances of
this line. Interestingly, he lost horribly with White in the 47th game of his first match with Kasparov
(which was cancelled after the 48th). Karpov is unrecognisable in this game, which demonstrates the
enormous fatigue both players must have experienced towards the end of the match: 5 ... Nbd7
(Kasparov transposes to the Cambridge Springs, which has been regularly relied upon by Carlsen,
amongst others) 6 e3 Qa5 7 cxd5 Nxd5 8 Qd2 N7b6 9 Nxd5 Qxd2+ 10 Nxd2 exd5 (a quiet variation
has produced a rather dull endgame; Kasparov’s play is an excellent example of how to play for a
win with Black in such structures, although one would not expect Karpov to go down without a fight
under normal circumstances) 11 Bd3 a5 12 a4 (with an outpost on b4, Black has no problems) 12 ...
Bb4 13 Ke2 Bg4+ 14 f3 Bh5 15 h4 (gaining space on the kingside seems logical, but Kasparov
quickly obtains the more comfortable position) 15 ... 0-0 16 g4?! (16 Nb3 looks more purposeful,
aiming for the c5-square) 16 ... Bg6 (one of Black’s main aims in the Carlsbad structure is to
exchange light-squared bishops, and here he does so on favourable terms) 17 b3 Bxd3+ 18 Kxd3
Rfe8 19 Rac1
Answer: 19 ... c5! (Kasparov demonstrates admirable tactical sharpness; 19 ... Nd7 was good too, manoeuvring the knight via f8 to
e6 where it supports the ... c6-c5 break) 20 Bf4 (after 20 dxc5 Nd7, Black regains the pawn with more active pieces; in addition, despite
the IQP on d5, White has at least as many pawn weaknesses, taken on during his aggressive kingside expansion – f3 and e3 are both
targets) 20 ... Rac8 21 dxc5 (I’m not a huge fan of this move, as it allows Black to activate his worst piece, the knight on b6) 21 ... Nd7!
22 c6?! (22 Bd6! was more tenacious; I still prefer Black after 22 ... Bxc5 23 Bxc5 Nxc5+ 24 Ke2, but the most likely result is a draw)
22 ... bxc6 (now Black has a solid advantage; the pawns on c6 and d5 would be potentially weak if they could be blockaded, but with the
support of the excellently-placed black pieces, these pawns will advance and create difficulties for the white king – presumably around
this point Karpov was wondering how he’d managed to mess up in such a level ending!) 23 Rhd1 Nc5+ 24 Kc2 (the king isn’t safe on
the e-file either: 24 Ke2 Ne6 25 Kf2 Nxf4 26 exf4 Bd6 and Black is better on all counts) 24 ... f6 25 Nf1 Ne6 26 Bg3 Red8 27 Bf2 c5
28 Nd2? (a remarkable oversight from such a master of prophylaxis; the cautious 28 Kb1 was better, though Black is still for preference)
28 ... c4! 29 bxc4 Nc5
(the lines are opening around the white king, so Karpov needs to be careful; ordinarily he would
bring such a position to safety without any problems but, as discussed, this game was played in
exceptional circumstances) 30 e4? (30 Ra1 is more solid) 30 ... d4! (the passed d-pawn creates
severe problems for White even against the best defence; Karpov collapses immediately) 31 Nb1??
d3+ 32 Kb2 d2 and White resigned since he’s losing at least a piece.
Somehow the finish reminds me of A.Karpov-G.Kasparov, Linares 1993 (0-1 in 27), though in that
game it was a passer on the c-file that did the damage.
5 ... Nbd7 6 Qc2
Karpov has made an outstanding contribution to this system, which he generally preferred to the
main line Meran with 6 Bd3. (An exception is his model win against the same opponent in Game 41.)
6 ... Bd6
Here 6 ... b6?! 7 cxd5 cxd5 leads to an Exchange Slav structure where the weaknesses on the c-file
created by ... b7-b6 should leave Black in a prospectless position (and the knight on d7 is misplaced).
A.Karpov-G.Kaidanov, online blitz 1996, continued 8 Bd3 Bb7 9 0-0 Be7 10 Bd2 0-0 11 Rfc1 a6 12
Ne2 Qb8 13 Qb3 Ne4?!
Answer: 14 Bb4! (trading dark-squared bishops is in White’s favour, since Black’s central pawns are fixed on light squares; 13 ... a5!
would have prevented this move) 14 ... Bd6 15 Rc2?! (15 Qa3 was better, increasing the pressure) 15 ... Rc8? (Black misses a tactical
opportunity: 15 ... Ndc5! with unclear play) 16 Rxc8+ Qxc8 17 Rc1 Qf8?! 18 Rc7! Bxb4 19 Rxb7 Nef6, and now 20 Qc2 would have
maintained a clear advantage for White.
7 Bd3
Karpov has essayed many moves here (including the quiet 7 b3), though never the Shabalov-Shirov
Attack (7 g4!?), which leads to irrational positions contrary to his tastes.
a) 7 e4 e5 (7 ... Nxe4 8 Nxe4 dxe4 9 Qxe4 e5 is the theoretically approved antidote, leading to a
fascinating queen sacrifice after 10 dxe5 0-0 11 exd6 Re8 as in, for instance, S.Mamedyarov-
A.Brkic, Istanbul Olympiad 2012, 1-0 in 32; advancing 10 c5?! is weaker: after 10 ... Bc7 11 dxe5
Nxc5 12 Qc4? Qe7! 13 Bg5 Ba5+ 14 Kd1 Qd7+, White was already in trouble in A.Karpov-
A.Grischuk, Odessa rapid 2008; 0-1 in 69) 8 cxd5 cxd5 9 exd5 exd4 10 Nxd4 0-0 11 Be2 Nb6 12 0-0
Nbxd5 13 Nxd5 Nxd5 and a draw was soon agreed in A.Karpov-G.Kasparov, World Championship
(33rd matchgame) Moscow 1984 (½-½ in 20).
b) 7 Be2 0-0 8 0-0 dxc4 (8 ... e5, accepting an IQP for active piece play, was successfully tried by
Karpov’s two greatest rivals, Korchnoi and Kasparov, in 1991) 9 Bxc4 is an important variation
which can be reached via both 7 Be2 and 7 Bd3. Karpov has a fine attacking victory to his credit
from this position: 9 ... b5 10 Be2 Re8 (here 10 ... Qc7 11 Bd2 Re8 12 Rac1 Qb8 13 Ne4 Nxe4 14
Qxe4 Bb7 15 Qh4 h6 16 Rfd1 a5 17 Be1 Qd8 18 Qxd8 Rexd8 19 Nd2 a4 20 Bf3 Ra6 21 Ne4 Be7 22
a3 Ra7 23 Bb4! Bxb4 24 axb4 led to a Karpovian endgame squeeze in A.Karpov-J.Van der Wiel,
Tilburg 1988; 1-0 in 44) 11 Rd1 Qc7 12 b3 e5 13 h3 Bb7 14 Bb2 a6 15 dxe5 Nxe5 16 a4!? (taking
some flexibility out of Black’s queenside pawns and making it more difficult to achieve the thematic
... c6-c5 break) 16 ... Rad8 (Black would do better to exchange: 16 ... Nxf3+ 17 Bxf3 Bh2+ 18 Kh1
Be5 with a solid position)
Exercise: How can White launch an attack?
Answer: 17 Ng5! (creating annoying threats; White plans to trade the black king’s main defender, namely the knight on f6, gaining an
automatic kingside attack) 17 ... Qe7 18 Nce4! Nxe4 19 Nxe4 (the queen on e7 forces this recapture, but White retains a strong kingside
initiative; in particular, the b7-bishop does not control key light squares such as f5) 19 ... Bb4 20 Ng3!? (the more positional 20 Bc3! was
available, trading bishops and playing for a bind on the dark squares; clearly Karpov was in an aggressive mood and was unwilling to
trade his attacking bishop on b2 so easily) 20 ... f6?! (20 ... g6 would keep the knight out of f5 and appears to be a stronger defence,
though Shirov’s reluctance to weaken the dark squares around his king is understandable) 21 Bxe5!? (a very simple continuation which
leads to a surprisingly venomous attack; 21 Nf5 Qf8 22 f4 would also lead to a continuing initiative) 21 ... Qxe5 22 Bd3
Exercise: How would White respond to 22 ... g6 - ?
Answer: 23 Bxg6! is an automatic move. While it is obvious that White has at least a draw, it soon becomes apparent that he has
more than this: 23 ... hxg6 24 Qxg6+ Kh8 (or 24 ... Kf8 25 Nf5 wins) 25 Qh6+ (not 25 Nf5? Bf8 or 25 Nh5? Qg5!, pouring cold water
on the attack) 25 ... Kg8 26 Nh5 f5 (now 26 ... Qg5 is impossible since 27 Nxf6+ is check) 27 Qg6+ Kf8 (or 27 ... Kh8 28 Nf6) 28 Nf6
and White wins material with a continuing attack.
Instead, the game continued 22 ... h6 23 Bg6 Rf8 24 Nf5 c5 (this allows White to include another
piece in the attack – to be fair, the alternatives don’t seem to help much either).
Exercise: Which is White’s worst piece and how can it be brought into play?
Answer: 25 axb5! axb5 26 Ra7! (Houdini gives the immediate 26 Nh4 as even stronger, but I think Karpov’s move would be played
by most GMs – mobilizing an inactive rook leads to such an obvious strengthening of the attack that it takes a lot of discipline even to
look at other moves!) 26 ... Qc7 27 Nh4 Rxd1+ 28 Qxd1 Ra8
Exercise: How can White end the game at once?
Answer: 29 Qg4 Qc6 (29 ... Rxa7 30 Qe6+ leads to mate) 30 Rxb7! Qxb7 31 Qe6+ Kh8 32 Be4 Black resigned in A.Karpov-
A.Shirov, Biel 1992. Karpov gives the attractive finish 32 ... Qa6 33 Ng6+ Kh7 34 Ne5+ Kh8 35 Nf7+ Kg8 36 Nxh6+ Kh8 37 Qg8+
Rxg8 38 Nf7 mate.
c) Another excellent demonstration of Karpov’s attacking prowess (all the more impressive
because it took place at a rapid time control) is his game against Mikhail Gurevich from Cap d’Agde
2000: 7 Bd2 0-0 8 Bd3 dxc4 9 Bxc4 e5 10 0-0 Qe7 11 h3 (a deeply instructive assessment of this
position was given by top English GM, Luke McShane: “It is barely perceptible, but Black is under
great pressure here. The problem is mainly on the light squares, which aren’t covered by the bishop
on c8.”) 11 ... Bb8 12 Bb3 h6 13 Nh4 Rd8 14 Nf5 Qe8 15 f4 exd4 16 exd4 Nf8 17 Rae1 Be6
Answer: 18 d5!! (a wonderful shot, decisively opening the position when Black’s rook on a8 can’t take part in the battle; 18 Nxh6+
gxh6 19 d5 leads to a transposition in some lines, but Karpov’s move order is more accurate since it gives Black fewer opportunities to
deviate; while 18 Nxg7 Kxg7 19 d5 offers Black chances of holding on after 19 ... Qd7 20 dxe6 Nxe6) 18 ... Nxd5 (after 18 ... cxd5,
White wins with 19 Ba4! N8d7 20 Nxh6+ gxh6 21 f5 or 19 ... N6d7 20 Nxd5 with too many threats) 19 Nxd5 cxd5 20 Nxh6+! (20 Nxg7!
was also strong) 20 ... gxh6 21 f5 (the white pieces are tremendously powerful, whereas the b8-bishop and a8-rook don’t participate in
the defence. McShane writes: “Now the point is clear. Karpov simply wanted to close the b3-e6 diagonal so that 21 ... Bxb3! 22 Rxe8
Bxc2 wouldn’t be possible here. Against the shattered kingside White has a raging attack.”) 21 ... Qc6 22 fxe6! Nxe6 (22 ... Qxc2 23
exf7+ Kg7 24 Bxc2 is hopeless for Black; or 22 ... fxe6 23 Rxf8+! with an overwhelming attack) 23 Qf5 Qc7
Exercise: How does White finish the game?
Answer: 24 Rxe6! (eliminating Black’s best defender) 24 ... Qh2+ (or 24 ... fxe6 25 Qxe6+ and mates) 25 Kf2 Qg3+ 26 Ke2 and
Black resigned – apparently it’s mate in ten.
7 ... 0-0 8 0-0
8 ... e5
Black has a wide choice here too:
a) 8 ... dxc4 9 Bxc4 transposes to 7 Be2 lines (note ‘b’ to White’s 7th move).
b) 8 ... Re8 9 e4 dxc4 10 Bxc4 e5 11 Rd1 Qe7 12 Bf1 a6 13 h3 exd4 14 Nxd4 Nc5 equalized in
A.Karpov-V.Ivanchuk, Monte Carlo (blindfold rapid) 1997 (½-½ in 62).
c) 8 ... h6 9 Rd1 Qe7 10 h3 b6 11 cxd5 cxd5 12 Nb5 Ba6 13 Bd2 Ne4 14 Be1 Bb4?! (exchanging
dark-squared bishops should favour White in principle; 14 ... Rfc8 was preferable, with approximate
equality) 15 Bxb4 Qxb4 16 a3 Qe7 17 Qa4 Bb7 18 Rac1 a6 19 Nc7 b5 20 Qa5 and White was
clearly better in A.Karpov-Y.Pelletier, Biel 1997 (1-0 in 51).
d) 8 ... Qe7 9 c5 Bc7 10 e4 seems nice for White; for instance 10 ... dxe4 (10 ... e5 11 exd5 cxd5
was A.Karpov-A.Kosteniuk, rapid match, Manises 2002, and now 12 Re1! would have been decisive
in view of 12 ... e4 13 Nxd5 Nxd5 14 Bxe4 and White wins material) 11 Nxe4 h6 12 Re1 Re8 13
Bd2 Nd5 (Black has less room, and only one of his knights can occupy the d5-outpost, so 13 ... Nxe4
is more sensible) 14 Ng3 with an enduring edge in view of White’s extra space in A.Karpov-
D.Banaszkiewicz, Radom (simul) 2000 (1-0 in 40).
9 cxd5 cxd5 10 e4
There is maximum tension in the centre, which is fully occupied by pawns. Black gets to move
first, but White’s better placed pieces (and the fact that his queen is already developed) mean that
Black must play precisely to equalize.
10 ... dxe4
Here 10 ... exd4 might be an improvement, as in their game a few months later. Following 11 Nxd5
Nxd5 12 exd5 h6 13 Nxd4 Qh4, Black had some compensation for the pawn in A.Karpov-V.Kramnik,
Las Palmas 1996 (½-½ in 38).
11 Nxe4 Nxe4 12 Bxe4 h6 13 Be3 exd4
The removal of all the central pawns leads to a situation where piece mobility is critical. White,
being better developed, always enjoys easier play in such situations; the question is whether Black
can equalize with precise play.
14 Bh7+!
A good novelty – previously White had just played 14 Bxd4.
14 ... Kh8 15 Bxd4 Nf6 16 Bf5
The king is slightly worse on h8 than on g8, so Black has not fully equalized.
16 ... Bxf5 17 Qxf5 Qc8 18 Qb5
18 Nh4! also comes into consideration, maintaining the queen on its strong square, and preparing
for an eventual Nf5.
18 ... a6 19 Qb6
Karpov gives 19 Qb3 Qe6 20 Qxe6 fxe6 21 Rfe1 as slightly better for White.
19 ... Qc7 20 Qxc7 Bxc7 21 Bxf6 gxf6 22 Rfe1 Rfe8 23 Rad1
White certainly has an edge in this endgame. His pawn structure is much superior, giving his knight
an excellent outpost on f5, and his pieces are better mobilized. Nevertheless, it takes remarkable
technique to increase the pressure from such a position against a defensive master like Kramnik.
23 ... Rxe1+ 24 Rxe1 Rd8 25 g3 Rd7 26 Re2!
A nice move, covering the second rank and freeing up the knight.
26 ... Kg7?!
It seems Kramnik would have been better off playing 26 ... f5; for instance 27 Nh4 (or 27 Kg2 Kg7
and the king comes to f6) 27 ... f4 28 g4 Kg7 29 Kg2 h5 with some counterplay, though I would still
prefer White.
27 Nh4
Now Black’s kingside pawns will be fixed, creating potential mating nets around the black king,
who is slightly short of squares.
27 ... Rd5
Preventing Nf5+, but allowing ...
28 Re7!
This position is very unpleasant for Black to defend.
28 ... Rc5 29 Rd7 b5 30 b4 Rc2 31 Nf5+ Kg6 32 Ne3
The knight creates a lot of threats.
32 ... Rc1+ 33 Kg2 Be5
Black has managed to find a relatively stable square for his bishop, but now his queenside pawns
are exposed.
34 Ra7!
34 ... Rc6
Question: Should Black take the opportunity to play into a

