Anatoly Karpov's

Best Games

Anatoly Karpov

Translated by Sarah J. Young

B. T. Batsford Ltd, London

First published 1 996
© Anatoly Karpov 1996

ISBN 0 7 1 34 7843 8

British Library Cataloguing-in-Publication Data.

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

All rights reserved. No part of this book may be

reproduced, by any means, without prior pennission
of the publisher.

Typeset by Petra Nunn

and printed in Great Britain by
Redwood Books, Trowbridge, Wilts
for the publishers,
B. T. Batsford Ltd,
4 Fitzhardinge Street,
London W l H OAH

A BATSFORD CHESS BOOK

Editorial Panel: Mark Dvoretsky, Jon Speelman
General Adviser: Raymond Keene OBE
Specialist Adviser: Dr John Nunn
Commissionin Editor: Graham Bur ess
Contents
Symbols 5
Introduction 6

1 Karpov - Gik, Moscow University Ch 1969 9

2 Karpov - Hort, Moscow Alekhine mem 1971 13
-\Vsmejkal - Karpov, Leningrad IZ 1973 18
-€}K arpov - Polugaevsky, Moscow Ct (6) 1974 25
5 Karpov - Spassky, Leningrad Ct (9) 1974 29
6 Karpov - Korchnoi, Moscow Ct (2) 1974 33
7 Karpov - Vaganian, Skopje 1976 37
8 Tatai - Karpov, Las Palmas 1977 41
9 Karpov - Korchnoi, Baguio City Wch (14) 1978 45
1 0 Karpov - Korchnoi, Baguio City Wch (32) 1978 49
1 1 Timman - Karpov, Montreal 1979 52
1 2 Korchnoi - Karpov, Merano Wch (9) 1981 55
1 3 Karpov - Korchnoi, Merano Wch (18) 1981 60
14 Karpov - Yusupov, USSR Ch (Moscow) 1983 65
1 5 Karpov - Kasparov, Moscow Wch (9) 1984/5 68
1 6 Karpov - Kasparov, Moscow Wch (27) 1984/5 73
1 7 Karpov- Kasparov, Moscow Wch (4) 1985 78
1 8 Kasparov - Karpov, Moscow Wch (5) 1985 83
,!.2_ Karpov - Kasparov, Moscow Wch (22) 1985 86
�Q·· Karpov - Beliavsky, Moscow tt 1986 88
2 1 Karpov - Kasparov, London/Leningrad Wch (5) 1986 91
22 Karpov - Kasparov, London/Leningrad Wch (1 7) 1986 94
2} Karpov - Kasparov, London/Leningrad Wch (19) 1986 98
241Karpov - Sznapik, Dubai OL (14) 1986 1 02
25 Karpov - A.Soko1ov, Linares Ct (10) 1987 105
26 Kasparov - Karpov, Seville Wch (2) 1987 1 10
CZ1';, Karpov - Farago, Wijk aan Zee 1988 1 13
2f Karpov - Timman, Brussels World Cup 1988 1 16
29 Karpov - Kasparov, Belfort World Cup 1 988 1 20
®> Karpov - M.Gurevich, USSR Ch (Moscow) 1988 1 25
 3J Karpov- Yusupov, USSR Ch ( Moscow) 1988 1 28
@ Karpov - Malaniuk, USSR Ch ( Moscow) 1988 131
3 3 Karpov - Hjartarson, Tilburg 1988 1 33
34 Hjartarson - Karpov, Seattle Ct (3 ) 1989 1 36
35 Karpov - P.Nikolic, Skelleftea World Cup 1989 1 38
� Karpov - Yusupov, London Ct ( 8) 1989 142
37,, Karpov - Timman, Kuala Lumpur Ct ( 4) 1990 145
38 Karpov - Anand, Brussels Ct ( 4) 1991 1 50
39 Karpov- Short, Linares Ct ( 7) 1992 153
40 Karpov - Kamsky, Moscow Alekhine mem / 992 1 57
(�l Kamsky - Karpov, Dortmund 1993 1 63
42 Karpov - Kramnik, Linares 1994 167
- ::41' Karpov - Beliavsky, Linares 1994 1 72
Symbols
1-0 White wins
0� 1 Black wins
lf2_lf2 Draw
+ Check
++ Double check
# Checkmate
! Good move
? Bad move
!! Outstanding move
?? Blunder
!? Interesting move
?! Dubious move
Ch Charnpionship
Cht Team championship
Echt European team championship
tt Team tournament
jr Junior Event
worn Women's event
rpd Rapidplay
Web World Championship
z Zonal
IZ Interzonal
Ct Candidates
OL Olympiad
Corr Postal game
(D) Diagram follows
(n) nth match game
Introduction
In 1968 I started at Moscow State University as a seventeen-year-old, and
in 1969 I won my first tournament in Moscow. You could say that the Mos­
cow University Championship was my first step in the struggle to reach
the top in chess. Immediately after it I was selected for the World Junior
Championship, won it, and six years later, in 1975, became 'adult' World
Champion. A quarter of a century has passed since 1 969, and quite re­
cently, in 1994, I won the ' super-tournament' at Linares - the 'World
Championship Tournament' , and what is more, I can say without false
modesty that I had a unique result - 1 1 points out of 13, and two and a half
points clear of Kasparov.
So, this anniversary of a quarter of a century in top-level chess, and also
my success at Linares, gave me the idea of compiling a collection of my
best games 1969- 1994.
In my already fairly long chess life I have taken part in around two hun­
dred tournaments and matches (not counting team competitions, rapid­
plays and blitz tournaments) and won well over a hundred of them.
Although it would be difficult to count up precisely, I think that all in all I
have won about a thousand games. Obviously, out of a quantity like that it
is not at all easy to pick out the very best games. I hope that all of them are
models, which the reader can judge, of the clearest examples of my work,
and many of them are also part of the true art of chess.
It stands to reason that this book includes all my most interesting victo­
ries from matches for the chess crown, from Candidates battles of various
years, and from prominent international tournaments. Many of the victo­
ries I have chosen won prizes for their beauty, best game prizes, and most
important theoretical game prizes in one of the most popular world chess
publications, lnformator. Incidentally, the author of this book has the
leading number of these lnformator prizes, way ahead of other grandmas­
ters.
Of course, over many years chess players remember fewer and fewer of
their old games, as they are displaced by fresher and more important du­
els. That is why this book in the main concentrates on my victories over
the last ten years.
 Introduction 7

I have also written new commentaries for the older games, or at least re­
worked the opening part, as theory has advanced so much since then.
In an overwhelming number of games in this collection I am playing
White. This is not surprising, as beautiful victories are most often gained
with the white pieces.
Let us quickly run through the composition of this collection. Which of
my victories have got into it?
The first game, dated 1969, was played when I was still a master, but the
next one is taken from a strong grandmaster tournament. Then my pro­
gress towards the Mount Olympus of chess is illustrated - here the reader
will find one game from the Interzonal tournament and all the Candidates'
matches. My match with Fischer did not take place, through no fault of
mine, and in 1975 I became the 1 2th World Champion. Then there are
some games from prominent international tournaments, and then two en­
counters from my duels with Korchnoi in Baguio and Merano. You could
say that my victory in our anniversary national championship ends the
first part of the book.
In 1984 a new era began in chess, linked with Karpov-Kasparov con­
frontation. Here the reader will find ten of my victories from my first four
duels with Kasparov. I also bring you some games played in between
these 'epoch-making' matches. After our battle in Seville, our dispute was
interrupted for three years. In this part of the book you will find my victo­
ries in the World Cup, our 55th national championships, and other con­
tests. After 15 years I again had to climb the Candidates ladder, and most
of the steps are illustrated with a victory, as well as games from my next
duel with Kasparov. This is where you could say the second part of my
book ends.
Soon I was joining the regular cycle of battles for the crown, which also
means regular games from new Candidates matches. Unfortunately, my
encounter with Short did not end happily for me, and, alas, my possible
sixth meeting with Kasparov did not take place. Therefore there are sev­
eral wins from tournaments in the early 1990s. In 1993 a schism occurred
in the chess world (it is scarcely worth dwelling on that question in this
book), as a result of which I was able to play another match for the chess
crown. Having won my match against Timman, I was again declared
FIDE World Champion (Kasparov and Short had played for the title of
PCA World Champion).
8 Introduction

I end the book with two victories from the 'super-tournament' at Li­
nares in 1994. Although after that I have won other interesting tournament
games, I think that the full stop in this book should be placed precisely af­
ter Linares, where I achieved one of the most convincing victories in the
entire history of chess.
In conclusion it remains for me to thank the chess master and writer
Evgeny Gik for his help in preparing the manuscript.

Anatoly Karpov
 Karpov - Gik 9

Game 1 The Dragon Variation, one of

Karpov - Gik the sharpest and most fascinating
Moscow University Ch 1969 in chess theory. Both sides' aims
Sicilian, Dragon are the same - getting to the enemy
king as quickly as possible. Be­
The Moscow University Champi­ cause of this, White attacks on the
onship was my first tournament in kingside, stopping at nothing, while
the capital, and I really wanted to Black in his turn attacks on the
win it. The champion's title was queenside. Games played in this
decided by this very game, and variation are almost always played
who would have thought that my in an open fighting spirit, and are
opponent in my first tournament in often awarded prizes for beauty.
the capital would later become co­ 1 1 h4 lLle5
author for some of my books? 12 �b3 l:r.fc8
1 e4 c5 Nowadays the queen's rook is
2 lLlf3 d6 most often placed on c8 (putting
3 d4 cxd4 the queen on a5 in that case is not
4 lLlxd4 lLlf6 obligatory), but at that time mov­
5 lLlc3 g6 ing the king's rook to c8 was more
6 �e3 �g7 popular.
7 f3 lLlc6 13 h5 lLlxh5
8 �c4 0-0 14 �h6
9 ifd2 'it'aS At the end of the 1 960s this posi­
10 0-0-0 �d7 (D) tion was subjected to thorough
analysis, and moreover there was a
lively discussion about the knight
check 14 . . . lLld3+. In this game I
was preparing to test one of my
own ideas on this theme, but my
opponent managed to avoid my
preparation.
14 ... �xh6
15 1Wxh6 .:xc3 (D)
The standard exchange sacrifice
in the Dragon - on the one hand
Black protects himself from the
w knight lunging onto d5, and on the
10 Moscow University Ch 1969

other he shatters the enemy king's

fortress.

B
18 g4 lLlf6
19 gS lLlhS (D)
w
16 bxc3 'if:xc3
The black queen here is occupy­
ing an ideal position to generate
threats to the white king, and it is
difficult to believe that this move
can already be a decisive mistake.
Either 16 . . .lLlf6 or 1 6 ....:r.c8 was
necessary.
17 lLle2! (D)
In the event of 1 7 ..t>b1 a5 Black
obtains fair counterchances. How­
ever, the modest knight retreat to w
e2 is extremely unpleasant for 20 .:r.xhS!
Black. The knight deals with the There is no time to lose. 20 lLlg3,
problem of ousting the queen beau­ which I almost played at the board,
tifully, and simultaneously joins in had to be rejected because at the
the attack on the kingside. last moment I spotted the effective
17 ••• 'iWcS 20... .i.g4 ! , which would exclude
Alas, after 1 7 ... lLld3+ 1 8 .:r.xd3 the white queen from the game.
'iWa1+ 19 ..t>d2 'iVxh1 20 g4 lLlg3 2 1 20 ••• gxhS
'iVxh 1 lLlxh 1 2 2 �e3 ! and 2 3 .:r.d 1 21 .:r.h1 'iWe3+
Black's knight is lost. 22 �bl!

In the Dragon even a slight inac­
curacy can spoil the whole game.
Thus, for example, 22 �b2? would
have given Black at least a draw:
22 . . . �d3+ 23 cxd3 (23 �bl loses
after 23 . . .'ifxf3 ! ) 23 ... 'ifxe2+ 24
�al 'ifxd3 and Black is guaran­
teed a perpetual check.
22••• 'i'xf3
The white knight is untouch­
able: 22... 'ifxe2 23 'fi'xh5 e6 24
'ifxh7+� 25 'fi'h8+ �e7 26 'iff6+
�e8 27 .::t h8#. 22 ... e6 also does not
work: 23 'ifxh5 'ifxf3 (23 . . .�g6 24
'ifxh7+ �f8 25 �g3 and 26 �f5)
24 'ifxh7+ �f8 25 �d4.
23 .::txhS (D)
Karpov - Gik 1 1
for White here was only revealed
ten years after this game. The key
move for White was ignored by the
world press, as many commentar­
ies only examined, after 24 "ii'xh7+
'ittf8, 25 .::th6 and 25 'fi'h6+:
a) 25 .::t h6 e6 26 .::t xg6 fxg6 27
it'xd7 "ii'xe2 28 "ii'xd6+ �g7 29
"ii'e7+ 'itth8 30 "ii'f6+ �h7 31 "ii'f7+
�h8 32 "ii'xg6 "ii'd l + 33 �b2
'ii'd4+ 34 c3 "ii'd2+ 35 .i.c2 'it'd7 !
and Black maintains equality.
b) 25 il'h6+ was also suggested
in my notes, with the following
variation: 25 . . .�e8 26 il'h8+ �f8
27 .::th7 and now:
bl) 27 . . . e6 turns out to be inac­
curate, as the bishop should be
placed on e6, not the pawn; after 28
g6 fxg6 29 �d4 'ifxe4 30 'ii'g 8
it'xd4 31 'iff7 + 'it>d8 32 'ifxf8+
'it>c7 33 il'xa8 the white king easily
steals away from the 'perpetual'
check: 33 ...'ii'dl+ 34 'it>b2 'ii'd4+
35 c3 'ii'f2+ 36 .i.c2 "ii'b6+ 37 �al
'ifgl+ 38 .i.bl .
b2) 27 ....i.e6 ! 28 g6 (28 �d4 is
not dangerous for Black either, as
after 28 . . .'ii'xe4 29 �xe6 fxe6 30
g6 .::tc 8 31 g7 il'e l + 32 �b2 White
is forced to submit to the perpetual
check; or 28 .i.xe6 fxe6 29 g6
'ii'xe2 and then . . . "ii'el+) 28 . . . fxg6
(28 . . . "ii'xe2 29 g7) 29 .i.xe6 "Yi'xe2
30 .:n 'ii'el+ 31 �b2 'ifb4+ 32
.i.b3 'i!fxb3+! (32 . . .�d7 33 .::t xf8
.::t xf8 34 'fi'xf8 'ifxe4 with three
12 Moscow University Ch 1969
pawns for the piece) 33 axb3 �xf7
with very unclear play. Moreover,
maybe Black does not have to give
up his queen.
It is amusing that old and, as we
shall see, not particularly precise
variations for some time caused
experts to change their attitude to­
wards 16 ...'ikxc3, and in some
theoretical publications this move
has been considered quite play­
able. But all the same, microscopic
analysis allowed me to return, and I
think definitively, to the old assess­
ment: taking on c3 loses for Black.
If after 23 ...lLlg6 24 'i!t'xh7+ �f8,
instead of 25 'ii'h6+ or 25 .:h6,
White should make the quiet rook
move, 25 l:hl ! ! , then Black would
be completely helpless.
Having thought of this rook ma­
noeuvre, it was easy to find sup­
porting variations:
c) 25 .:hi ! ! (D) and then:
B Black ! The elegant geometrical
c l ) 25 . . ..i.e6 26 lLld4 (but not
the continuation 26 .i.xe6 fxe6 27
lbd4 'ii'f7 !) 26 ... 'ii'xe4 27 lLlxe6+
fxe6 28 :Cl+ 'iti>e8 29 'ii'g8+.
c2) 25 ...e5 26 lLlg3 ! 'ifte8 27 .:n
'itxg3 28 'i!t'g8+ lLlf8 29 'ii'xf7+
�d8 30 'ilixf8+ .i.e8 31 'ii'xd6+
and Black is forced to lay down his
arms.
c3) 25 . . . e6 26 ltJd4 'iixe4 (al­
ternatively, 26 .. .'iff4 27 lLlf5) 27
.:n .i.e8 28 lLlxe6+ 'ifte7 29 lLlc7.
24 g6!
The courageous pawn sacri­
fices itself, guaranteeing a breach
in the enemy defences. The hasty
24 ifxh7+ would have left the
black king at large: 24 . . . �f8 and
now not 25 ii'h8+ (nor 25 lLld4 ow­
ing to 25 . . .'ilidl + 26 'iti'b2 'ii'xd4+)
25 .. .ri;e7 26 'ii'xa8 due to 26...'ii'xh5.
24 .•• lLlxg6
Not 24 ... fxg6 (24 . . . hxg6 is even
worse: 25 'ilih8#) 25 'ii'xh7+ 'ifilf8
26 'ii'h 8+ �e7 27 .:h7+ lLlf7 28
'iVxa8.
However, if, as in the game,
Black takes on g6 with the knight,
it seems that everything is in order
for him: White's attack has been
repulsed, and furthermore he is
two pawns down.
25 'i!t'xh7+ 'itf8
26 .:rs ! ! (DJ
This unexpected rook move was
like a thunderclap in a clear sky for
 Karpov - Hort 13

.I �
-
� �
� -�� - - �
­�
�' -
• �
� �
� ... .�
w�' �
-
��\Wr
w
. . -�·
• • •••
.
- -
�
� � .
�� � -
·�· . .
-�· . .
.<it?. • •
B B
idea immediately decides the fate 31 'ii'xg6+ ri;e7
of the game. Two lines - the a2-g8 32 'ilg5+!
diagonal and the f-file - intersect After 32 exf5 .:fs Black could
on one critical square - f7. The still resist.
threat is 27 'ifxf7#, and moreover, 32 ri;e8
the rook is supporting the queen 33 exfS .:c8
along the file, while the bishop - in 34 'ii'g8+ q;e7
the event of 26 . . .exf5 - is support­ 35 'ii'g7+ 1-0
ing her on the diagonal. Because of
this Black is forced to part with his Game 2
own queen. Karpov - Hort
26 'ilxb3+ Moscow Alekhine mem 1971
27 axb3 exfS Sicilian, Keres Attack
28 lbf4! (D)
Another elegant manoeuvre. 1 e4 cS ,
The aS-rook is not defended, and 2 lbf3 d6
thanks to this White can com­ 3 d4 cxd4
pletely destroys his opponent's de­ 4 lbxd4 lbf6
fences. 5 lbc3 e6
28 ... .:d8 6 g4 (D)
29 'ii'h6+ The Keres Attack - my favourite
A last finesse - the g6-pawn will weapon against the Scheveningen
drop with check. in the 1970s and 1980s. This vic­
29 �e8 tory is perhaps one of the clearest I
30 lbxg6 fxg6 gained in this variation, although
14 Moscow Alekhine mem 1971
B dentally, does not work, as after the
almost a quarter of a century has
passed since .
6 ... 4Jc6
It is interesting that I used the
Keres Attack in the very first game
of my long marathon against Kas­
parov (Moscow 1984/5). True, we
did not play it again, except a
modified form of it in game 14 of
our second match, when I chose an
unusual move-order: 2 . . .e6 3 d4
cxd4 4 4Jxd4 4Jc6 5 4Jc3 d6 6
g4 !?, not then expecting the knight
to go to f6. After a sharp battle the
matter ended peacefully. In the first
game of our first match Kasparov
chose 6 . . . h6 in the diagram posi­
tion, a more popular continuation
in recent years (6. . . a6 is a third pos­
sibility). After 7 h4 4Jc6 8 :g 1
h5 ! ? 9 gxh5 4Jxh5 10 .ig54Jf6 1 1
'ilfd2 I gained somewhat the better
chances, but that encounter also
ended in a draw.
7 gS 4Jd7
8 f4
8 .ie3 a6 (two other popular
lines are 8 ...4Jb6 and 8 ....ie7) 9 h4
(theory has also minutely studied 9
a4, 9'tlfd2 and 9 :g1) is an impor­
tant alternative ; 8 4Jdb5 4Jb6 9
.if44Je5 10 'ilih5 has been played
in its time; now 10 .. 4Jg6 ! (after
.
1o... g6 1 1 it'h3 .id7 12 o-o-o it'b8
13 it'g3 Black is defenceless) 11
.ig3 ( 1 1 .ixd6 .ixd6 1 2 :d1, inci­
cold-blooded 12 . 0-0 13 4Jxd6
..
'ii'e 7 ! , the initiative is on Black's
side in spite of his lack of pawns)
l l . . .a6 1 2 4Jd4 .ie7 (12 . . . d5 ! is
also interesting) 13 0-0-0 .ixg5+
14 'it>b1 0-0 is unclear.
8
••• a6
9 .ie3 .ie7
9 . . . h6? is very risky: 1 0 4Jxe 6 !
fxe6 1 1 'ii'h5+ �e7 1 2 .ic4 'ii'e 8
13 1li'h3. However, 9 ...'tlfc7 de­
serves attention.
10 :g1 (D)
10 'tlfd2 4Jxd4 11 .ixd4 e5 1 2
.ie3 exf4 1 3 .ixf44Je5 1 4 0-0-0
.ie6 15 h4 'i!i'a5 16 a3 :c8 174Jd5
led to an even game in Tarjan-Hu­
lak, Indonesia 1983.
10
••• 4Jxd4
Black does not usually hurry
with this exchange , but here it is
not easy for him to find another
move : IO. .'i'c7 looks too sluggish;
.
deciding on 10 ...0-0 when one is
directly under attack is not easy;