rook endgame after 34 ... Bd4 35 Rxa6 Bxe3 36 fxe3 - ?
Answer: One of the hardest problems in chess is choosing between two bad positions. Here, I think Kramnik made the wrong call.
Rook and pawn endgames are notoriously drawish and it isn’t clear (at least to me) how White can easily transform his advantage into a
win here. Black’s pawn structure is certainly weakened, but so is White’s – he currently has weaknesses on a2, b4, e3 and h2, and
covering them all is not trivial.
Moreover, whatever the assessment of the rook endgame, choosing to play with the current
material balance and passive pieces means that White can seek to improve his position at leisure. The
white pieces are much more active than their counterparts which, coupled with Black’s inferior
structure, leads to a situation which should be avoided at any cost when playing against Karpov.
35 Nd5 Bd6 36 a3 Kf5 37 Ne3+ Kg6 38 Kf3!
Good technique, activating the king.
38 ... Be5
Again, Kramnik refuses to touch his kingside pawns, though 38 ... f5 made a lot of sense,
restricting the white king.
39 Nd5 Kg7 40 Ne7!
Forcing Black into a tough decision.

40 ... Rc3+?
Very tempting, not least because it wins a pawn. I don’t think Kramnik was under any illusions
about winning the game, but he probably aimed to liquidate the queenside completely, which would
dramatically increase his drawing chances (for instance, if he could give up his bishop for the white
kingside pawns, a drawn rook + knight vs. rook endgame would arise). However, this only makes
White’s task easier. 40 ... Re6 was more tenacious.
41 Kg4!
The correct square, controlling f5, g5 and h5.
41 ... Rxa3 42 f4!?
42 h4! was also extremely strong. The threat is 43 Nf5+ (and clearly 43 ... Kg6 is useless in view
of 44 h5+), so Black has nothing better than 42 ... h5+ 43 Kxh5 Rf3 44 Kg4! Rxf2 45 Nf5+ Kg8 46
Rxa6, when the black king remains in serious danger of getting mated. Houdini assesses the position
as winning for White.
42 ... Bc3!
Eliminating the queenside is the only chance. Now White has to play for mate.
43 Kh5
Taking the g6-square under control. 43 Nf5+ allows 43 ... Kg6; but 43 h4 was still strong, with
serious winning chances after 43 ... h5+ 44 Kxh5 Bxb4 45 Nf5+ Kg8 46 g4.
43 ... Bxb4 44 Nf5+ Kg8 45 Ra8+ Kh7 46 Ra7 Kg8 47 Nxh6+ Kf8 48 Rxf7+ Ke8 49 Kg6
The position is extremely complicated, and even computers have some difficulty working out
what’s going on!
49 ... Bc3
This loses. The best defence, which would be almost impossible to find over the board, seems to
be 49 ... Kd8 50 Nf5 Be1, aiming to restrict the advance of the white pawns.
50 Nf5! b4
Exercise: White to play and win.