B B
10 . . . lDc5 1 1 lDxc6 bxc6 1 2 �xc5
gives White a clear advantage.
1 1 'ilt'xd4 e5
12 'ilt'd2 exf4
13 �xf4 lDe5
1 3 . . . 'ilt'b6 14 llg3 'ilt'xb2 1 5 llb1
Wa3 1 6 lDd5 is obviously bad for
Black.
14 �e2 �e6
Black could not generate any ac­
tivity with 1 4 . . . 'ilt'a5 as the end­
game which follows is clearly in
White's favour: 1 5 lDd5 'ilt'xd2+ 16
'itxd2 �d8 17 llad 1 �e6 18 �c 1 .
15 lDd5
Immediately ! Otherwise Black's
queen could take up an active posi­
tion, e.g. 15 0-0-0 'ilt'a5 ! .
1 5 ... �xd5
16 exd5 (D)
As a rule, you should try to oc­
cupy a blockading square with a
piece. Actually, 16 'ilt'xd5 would
also have left me with an advan­
tage, as the d6-square demands
Karpov - Hort 15
constant defence by Black. But
then my e4-pawn would also in
some cases need watching, which
might restrict my light-squared
bishop. Now this bishop is free for
its own action, all the more so as its
black counterpart is no longer on
the board.
16••• lbg6
Black's position is appreciably
worse, so Hort seeks a tactical so­
lution to his problems. Castling on
either side ( 16 0-0 or 16 ...'ilt'c7
...
followed by 16 ...0-0-0) would con-
demn Black to passive defence.
17 �e3 h6!?
It is difficult to give a move like
this a symbol. It shows a fighting
mood and a desire to carry out the
battle in a tactical key. But it also
represents a precise assessment of
the position with a clear under­
standing of its disadvantages. In
short, it is risky for both sides !
18 gxh6 �h4+
16 Moscow Alekhine mem 197 1
1 9 �d1 (D)
B The h2-pawn is disturbing no­
The white king is not troubled
by the loss of his castling rights. At
the same time, the black king
maintains that possibility right to
the end of the game, but he does
not manage to make use of it.
19 ... gxh6
20 �xh6 �f6
20 ...Wf6 would probably not
have worked out for Black in view
of the fact that the h4-bishop has
perceptibly lost its mobility. Hort
was intending idealistically to re­
build his forces, but in chess there
are two sides, and I in my turn took
some necessary countermeasures.
21 c3 �eS (D)
Black seems to have achieved
his aim - he is threatening to play
22 ...Wh4+, while after 22 �g5 "ilfb6
23 �e3 Wc7 my achievements will
not be that great. But White finds a
very strong retort.
22 l:.g4!
one (for the time being !). Now the
main thing is not allowing the en­
emy queen an active position on
the kingside, where my pieces are
rather precariously arranged.
22 ... 'iVf6
22 ...�xh2 would have been a
relatively better choice, restoring
the material balance. But I can un­
derstand my opponent; he wanted
to develop.
23 h4!
Now not 23 ... li)xh4? because of
24 �g7. In general it will be diffi­
cult for Black to win back this until
recently so helpless pawn, which is
now gradually finding its strength.
23 . ..0-0-0 was also impossible in
view of 24 �g5.
23 ... 'ii'fS
Defending against the threat­
ened 24 h5, Black prepares to cas­
tle long.
 Karpov - Hon 17

24 l:tb4! (D)

w
nothing after 27 ......xh5 28 l:txf6
B 'iVh l + 29 .tfl (the rook defends
A beautiful place for the rook ! both bishops) 29 ... t[jg8 30 •e l+ ,

In the first place 24 0-0-0 25 .ig4

... and White wins.
is impossible, and in any case the 27 ... t[jxd5
b7 -pawn is under threat of capture. 28 l:td3 l:txb6
24 .•• .tf6 Nothing better is apparent - if
25 h5 t[je7 28 ... t[je7, then 29 .if4 is decisive.
Of course, this retreat is less val­ 29 l:txd5 "ii'e4 (D)
iant than 25 . . . �5. but then his
pieces would not be able to gener­
ate any real activity (after 26 l:tf4) .
I should note in passing that during
all this time it has not been feasible
for Black to sacrifice the exchange
with ...l:txh6.
26 l:tf4 'iWe5 (D)
27 l:tf3
The rook, which so often lacks
agility, in this position shows re­
markable manoeuvrability. It cre­
ates one threat after another, and
moreover is acting productively
not only in attack, but also in de­
fence. Thus, Black will now gain
18 Leningrad IZ 1973

A hymn to the rook! The chess

player's consciousness is used to
the working potential of all his
pieces, but, you will agree, with
rooks this mainly applies to the
endgame.
30 ••• 'ii'h 1+
Black is forced to go into this by
no means equal exchange.
31 �c2 'iVxa1
32 'ifxh6 i.eS
33 'iWgS w
Preventing queenside castling 9 lba4
and creating irresistible threats to One of the most popular posi­
the exposed king. In this desperate tions in the Paulsen System, which
position Hort lost on time. was established long ago. The b6-
1-0 square is weakened, and White dis­
The game was nominated as the patches his knight to it. As well as
best at the Alekhine Memorial, and this he makes use of the fact that
then lnformator included it in their taking the e4-pawn is dangerous:
best games. Although many years 9 . . . lbxe4 1 0 lbxc6 'iVxc6 1 1 ltlb6
have passed, I will confess that I .:.bs 1 2 'i!Vd4 i.f8 1 3 i.f3 d5 (or
still love this game. 1 3 ...f5 14 .:.ad 1 ) 14 c4.
9 .
.. 0-0
Game{!) Black has a wide choice: 9 ...b5 ?
Smejkal Karpov
- or 9 .. d5?; retreating the bishop to
.

Leningrad IZ 1973 d6 or e7, or even 9 . . lbe7. But cas­

.

Sicilian, Paulsen Hing is the most reliable way of fin­

ishing his development.
1 e4 c5 10 lbxc6 bxc6
2 ltlf3 e6 Theory shows a clear-cut way
3 d4 cxd4 for White to gain an advantage af­
4 lbxd4 lbf6 ter either 10...dxc6 or 10...'iVxc6.
5 lbc3 a6 1 1 lbb6
6 i.e2 'fic7 The most logical, although 1 1
7 0-0 lbc6 c4 and 1 1 f4 are also known.
8 i.e3 i.b4 (D) 1 1 ... .:.bs (DJ
 Smejkal - Karpov 19

-�· ••• . . . ••
�-
� �.....
- · - .
· �·····
- · -
·�···- . �-···- .
• • • • • • • •
·- �·
- - .
- . --��-�
� .
-
- . . - . . . .
�"�
U �-�U "
��,� U ·
�"� "�rf.�
� U " ,
�
- -
�
��-
� � :�
� �
- -�-
� � :�
�
w w
12 tbxc8 l:l.fxc8 game after 24 'ifxf7+. As you can
Twenty-odd years after this game 'I
see, if the rook were on e8, Black
I decided to test another capture- would have had to resign immed.i-
1 2 ... 'ifxc8 (Salov-Karpov, Buenos ately.
Aires, Sicilian theme tournament 14 i.d3
1 994) 1 3 e5 ltJd5 14 i.c 1 i.c5 and 14 i.g5 has apparently not been
now White, instead of the theoreti- tested in practice, for example,
cal recommendations 1 5 i.d3 or 14 .. .'ii'e5 1 5 i.h4 'ii'xe4 1 6 .i.g3
1 5 c4, decided on an aggressive l:l.a8 17 i.d3 leads to a complex po­
policy - bringing his queen over to sition.
the kingside with the standard Si- 14 ... i.d6
cilian manoeuvre 'ii'd 1-d3-g3: 1 5 15 �h1
'ii'd 3 0-0 16 'it'g3l:l.e8!? 17l:l.dl aS 15 f4 has also been seen more
1 8 b3 a4 1 9 i.h6 i.f8 20 h4 'f!lc7. than once. If 1 5 ... e5 16 f5l:l.xb2 17
Thus, pressure on the g7-point is g4, then 1 7 . ..'ifa5 18 �h 1 i.c5 19
parried, and the ending arising is i.c 1 l:l.xa2 20 l:l.xa2 "ii'xa2 21 g5
highly favourable for Black. ltJe8 22 'ifh5 d5 ! 23 g6 lLlf6 is
13 i.xa6 l:l.d8! (D) good for Black.
At the time when this game was 15 .i.e5
played, the rook move to d8 was 16 c3 l:l.xb2
presented as a novelty, and more- 17 'it'cl ltJg4! (D)
over a very important one. It looks 18 f4
as though the rook is no worse on Of course, 1 8 'W'xb2 loses after
e8, but run ahead for a second and 1 8 ...i.xc3 ! .
have a look at the situation i n the 18••• ltJxe3
20 Leningrad IZ 1973

21 ..• eS
Later it was established that here
the best move is 2 l . .. g5 !? with
roughly the following variation: 22
g3 'ifd6 ! (an important intermedi­
ary manoeuvre) 23 i.e2 i.e5 24
'ifxf7+ Wh8 25 l:td l 'it'c7 26 'ife7
(26 .tg4 d5 ! 27 'ifxe6 i.xc3 28
exd5 cxd5 29 i.f3 We5 30 Wxe5+
i.xe5 with a draw) 26 ... i.xc3 27
i.g4 W'c8 ! 28 i.xe6 l:te8 29 Wc5
w l:txe6 30 Wxc3+ �g8 3 1 'ii'c4 'ii'b 8
19 'li'xbl 32 a4 'ife5 Razuvaev-Matulovic,
1 9 fxe5 is no good in view of Tbilisi 1 973.
1 9 'ifb6 with a clear advantage.
... 22 g3 'ifd6!
19 ... i.xf4 23 i.e2 i.gS
But not 19 ... lLlxfl ? 20 fxe5 l2Je3 24 'ilfxf7+ �h8
2 1 W'e2 'it'b6 22 :tel . 25 a4 (D)
20 W'f2 Incidentally, 25 i.g4 is also a
He has to return the exchange, promising coninuation for White,
because after 20 l:tf3 lLlg4 Black's for example, 25 ...i.f6 26 �g2
chances are better, as 2 1 h3 is im­ l:tf8 27 'ii'xd7 'ii'c5 28 �h3 Wc4 29
possible due to 2 1 ...i.c 1 ! !. l:tf2 preserving an advantage for
20 .•. lLlxfl White; Adorjan-Matulovic, Novi
21 :xn (DJ Sad 1 973.

B
 Smejkal - Karpov 21

Strangely enough, I had this po­ Wh4 'ii'f6+ 35 'ii'xf6 gxf6 immedi­
sition on the board when I was pre­ ately loses a pawn to 36 Jld7 and
paring for the game. True, it turns the white king will easily steal his
out to be weak for Black. I had to way through the holes in the pawn
apply more than a little force to ex­ barrier) 34 'i!fe7 ! Jle3 ! (the strongest
tricate myself from it, and then I continuation because co-operative
managed to swindle my opponent. variations such as 34 ... h6 35 i.e6
25 Jle7 \i'fl + 36 �h4 lead to victory for
26 aS l:.f8 White), and now Black, by defend­
27 'i'c4 l:.xfi+ ing against the direct threat, creates
28 Jlxfi 'ili'f6! counterplay:
The only way for Black to or­ a) 35 Jle6 'ii'f l + 36 �g4 'i'e2+
ganize his position. The bishop be­ (but on no account 36 ... h5+?? be­
longs on c5, and so needs support cause of 37 �h4) 37 �h3 'ii'f l + 38
from the queen on f8. Wh4 g5+ 39 �g4 (after 39 'itth 5?
29 �gl 'ili'f8 Black forces mate: 39...'ii'e2+ 40
30 Jle2 Jlc5 Jlg4 'fi'xh2+ 4 1 Jlh3 'ii'xh3#)
31 Jlg4 'iif2+ 39 ...'ii'e2+ 4(1 'itth 3 'iWfl +, etc., is a
32 �h3 d6 (D) perpetual check.
b) White cannot even advance
his passed pawn with 35 a6 since
after 35 . ..h6 threatening ...'iWfl +
and ...Jlg5+, he would be forced to
give perpetual check himself.
c) 35 Jle2! (not 35 ...'ii'xe2?? 36
ir'f8#).
d) 35 'i!fxd6 also leads to per­
petual check after 35 ... 'iWfl+ 36
�h4 g5+ 37 �h5 'iWf7+ 38 'itth6
ii'g7+ 39 �h5 'i'f7+.
Smejkal was counting on win­
w ning the c6-pawn and quietly going
The critical moment. about attaining an advantage. The
33 Jld7 move played is perhaps the correct
33 'ii'e6 g6 (Black cannot move practical decision.
into an ending with opposite-col­ 33 g6
oured bishops, since 33 ... 'iif l+ 34 34 Jlxc6 'ittg7
22 Leningrad IZ 1973

35 .i.bS 'ifb2! 38 .i.c4 Wet

A fine move, the idea of which 39 Wn ?
Smejkal had apparently over­ 39 �g2 was correct. Now Black
looked. A threat to the h2-pawn has achieves an advantage.
suddenly arisen, and both the c3- 39 Wh6+
pawn and the b5-bishop find them­ 40 �g2 Wxh2+
selves under attack. White has to 41 �f3 Wh5+
part with the c3-pawn, but in the 42 �g2 (D)
first place he has a very dangerous Here the game was adjourned.
passed pawn. The misfortune for Analysis shows that with precise
Black lies in the fact that he can play White has chances for a draw,
never transfer play into an oppo­ but to achieve this he will have to
site-coloured bishop endgame, overcome great difficulties.
because White, by threatening an
exchange, can at any moment
'question' the black queen from al-
most any square.
36 a6 .i.gl
37 We2 Wxc3 (D)

• • • •
-� �
�-� -� �-�
:� -� .,.
-
�- . . • •
• .i. . • • B
. ·�· . 42 Wh2+
•
. �tlli
-
• �
. �� 43 �f3 .i.d4
•
• �
-
�..wrll.� "
�,@-� u 44 .i.dS .i.cS
45 .i.c6 .i.d4
.
. �-�
� �
� -
� - � 46 .i.b7? (D)
w White has fallen into the trap I
Precise play by Black has al­ set for him and made my problems
most equalized the position, but the easier. 46 .i.d5 ! was correct. I do
presence of queens on the board not at all want to say that after this
makes the situation dangerous for move everything is in order for
both sides. White, because that is really not
 Smejkal - Karpov 23

the case. In the adjourned position Black to begin his pawn movement
Black can try various ways of with his bishop on c5 or d4, while it
achieving victory, but White has is better for White to keep his own
defensive resources for every one bishop on c6 or dS. As you can see,
of them. a system of corresponding squares
has arisen. Smejkal feared this and
accordingly spent more time ana­
• • • • lysing a different continuation, so
�· -·
.i.. �• ��
he did not find this area of corre­
- - �··
- - spondence, and took his bishop too
far away. I will show the difference
. .
- - �
- .
-
•
• - A �-� •
-L:l· •
in later notes .
46 gS!
• • •�o 47 �g4 hS+
�
-
� .
� • .
• mu - 48 �rs
The only move. After 48 �xg5
. . . .. . 'iVxg3+ 49 �xhS (49 �fS 'ii'g4#)
B .
49 . . ..i.f2! White has no defence
The position which has arisen against mate.
on the board is a kind of zugzwang 48 'iVxg3
for White. The king cannot go to 49 �e6 'iif2 ! (D)
the only free square, g4, in view of
the mate. The white queen cannot
move to either g2 or e2, as in that
case he will be forced to advance
the g3-pawn by means of ...'iVhS+,
and definitively compromise the
position of the king. In fact, only
the bishop can move. In his turn,
Black has to find a plan to
strengthen his position. As we saw
earlier, simple checks here achieve
nothing. I chose a plan linked with
advancing my kingside pawns.
Analysis revealed that this con­
tinuation is not the only one, but it
is effective enough. It is better for
24 Leningrad IZ 1973

his advantage: 50 'ii'xf2 .i.xf2 5 1 Black has successfully arranged

�xd6 g4 ! and the bishop does not his forces and easily parries all
manage to stop the pawn from White's attempts to create counter­
reaching f1 , whilst in the event of play.
52 .i.c8 �f6 Black advances the 61 'ii'al 'ii'fl
king. The difference lies in the fact 62 'iVbl g3
that if the bishop were not on b7, 63 .th3 (D)
but on d5 instead, White could
have continued 50 "iixf2 .txf2 5 1
�f5 ! g4 52 .tc4 ! and the passed • • • •
pawn is held back, while the black . . - .
king cannot move forward. ��
- � - .
-� . .
50 'ii'f6+ -
• - w;@
.W%:-?1
0. ?&l.&
.0:?;!; "
� "'• � •
51 �dS g4
A.
,__ '

52 .tc8 "fle7
•
• • .0 • • .
53 .trs �h6 • • • .i.
• • mu •
-
54 "ii'n "ilc7
. . .
SS 'ii'el 'ii'cS+
56 �e6 (D) B'ii'B • B
B
White has occupied the last line
of defence, but this blockade is
only temporary.
63 .•• �h4
64 .tgl 'iVgl
Black forces an exchange of
queens, after which the road to vic­
tory becomes quite short.
65 'ii'xgl .i.xgl
66 �d6
Black would have achieved an
B effective finale after 66 �f6 .td4
56 �gS 67 .tfl g2! 68 .txg2 �g3 fol­
57 wn 'ii'a3 lowed by ...h4.
58 'ii'el .tcs 66 .i.d4
59 'ii'dl+ 'ii'e3 67 a7 .txa7
60 'ii'aS .tb6 68 �xeS �g4
 Karpov - Polugaevsky 25

69�dS h4
70eS h3 •• • •••
71.txh3+ �xh3 --
� ·�
� - -�.�. -
72e6 .tcS! •• ·-*-- •
0-1
This encounter won the regular
. . - .
competition in lnformator for the � • • �a •
most important theoretical game. -
�!@�� -
��� � � ­�
Game@
" . i. -
u �· �� � "
� u
Karpov - Polugaevsky a ••• •<att
Moscow Ct (6) 1974 w
Sicilian, Najdorf .:td 1 .:te8 17 tbf5 .td8 1 8 tbd4
tbg6 1 9 .:tffl tbe5, Black had a
1 e4 cS lovely game. Later on both 14 'ii'd2
2 tbf3 d6 and 14 tbd5 were tested several
3 d4 cxd4 times. The second of these moves
4 tbxd4 tbf6 deserves special attention, for ex­
5 tbc3 a6 ample in the encounter Hulak-Por­
6 .te2 eS tisch, Indonesia 1 983, where after
7 lbb3 .te7 14 tbd5 .txd5 1 5 exd5 tbfd7 1 6
8 0-0 .te6 .:tb4 .:tfe8 1 7 a5 .tf6 1 8 .t g1 .tg5
9 f4 'fic7 19 tbd2 tbf6 20 tbfl g6 White
10 a4 lbbd7 could have gained an advantage by
1 1 �h1 0-0 attacking on the queenside: 2 1 c3
12 .te3 exf4 .:te7 22 ifb3 h5 23 h3 lbh7 (pre­
13 .:txf4 tbeS (D) paring ... f5) 24 .:tb6 followed by 25
14 aS .:ta4.
Fourteen moves have gone and I 14 ... tbfd7
have not yet made a single com­ It was scarcely worth moving
ment. This is not surprising - this this knight away from the centre.
position is one of the standard pat­ The back rank is free of pieces, and
terns in the Najdorf Variation, and an initial manoeuvre by one of
it has been seen in practice many Black's rooks suggests itself -
times. In the fourth game of the ....:tfe8, ....:ae8, or ....:ac8. For ex­
match I played the weaker 14 tbd4, ample, after 14 . . . .:tac8 1 5 tbd4
and after 14 ....:tad8 1 5 'ii'g 1 .:td7 1 6 tbfd7 16 'ii'd2 .:tfe8 17 tbf5 .tf8 1 8
26 Moscow Ct (6) 1974