Answer:
51 Rb7!!
The only winning move. The co-ordination between Karpov’s three pieces is extraordinary.
51 ... Ra2
Or 51 ... b3 52 Nd6+ Kd8 53 Kf7!! and Black has to shed material to meet the threat of 54 Ke6 and
55 Rd7 mate.
52 h4!
The h-pawn is unstoppable.
52 ... a5 53 h5 a4 54 h6 Rh2 55 h7 Kd8 56 Nh4 f5 57 Rxb4!
57 Rg7 gains a queen but needs some calculation. Karpov’s move wins with no fuss.
57 ... Rh3 58 Rxa4 Rxg3+ 59 Kxf5 1-0
Chapter Four
Linares 1994
Karpov’s victory at Linares 1994 was the greatest tournament performance of all time, in my opinion
(even including Caruana in the 2014 Sinquefield Cup). Aside from the games I’ve covered below,
Karpov had a fair dose of luck. However, the consistency and level of his play throughout the
tournament, coupled with his astonishing score against such a field, makes Karpov’s performance one
for the ages.
Game 36
J.Lautier-A.Karpov
Linares (round 1) 1994
English Opening
1 c4
Throughout his career, Karpov made occasional use of the English with White, scoring heavily. He
was also very successful with Black, preferring active systems with healthy development and good
central control.
1 ... e5
In many ways, this is Black’s most ambitious response to the English. White now plays a reversed
Sicilian with an extra tempo.
Karpov played one of the classic games in the Symmetrical English, against Ribli at the IBM
tournament in Amsterdam 1980: 1 ... c5 2 Nf3 Nf6 3 Nc3 Nc6 4 g3 d5 (one of Black’s more
ambitious options, eliminating the c4-pawn and gaining a space advantage; the downsides are
twofold: firstly, the queen on d5 will be exposed to attack; secondly, the h1-a8 diagonal is being
cleared, so the bishop on g2 will be very powerful) 5 cxd5 Nxd5 6 Bg2 g6 7 0-0 Bg7 8 Nxd5 Qxd5 9
d3 0-0 10 Be3 and now:
a) 10 ... Bxb2 looks critical. After 11 Rb1 Bf6 12 Qa4, Black can play the precise 12 ... Qd7 13 Bxc5 (13 Qb5 b6 14 Nd2 Bb7 15
Ne4 Rab8 16 Bf4 Qc8 17 Bxb8 Nd4 18 Nxf6+ exf6 19 Qb2 Bxg2 20 Kxg2 Qc6+ 21 f3 Rxb8 left Black with reasonable compensation
for the exchange in D.Jakovenko-I.Cheparinov, Poikovsky 2013, since he has one pawn for the exchange, a strong knight on d4, and
there are no open files; ½-½ in 31) 13 ... b6 14 Bxb6 Nd4 15 Qd1 axb6 16 Nxd4 Rxa2, White elected to force a repetition with 17 Qb3
Rd2 18 Qb4 Ra2 19 Qb3 Rd2 20 Qb4 Ra2 in H.Nakamura-P.Leko, Wijk aan Zee 2010 (and two subsequent games).
b) Instead, Karpov-Ribli continued 10 ... Bd7 11 Nd4 Qd6 12 Nxc6 Bxc6 13 Bxc6 Qxc6 14 Rc1
Qe6 15 Rxc5 Qxa2 16 Rb5 b6 17 Qa1 (17 Qb3 is recommended by Marin, avoiding the proposed
improvement to Black’s next move) 17 ... Qxa1 (here Marin prefers 17 ... Qe6, “leaving the white
queen rather passive”, though White has scored well from this position too) 18 Rxa1 Rfb8 19 Ra6
Kf8
Exercise: How might White force a weakening of Black’s queenside pawns?
Answer: 20 Rb4! Be5 21 Rba4! (the threat of Bxb6 forces a concession) 21 ... b5 (otherwise 21 ... Bxb2 22 Bxb6 wins a pawn) 22
Ra2 Rb7 23 b3
Exercise: How should Black try to defend?
Not like this: 23 ... Bb8? was Ribli’s choice, after which Karpov finished smoothly: 24 Bc5 Ke8
25 d4 Kd7 26 e4 e6 27 b4 Kc8 28 d5 exd5 29 exd5 Rd7 30 d6 Rd8 31 Kg2 Kd7 32 Re2 Kc8 33 Re7
Rd7 34 Ra2 a5 35 Rc2 and Black resigned.
Answer: Often in a difficult position, one is faced with the choice of playing passively or actively (the latter course will frequently
involve a sacrifice). While no hard and fast rules can be prescribed as to which approach is better, it is certainly the case in this position
that blocking all the black pieces in the top left-hand corner of the board is hopeless. Instead, 23 ... Rc8! offers Black activity and some
drawing chances. Now ... Bb8 is a threat, so White should go for 24 Rxa7 Rxa7 25 Rxa7 Rc2, when the position is not yet completely
clear, though my money would be on Karpov.
2 Nc3 Nf6 3 Nf3 Nc6 4 g3
The main continuation.
A real classic occurred after 4 e3 in J.Timman-A.Karpov, Montreal 1979: 4 ... Be7 5 d4 exd4 6
Nxd4 0-0 7 Nxc6 bxc6 8 Be2 d5 9 0-0 Bd6 10 b3 Qe7 11 Bb2 dxc4!
This was an idea of Igor Zaitsev’s, suggested during preparation for the Baguio match against
Korchnoi. Black ruins his structure in order to deflect the white pieces from the kingside, and starts a
strong attack. While modern engines immediately suggest this as Black’s strongest line, at the time of
the idea’s development it was a real analytical achievement.
Now 12 Bxc4 looks scary in view of 12 ... Ng4 13 g3 (not 13 h3?? Qe5 14 g3 Nxe3! and wins) 13
... Nxh2, but in fact Black has nothing more than a draw: 14 Kxh2 Qh4+ 15 Kg1 Bxg3 16 fxg3 Qxg3+
17 Kh1 Qh3+ with perpetual, as in Mu.Smith-T.Reilly, Sydney 1991. 13 ... Ne5!? would keep the
game going.
Instead, Timman played 12 bxc4? Rb8 13 Qc1 Ng4 14 g3 Re8 15 Nd1 (in a more recent GM
encounter, White was tricked into this line and tried 15 Bf3, but didn’t fare any better: 15 ... Qf6 16
Bg2 Qh6 17 h4 g5 18 c5 gxh4 19 e4 Qh5 20 cxd6 hxg3 21 Rd1 gxf2+ 22 Kf1 Ba6+ 23 Ne2 Rxb2 24
Qxb2 Qh2 and White resigned in D.Khismatullin-E.Tomashevsky, Serpukhov 2004).
Answer: 15 ... Nxh2! (other moves kept Black in the driving seat, but this ends the game) 16 c5 (now 16 Kxh2 gives Black a
decisive attack: 16 ... Qh4+ 17 Kg1 Bxg3 18 fxg3 Qxg3+ 19 Kh1 Re4 20 Rf4 Bh3 and wins) 16 ... Nxf1! (again strongest) 17 cxd6
Nxg3! (the completion of the tactical operation launched by Karpov at move 12; Black has a material advantage and a decisive initiative)
18 fxg3 (or 18 dxe7? Nxe2+ etc) 18 ... Qxd6 19 Kf2 Qh6 20 Bd4 Qh2+ 21 Ke1 Qxg3+ 22 Kd2 Qg2 23 Nb2 Ba6 24 Nd3 Bxd3 25 Kxd3
Rbd8 26 Bf1 Qe4+ 27 Kc3 c5 28 Bxc5 Qc6 29 Kb3 Rb8+ 30 Ka3 Re5 31 Bb4 Qb6 and White resigned.
4 ... Bb4
Karpov favoured classical development schemes against the English. Playing a pawn to e5,
developing the knights to their best squares and the bishop to b4 or c5 leaves Black with a fair share
of space and excellent development.
He opted for 4 ... Bc5 against GM Marin (an English expert, who in fact dubbed this 4 ... Bc5 line
the “Karpov Variation” in his excellent series of repertoire books): 5 Bg2 d6 6 0-0 0-0 7 d3 h6 8 a3
a5 9 e3 Re8 10 b3 Bf5 11 Bb2 Ba7 12 h3 Qd7 13 Kh2 Bh7 14 Ng1!? (starting a thematic regrouping
of the knight to e2, so that White can advance on the kingside with f2-f4) 14 ... Nd8 15 Nge2 c6 16 e4
Ne6 (Karpov later proposed 16 ... b5 instead) 17 f4 b5! (creating queenside counterplay) 18 cxb5
cxb5 19 fxe5 dxe5 20 Rxf6! (effectively forced, as otherwise Black would have an excellent
position) 20 ... gxf6 21 Nd5 Qd8 22 Qd2 Nc7 23 Nef4 Nxd5 24 Nxd5 Re6, when Black defended
against the immediate threats and went on to win a long game in M.Marin-A.Karpov, Spanish Team
Championship 2003 (0-1 in 80).
Alternatively, 4 ... d5 5 cxd5 Nxd5 6 Bg2 Nb6 leads to the “Reversed Dragon”, in which Karpov
played an excellent game with White: 7 0-0 Be7 8 a3 Be6 9 b4 0-0 10 Rb1 f6 11 d3 Qd7 (the main
lines are 11 ... a5 or 11 ... Nd4) 12 Ne4 Nd5 13 Qc2 b6 14 Bb2 Rac8 15 Rbc1 Nd4 16 Bxd4 exd4 17
Qc6 Qxc6 18 Rxc6 Bd7
Exercise: How would you continue with White?
Answer: 19 Nxd4! – at its most basic, the English Opening can be seen as an attempt to win control of the central light squares
(White puts his pawn on c4 and his bishop on g2). Therefore an exchange sacrifice to gain more control is always thematic. Here White
immediately obtains two pawns for the exchange, and his g2-bishop will face no competition. It is also important that there are no open
files, so the black rooks struggle to assert themselves. After 19 ... Bxc6 20 Nxc6 Rce8 (Black can’t hold on to the a-pawn, since 20 ...
a6? 21 Nd2 sees his position collapse immediately) 21 Rc1 f5 22 Nd2 Nf6 23 Nxa7, White went on to win in A.Karpov-J.Hjartarson,
Candidates (2nd matchgame), Seattle 1989 (1-0 in 45).
5 Nd5
The fourth World Championship match between Karpov and Kasparov (Seville 1987) contributed
significantly to the basic theory of the 5 Bg2 0-0 6 0-0 variation, which can lead to great
complications after, for instance, 6 ... e4.
5 ... Bc5 6 Bg2 d6 7 0-0 0-0
8 e3
A very natural move, limiting the action of the bishop on c5 and preparing to play d2-d4 in some
lines. Instead, R.Hübner-A.Karpov, Bad Kissingen 1980, saw 8 d3 h6 9 e3 a5 10 b3 Nxd5 11 cxd5
Ne7 12 d4 exd4 13 Nxd4 Qd7 14 Bb2 c6 15 dxc6 bxc6 16 Qc2 Bb7 17 Rfd1 Rfd8 18 Rac1, when
White had a very pleasant position (½-½ in 35).
8 ... Bg4
Fighting against d2-d4. After 8 ... a6 9 d4 Ba7 10 dxe5 Nxe5 11 Nxe5, a draw was agreed in
A.Adorjan-A.Karpov, Oslo 1984.
9 h3 Bh5
An improvement on the 15th game of Karpov’s World Championship match with Korchnoi in
Merano in 1981, which proceeded 9 ... Bxf3 10 Bxf3 Nxd5 11 cxd5 Ne7 12 b3 Qd7 13 Bg2 c6 14
dxc6 Nxc6 15 Bb2 d5 16 Bxe5 Nxe5 17 d4 Bd6 18 dxe5 Bxe5 19 Rc1 d4 20 Rc5 Bf6 21 Rd5 Qc7 22
exd4 and Karpov had to struggle to draw (½-½ in 40).
10 d3 a5
Controlling b4 and giving the bishop a retreat on a7.
10 ... Bb6 is a valid alternative, which Karpov also played that year: 11 a3 (I think I would prefer
to retreat the d5-knight to c3, as Lautier does in the main game) 11 ... Nxd5! 12 cxd5 Ne7 13 Qb3 Qd7
14 Nd2 c6 15 dxc6 bxc6 (Black has a space advantage in the centre and healthy development) 16 Nc4
Bc7 17 Qc3 a5 18 b3 Rfb8 19 Bb2 f6 20 Nd2 d5 21 Rac1 Bd6 22 e4 Bf7 and Black’s central
dominance assured him of the better game in M.Petursson-A.Karpov, Reykjavik (rapid) 1994 (0-1 in
60).
11 a3 Ba7
Now that d2-d4 will no longer come with tempo, Black threatens 12 ... Nxd5 13 cxd5 Ne7, when
14 d4? can be met by 14 ... e4.
12 Nc3 Re8 13 Qc2 Nd7 14 Rb1 Ne7
Another typical regrouping – the knight retreats from c6 (where it would have been subject to
attack by b4-b5), clearing the way for ... c7-c6, limiting the g2-bishop and preparing play in the
centre with ... d6-d5.
15 b4 axb4 16 axb4 c6 17 Qb3 Bg6 18 Rd1 h6 19 Nh4 Bh7 20 Ne4 Bb8 21 Ba3 Nf8 22 b5 Ne6
23 Ra1 Qd7 24 Nc3
24 c5 d5 is good for Black, since 25 Nd6? loses material to 25 ... Nxc5!.
24 ... Bc7 25 d4
Lautier plays ambitiously. Seizing space in the centre with d3-d4 is thematic and can hardly be
bad, but now the black pieces gain additional squares – in particular, the rook on e8 is activated on
the e-file and Black can try to use the e4-square for his minor pieces.
Otherwise, White could have maintained the balance with 25 Nf3.
25 ... exd4 26 exd4 Ng5 27 Kh2 Ba5!?
Activating the bishop and creating an idea of ... Bxc3, which would weaken the b5-pawn, as well
as the e2- and e4-squares.
28 Bb2
28 Bb4 would keep the position level.
28 ... Bxc3!?
Karpov believes in the power of his knights in the resulting position. This works out well in the
game, though I’m not completely sure of the move’s objective merits.
29 Bxc3 Be4
This was Karpov’s idea. White is reluctant to exchange the g2-bishop, which is his king’s main
defender, and so f2-f3 is more or less forced, which slightly weakens the kingside and the seventh
rank.
30 f3 Bh7 31 Rxa8 Rxa8
32 Re1?!
This allows Black to close the position and gain good squares for his knights.
Answer: A good general rule when playing with the bishop pair is to emphasize the power of the one which has no opponent. Here,
this means the dark-squared bishop – if it can be brought powerfully into play, there is no black piece which can effectively compete with
it. Therefore 32 d5! is what White wants to play, and there is no tactical reason to refrain from this move. The pressure on g7 will force
... f7-f6, when White seems to have some advantage; for instance, 32 ... cxd5 33 cxd5 f6 34 Re1 Ng6 35 Nxg6 Bxg6 36 f4 Nf7 37 Bd4
with the better chances.
32 ... d5!
Of course, Karpov doesn’t miss the chance to fix the pawn on d4. While, strictly speaking, this
move could be said to weaken the dark squares, it is very difficult for the c3-bishop to find useful
work on the a3-f8 diagonal and, even if it could get to e5, then ... f7-f6 would limit the bishop quite
effectively.
An interesting alternative is 32 ... cxb5 33 cxb5 Ne6, also aiming to exploit White’s omission of
d4-d5, since that move would now give the black knight a wonderful outpost on c5.
33 bxc6 bxc6 34 Bb4
Lautier tries to activate the bishop but, as pointed out earlier, a3-f8 is a less important diagonal
than a1-h8.
34 ... dxc4!
Karpov is always tactically alert.
35 Qxc4
Ribli notes that 35 Rxe7?? is impossible in view of 35 ... cxb3 36 Rxd7 b2, queening.
Alternatively, Black might simply play 35 ... Qxe7, winning an exchange.
35 ... Nd5 36 f4 Ne6 37 Bd2
The immediate 37 Nf3 would also be met by 37 ... f6!?, when I prefer Black’s practical chances
(since his pieces are more active and his king is more secure), although the position must be very
close to equality.
37 ... Re8 38 Nf3 f6 39 Rc1 Be4 40 Qxc6?
A mistake on move 40, as happens so often.
Exercise: How can Black exploit White’s error?
Answer:
40 ... Qxc6 41 Rxc6 Ndxf4!
This discovered attack on the c6-rook leaves White struggling to defend an inferior endgame. In
view of the reduced material and the opposite-coloured bishops it would seem that White should be
able to defend without too many problems, but Black retains a strong initiative (which is actually
helped by the opposite bishops) and Lautier collapsed quickly.
42 Rd6 Nxg2 43 Kxg2 Nc7 44 Kf2 Ra8! 45 Be3 Nb5 46 Rb6 Nc3 47 Rb3??
A horrible blunder. Objectively, White should be able to hold after 47 Bd2, though it would be a
grim defensive task against Karpov.
Exercise: How can Black win on the spot?
Answer:
47 ... Nd1+ 48 Ke2 Bd5!
White can’t defend all his weaknesses.
49 Rd3
49 Kxd1 Bxb3+ is equally hopeless.
49 ... Bc4 0-1
Karpov’s win in round 2 was bizarre. Bareev, in a completely equal position, blundered into mate
in one while hanging a rook (thus arguably worse than Kramnik’s blunder against Deep Fritz) on
move 35.
Game 37
A.Karpov-E.Bareev
French Defence
1 e4 e6 2 d4 d5 3 Nd2 c5 4 exd5 exd5 5 Ngf3 Nf6 6 Bb5+ Bd7 7 Bxd7+ Nbxd7 8 0-0 Be7 9 dxc5
Nxc5 10 Nd4 Qd7 11 N2f3 0-0 12 Bf4 Rfe8 13 Re1 Bf8 14 Ne5 Qa4 15 c3 Qa6 16 Qe2 Qxe2 17
Rxe2 Bd6 18 Nd7 Bxf4 19 Rxe8+ Rxe8 20 Nxc5 Bc7 21 Nd3 Bb6 22 Nb3 Kf8 23 Rd1 a5 24 Kf1
Rc8 25 Nd2 a4 26 a3 g5 27 Nf3 g4 28 Nh4 d4 29 cxd4 Bxd4 30 Nf5 Bb6 31 Nb4 Ne4 32 f3 gxf3
33 gxf3 Nc5 34 h4 Rd8 35 Rd5
A perfectly normal position, in which Black is at least equal after 35 ... Rxd5. Instead, Bareev, in
horrible time trouble, found the worst move on the board:
35 ... Ba7?? 36 Rxd8 mate
Game 38
M.Illescas Cordoba-A.Karpov
English Opening
1 Nf3 Nf6 2 c4 b6 3 g3 Bb7 4 Bg2 e6 5 Nc3 Bb4 6 0-0 0-0 7 Qc2 Re8 8 d4 Bxc3 9 Qxc3 d6 10 b3 Nbd7 11 Bb2 Be4 12
Rac1 Rc8 13 Rfd1 c6 14 Qb4 Qc7 15 Qd2 Qb7 16 Qf4 d5 17 Bf1 b5 18 cxb5 cxb5 19 Ne1 Qa6 20 a3 h6 21 Rxc8 Rxc8 22
Rc1 Nb8 23 e3 Rxc1 24 Bxc1 Qb6 25 Bd2 Nbd7
The position looks (and is) balanced – White’s bishop pair doesn’t count for much in this structure.
After 26 g4 or 26 h4, it’s hard to see a different result than a quick draw.
26 Bb4?
Seemingly logical, activating the bishop, but in fact walking into a tactic.
Exercise: How can Black gain the upper hand?
Answer:
26 ... a5! 27 Be7
27 Bd2? g5 wins the queen.
27 ... e5!
Cutting off the queen’s access to the d6-square and thus threatening ... g7-g5 again. Full marks, too,
for 26 ... e5! immediately.
28 Qh4 exd4 29 Qf4 dxe3 30 Qxe3 d4
Karpov was a clear pawn up and smoothly converted in 63 moves.