.:tf2 �h8 in Matanovic-Polugaev­ ends up in a difficult positional

sky, Moscow 1 977 a tense, com­ bind.
plex struggle arose. 18 lbd4 'ji'xb2
1s .:r.n ..tf6 Not going the whole hog (even
16 lbdS! ..txd5 if that is what Black wanted to do)
17 •xd5! ? would lead to a position with
After the quiet 17 exd5 Black material equality and a noticeable
should continue 17 ...lbc4 1 8 ..txc4 advantage to White. For example,
'ifxc4, and White's opening initia­ after 1 8 . . ...c5 1 9 lbf5 ! 'ji'xd5 20
tive is exhausted. In order to pre­ exd5 the black pawn on d6 cannot
serve it, an idea came to me during be saved.
the game of sacrificing two pawns 19 .:tabl 'ji'cJ
straight away. 1 9 . ....a3 is worse - the black
17 "ikxc2 (D) queen finds herself on the edge of
the board and totally out of the
game.
20 lbfS (D)
The aim of this move is to de­
fend the bishop and not allow the
queen to c5, which might happen
in the variation 20 .:tb3 'ji'cS 2 1
lbf5 'ji'xd5 2 2 exd5 lbc5, when
Black can defend himself.

w
Polugaevsky spent more than an
hour wondering whether to take or
not. He might have been thinking
that if I was prepared to sacrifice, it
meant that I had analysed every­
thing thoroughly at home, but on
the other hand, he could not find the
forced loss at all, so he kept search­
ing through it again and again. If
Black declines the sacrifice he
 Karpov - Polugaevsky 27

Polugaevsky is manoeuvring 24 ...t:he5 25 lhxb7, but in both

magnificently. If Black procrasti­ cases White has the initiative.
nates, White can improve the inter­ 25 'ii'xb7 (D)
action of his pieces: 2 1 J.d4 'ii'd2 I was short of first two, and then
22 l:.b2 or 2 1 l:.fc l Wa3 22 l:.al one pawn, and perhaps I was there­
'i!i'b2 23 l:.a2 'i!i'b4 24 J.d2 ! trap­ fore unconsciously trying to liqui­
ping the queen. With the move in date my 'material disadvantage' .
the game Black avoids this. However, 25 'i!i'h5 would have cre­
21 l:.be1 ated the threat of 26 l:.xf6 gxf6 27
Imprecise. After 2 1 l:.fe 1 White 'ili'g4+ c;i;lh8 28 J.h6, which would
would have gained a big advan­ not. have been easy to repulse
tage. Now Black can construct a (25 ...g6 26 'ii'c 5).
defence.
21 t:hc5
22 t:hxd6 lhcd3 •• • •••
23 J. xd3 t:hxd3 (D) .'ii' . • .......
.... � . .
.1.•�$! •· �··
- - �
u . .
• • .
•
..... ··-· -
- .,q�
.0. •
•
� �
. !?L.X
... • -
.
- • • m •
" •••.
� � � .� •v• •�o
. -�· . . •.:• : •�
. ··� . B

••• •�n 25 .••

• • a.:•�
w is no good either: 27 'i!i'c7t:hb4 28
24 l:.ed1 t:hb4
This is the essential difference.
If on my 2 1 st move I had played
l:.fel, the white rooks would now
be on bl and d l. Consequently,
Black would only have two reason­
able possibilities: 24. . . lhf2+ 25
J.xf2 'i!i'xf2 26 e5 J.e7 27 l:.fl , and
26 'ii'a7
l:.ab8
Wc6
The decisive mistake. 26 ...t:hc6
'ti'xc2 lhxc2 29 J.b6 with advan­
tage to White. But 26 ...'ii'e2 would
have given Black definite counter­
play, although here too after 27
J.b6 White has an advantage.
27 J.f4
Later I discovered that in the
press centre Furman was at this
28 Moscow Ct (6) 1974

very moment proving that an ex­ Otherwise e4-e5 with a decisive

change sacrifice leads to victory attack. However, the attack has not
for White: 27 llxf6 ! gxf6 28 .i.h6 been stopped.
(threatening 29 'ii'e 3) 28 . . . 'iic2 (or
28 . .. tbd3 29 .i.xf8) 29 .U.c l 'ii'd3 30 �W% - •••
1t'c5 ! . But when I played 27 .i.f4 • ta'i - -
(threatening 28 e5) my trainer con­ • ,� /,ffl i . i
tentedly commented, 'and that is •• m • •
good as well' . In fact, the threat of
e4-e5 contains Black's activity,
"
iQ)i ·
,W£i �}A
�,«, -
R
while the b4-knight finds itself out ···
,..."- r� � '·'
- �;� . 8
• • ��1W illti
of the game. . . ����
i{�:y,, � u
.\Ulr. · "
27 .:.as
--"�
28 1ifl .U.ad8 (D)
. . • :•�
w
32 .i.h6
The fire is now directed at two
points at once, f7 and g7.
32 .•• tbc6
In the hope of complicating the
game after 33 tbxf7li'c4 ! [editor's
note: but then 34 .i.xg7 .i.xg7 (af­
ter 34 . . .'Wxfl + 35 .U.xfl .i.xg7 36
tbh6+ �h8 37 .U.xf8+ .U.xf8 38 tbf5
White has a decisive material ad­
w vantage) 35 1ixg7+ rJilxg7 36 .U.g3#
29 'Wg3! is mate].
The threat is stronger than its 33 tiJfS! (D)
execution ! It is a paradoxical prin­ 33 ••• 'ii'b2
ciple, but it is valid surprisingly The game would have had a cu­
often. White does not hurry with rious finish in the event of 33 ....i.e5
the advance e4-e5, and increases 34 .i.xg7 .i.xg3 35 .U.xg3, when
the pressure. Black cannot prevent mate.
29 'ii'c3 34 .i.cl!
30 .U.f3 'fi'c2 White frees h6 for the knight
31 .U.dfi .i.d4 (D) with tempo. There is no defence.
 Karpov - Spassky 29

B B
34 ti'b5 The so-called Modern Schev­
35 lbh6+ Wh8 eningen is seen these days only
36 lbxf7+ l:.xf7 rarely, and this match with Spassky
Or 36 ...�g8 37 lbh6+ �h8 38 did not play the final role in this. In
.:.Xf8+ and mate. my duels with Kasparov in 1 984
37 l:.xf7 ..tf6 and 1 985, where the Sicilian De­
38 ti'f2 Wg8 fence arose regularly, my opponent
39 l:.xf6 gxf6 preferred the Classical Scheven­
40 ti'xf6 1-0 ingen with a more flexible move­
order - first ... a7-a6 and ...ti'd8-c7,
Game 5 then ... lbb8-c6 and ...l:.f8-e8, and
Karpov Spassky
- only then .....tc8-d7. Premature de­
Leningrad Ct (9) 1974 velopment of this bishop is playing
Sicilian, Scheveningen into White's hands. The alterna­
tives are 9 ...ti'c7 or 9 ...e5.
1 e4 c5 10 lbb3
2 lbf3 e6 Black was preparing to ex­
3 d4 cxd4 change knights in the centre and
4 lbxd4 lDf6 advance the bishop with .....tc6.
5 lbc3 d6 Thus this knight retreat is quite ac­
6 ..te2 ..te7 ceptable and more precise than the
7 0-0 0-0 other well-known moves 10 �h1
8 f4 lbc6 and 10 'fi'el .
9 ..te3 (D) With the bishop on d7, Black's
9 ..td7 counterplay on the queenside falls
30 Leningrad Ct (9) 1974

behind in comparison with White's White a clear positional plus in any

threatening pawn storm on the case.
kingside, all the more so as the cen­ 12 .tf3 .tc6 (D)
tre is fully under his control. Returning the knight to d4 is not
10••• aS?! (D) really to Black's liking, but 1 2 ... e5,
Although White also has the with the passive bishop on d7, is
initiative after the waiting moves not very attractive either.
1 0 ... a6 or 10 ...'ifc7, playing like
this is even less dangerous. The a­
pawn's march is superficially ac­
tive, but in strategic terms it is
dubious. White gains the excellent
b5-square for the knight, and hin­
ders freeing manoeuvres by the
black pieces. This game may have
been played twenty years ago, but
it is still a good model of how
White should play in situations like
this.
w
13 t2Jd4 g6
Or 1 3 . . . 'ii'b 8 14 'ili'e2 l:l.e8 1 5
l:l.ad 1 '1th8 1 6 lbdb5 l:l.a6 1 7 l:l.d2
l:l.d8 18 l:l.fd 1 d5 19 e5 lbd7 20
'iif2 ! and White had a clear advan­
tage in Klovans-Vasiukov, Mos­
cow 1968.
14 l:l.f2 e5
15 tbxc6 bxc6
16 fxe5 dxe5 (D)
17 'ii'f l!
w The battle is flaring up around
1 1 a4! lbb4 the c4-square, which White is
It is more precise to play 1 l ...e5 counting on occupying with one of
1 2 'ith1 and only then 1 2 ... lbb4, al­ his pieces. If Black managed to
though 1 3 .tb5 ! .tc6 14 .txc6 prevent this, things would not look
bxc6 15 fxe5 dxe5 16 'ii'e2 gives at all bleak for him. It makes no

w 21 'i'c4 (D)
sense to move the rook from the f­
file ( 17 l:td2), as it is still unclear
which file will be the centre of ac­
tivity.
17 ••• 'ii'c8
18 h3 (D)
Of course, there was no point al­
lowing an exchange of knight for
bishop ( 1 8 ... lLlg4).
8 I was not tempted by the chance
18 ••• lbd7
In the event of 1 8 . . . 'Mi'e6, White
could have developed the initiative
Karpov - Spassky 31
with 19 l:tcl (or 19 g4 followed by
g5 and .i.g4) 1 9 . . . l:tfd8 20 .i.e2
l:td4 21 b3 (in order to establish the
bishop on c4).
19 .i.g4 h5
This severely weakens the posi­
tion of the black king. He should
have chosen the simple 19 ...'i'c7,
escaping from the pin and connect­
ing the rooks along the back rank.
20 .txd7 'i'xd7
So, the queen has arrived at her
appointed destination.
8
21 ••• .th4
An unpleasant endgame awaits
Black after 2l ...'i'e6 22 1fxe6 fxe6
23 l:taf l .
22 l:td2 'i'e7
23 l:tO!
of winning the exchange; after 23
.i.c5 'ii'g 5 24 l:td7 lLlxc2 25 i..xf8
l:txf8 Black has active counterplay.
32 Leningrad Ct (9) 1974
The d-file does not matter; the
d8-square is securely covered.
Therefore I have to generate ac­
tivity in other directions.
23 l:l.fd8 (D)
w
24 llJbl!
The craftiness of this move lies
in the fact that I have managed to
choose the most suitable moment
to move the knight into a more ac­
tive position (this can happen par­
ticularly quickly if Black swaps
rooks). It is amusing that conse­
quently this unexpected knight
retreat has become an almost char­
acteristic illustration of my work.
24 ••. 'iiib7
25 'itth 2!
It is a rare occurrence when, in
the middlegame, the mobility of
your king limits the activity of an
enemy bishop.
25 �g7
26 c3
The time has come to invite the
black knight to move away from its
familiar place.
26 llJa6
27 l:l.e2! (D)
White does not intend to ex­
change rooks yet; the major pieces
are needed for an attack on the f­
file. This also frees a square for the
knight transfer llJd2-f3-g5, and
moreover threatens 28 g3 �f6 29
l:l.ef2 l:l.d6 30 �g5.
27 .:r.rs
28 llJd2 �d8
29 llJf3 f6
In defending the e5-pawn, Black
is simultaneously trying to cover
the f-file. But White's attack is al­
ready unstoppable.
30 l:l.d2!
The white rook's 'hesitation'
might seem illogical. First he occu­
pied the d-file, then abandoned it,
and now returns in decisive fashion
 Karpov - Korchnoi 33

(that is what is important!) onto the . -

d - �
- d .
open file.
30 �e7 (D)
.\Ull • d
dwd • �•
••••••••
.•.

After 30 ... o!Db8 3 l.!Dg5 ! I would

have won straight away. - -. -� ••
- -
�· ·�- .
. �
d � ��-�
-�d
."•.:•
u .
d �-"��
� �
.
B
immediate threats, but Spassky's
position is worsening with every
move. The problem is that the
black pieces are disconnected and
cannot come to each other's help.
w Now, for example, it is impossible
When Spassky made his move, to defend the seventh rank by
I lost my head a little. At first I means of 33 . l:tt7 due to the hang­
..

thought that I was winning in all ing bishop on d8.

variations, but then I looked and 33 o!Db8
could not find a win. Fortunately, 34 �cS llh8
my confusion only lasted for a mo­ 35 l:txd8! 1·0
ment. After 35 . . .l:txd8, 36 �e7 ! is de­
31 W'e6 l:tad8 cisive.
This loses by force. Dragging
out resistance is only possible by Game 6
means of 3 1 ....!Db8, as if setting up Karpov - Korchnoi
the pieces for a new game. Moscow Ct (2) 1974
32 .:.Xd8 �xd8 Sicilian, Dragon
If 32...l:txd8, then 33.!Dxe5 'ilc7
34 'ilt7+ �h8 and 35 'ilxe7 'ilxe5+ 1 e4 cS
36 'ilxe5 fxe5 37 l:tf6. 2 o!Df3 d6
33 lld1 (D) 3 d4 cxd4
Material is level on the board. 4 o!Dxd4 .!Df6
Black's king has avoided White's s o!Dc3 g6
34 Moscow Ct (2) 1974
In my long chess career the
Dragon Variation has been used
against me about 20 times, and I
think only once has my opponent
managed to gain a draw. But al­
though many years have passed, to
this day this encounter with Korch­
noi remains one of the clearest ex­
amples in the whole history of the
variation. Possibly the level at
which his game was played had a
part in that.
6 .te3 .tg7
7 f3 lh c6
8 'ifd2 0-0
9 .tc4 .td7 (D)
w
10 h4 .:cs
In Game 1 examined above, the
then fashionable line .. .'i!i'a5 and
....:tfc8 was played. Transferring
the f8- rook to c8 was later totally
displaced by putting the queen's
rook there. Besides, in recent years
Black has tended to meet h4 with
the symmetrical . . .h5 (now or on
the next move). The basic develop­
ment of contemporary theory in
this variation has also gone in pre­
cisely that direction.
1 1 .tb3 lheS
12 0-0-0
The immediate 1 2 h5 is also
possible.
12 ... lhc4
This was Black's last chance of
averting the dangerous opening of
the h-file by means of 12 . . .h5.
13 .txc4 .:txc4
14 hS
14 g4 also leads to a sharp game.
14 lhxhS
15 g4 lh f6 (D)
w
A very topical position in those
days. Here White has many possi­
bilities: 1 6 .:tdg1, 16 lhd5, 16 .th6,
or 16 e5. However, I chose the
modest knight retreat to e2, as I had
in mind one new idea which I had

prepared specially for this match. I
should note that a detailed investi­
gation of all the finesses of even
one of the above-mentioned lines
of the Dragon would take many,
many pages.
16 lbde2! ?
The idea o f this move i s under­
standable - to reinforce the c3-
square, as exchange sacrifices such
as . . Jhc3 often let B lack gain a
powerful attack; indeed the Dragon
Variation is underpinned by this
blow. However, from e2 the knight,
when the occasion arises, can eas­
ily be transferred over for a direct
attack against the enemy king.
16 ... 'iia5
1 6 . . . lte8 seems safer for Black,
as 1 7 �h6 ( 1 7 eS? is no good be­
cause of 17 . . .lbxg4 ! 18 fxg4 �xg4
19 'ii'd3 't!l'c8 with advantage for
Black; Bemai-A.Schneider, Hun­
gary 1 976) is met by the retreat
17 ...�h8. The game Klovans-Beli­
avsky, USSR 1 977 continued 1 8 eS
lbxg4 1 9 fxg4 �xeS 20 �f4 'ifaS
2 1 �xeS 't!l'xe5 22 lbd5 ltxg4 with
roughly even chances.
17 �h6 �xh6
Here as well it was possible to
play 1 7 . . . �h8, in the spirit of GM
Simagin, sacrificing the exchange
but preserving Black's long-range
bishop, which proves useful for
Black in both attack and defence .
On the other hand, 17 . . . ltfc8 leads
Karpov - Korchnoi 35
merely to a transposition of moves
after the continuation 1 8 �xg7
rj;xg7 19 'ii'h6+ rj;gS.
18 'iixh6 ltfc8
19 ltdJ! (D)
B
Only this move at that time was
a novelty - by overprotecting c3,
White frees the e2-knight to par­
ticipate in the attack. The theoreti­
cal paths 19 ltdS or 1 9 g5 bring
White no great advantage, and fur­
thermore Korchnoi was certainly
ready for them. But after the mod­
est rook transfer he plunged deep
into thought.
19
... l14c5
This move apparently leads to a
forced win for White. However, af­
ter 1 9 ... �e6 20 gS lbh5 2 1 lbg3
'it'eS 22 lbxh5 gxh5 23 'ii'xhS as
well, he has an appreciable ad­
vantage. Botvinnik's continuation
19 ...'ili'd8 was the most stubborn.
20 g5
36 Moscow Ct (2) 1 974

The knights on c3 and f6 are

both defending their kings, and - • • •••
- .. �� - .. �� ..
therefore it is precisely these � .. � �
.. � ..
pieces which are exposed to the •.t�
. - -·- �
greatest danger (if the black knight illit �� )";-. • �;l
_m
moved from f6, it would immedi­
- .e.z.J. •
ately invite a white invasion on the
• ·
• -0 A�� - !!!f}�
.?:.'.
square d5). - .� ·· � ·
20 ••• llxgS (D) " � �
�C j� �
w��� •
•
• m
� •
• .
lW� �
....