31 Qf4 Qc6 32 Bxf6 Nxf6 33 Qb8+ Kh7 34 Qxb5 Qc1 35 Nd3 Qd1 36 Nc5 Bg6 37 Kg2 Ne4 38 Be2 Qe1 39 Nxe4
Bxe4+ 40 f3 Bg6 41 h4 h5 42 f4 Be4+ 43 Bf3 g6 44 Bxe4 Qxe4+ 45 Kf2 Qe3+ 46 Kg2 d3 47 Qc4 Qe2+ 48 Kg1 Qd1+ 49
Kf2 Qd2+ 50 Kf1 Qd1+ 51 Kf2 Qe2+ 52 Kg1 Kg8 53 Qc8+ Kg7 54 Qc3+ Kf8 55 Qc5+ Ke8 56 Qc6+ Ke7 57 Qc5+ Ke6 58
Qf2 Qd1+ 59 Kg2 Qxb3 60 Qe1+ Kd7 61 Qxa5 Qc2+ 62 Kh3 d2 63 Qd5+ Kc8 0-1
Game 39
A.Karpov-V.Topalov
English Opening
This is a fabulous game, which Topalov continues to be impressed by (see his masterclass at Gibraltar 2015).
1 d4 Nf6 2 c4 c5 3 Nf3 cxd4 4 Nxd4 e6 5 g3 Nc6 6 Bg2 Bc5 7 Nb3 Be7 8 Nc3 0-0 9 0-0 d6 10
Bf4 Nh5 11 e3!?
A strong novelty, though it should be noted that this is not necessarily the best move – in 2001
Kasparov was successful with the more straightforward 11 Be3.
11 ... Nxf4
Black has nothing better. The tournament bulletin gives 11 ... g6 12 Bh6 Re8 13 e4 and 11 ... Nf6
12 e4, both with an edge for White.
12 exf4
Exercise: What does White get in exchange

for his doubled pawns and Black’s bishop pair?
Answer: The first thing to note is that the white pawn structure does a great job of restraining any pawn breaks by Black. The c4-
pawn, as usually occurs in the Maróczy Bind, restrains the ... b7-b5 and ... d6-d5 breaks (in the game, Topalov spends a lot of time
preparing ... b7-b5, which is time Karpov uses to create a strong kingside attack), while the f4-pawn controls the e5-square. The bishop
on g2 is a monster and a path has been opened to the kingside for the white queen.
12 ... Bd7 13 Qd2 Qb8 14 Rfe1 g6
Black was clearly concerned about the possibility of 15 f5, and decides to put a stop to it.
Exercise: Find another pawn break for White.

Answer:
15 h4!
A standard response to a kingside fianchetto. The h-pawn marches up the board, opening the h-file
and softening up the black kingside. While the white king might theoretically miss his h-pawn too, in
this position White has a significant central and kingside space advantage, so only the black king is
realistically likely to come under attack.
15 ... a6
Unsurprisingly, Topalov is unwilling to sit passively and seeks queenside counterplay.
16 h5 b5? 17 hxg6 hxg6
Answer:
18 Nc5!
Exploiting the hanging positions of all of Black’s minor pieces. 18 cxb5 axb5 19 Nc5! was equally
strong.
18 ... dxc5?!
On 18 ... Be8, Karpov intended 19 Nxe6! (after 19 Nxa6 Rxa6 20 cxb5 Rb6 21 bxc6 Bxc6, Black
has decent compensation for the pawn) 19 ... fxe6 20 Rxe6 with a decisive attack.
Black should have defended his bishop with 18 ... Qc8, when White retains the advantage by
taking on d7 or else playing 19 N5e4.
19 Qxd7 Rc8
Exercise: Find a tactic to crown Karpov’s masterful strategic play.
Answer:
20 Rxe6!! Ra7
White walks straight in after 20 ... fxe6 21 Qxe6+ Kg7 22 Bxc6 and Black must return the exchange
with a serious pawn deficit, since 22 ... Ra7 23 Be4 Bf6 24 Qg4 leads to disaster on the kingside.
21 Rxg6+
21 Qd5 fxe6 22 Qxe6+ transposes.
21 ... fxg6 22 Qe6+ Kg7 23 Bxc6
Topalov has gained a tempo over the variation given to Black’s 20th, but his position is still awful.
23 ... Rd8 24 cxb5 Bf6 25 Ne4 Bd4 26 bxa6
White already has a lot of pawns for the exchange, and his initiative has not yet been extinguished.
26 ... Qb6 27 Rd1 Qxa6
Exercise: What is the most convincing way to end the game?
Answer:
28 Rxd4!!
Karpov hasn’t shown much respect for his rooks in this game! Several moves won here, including
28 Nxc5 and 28 f5, but eliminating Black’s best piece leads to an unstoppable attack. Actually, it’s
slightly surprising not to see a computer declaring mate in 22 in such a position, but “+13” (as
Houdini assesses it) is enough.
28 ... Rxd4 29 Qf6+ Kg8 30 Qxg6+ Kf8 31 Qe8+ Kg7 32 Qe5+ Kg8 33 Nf6+
Not the strongest move according to the computer, but with Karpov’s technique five(!) pawns for
the exchange is quite sufficient.
33 ... Kf7 34 Be8+ Kf8 35 Qxc5+ Qd6 36 Qxa7 Qxf6 37 Bh5 Rd2 38 b3 Rb2 39 Kg2 1-0
Round 5 saw a win with Black against Ivanchuk, where the Ukrainian genius committed a colossal
blunder in a level position on move 28, resigning in frustration two moves later.
Karpov’s win against Polgar in the next round was a smooth technical display.
Game 40
A.Karpov-J.Polgar
Sicilian Defence
1 e4
Avoiding Polgar’s King’s Indian (as wonderfully described in the second volume of her best
games collection), Karpov returns to his old favourite.
1 ... c5 2 c3
An unusual choice for Karpov, but Polgar sometimes makes strange selections with Black in
sidelines (for instance, her response to the Trompowsky is often quite dubious), and Karpov must
have seen something he liked in her play against the c3-Sicilian.
2 ... e6
Already a slightly strange decision. Rather than the known equalizing lines after 2 ... Nf6 and 2 ...
d5, Polgar steers for a type of position in which Karpov has enormous experience.
3 d4 d5 4 exd5 exd5 5 Nf3 Nc6 6 Bb5 c4?
Perhaps fearing Karpov’s ability in the IQP positions after d4xc5, which are highly similar to (and
sometimes directly transpose to) those arising from the Tarrasch French, Polgar opts for a different
structure. Unfortunately, her move looks and is dubious: Black abandons the central tension and
leaves her structure vulnerable to b2-b3. A known variation proceeds 6 Be3 c4, but here White has a
clearly superior version since the bishop is very active on b5.
7 Ne5
This already looks dire for Black, who is underdeveloped and whose only asset, namely more
queenside space, is about to result in chronic weaknesses.
7 ... Qb6
Exercise: How can White develop his initiative?
Answer:
8 Bxc6+!
Simple chess. White compromises the black queenside and continues development.
Ftacnik thought 8 Qa4 Nge7 9 b3 was better, but this allows Black to free herself with 9 ... f6 10
Nf3 a6!; for instance, 11 bxc4 Bf5 12 0-0 Rc8 13 Bxc6+ Qxc6 14 Qxc6+ Rxc6 15 c5 b6 16 cxb6
Rxb6 with some compensation.
8 ... bxc6 9 0-0 Bd6 10 b3!
Targeting the queenside weaknesses and preparing Ba3. This is stronger than tempting attacking
alternatives (such as 10 Re1 or 10 Qh5) since Black’s only good piece (the dark-squared bishop)
will be traded in the process.
10 ... cxb3 11 axb3 Ne7 12 Ba3
Giving Polgar an unpleasant choice.
12 ... Bxe5?!
Abandoning any contest for the dark squares. 12 ... Bxa3 13 Rxa3 0-0 is very pleasant for White,
but his advantage is less than in the game.
13 dxe5 Be6 14 Qd4!
A strong move. Again, it was tempting (and promising) to play for an attack against the black king
in the centre with 14 Bxe7 or 14 Nd2. However, the endgame is deeply unpleasant for Black.
14 ... Nf5
Exercise: Can Black take the pawn with 14 ... Qxb3 - ?

Answer: Opening the b-file and falling behind in development is fatal in this position, where White’s initiative is greatly enhanced by
the opposite-coloured bishops and the black king being in the centre. In some old notes Ftacnik gives 15 Nd2 Qb6 16 Rfb1 with an attack,
but with modern engines we can assess the position as winning for White after 15 Bxe7! Kxe7 16 Qc5+ Kd8 (or 16 ... Kd7 17 Qd6+ and
the c6-pawn drops with check) 17 Qxc6 Qb8 18 Nd2 with level material and a decisive attack.
Exchanging queens is also unappealing, as after 14 ... Qxd4 15 cxd4, the weakness of the a7-pawn
and the c5-square remain, while the c6-pawn can now be attacked by the white rooks; for example,
15 ... f6 16 exf6 gxf6 17 Bc5 a6 18 Re1 Kf7 and White is on top, though perhaps less so than in the
game.
Exercise: How should White proceed now?
Answer:
15 Qc5!
Using the c5-outpost. Sooner or later Black will be forced to trade queens, resulting in a miserable
endgame.
Of course not 15 Qxb6?? axb6, when Black’s structure has been fixed and she has no problems.
15 ... h5 16 Nd2 f6
It is natural to trade pawns in order to reduce White’s space advantage and gain some counterplay.
17 exf6 gxf6 18 Bb4! Kf7 19 Ra4!
An elegant regrouping. White threatens to double on the a-file so, in order to defend her a-pawn,
Polgar must trade queens on c5, which she has been trying to avoid for the past few moves.
19 ... Qxc5 20 Bxc5 a6 21 f3 h4 22 Rfa1 Bc8 23 Kf2 Rb8 24 b4 Re8
For the moment Black is holding her queenside weaknesses, at the cost of severe passivity for (at
least) the c8-bishop and the b8-rook. If White could quickly transfer his pieces to the centre or the
kingside, this would likely lead to a strong initiative (enhanced by the presence of opposite-coloured
bishops), but that isn’t so easy to achieve, in particular in view of the outstanding black knight on f5,
which guards the h4-pawn and the e7-square while somewhat limiting White’s dark-squared bishop.
Accordingly, Karpov elects to bring another piece into the assault on the queenside.
25 Nb3! Bb7
Transferring the bishop to a8 is clearly hopeless, but it’s hard to give Black good advice at this
point.
26 Na5 Ba8 27 R4a2 Re6 28 Ba7!
Vacating the c5-square for the knight with tempo.
28 ... Rb5
It’s hard to imagine a worse square for the rook!
Polgar’s last chance to put up any resistance was the natural 28 ... Rbe8 29 Nb3 and now 29 ... c5!;
for instance, 30 Nxc5 (not 30 Bxc5? Bc6, when Black’s position will be greatly improved by the
bishop landing on b5, even at the cost of a pawn) 30 ... Re3 and Karpov would need to play well to
keep control.
29 Nb3 Bb7
Exercise: Find the most precise way to end Black’s resistance.
Answer:
30 g4!
While c5 is an excellent square for the knight, d4 is even better.
30 ... hxg3+ 31 hxg3 Re8 32 g4 Nd6 33 Nd4 1-0
The rook is trapped, though White doesn’t even need to take it since it can’t run away.
It’s amazing to beat a player of Polgar’s calibre in such a simple way.
In round 7, even Kasparov with the white pieces appeared uncertain, committing some
uncharacteristic errors and only saving the game with precise defence.
In the next round, Karpov was in some trouble with Black against Gelfand, but turned the tables
and won.
In round 9, Karpov drew with Shirov in 67 moves; while in round 10, Karpov’s Petroff netted him
an extra pawn and some chances against Kamsky, though the game ended in another 67-move draw.
Game 41
A.Karpov-V.Kramnik
Semi-Slav Defence
1 d4 d5 2 c4 c6 3 Nf3 Nf6 4 Nc3 e6

Karpov put on a masterclass against the true Slav with 4 ... dxc4 at Tilburg 1988: 5 a4 Bf5 6 Ne5
e6 7 f3 Bb4 (the traditional main line, preparing a piece sacrifice; Kramnik, following Smyslov,
subsequently revolutionized the line with 7 ... c5 8 e4 Bg6, leading, after 9 Be3 cxd4 10 Qxd4 Qxd4
11 Bxd4 Nfd7, to an endgame where Black is extremely solid) 8 e4 Bxe4 9 fxe4 Nxe4 10 Bd2 Qxd4
11 Nxe4 Qxe4+ 12 Qe2 Bxd2+ (the rare continuation 12 ... Qh4+ 13 g3 Bxd2+ 14 Kxd2 Qe7 15 Qe3
Na6 16 Bxc4 Nb4 17 Rad1 Nd5 18 Bxd5 cxd5 19 Nd3 0-0 20 Rc1, as in Karpov-Timman (1-0 in
41), would have given Black a satisfactory position after 20 ... b6) 13 Kxd2 Qd5+ 14 Kc2 Na6 15
Nxc4 0-0-0 16 Qe5 f6 17 Qe3!
(this novelty was Karpov’s key idea; rather than allowing Black a strong mass of central pawns
after 17 Qxd5, White waits for a more favourable opportunity to exchange queens) 17 ... c5 (theory
subsequently focused on 17 ... Kb8) 18 Kb3 (after 18 Be2 Nb4+ 19 Kb3 Nc6 20 Kc3 Nd4 21 Bf3
Nxf3 22 gxf3 Qd4+ 23 Qxd4 Rxd4 24 b4! Rd5 25 bxc5 Rxc5, White had an edge but Black held on in
Karpov-Hübner; ½-½ in 37) 18 ... Nb4 19 Rc1 Nc6 20 Ka3 Nd4 21 Na5 e5 22 Qc3 b6 23 Nb3 Nxb3
24 Qxb3 Qxb3+ 25 Kxb3 and the bishop was more valuable than the pawns in Karpov-Hjartarson (1-
0 in 43).
5 e3
With his solid positional style and distaste for lengthy opening analysis, Karpov has never been a
fan of the complex gambit lines beginning 5 Bg5.
5 ... Nbd7 6 Bd3
Leading to the sharp Meran Variation. As we noted in the previous chapter, Karpov’s main weapon
here has been 6 Qc2 (see Game 35).
6 ... dxc4 7 Bxc4 b5 8 Bd3 a6
Shirov had selected the alternative 8 ... Bb7 in round 9.
9 e4 c5 10 d5
10 ... c4
Kramnik subsequently varied against Karpov in Monte Carlo (rapid) 1994. After 10 ... Qc7 11 0-0
Bb7 12 dxe6 fxe6 13 Bc2, rather than transpose into the current game with 13 ... c4, he opted for an
interesting exchange sacrifice: 13 ... Be7 14 Ng5 Qc6 15 Qf3 h6 16 Qh3 hxg5 17 Qxh8+ Kf7 18 Qh3
g4 19 Qh4?! Ne5 20 f3 gxf3 21 gxf3. While Karpov laconically wrote that “Black did not gain any
compensation for the lost material”, it is notable that he went on to lose in 40 moves and the variation
with 16 ... hxg5 became popular for a while.
Reynaldo Vera notes that 19 Qe3!? is an improvement, observing: “At first White kept his queen on
the kingside with 19 Qh4 or 19 Qg3, but it was harassed by Black’s minor pieces. With the queen on
e3 White plans to consolidate the defence of his king with the manoeuvre Ne2-g3 and, if necessary,
Bd1-e2, which ensures him some advantage, as shown in Hübner-Vallejo Pons, French Team Ch.,
Port Barcares 2005.”
11 dxe6 fxe6 12 Bc2 Bb7 13 0-0 Qc7
One of the most well-analysed theoretical positions in chess, with heavyweights like Kasparov
and Kramnik frequently appearing on both sides. Karpov selects one of the most solid lines, playing
for an endgame which Black should hold.
14 Ng5
After 14 Qe2 Bd6 15 Nd4 Nc5 16 f4 e5, Karpov uncorked a dangerous novelty against Anand: 17
Ndxb5!? (previous theory had focused on 17 Nf5) 17 ... axb5 18 Nxb5 Qb6 19 Nxd6+ Qxd6 20 fxe5
Qxe5 21 Rf5 Qe7! 22 Qxc4 Rc8! (for the time being Vishy finds all the best defensive moves) 23
Qb5+ Ncd7 24 Qxb7 Rxc2 25 Bg5 Qd6 26 Qa8+ Kf7?? (here 26 ... Qb8, as played in Li Shilong-
R.Robson, Moscow 2009, holds the balance; ½-½ in 47) 27 Qxh8 Qd4+ 28 Kh1 Qxe4 29 Rf3 Rxg2
30 Kxg2 Ne5 31 Qxg7+! and White went on to win in A.Karpov-V.Anand, FIDE World Championship
(1st matchgame), Lausanne 1998 (1-0 in 108).
14 ... Nc5 15 e5 Qxe5 16 Re1 Qd6 17 Qxd6 Bxd6 18 Be3 0-0 19 Rad1 Be7 20 Bxc5 Bxc5 21
Nxe6 Rfc8
Exercise: How would you continue with White?
Answer:
22 h3!!
A wonderful idea. Straightforward continuations such as 22 Nxc5 don’t promise White anything.
Instead, Karpov prepares to use his kingside majority with g2-g4, which will gain space and threaten
the stability of the f6-knight. If this knight is forced to move, White will have access to fantastic
squares on e4 and d5 for his minor pieces and d7 for his rooks.
22 ... Bf8
Surprised by Karpov’s move, Kramnik selects the most solid continuation. From this point on,
however, White’s initiative grows surprisingly strong and, aside from G.Hanssen-G.Serper, Oslo
1994 (in which Black agreed a draw here despite his 300-rating point advantage), no-one has sought
to play 22 ... Bf8 again. Modern theory has favoured 22 ... Rab8, with numerous games between
strong GMs.
23 g4 h6
Answer:
24 f4!
Consistent with the objectives of gaining kingside space and harassing the f6-knight, as discussed
in the earlier note. The following sequence of play is not forced but both sides’ moves are logical.
24 ... Bf3 25 Rd2 Bc6 26 g5 hxg5 27 fxg5 Nd7 28 Nxf8 Nxf8 29 Rd6 b4 30 Ne4 Be8 31 Ng3
Rd8 32 Nf5 Rxd6 33 Nxd6 Bg6?!
This move was criticized subsequently, with 33 ... Bd7 (Ftacnik) and 33 ... Bf7 (Ernst) suggested
as improvements.
34 Bxg6 Nxg6 35 Nxc4
Karpov has clipped a pawn and Black remains with weaknesses on a6 and b4. Black has some
hopes for a draw (in particular against inaccurate technique, which he might expect from a lesser
opponent) based on the limited material remaining, but Karpov manages to break Kramnik’s
resistance in only five moves.
35 ... Rd8?!
The most natural move, activating the rook, but Karpov’s accurate response kills all counterplay.
35 ... Nh4 was a better try, going after the g5-pawn. Lines like 36 Kf2 Rf8+ 37 Kg3 Nf3 or 36 Rf1
Rc8 give Black counterplay, if not enough to balance White’s extra pawn completely. Nevertheless,
perhaps White should go for the queenside with 36 Re6 Nf3+ 37 Kg2 Nxg5 38 Rb6 a5 39 Rb5 Rc8
40 b3, when I would view White’s winning chances as roughly equivalent to Black’s drawing
chances.
Exercise: Returning to Kramnik’s 35 ... Rd8?!, how should White proceed?
Answer:
36 Re4!
Such a typical Karpov move: the g6-knight is completely dominated – and, for good measure, the
c4-knight is protected and the b4-pawn pressurized.
36 ... b3 37 axb3 Rd3 38 Kg2 Rxb3
At first sight, Kramnik’s plan seems very logical. He has traded off a pair of pawns (every pawn
exchange tends to favour the defender in such positions) and has activated his rook.
Exercise: So what’s the problem?
Answer:
39 h4!
A beautiful continuation – both on the theme of dominating the black knight, and of White’s more
general strategy of a kingside pawn advance.
39 ... Nf8 40 Re8 1-0
Kramnik lost on time here. After 40 ... Kf7 41 Re3 Rb4 42 Re4, White maintains winning chances
by carefully advancing his kingside pawns, though whether Black can make a draw with best play (for
instance, by trading the a- and b-pawns and sacrificing his knight for the kingside pawns to leave a
drawn rook + knight vs. rook position) is difficult to say.
As a postscript to Karpov-Kramnik, I would like to mention the final game of the 1981 World
Championship match between Karpov and Korchnoi. Despite featuring an entirely different opening,
the similarities are striking: excellent opening preparation, resulting in an endgame with a strong
initiative for White, who advanced his kingside pawn majority to harass the black knight and gain
additional squares for his pieces.
Game 42
A.Karpov-V.Korchnoi
World Championship (18th matchgame), Merano 1981
Ruy Lopez
1 e4 e5 2 Nf3 Nc6 3 Bb5 a6 4 Ba4 Nf6 5 0-0 Nxe4 6 d4 b5 7 Bb3 d5 8 dxe5 Be6 9 Nbd2 Nc5 10
c3 d4 11 Bxe6 Nxe6 12 cxd4 Ncxd4
13 a4
Karpov also extensively tested the position after 13 Ne4 Be7 14 Be3 against Korchnoi and
Yusupov, two of the best Open Spanish players in history; while 13 Nxd4 Qxd4 14 Qf3 Rd8 15 a4
Qd5 16 Qxd5 Rxd5 led to a draw in the classic encounter J.R.Capablanca-Em.Lasker, St Petersburg
1914 (½-½ in 100).
13 ... Be7 14 Nxd4 Nxd4 15 Ne4 Ne6 16 Be3 0-0 17 f4 Qxd1 18 Rfxd1 Rfb8?!
Aiming for queenside counterplay, but Karpov deals with it convincingly. Instead, 18 ... Rad8 19
axb5 axb5 20 f5 Rxd1+ 21 Rxd1 Rd8 22 Rc1 Nd4 23 f6 Ne2+ 24 Kf2 Nxc1 25 fxe7 Re8 26 Bxc1
Rxe7 reached a balanced endgame in M.Adams-W.Unzicker, German League 1996 (1-0 in 66).
19 Rd7 Bf8
19 ... Bd8 is no improvement. After 20 a5 b4 21 b3 Nf8 22 Rd5 c6 23 Rc5 Rc8 24 Rc4, White
was clearly better in view of his more active pieces and Black’s queenside weaknesses in T.Luther-
E.Steflitsch, Austrian Team Championship 2000 (1-0 in 32).
20 f5 Nd8 21 a5!
Now the black rooks look ridiculous in the top left hand corner. 21 Rxc7 bxa4 22 Bd4 is also
clearly better for White, but Karpov characteristically wants to restrain Black’s counterplay.
21 ... Nc6
Exercise: How should White meet the threat to his e-pawn?
Answer:
22 e6!
No points for 22 Bf4 Nxe5! 23 Bxe5 Re8, when White has no more than equality after 24 Nf6+
gxf6 25 Bxf6 Bd6. The white rooks can’t reach the g3 or g4 squares, so Black’s king is relatively
safe.
22 ... fxe6 23 f6!
This was the whole point. 23 fxe6? Re8 just leaves White with a weak e-pawn.
23 ... Ne5
The computer’s top choice of 23 ... h6 isn’t a particular improvement; for instance, 24 Rc1 Ne5 25
Rdxc7 Nc4 26 Bc5 and Black is struggling to hold his position together.
24 Rxc7 Rc8
Exercise: How can White increase his initiative?
Answer:
25 Rac1!
Simple and extremely strong: White will be left with a rook on c7 against a rook on a8. His
advantage in mobility is decisive – Black’s only active piece is the knight on e5, and it’s not clear
what it can achieve on its own.
25 ... Rxc7 26 Rxc7 Rd8 27 h3 h6 28 Ra7?!
28 b3! would have been an excellent move, preventing ... Nc4. The knight on e5 is then limited in
its movements and Black is in serious trouble. Trying to create counterplay with 28 ... Rd1+ 29 Kf2
Rb1 leads to a lost position after 30 Bd4!.
28 ... Nc4?!
Korchnoi misses a good opportunity for counterplay with 28 ... Rd1+ 29 Kf2 (here 29 Kh2?! Re1 30 Bc5 Rxe4 31 f7+ Nxf7 32 Ra8
Ne5 is unclear) 29 ... Rb1. However, 30 Bd4! still seems to win; for example, 30 ... Nc6 31 f7+ Kh7 32 Ra8 Nxd4 33 Rxf8 Rxb2+ 34
Kg1 Rb1+ 35 Kh2 Rf1 36 Nd6 etc.
29 Bb6 Rb8 30 Bc5
An interesting decision, giving up White’s pride and joy (the f6-pawn) in order to force matters on
the queenside. Otherwise, 30 f7+ Kh7 31 Bc7 Rc8 32 b3 would have left White in a dominant
position.
30 ... Bxc5+ 31 Nxc5 gxf6 32 b4 Rd8 33 Rxa6 Kf7 34 Ra7+ Kg6
Material is equal but Black will struggle to cope with the passed a-pawn.
Exercise: What is the most accurate way to win the game?
Answer:
35 Rd7!
As always, Karpov is deeply attentive to his opponent’s possibilities for counterplay. Instead, 35
Nxe6 Rd1+ 36 Kh2 Kf5 should be winning for White, but the black pieces gain a lot of activity and,
after ... Ne3, create strong kingside threats. Karpov doesn’t see any reason to give his opponent any
chances.
35 ... Re8
35 ... Rxd7 36 Nxd7 is straightforward – knights are notoriously bad at stopping rook’s pawns.
After 36 ... Nd6 37 Nb6! (preventing ... Nc8), Black will have to give up his knight when the pawn
promotes.
36 a6 Ra8 37 Rb7 Kf5 38 Rxb5 Ke5 39 Rb7 Kd5 40 Rf7 f5 41 Rf6 1-0
White wins another pawn while retaining all the advantages of his position. Korchnoi resigned
without resuming the game.
In round 12, Anand repeated the first nine moves of Kasparov’s opening from round 7, but without
disturbing the solidity of the Caro-Kann: draw in 26 moves.
Game 43
A.Karpov-A.Beliavsky
Catalan Opening
1 d4 Nf6 2 Nf3 d5 3 c4 e6 4 g3 Be7 5 Bg2 0-0 6 0-0 dxc4 7 Qc2 a6 8 a4