w 24 eS!
21 lidS!
Not, of course, 21 lt:Jd5 llxd5 and
the main guardian of Black's for­
tress, the f6-knight, remains alive.
21 ... llxdS
22 lt:JxdS lle8
Now the queen can no longer
manage to return to her camp:
22 . . .'ifd8 23 lt:Jef4 'ii'f8 24 lt:Jxf6+
exf6 25 'fi'xh7#; if 22 ... lbh5, then
White wins by 23 lbxe7+ �h8 24
lt:Jxc8.
23 lt:Jef4 ..tc6 (D)
It is necessary to control the d5-
square, as otherwise lt:Jxf6+ and
w
lt:Jd5, mating. After the alternative
23 . . . ..te6, I had prepared 24 lt:Jxe6
fxe6 25 lt:Jxf6+ exf6 26 ifxh7+
r.fi>f8 27 ifxb7 'fi'g5+ 28 �bl lle7
29 'ii'b8+ lle8 30 'ii'xa7 (but not 30
llh8+?? <l;g7 and Black even wins,
due to the threat of 3 1 . . .'ii'g l#)
30 . . .lle7 3 1 'ili'b8+ lle8 32 'ili'xd6+
with a form of 'windmill' .
Cutting off everything on the
fifth rank. I was almost dazzled by
the wealth of apparently effective
possibilities, but only this continu­
ation appears to be decisive. The
straightforward 24 lt:Jxf6+ exf6 25
lt:Jh5 'ili'g5+ (the point !) 26 'ifxg5
fxg5 27 lt:Jf6+ �g7 28 lt:Jxe8+
..txe8 would not have won.
24 ... ..txd5
After 24 ... dxe5 25 lt:Jxf6+ exf6
26 lt:Jh5 mate is inevitable.
25 exf6 exf6
The main thing is to be strong to
the finish ! It is still not too late to

lose: if 26 lLlh5 (to meet 26 . . . gxh5
by 27 l:tg1 + and 28 "ilg7#) then the
sobering 26 ...l:te1+.
26 "ii'xh7+ <iitf8
27 "ilh8+
After 27 .. .rt;e7, 28 lLlxd5+ ii'xd5
29 l:te1+ would be decisive.
1-0
In the second half of 1 974 this
game won the best game prize in
lnformator, with 89 marks out of
90 ! In the almost thirty-year his­
tory of this contest, never have the
grandmaster jury been so unani­
mous!
Game 7
Karpov - Vaganian
Skopje 1976
French, Tarrasch
1 e4 e6
2 d4 dS
3 lLld2 cS
4 exdS exdS
5 iLlgf3 a6
In those days I almost always
chose the variation with 3 lLld2 in
this opening. And my opponent,
for whom the French Defence was
then and is to this day a fundamen­
tal weapon for Black in reply to 1 e4,
had prepared a quite rare continu­
ation. The aim of it is to prevent the
bishop appearing on b5. Besides
that, when the occasion arises he is
prepared to play . . . c5-c4, and its
Karpov - Vaganian 37
support, . . . b7-b5, has already been
prepared. But this rook's pawn
move has one essential defect - it
does little to assist piece develop­
ment.
6 dxcS
Apparently the simplest solu­
tion to the problem. However, 6
.te2 also gives White reasonable
prospects.
6 •.. .txcS
7 lLlb3 .tb6 (D)
The bishop has many possible
retreats - besides this, it could also
go to e7, d6 and a7. It is difficult to
say which one of these is the safest.
w
8 .td3
8 .tg5 gives White nothing:
8 ... lLle7 9 'ii'd2 lLlbc6 10 .te2 0-0
1 1 0-0 h6 12 .te3 l:te8 1 3 l:tad1
lLlf5 ! 14 .txb6 'ii'xb6 15 l:tfe 1
.te6.
8 lLle7
9 0-0 lLlbc6
38 Skopje 1976
10 :et J. g4
Both sides are making their own
moves, seemingly with little regard
for the other. The battle is being
fought around the critical d4-
square. If Black manages to ad­
vance his central pawn to d4, he
will have a reasonable game. If, on
the other hand, White can guaran­
tee control of this important square
for his pieces, his advantage from
the opening will not be in doubt.
In connection with this, 1 1 J.e3
would now be a mistake because of
1 l . ..d4.
1 1 c3 (D)
Immediately pushing back the
enemy light-squared bishop by 1 1
h3 makes no sense, as Black can­
not castle at the moment in view of
the sacrifice on h7.
B game takes on a forcing character.
11 h6
12 h3 J.hS
13 J.e3 0-0
Exchanging dark-squared bish­
ops is in principle not to Black's
advantage, although, for example,
in the event of 1 3 . . . J.c7 he is even
further behind in development.
14 J.xb6 'ifxb6
15 'ife2
This does not look very logical ­
the queen is standing in front of the
rook on an open file. Nevertheless,
Black does not have time to make
use of this fact: after 1 5 . . .l:tfe8 his
rook is insufficiently defended.
15 l:tfd8
16 l:tad1 aS (D)
w
Vaganian is trying to become ac­
tive on the queenside by forcing
the white knight from b3. He prob­
ably underestimated the ensuing
pawn sacrifice, after which the
17 J.b1
It may seem superfluous to re­
peat that White's main idea is to

control d4, but he has to watch it
constantly. Therefore the move
'it'e3, which enters into White's
plans, would in this specific case
be unsuccessful: 1 7 'it'e3 ti'xe3 1 8
l:.xe3 i.xf3 1 9 l:.xf3 a4, and Black
has everything under control.
17 i.xf3
18 'it'xf3 a4
19 lbd4 'it'x b2
Not 19 ...lbxd4 20 l:.xd4 'it'xb2,
as after 2 1 l:.b4 Black loses a piece
(the rook not only defends the
bishop, but also blocks the a3-f8
diagonal).
20 lbxc6
Black's final piece is diverted
away from defending its king.
20 lbxc6 (D)
w 24. . .f4 (24 . . . lbe7 25 l:.xe7) 25
21 'ii'fS g6
22 'it'f6
The white pieces are very har­
moniously placed (in particular,
the queen is not only defending the
Karpov - Vaganian 39
c3-pawn, and keeping the major
black pieces out of the game, but is
also creating the threat of a bishop
sacrifice on g6, with a mating at­
tack).
22 ... l:.d7
After 22...l:.e8 I was planning to
continue 23 i.xg6 fxg6 24 ti'xg6+
'it>f8 25 ti'xh6+ 'it>g8 26 l:.e6.
23 i.fS! (D)
The point!
B
23..• l:.e7
After 23 . . .l:.c7 White would
continue with the simple 24 l:.xd5;
23 ... gxf5 loses due to 24 l:.d3 (it
must be precisely this rook, cover­
ing the b1 -h7 diagonal, as the other
white rook has to control the e-file)
'ii'xf4 ti'c2 (White was threatening
26 'ilfg4+, but 25 . . .f6 is probably
more stubborn) 26 l:.g3+ 'it>h7 27
'ilff6 l:.g8 28 l:.xg8.
24 l:.xe7
40 Skopje 1976

I could also have gained an ad­ 28 'it'xh6

vantage by means of 24 .i.xg6 fxg6 The immediate 28 lle3 sug­
25 llxe7 li:Jxe7 26 'iixe7, but I gested itself, but then Black would
wanted more. have had an interesting way of
24 ... li:Jxe7 gaining a draw due to his cunning
25 .i.d3 li:JfS trap: 28 .. .f4 29 'iixf4 'iVbl + 30 �h2
The only move, as there are no a3 ! 3 1 llg3+ 'tig6 32 .:txg6+ fxg6,
other possibilities for the knight and here White has nothing better
(25 . . . li:Jc6 26 .txg6); 25 ....:te8 26 than a perpetual check, as his
.:tel 'iia 3 27 .i.b5 is no good either. queen cannot even approach the
Defending the knight with the king black rooks.
is also unsuccessful: 25 ... �f8 26 28 ... a3
llbl 'iixa2 27 .:txb7, and now both 29 'figS+
27 ... lle8 and 27 ...1\Va3 should be The queen is heading for f6 with
met by 28 .txg6. tempo.
26 .i.xfS gxfS (D) 29 'i1i>f8
30 'iif6 'it>g8
31 'iixfS 'ifd2 (D)

27 .:tel!
The black rook must not be al­ w
lowed onto his third rank, a s a fter 32 .:te7!
....:ta6-g6 he could gain reasonable A final finesse. The black rook
counterchances linked with the is now diverted from the a-file.
dangerous passed a-pawn, by giv­ 32 .:tf8
ing up some pawns. 33 'ifg4+ 'i1i>h7
27 ... 'iixa2 34 .:teS 'li'h6
 Tatai - Karpov 41

35 l:th5 :as sufficient compensation for the

36 'ii'f5+ <j;g7 pawn, as in the game Kaiszauri­
37 l:txh6 <j;xh6 T.Georgadze, Sukhumi 1977.
38 'iff6+ <j;b7 8 ... e6
39 'ii'xf7+ <j;bS 9 ltJge4 ltJb6! (D)
40 Wxb7 1-0 It is ridiculous to defend the
pawn with the bishop - 9 . . i..f8,
.

Game S while 9 'ii'e7 loses straight away

...

Tatai Karpov
- in view of 10 ltJxdS exd5 1 1 ltJc3.
Las Palmas 1 977
English Opening

1 ltJf3 c5
2 c4 ltJf6
3 ltJc3 d5
4 cxd5 ltJxd5
5 g3 g6
6 ..tg2 ..tg7
The so-called Modem Variation
of the English Opening. The stand­
ard continuation here is 7 0-0, but. ..

7 ii'a4+
An entertaining idea. White
strives to make immediate use of
his small advantage in develop­
ment and the lack of harmony
among the black pieces.
7 ••• ltJc6
After 7 .....td7 8 Wc4 ltJb4 9 0-0
ltJ8c6 10 'ii'xc5 it is not easy for
Black to prove that he has compen­
sation for the pawn.
8 ltJg5
In the event of 8 Wc4 ltJdb4 ! 9
0-0 'ilfaS ! 1 0 ltJe4 ii'a6 ! 1 1 'ilfxc5
b6 1 2 'ile3 0-0 1 3 ltJe1 ..te6 14
ltJc3 l:tac8 B lack has more than
w
10 'ii'b5 c4
1 1 ltJa4
A logical continuation of the
plan which was begun on move 7 .
1 1 'iWcS is repulsed by means of
1 l . . . ..tf8, while 1 1 ltJcS 0-0 1 2
..txc6 bxc6 1 3 'ii'xc6 e5 would give
Black a dangerous initiative.
11 0-0
12 ltJxb6 axb6
13 ii'xc4 (D)
The critical position. White has
got his way, having won the c­
pawn. Black meanwhile has man­
aged to remove his king from the
42 Las Palmas 1 977
B fence 16 e3 ! lbb3 17 l:[bl lbxcl 1 8
centre and is ready to get down to
action. Incidentally, the fact that
the a-file has been opened is un­
doubtedly to his advantage (the
weakness of the doubled pawns
will only tell in the endgame, but
that is still a long way off). Be­
sides, the white queen has made an
early exit from the centre, and now
one of Black's problems is how to
use this fact wisely, and this, in
conjunction with other threats,
should guarantee him a definite ad­
vantage.
13
... e5!?
1 3 . . . .i.d7 also looked tempting,
in order to move the rook to the c­
file as quickly as possible. In this
case White has to be on the look­
out for danger with every step.
However, he does have an interest­
ing resource:
a) Not 14 'ii'c 2? lbd4 15 'ii'b 1
.i.a4 16 b3 .i.xb3 17 axb3 l:txa1 1 8
'ii'x a1 lbc2+.
b) 14 0-0 lbd4 1 5 lbc3 (the only
defence against 1 5 ....i.b5) 15 . . . b5
1 6 'ii'd3 b4 17 e3 lbb3 is no use to
White either.
c) White can try to take the d4-
square away from the knight by
means of 14 e3, but then 1 4...e5 is
better for Black.
d) 14 lbc3 ! lbd4 15 'ii'd3 ! .i.c6
( 1 5 . . .b5 looks promising, but here
as well there is a satisfactory de­
:xc 1 b4 19 lbe2 lha2 20 'ii'b3 fol­
lowed by 2 1 d4) 1 6 .i.xc6 ( 1 6 0-0 is
bad due to 16 . . .lbb3 ! 17 'ii'xd8
l:[fxd8 1 8 l:[b1 lbxd2 19 .i.xd2
:xd2 with an advantage for Black)
1 6 . . . bxc6 1 7 e3 ! - the point! -
17 . . . lbb3 (White has removed the
important knight, without drop­
ping his the queen: 1 7 . . . lbf3+ 1 8
�e2), and 1 8 'ii'xd8 l:[fxd8 19 l:[b1
lbxc 1 20 :xc 1 .i.xc3 2 1 bxc3
:xa2 22 d4 gives rise to an equal
ending.
It is not surprising that it took
me an hour to decide between
1 3 . . ..i.d7 and the move played in
the game.
14 'ii'c2
The only move. 14 0-0 is bad
due to 14....i.e6, when the queen
has no good square. 14 lbc3 is also
weak in view of 14 ... .i.e6 15 'ii'e4
( 1 5 .i.d5 is answered by 15 ... b5 !)
1 5 ....i.f5 followed by 16...lbd4 .
14 ... lbd4
 Tatai - Karpov 43