Considering how well the resulting positions seem to suit his style, Karpov uses the Catalan
surprisingly infrequently. When he reaches the position after 7 ... a6 with White, the text is his normal
choice, keeping more tension in the position than the immediate recapture on c4.
8 ... Bd7 9 Qxc4 Bc6
10 Bg5
In a game earlier that year, Karpov proceeded with 10 Nc3 a5 (as Black I would take the
opportunity to play 10 ... b5) 11 Re1 Ne4 12 Qd3 Nxc3 13 bxc3 Nd7 14 c4 Nb6 15 c5 Nd7 16 Qc2
e5 17 dxe5 Bxf3 18 exf3 Nxc5 19 Rd1 Qe8 20 Be3 b6 21 f4 and White’s strong, mobile kingside
pawn mass, coupled with the bishop pair, gave him a clear advantage in A.Karpov-I.Morovic
Fernandez, Las Palmas (5th matchgame) 1994 (1-0 in 37).
10 ... Bd5
After 10 ... a5 11 Nc3 Na6 12 Bxf6 Bxf6 13 e4 Nb4 14 Rfd1, White might have an edge but Black
is extremely solid. Karpov made draws from this position with both White (against Milos and Polgar)
and Black (against Kaidanov).
11 Qd3 c5 12 Nc3
Despite his success in this game, Karpov subsequently switched to 12 dxc5. After the logical
continuation 12 ... Nbd7 13 Nc3 Nxc5 14 Qe3 Qa5 15 Bxf6 Bxf6 16 Nxd5 exd5 17 Qa3, White hopes
for pressure on Black’s IQP: 17 ... Qb6 (Gelfand subsequently improved with 17 ... Ne4 18 Rad1
Rfd8 19 e3 Rac8 20 Nd4 Rc4, when Black had more than enough activity to balance the theoretical
weakness of the d5-pawn in A.Karpov-B.Gelfand, Monte Carlo rapid 2001; ½-½ in 28) 18 Rab1
Nb3 (I don’t like this whole plan: b2-b4 doesn’t seem to be such a big threat that it is necessary to
misplace the knight; Black would have a comfortable game after 18 ... Ne4) 19 Rfd1 Rfe8 20 e3 Rac8
Answer: 21 Nd2! (simplification generally favours the side playing against the IQP: now the weakness of d5 is felt; 21 Rxd5 Be7 22
a5 Qg6 23 Qa2 allows an amusing repetition with 23 ... Nc1! 24 Qa1 Nb3 25 Qa2 Nc1 and a draw) 21 ... Nxd2 22 Rxd2 d4 23 exd4
Bxd4.
Exercise: Isn’t this position just a dead draw? How should White continue?
Answer: Black’s defence is actually extremely difficult. Despite the simplified position and symmetrical pawn structure, the real
difference is in the power of the bishops – the black bishop on d4 hits well-covered pawns on b2 and f2, while White’s g2-bishop targets
the pawn on b7.
The game continued 24 a5! (fixing Black’s queenside and forcing the black queen away from the
defence of the pawns: slower play with 24 Rbd1? would allow Black to equalize with 24 ... Rcd8
since the threat of ... Bxf2+ forces White into defensive measures; for instance, 25 Qf3 Rd7, when b7
is well defended and Black should make a comfortable draw) 24 ... Qg6 25 Rbd1 (25 Qd3 was strong
as well, since the exchange of queens reduces Black’s chances for counterplay and will leave the a6-
pawn weak after the inevitable Bxb7) 25 ... Bc5 26 b4 Bf8 27 Bxb7 Rc2 (27 ... Rc7 28 Bg2 is also
much better for White) 28 Rd4! (wonderful technique from Karpov, keeping the b4-pawn covered and
refusing to activate Black’s queen by capturing on c2) 28 ... Rce2 29 Qf3 Re1+ 30 Kg2 Rxd1 31
Qxd1 Rb8 32 Qf3 Be7
Exercise: How can White finish the game?
Answer: 33 b5! (White retains an easily winning position after 33 Re4, but Karpov’s move is much more incisive: the a-pawn is
impossible to stop without decisive loss of material, in particular when combined with Black’s weak back rank) 33 ... axb5 34 a6 Bc5 35
Rd5 Qc2 (or 35 ... Bb6 36 Rd7 Qf6 37 Qxf6 gxf6 38 Bd5 and White wins most of the black kingside before collecting the bishop with
a6-a7) 36 Rd7 Kh8 37 Be4 Qa2 38 Qf5 and Black’s best move is the rather sad 38 ... Qxf2+, so he resigned here in A.Karpov-G.Milos,
Buenos Aires 2000.
12 ... cxd4
Theory has subsequently favoured the more solid 12 ... Bc6, as played repeatedly (and
successfully) by Karjakin, Kasimdzhanov and Jakovenko, amongst others (including Beliavsky).
13 Nxd5 Qxd5
Black has scored reasonably in practice with the more conservative 13 ... Nxd5 14 Bxe7 Qxe7 15
Nxd4 Nc6. I can’t imagine that Karpov would have objected to playing such a position with White,
when his strong g2-bishop (and slightly superior pawn structure after 16 Nxc6) should still count for
something, even if he did have no trouble drawing it with Black in R.Hübner-A.Karpov, Tilburg 1979
(½-½ in 23).
14 h4!
The only move to keep some prospects of an advantage. 14 Qxd4 Nc6 15 Qxd5 Nxd5 gives Black
a very solid position, as in Zhao Xue-Ding Yixin, Jiangsu Wuxi 2011 (½-½ in 121).
14 ... Nbd7?
14 ... Nc6 is critical, making White work a little to regain his pawn: 15 Bxf6 gxf6! (a number of
commentators on the Karpov-Beliavsky game, including Karpov himself, only considered 15 ...
Bxf6?, which is markedly weaker) 16 Rfd1 (or 16 Ng5 Qf5 17 Be4 Qc5 and Black holds the balance
in the complications after, for instance, 18 Bxc6 fxg5 19 Bxb7 Rab8 with counterplay) 16 ... Qh5!
(better than 16 ... f5 17 Nxd4 Qxd4 18 Qxd4 Nxd4 19 Rxd4 Rfd8 20 Rxd8+ Rxd8 21 Bxb7 with some
winning chances in the endgame, though a draw was soon agreed in A.Shimanov-A.Mammadov,
World Junior Championship, Porto Carras 2010; ½-½ in 24) 17 Nxd4 Nxd4 18 Qxd4 Qxe2 and ½-½
in Etaoin Shrdlu-Kaputtze, online rapid 2007.
15 Nxd4
The two active white bishops give White a clear advantage, although one would hardly have
expected a collapse from Beliavsky within five moves.
15 ... Qd6 16 Rfd1 Nc5 17 Qc4 Rfd8
It is hard to give Black good advice here, and no clear improvements suggest themselves over the
past few moves.
Answer:
18 b4!
Undermining the knight and threatening Black’s stability in the centre, while driving away a
defender of the weak b7-pawn.
18 ... Nxa4?
A blunder – although in such a position no moves hold much hope of saving the game. Even the
most tenacious 18 ... Nce4 leaves White with excellent winning chances after 19 Bxf6 Nxf6 20 Bxb7
Rab8 21 Nc2 Qe5 (sacrificing the queen doesn’t help: 21 ... Qxd1+ 22 Rxd1 Rxd1+ 23 Kg2 and
Black is collapsing on the queenside) 22 Rxd8+ Rxd8 23 Rb1.
19 Qb3 Qb6
19 ... Nb6 20 Bxb7 is equally hopeless.
20 e3! 1-0
White wins the knight on a4 and Beliavsky saw no point in continuing.
Chapter Five
Recent Battles
While the bulk of Karpov’s great games will remain those he played during the 70s, 80s and 90s, he
remains an active competitor in rapid and blitz events and has continued to participate in some team
tournaments, as well as simuls and exhibitions. Moreover, even in games at classical time controls
against strong professional players, he manages to display some of the traits which made him such a
great champion.
In a rare Bundesliga outing Karpov scored a convincing victory with Black against a very strong
grandmaster. It also gives us a chance to discuss his play with Black against the Catalan, which
remains a model for GMs even today.
Game 44
M.Rodshtein-A.Karpov
German League 2014
Catalan Opening
1 d4 Nf6 2 c4 e6 3 Nf3 d5 4 g3 Be7 5 Bg2 0-0 6 0-0 dxc4
7 Qc2
The main line. Kiril Georgiev was successful with 7 Ne5!? Nc6! (this pawn sacrifice is Black’s
best response, whereby he obtains active play) 8 Bxc6 bxc6 9 Nxc6 Qe8 10 Nxe7+ Qxe7 in
Ki.Georgiev-A.Karpov, FIDE Grand Prix (rapid), Dubai 2002 (1-0 in 43). Anyone interested in the
assessment of this position should refer to Kramnik’s games as White: at Dortmund 2010 (vs. Leko;
½-½ in 47) and from the Candidates Tournaments in 2013 (vs. Carlsen, ½-½ in 41) and 2014 (vs.
Anand, ½-½ in 31).
7 ... a6 8 Qxc4
White recaptures the pawn at once. For the preliminary 8 a4 see Game 43 (Karpov-Beliavsky).
8 ... b5 9 Qc2 Bb7 10 Bd2
This is the modern preference, as endorsed by Vladimir Kramnik. Karpov has defended the
position after 10 Bg5 Nbd7 11 Bxf6 Nxf6 12 Nbd2 Rc8 13 Nb3 several times, with no difficulties.
White’s entire concept is based around preventing ... c7-c5.
Answer: 13 ... c5! (the cleanest equalizer – it’s always a good sign when you can execute a pawn break which your opponent has
spent a lot of time trying to prevent; Karpov’s initial response was 13 ... Be4, leading to a quick draw after 14 Qc3 Nd5 15 Qc1 c5 16
Nxc5 Qb6 17 Qd2 Bxc5 18 dxc5 Rxc5 19 Rfc1 Rfc8 20 Rxc5 in U.Andersson-A.Karpov, Moscow 1981, but the immediate 13 ... c5 is
even easier) 14 dxc5 (14 Nxc5?? fails to 14 ... Bxf3 15 Bxf3 Qxd4, when the best White can get is a miserable pawn-down endgame
after 16 Nb3 Qxf2+ 17 Rxf2 Rxc2) 14 ... Bd5 (White can’t hold the c5-pawn) 15 Rfd1 (15 Rac1 isn’t an improvement: 15 ... Nd7 16
Qd3 Nxc5 17 Nxc5 Rxc5 18 Rxc5 Bxc5 and Black was certainly not worse in V.Cmilyte-A.Karpov, Marianske Lazne 2008; ½-½ in 31)
15 ... Bxb3 (15 ... Qc7 first is slightly more flexible) 16 Qxb3 Qc7 17 a4 Qxc5 18 axb5 axb5 19 Nd4 b4 20 e3 Rfd8 21 Rd2 Qb6 and a
draw was agreed in G.Kasparov-A.Karpov, World Championship (20th matchgame), London/Leningrad 1986.
The other bishop development, 10 Bf4, has been used by several strong GMs (successively
Portisch, Andersson, Agzamov, Seirawan and Piket) to make draws with Karpov. However, at
Amsterdam IBM 1980, Ribli managed to win in almost “Karpovian” fashion: 10 ... Nd5
(subsequently Karpov opted for either 10 ... Nc6 or 10 ... Bd6 here) 11 Nc3 Nxf4 12 gxf4 Nd7 13
Rfd1 Qc8 14 Ne4 c5 15 dxc5 Nxc5 16 Nxc5 Qxc5 17 Qxc5 Bxc5 18 Rac1 Rfc8 19 Ne5 Bxg2 20
Kxg2
Question: Surely this position is just equal?
Answer: At first glance, Black appears to be doing fine. He has no visible weaknesses (indeed, White’s structure appears statically
inferior, with doubled f-pawns) and it seems as if the bishop should be at least as good as a knight in a relatively open position. But in fact
this endgame is something close to White’s dream in the Catalan! The knight is actually the stronger piece, since it can attack the a6- and
b5-pawns (whereas the bishop has nothing to attack and can even be restricted by white pawns on f2, e3 and f4), while White’s rooks
are perfectly placed on the open files. As Yermolinsky notes, it is also easier for White to bring his king into the centre, since the black
king must avoid knight forks (most obviously 20 ... Kf8?? 21 Nd7+).
The game continued 20 ... f6 (Black cannot tolerate the knight on e5 forever, but now a slight
weakness is created on e6 and the seventh rank is opened; in addition, I’m sure that Karpov was
aware of the strategic dangers inherent in putting a pawn on a dark square, where it – at best –
duplicates the functions of the dark-squared bishop) 21 Nf3 Bf8 22 e3 g6 23 b3 Bb4 24 h3 Kf8 25
Nd4 Kf7 26 a4 bxa4 27 bxa4 Bc5 28 Rc4 (Black now has a choice of endgames, all of them
unpleasant) 28 ... Ba3 (Karpov didn’t fancy the pure rook and pawn endgame after 28 ... Bxd4 –
indeed, this looks like a clear advantage for White) 29 Rxc8 Rxc8 30 Rb1 Rc4 31 Rb7+ Be7 32 Ra7
e5 33 fxe5 fxe5 34 Nf3 Rxa4 35 Nxe5+, when Black still hadn’t equalized and White went on to win
a wonderful game (1-0 in 55). (If you have a spare half hour, check out GM Alex Yermolinsky’s
commentary in his outstanding web series Every Russian Schoolboy Knows.) Karpov soon got his
revenge, beating Ribli in the second cycle (as we saw the notes to Game 36).
10 ... Be4
Karpov suffered a debacle after 10 ... Bd6 11 Bg5 Nbd7 12 Nbd2 c5?! 13 Bxf6 Qxf6??
Answer: 14 Ne5! Bxg2 15 Nxd7 Qxd4? (but 15 ... Qd8 16 Nxf8 Bxf1 17 Qxh7+ Kxf8 18 Rxf1 cxd4 19 Qh8+ Ke7 20 Qxg7 leaves
White a clear pawn up) 16 Kxg2 Rfd8 17 Nf3 and Black resigned in V.Kramnik-A.Karpov, Zürich (rapid) 2009.
11 Qc1
11 ... Bb7
This bishop retreat was Karpov’s usual recipe. Of course, White can repeat the position with 12
Qc2 if he wishes, but Karpov didn’t shy away from such move repetitions with Black, as shown by
his repeated reliance on the Zaitsev Ruy Lopez: 1 e4 e5 2 Nf3 Nc6 3 Bb5 a6 4 Ba4 Nf6 5 0-0 Be7 6
Re1 b5 7 Bb3 d6 8 c3 0-0 9 h3 Bb7 10 d4 Re8, when a number of Karpov’s games have ended in
short draws after 11 Ng5 Rf8 12 Nf3 Re8 13 Ng5.
Interestingly, Karpov played differently in early games against Sosonko, equalizing easily and
drawing twice with 11 ... Qc8 12 Bg5 c5 (at Tilburg 1982 and 1983), and even won in 43 moves after
11 ... b4 (at Waddinxveen 1979).
12 a3
This little move is surprisingly useful in the Catalan. One of White’s ideas (after ... Nbd7) is to
restrict ... c7-c5 by the surprising Ba5 and b2-b4, often following up with a2-a3. In addition,
controlling the b4-square is generally helpful, preventing ... b5-b4 in some variations.
The position after 12 Bf4 Nd5 13 Nc3 Nxf4 14 Qxf4 has been extensively tested by Karpov, and
led to draws against Grischuk, Khalifman, Leko, Piket and Ubilava.
Another game saw 12 Be3 Nd5 13 Nc3 Nd7 14 Rd1 Rc8 15 Nxd5 Bxd5 16 Ne1 c6 17 Nd3 Qb6
18 Qc3 b4 19 Qd2 a5 20 Rdc1 ½-½ G.Kasparov-A.Karpov, World Championship (8th matchgame)
Moscow 1984.
12 ... Qc8 13 b4 Nbd7 14 Bc3 Ne4 15 Qc2
Varying from 15 Nbd2, as in L.Aronian-Ma.Carlsen, Linares 2009 (½-½ in 37).
15 ... a5!
A really strong move, which solves all of Black’s problems.

16 Nbd2
The key point is that 16 bxa5 is excellently met by 16 ... c5!. Generally, if Black achieves ... c7-c5
in the Catalan, he should have equalized. Accordingly, the only critical continuation here is 17 Be1,
trying to hold on to the pawn, but Black has excellent compensation after 17 ... Bd5, with superior
development and ideas of ... Bf6.
16 ... Nxc3 17 Qxc3 Nb6 18 Ne5?!
White shouldn’t let Black get ... Nd5 in, attacking the b4-pawn. Both 18 bxa5 and 18 e4 are better,
with approximate equality in either case.
18 ... axb4 19 axb4 Bxg2 20 Kxg2 Nd5 21 Qc6 Bxb4
White has dropped a pawn and is struggling to equalize.
22 Ndf3 Bc3!
Excellent tactical awareness from Karpov.
23 Rxa8 Qxa8 24 Qxb5
Exercise: What should Black play now?
Answer:
24 ... Bxd4!
This shot regains the material and leaves Black a healthy pawn ahead.
25 Nc6
Of course not 25 Nxd4?? Nf4+ 26 Kg1 Nh3 mate.
25 ... Bb6 26 Rc1 Qa3 27 Rc4 Ra8 28 Nce5 h6!
Having mobilized all his pieces, Karpov now creates some luft for his king.
29 Nd3 Qa6 30 Qc6 Ne7 31 Qe4 Rd8 32 Qh4 Rf8 33 Ra4 Qb5 34 Rb4 Qe8
The black pieces have temporarily been driven back, but his position remains solid and he retains
the extra pawn.
35 Nde5 f6 36 Nd3 e5 37 Qe4?! Qc6!
Now White will struggle to avoid the exchange of queens.
38 g4 Ra8 39 Qxc6 Nxc6 40 Rc4 Ne7 41 e3 Kf7 42 Nfe1 Nc8 43 Nb2 Nd6 44 Rc2 Ke6
Excellent technique, centralizing the king and preparing to support the passed c-pawn.
45 Ned3 Ra2
One of the recurring themes in Karpov’s games is his use of pins when trying to convert a superior
position. Such pins certainly restrict his opponent’s active possibilities. Here the knight is kept on the
passive b2-square.
46 Nb4 Ra3 47 Nc6 g6 48 h3 h5 49 gxh5 gxh5 50 Kg3 Ra8 51 Kf3 Ra2 52 Nb4 Ra5 53 Kg3
Nf5+ 54 Kf3 c5 55 N4d3 Nd6 56 Ke2 Kd5 57 f3 Ra2 58 Rd2 Kc6 59 Rc2 Kb5 60 Nc1 Ra3 61 Rd2
Bc7 62 Nd1 Kc6 63 Rc2 c4 64 e4 Bb6 65 Na2 Bd4 66 Nb4+ Kc5 67 Nd5 f5 68 N1c3 fxe4 69 fxe4
h4 0-1
Fixing the h3-pawn as a weakness. White is almost in zugzwang and threw in the towel here.
In the last three years (2012-14) Karpov has played successfully in the eponymous “Anatoly Karpov Trophy” at Cap d’Agde,
demonstrating that, at rapid time controls, he can still compete with the younger generation.
Game 45
M.Sebag-A.Karpov
Cap d’Agde (rapid) 2014
Scandinavian Defence
1 e4 d5!?
In recent years I would imagine that study of opening theory hasn’t been top of Karpov’s to-do list.
Accordingly, it is unsurprising that he avoids the lines he relied on during his active career. For
instance, the Caro-Kann, previously thought to be a dull and dry variation, leads to some of the
sharpest positions in modern chess after 3 e5. The Scandinavian is therefore a great choice. First, it
requires less preparation than other openings since it is quite “young” (theoretically speaking) and
White’s options are relatively limited (2 exd5 is the only serious reply, and after 2 ... Qxd5, 3 Nc3
and 3 Nf3 are overwhelmingly the most-played moves). Second, and perhaps more important, it leads
to the same pawn structure as the Caro-Kann (or the Rubinstein French), so Karpov can draw on his
years of experience while avoiding unpleasant opening surprises.
2 exd5 Qxd5 3 Nc3 Qd6
Previously 3 ... Qa5 was the main line, but nowadays 3 ... Qd6 has a number of regular GM
advocates, chief amongst them Sergei Tiviakov, who has a fabulous score in this variation.
4 d4 c6
5 Bc4
In the blitz section of the same event, Romain Edouard tried 5 h3 against Karpov: 5 ... Bf5 6 Nf3
Nd7 7 Be3 e6 8 Qd2 Ngf6 9 0-0-0 Ne4 10 Nxe4 Bxe4 11 Ng5 Bg6 12 Bd3 Qb4 13 Qe2 Be7 14 Bxg6
hxg6 15 h4 Nb6 16 Rd3 Qb5 17 Qf3 Rf8 18 Rb3 Qd5 19 Qg3 0-0-0 20 Rd1 Nc4 21 Nh3? Nxe3 (21
... Na5! wins on the spot) 22 fxe3 Qh5 23 Ng5 Bxg5 24 hxg5 Rd5 25 c4 (here 25 Qe1! Rfd8 26 c4
Rxg5 27 Qb4 b6 28 Ra3 offers sufficient compensation) 25 ... Qxg5 26 Qf3 Rd7 and Karpov went on
to convert his extra pawn (0-1 in 72).
5 ... Nf6 6 Nf3 Bg4 7 h3
A dubious gambit variation which scores badly. Instead, 7 Be3 has been tried by serious players,
including Grischuk.
7 ... Bxf3 8 Qxf3 Qxd4 9 Bb3 e6
Question: How would you assess this position?
Answer: White has a lead in development and the bishop pair, which must count for something (Houdini, for instance, assesses this
position as equal). Against that, Black has an extra central pawn and no weaknesses. In particular, I would be concerned by White’s lack
of pawn breaks – it takes a long time to arrange f4-f5, and a successful sacrifice on e6 seems unlikely. I think which side you prefer is a
matter of taste and, were Kasparov handling the white pieces at a classical time control, White’s attacking resources might well look
quite threatening. However, at a rapid time control, the French GM struggles to demonstrate anything concrete, while Karpov manages to
improve his position and consolidate his extra pawn with precise and natural moves.
10 Be3
After 10 Bf4 Nbd7 11 0-0 Bd6 12 Ne2 Ne5 13 Qg3 Qe4 14 Rfe1 Nh5 15 Qe3 Qxe3 16 Bxe3 0-0,
White was struggling to demonstrate compensation in C.Ponizil-M.Konopka, Czech League 2014
(though 1-0 in 54).
10 ... Qd8 11 Ne4 Be7
A good practical decision. 11 ... Nbd7 gives White the option of 12 0-0-0, sharpening the play.
12 Rd1
A slight inaccuracy. Had White castled before occupying the d-file, she would not have been
forced into playing the semi-useful c2-c3.
12 ... Qa5+!
Of course. Black deals with the attack on his queen with tempo.
13 c3 Nbd7!
Again precise. After 13 ... 0-0 14 Nxf6+ Bxf6, the development of Black’s knight is more
awkward.
14 0-0 Nxe4 15 Qxe4
15 Rxd7? is nothing: 15 ... Kxd7 16 Qxe4 Rad8 and Black can castle “by hand” with ... Kc8.
15 ... Nf6 16 Qh4 0-0 17 Bc2 h6 18 Rd3?
A natural move but after Karpov’s response, Black is on top.
Instead, 18 Qg3! virtually forces a draw: 18 ... Nh5 (18 ... Kh8 19 a4 prepares a queenside
expansion with b2-b4, and if 19 ... c5 20 Qe5 Rac8 then 21 Bxh6 gxh6 22 Rd7 Bd8 23 Qf4 Kg7 24
Qg3+ Kh8 25 Qf4 Kg7 26 Qg3+ and the game ends in repeated moves) 19 Qf3 Nf6 (leaving the knight
on h5 doesn’t help; for instance, 19 ... Rfd8 20 b4 Qe5 21 Bd4 Qg5 22 Be3 Qe5 23 Bd4 and White
can force a draw if she wants) 20 Qg3 Nh5 21 Qf3 Nf6 with a draw by repetition.
Exercise: Which rook should Black place on the d-file?

Answer:
18 ... Rfd8!
Exactly the type of move which goes unnoticed. Karpov, through intuition or calculation, manages
to counter White’s attack by giving his king the f8-square.
Instead, 18 ... Rad8 would be inaccurate due to 19 Bd4!, when continuing as in the game is
impossible in view of 19 ... Qxa2?? (other moves, such as 19 ... Rd5, are more sensible) 20 Rg3 with
a quick mate; for instance, 20 ... Kh8 is crushed by 21 Rxg7! Kxg7 22 Qg4+ Kh8 23 Bxf6+ Bxf6 24
Qe4 etc.
19 Bd4
19 Rfd1 is more tenacious, though Black can slowly consolidate after 19 ... Rxd3 20 Rxd3 Qe5.
19 ... Qxa2!
A great move; Karpov isn’t afraid of ghosts. He is also aware that, while White’s will always
have some compensation for one pawn with the bishop pair, two or three pawns is too much.
20 Rg3 Kf8!
Showing the difference from the 18 ... Rad8 variation above. White has no breakthrough.
21 Bxf6 Bxf6 22 Qb4+ Kg8 23 Qxb7
White has gone for a forcing continuation to regain one of her lost pawns, at the cost of making her
pieces less co-ordinated.
23 ... Rab8!
Of course, Karpov brings the rook crashing into the game.
24 Qxc6 Rxb2
Now Black is a pawn up with the initiative.
25 Be4 Rbd2
Not forced, but it’s nice to have the pieces overprotected (a constant theme in Karpov’s games)
and to keep the option of ... Rd1 in reserve.
26 Bf3 Qa5 27 Bg4 h5!
Calling White’s bluff and winning more material.
28 Bf3
28 Bxe6 is hopeless, since 28 ... fxe6 29 Qxe6+ Kh8 leaves White without a follow-up.
Exercise: How can Black win further material?
Answer:
28 ... Bh4!
Typical tactical alertness from Karpov, adding an exchange to his treasure chest. 28 ... Be5 29 Rg5
Bh2+ 30 Kxh2 Qxg5 is equally effective.
Sebag could have resigned here (at any time control) but elects to prolong the agony for a few
moves.
29 Rb1 Bxg3 30 fxg3 Qe5 31 c4 g6 32 Qa6 R8d7 33 Qc8+ Kg7 34 Qa8 Qxg3 35 c5 Qf2+ 36 Kh1 Qxc5 0-1
Game 46
A.Karpov-M.Mchedlishvili
German League 2014
Grünfeld Defence
1 Nf3 Nf6 2 c4 g6 3 Nc3 d5 4 cxd5 Nxd5 5 Qb3 Nb6 6 d4

Even the Anti-Grünfeld has developed a substantial body of theory, but it is unsurprising that
Karpov goes for this rather than one of the Grünfeld main lines. He quickly reaches a position in
which his superior experience and strategic understanding tell.
6 ... Bg7 7 Bf4 Be6 8 Qa3 0-0 9 e3 Nc6 10 Bb5 Nd5 11 Bg3 Ncb4 12 Rc1
12 ... Nxc3
Other moves:
a) 12 ... c5 is in the spirit of the Grünfeld, but after 13 Nxd5 Qxd5 14 0-0 (14 Bc4 allows 14 ...
Qxc4!? 15 Rxc4 Bxc4 with reasonable compensation for the queen) 14 ... Rfc8 15 Bc4 Qd7 16 dxc5
Bxc4 17 Rxc4, Black still needed to demonstrate full compensation in D.Jakovenko-M.Vachier
Lagrave, European Team Championship, Heraklio 2007 (½-½ in 66).
b) 12 ... c6 13 Be2 (13 Nxd5 was A.Obukhov-A.Gabrielian, Alushta 2006, 1-0 in 43; and now 13
... Bxd5! is fine for Black in view of 14 Qxb4 Qb6, regaining the piece safely) 13 ... a5 is a better
approach than Mchedlishvili’s, despite the result of B.Lalith-I.Aldokhin, Moscow 2012 (1-0 in 44).
13 bxc3 Nd5 14 Be2 Nf6 15 Be5 b6?!
Aiming for ... c7-c5 again, but Black doesn’t have sufficient space for his pieces in the resulting
structure. The less ambitious 15 ... Ne4 was preferable; for instance, 16 Bxg7 Kxg7 17 0-0 Qd6 18
Qb2 c5 and White has only a slight advantage.
16 0-0 c5 17 Rfd1
The threat of d4xc5 forces a concession.
17 ... Qc8
If 17 ... cxd4 18 Rxd4!, White either wins the e7-pawn or remains dominant after 18 ... Qe8 19 c4,
followed by c4-c5.
18 c4 cxd4
18 ... Bf5 would still keep White’s advantage within manageable limits.
19 Nxd4!
Increasing the tension and emphasizing White’s development lead. While 19 exd4 was good too,
Black’s position quickly collapses after Karpov’s choice.
19 ... Bg4
Mchedlishvili aims to prevent Bf3, but the loss of time allows White to break open the position.
Houdini insists upon 19 ... Qc5 20 Qxc5 bxc5 21 Nxe6 fxe6, which would be a horrible position to
defend.
20 f3 Bd7
Answer:
21 c5!
Decisively opening lines. White’s lead in development is enormous and he will gain additional
time on Black’s queen and light-squared bishop. Karpov isn’t going to allow his opponent a chance
like 21 Qxe7? Re8 22 Qd6 Qc5!, when White’s advantage evaporates; e.g. 23 Qxc5 bxc5 24 Bxf6
Bxf6 25 Nb3 Rxe3 26 Rxd7 Rxe2 27 Nxc5 Rxa2 with equality.
21 ... bxc5 22 Rxc5 Qe8 23 Nb5!
Not the only winning move but a very efficient one, clearing the d-file for the rook and penetrating
to the c7-square.
23 ... Rc8 24 Nc7 Qd8 25 Bb5!
White wins a piece.

25 ... a6 26 Ba4 h5 27 Bxf6 Rxc7 28 Bxg7 1-0
Index of Complete Games
Alterman.B-Karpov.A, European Cup, Tyniste 1995
Andersson.U-Karpov.A, Nyköping (rapid; 1st matchgame) 1995

Bisguier.A-Karpov.A, S kopje Olympiad 1972
Illescas Cordoba.M-Karpov.A, Linares (round 3) 1994
Kamsky.G-Karpov.A, FIDE World Championship (4th matchgame), Elista 1996
Karpov.A-Adianto.U, Bali 2000
Karpov.A-Bareev.E, Linares (round 2) 1994
Karpov.A-Beliavsky.A, Linares (round 13) 1994
Karpov.A-Beliavsky.A, Moscow 1981
Karpov.A-Campora.D, S an Nicolas (2nd matchgame) 1994
Karpov.A-Dolmatov.S , Amsterdam IBM 1980
Karpov.A-Dorfman.J, US S R Championship, Moscow 1976
Karpov.A-Gelfand.B, FIDE Candidates final (4th matchgame), S anghi Nagar 1995
Karpov.A-Hansen.C, Biel 1992
Karpov.A-Hort.V, Budapest 1973
Karpov.A-Kasparov.G, World Championship (4th matchgame), Moscow 1985
Karpov.A-Khalifman.A, Linares 1995
Karpov.A-Korchnoi.V, Candidates final (2nd matchgame), Moscow 1974
Karpov.A-Korchnoi.V, Candidates final (8th matchgame), Moscow 1974
Karpov.A-Korchnoi.V, World Championship (14th matchgame) Baguio 1978
Karpov.A-Korchnoi.V, World Championship (18th matchgame), Merano 1981
Karpov.A-Korchnoi.V, World Championship (32nd matchgame), Baguio 1978
Karpov.A-Kramnik.V, Linares (round 11) 1994
Karpov.A-Kramnik.V, Vienna 1996
Karpov.A-Ljubojevic.L, Linares 1989
Karpov.A-Malaniuk.V, US S R Championship, Moscow 1988
Karpov.A-Mchedlishvili.M, German League 2014
Karpov.A-Miroshnichenko.E, Russian Team Championship 2005
Karpov.A-Polgar.J, Las Palmas 1994
Karpov.A-Polgar.J, Linares (round 6) 1994
Karpov.A-S alov.V, Linares 1993
Karpov.A-S ax.G, Linares 1983
Karpov.A-S passky.B, Candidates semi-final (9th matchgame), Leningrad 1974
Karpov.A-S passky.B, Moscow 1973
Karpov.A-Timman.J, Montreal 1979
Karpov.A-Topalov.V, Linares (round 4) 1994
Karpov.A-Uhlmann.W, Madrid 1973
Karpov.A-Yusupov.A, Candidates semi-final (7th matchgame), London 1989
Karpov.A-Yusupov.A, US S R Championship, Moscow 1983
Kasparov.G-Karpov.A, World Championship (22nd matchgame), London/Leningrad 1986
Korchnoi.V-Karpov.A, World Championship (9th matchgame), Merano 1981
Lautier.J-Karpov.A, Linares (round 1) 1994
Portisch.L-Karpov.A, Milan (5th matchgame) 1975
Rodshtein.M-Karpov.A, German League 2014
S ebag.M-Karpov.A, Cap d’Agde (rapid) 2014
S myslov.V-Karpov.A, US S R Championship, Leningrad 1971

Karpov - Move by Move

Uploaded by

Copyright:

Available Formats

Karpov - Move by Move

Uploaded by

Document Information

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

Karpov - Move by Move

Uploaded by

Copyright:

Available Formats

First published in 2015 by Gloucester Publishers Limited, Northburgh House,

10 Northburgh Street, London EC1V 0AT

Copyright © 2015 Sam Collins

The right of Sam Collins

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

British Library Cataloguing-in-Publication Data

ISBN Kindle: 978-1-78194-230-7

Distributed in North America by National Book Network,

Distributed in Europe by Central Books Ltd.,

All other sales enquiries should be directed to Everyman Chess,

Everyman Chess Series

Typeset and edited by First Rank Publishing, Brighton.

Also by the Author

Index of Complete Games

A Taste of what is to Come

1 e4 c5 2 Nf3 e6 3 d4 cxd4 4 Nxd4 Nf6 5 Nc3 d6 6 Be2

Exercise: What is Black’s intention and how should White respond?

Exercise: What should White play now?

Exercise: How should White continue?

Exercise: How should White proceed now?

Exercise: Find a tactical blow to end the game.

Answer: There are at least two good moves here!

Winning more material.

Exercise: How should White continue?

Exercise: How can White respond to the immediate 24 ... Na5 - ?

Exercise: How should Black continue?

Exercise: White to play and win!

Exercise: What is Black’s threat, and how can it be stopped?

Opposite-Coloured Bishops in the Middlegame

1 d4 d5 2 c4 e6 3 Nc3 Be7 4 Nf3 Nf6 5 Bg5 h6 6 Bxf6 Bxf6 7 e3 0-0 8 Qc2

Exercise: What is the cleanest way for Black to equalize?

Exercise: How should White continue?

Exercise: What is your assessment of this position?

28 ... b5 29 Bg6 Bd8 30 Bd3 b4 31 Qg4 Qe8 32 e4 Bg5 33 Rc2 Rxc2?

Exercise: How should White continue?

Exercise: How can White improve his position?

1 d4 Nf6 2 c4 g6 3 Nf3 Bg7 4 g3

My thanks to John Emms for drawing my attention to the next game.

1 d4 Nf6 2 c4 e6 3 Nf3 b6 4 g3 Ba6 5 b3 b5

Exercise: And now what?

1 d4 Nf6 2 c4 g6 3 Nc3 Bg7 4 e4 0-0 5 Be3 d6 6 h3 e5 7 d5 a5 8 c5!?

Exercise: How should White proceed?

Exercise: How should White proceed?

Answer: Let’s look at the positional features:

Exercise: How should White continue?

Exercise: Find the best continuation for White.

Exercise: Find the mate in four.

Exercise: How should White continue?

Exercise: How should White continue?

Exercise: How can White make progress?

The Isolated Queen’s Pawn

Exercise: How can Black activate his game?

Exercise: How should White proceed?

Exercise: How should White play?

Exercise: How should Black respond?

Exercise: What would be White’s response to 16 ... g6 - ?

Exercise: And what now?

Exercise: How should White continue?

Exercise: How can Black put up any resistance in this position?

Exercise: How can Black force a draw?