15 ii'b1 fS 21 ••• 'it'xd4 (D)

16 ltlc3 e4 (D)
1 6 . . .b5 allows White the ade­
quate response 17 e3, but perhaps
I should have paid more attention
to the preparatory 16 ... -te6, threat­
ening 17 .. .-txb3.

w
22 a3
22 0-0 allows Black a pleasant
choice between the two possibili­
tie s 22 ... -tg4 and 22 . . . ii'd2. The
latter is more promising, as White
w is obliged to go for the variation 23
17 d3 -txe4 -th3 24 -tg2 -txg2 25 <iPxg2
1 7 e3 gives White a hopele ssly .:.xe2 26 'ilfc1 'ii'd5+ 27 'ittg 1 -td4
weak position after either 17 ... ltlc6 with good attacking prospects for
followed by 1 8 . . . ltle5 or 1 8 ... ltlb4, Black.
or 17 ... ltlf3+ 1 8 -txf3 exf3. 22 ••• -tg4
17 ••• bS Black guesses his opponent's
18 -te3 crafty little idea: 22 ... bxa3 23 0-0,
. Again the only move. 18 e3 ltlf3+ making use of the undefended po­
1 9 -txf3 exf3 20 ltlxb5 �a5+ 2 1 sition of the aS-rook.
ltlc3 b5 serves no purpose at all. 23 �c2 (D)
18 b4 Ruling out the threat 23 . . .-tf3,
19 ltld 1 .:.es because of 24 -txf3 exf3 25 �b3+
20 dxe4 fxe4 and 26 �xf3 .
21 -txd4 23 ••• 'it'd3!
But not 21 -txe4? in view of the It is rare that one can place the
reply 2 1 . . . .:.xe4 22 'it'xe4 -tf5 and queen under attack from a pawn .
... ltlc2+. 24 exd3
44 Las Palmas 1977

�c4 l:lxc2+ 29 �xd5 followed by

29 ....i.f3+.
27 l::txc2+
28 �xb4 l:lcd2 (D)

- - -·­
• •• • • •
- - • ••
B B B B
B �
� .
� . ...
. ... .
.
Allowing the attack to develop �
�f . �g� .- ;,- -� ��
�
��-� --- D�.i.r,;� ;;;_
,, �

elegantly. 24 lbe3 would have lost ��

/, �
0
,
�
' /
r:' !;;;: "
�

immediately to 24...'ilfxc2 25 lbxc2

.i.xb2. 24 .:le 1 would not have a •ttJ• a :
brought White any relief due to the w
simple 24 ...bxa3, but even the su­ Again White's knight is the key
perior 24 'ii'd2 would have left piece ! It contains the action of all
B lack a number of promising pos­ the black pieces, and must be at­
sibilities, e.g. 24 ... 'ii'xd2+ 25 �xd2 tacked immediately.
.:ladS+ 26 �e 1 , and now either 29 f3
26 . . J::tc 8 threatening 27 ...l:lc2, or Forced, as otherwise the b2-
the immediate 26 ... .i.f3 27 l:lgl pawn will be lost.
(both 27 exf3 exf3+ 28 �fl fxg2+ 29 ••• .i.f8+
29 �xg2 b3, and 27 .i.xf3 exf3 28 30 �aS
e3 lead to a difficult endgame for After a retreat to any other
White) 27 ....i.xg2 28 l::txg2. square the g4-bishop jumps aside
24 ••• exd3+ with check and the g2-bishop is
25 �d2 l:le2+ lost.
It is possible that White was not 30 ••• .i.d7!
expecting this check when he was The white bishop is under at­
considering his own 24th move. tack, and if 3 1 .i.fl (or 3 1 lbe3),
26 �xd3 l:ld8+ then 3 1 ... .i.c5, and there is no way
27 �c4 of saving White from mate by the
Black emerges with an extra rook.
piece after 27 .i.d5+ l:lxd5+ 28 0-1
 Karpov - Korchnoi 45

Garne 9
Karpov - Korchnoi
Baguio City Wch (14) 1978
Spanish, Open Variation

1 e4 e5
2 .!iJf3 .!lJc6
3 .i.b5 a6
4 .i.a4 .!iJf6
5 0-0 .!lJxe4
The Open Variation of the Span­
ish, one of the most popular open­ w
ings in both my matches with I thought of an interesting idea
Korchnoi for the world chess crown. linked with transferring into the
One could almost write a whole sort of endgame where opposite­
book about the games we played coloured bishops do not diminish,
with this opening. In this encoun­ but on the contrary even increase,
ter, for the first time in our theoreti­ White's advantage.
cal dispute I managed to use an 15 .i.xe4 dxe4
extremely valuable novelty in this 16 .!lJxc5 exf3
variation. 17 .i.f4! (D)
6 d4 b5
7 .i.b3 d5
8 dxeS .i.e6
9 c3 .i.cS
10 .!iJbd2 0-0
1 1 .i.c2 .i.f5
12 .!iJb3 .i.g4
1 2 . .i.g6 has the same value.
. .

13 h3 .i.h5
14 g4! ? .i.g6 (D)
In opening manuals at the time
this position was said to be totally
in Black's favour. In fact, if you are B
referring to the rniddlegame, the If White exchanges queens im­
open position of the white king is mediately by 17 Wxd8, then he
the basis for this assessment. But loses a tempo and Black has time
46 Baguio City Wch (14) 1978

to organise counterplay against the 19 .:.d7 .!Lle6

e5-pawn. The point of my play lies 20 .!Llxe6 fxe6
in the fact that after the move 21 .i.e3 .:.ac8
played Black is forced to exchange Exchanging one pair of rooks
queens, which gives White two ex­ (2 l ... .:.f7) only weakens the possi­
tra tempi. bilities for defence. After 22 .:.fd 1
17 ••• 'ii'x d1 White has enough squares to main­
17 ...'Wh4 is insufficient in view tain the rook on the seventh rank.
of 1 8 'ti'xf3, defending the h3- 22 .:.rd1 (DJ
pawn and attacking the c6-knight. 22 .i.cS immediately deserved
After 17 ..."ile7 White wins a pawn: attention, so that after 22 . . . .:.fe8
1 8 'ii'd5 .!Lla5 1 9 b4 .!Llc4 20 'ii'xf3, (if 22....:.fd8 { or 22 . . ..:.f7 } . then 23
and 20 . . ..!Llxe5 is impossible due to .:.fd 1 ; 22 ....:.f4 leads to unclear
2 1 .i.xe5 'ti'xeS 22 .!Lld7. consequences) White can play 23
18 .:.axd1 .!Lld8! (D) .:et .
Korchnoi finds the best response
to White's unexpected move. Ob­
viously, Black cannot tolerate the • • • • ••
beautifully placed white knight on . �
�1 -;;; : •,� � -.
cS. Exchanging off into a rook end­ •• . . .... .
ing with opposite-coloured bish­
ops promises him some chances
•
�- - . •
0%f; �"QJ� -
��
-�

for defence. Naturally, I had fore­ • E ·�·

seen this possibility. .
- "
�;,: � ·�
- -��
��� •
:«�� ;r�"' � •
� . 'I'll' . m1
p
• . t=. . �
B
22 ... .i.e4
Moving the bishop to d5 in an
attempt to disturb the co-ordina­
tion of the white rooks is Black's
only realistic defensive possibil­
ity. If 22 . . .h5, then 23 gxhS .i.xh5
24 .:.e7, invading with the second
rook.
w 23 .i.cS .:res

Here is the whole point of
B lack's play; he does not put his
rook on either f4 or f7, offering to
exchange, but takes the e7 -square
under his control, and the white
rook becomes short of space on the
seventh rank.
24 ll7d4 .i.dS
25 b3 aS
Again the only possible attempt
to create counterplay, as otherwise
nothing will oppose White's gen­
eral advance on the kingside.
26 �h2 lla8
If 26 ... a4, then 27 c4 bxc4 28
bxc4 .i.c6 29 a3 .
27 �g3 lla6
27 ... a4 would have been more
logical, and in reply to 28 c4 both
28 ...bxc4 29 bxc4 .i.c6 (now White
does not have time for 30 a3 in
view of 30 ... lla5 when the e5-
pawn perishes), and 28 ... .i.c6 are
possible, in order to exchange a
few more pawns, although in this
case as well White has a clear posi­
tional advantage.
Korchnoi 's move is part of an in­
correct plan.
28 h4 llc6 (D)
Continuing the same mistaken
plan. Korchnoi clearly underesti­
mated his opponent's idea, which
is described after the next move.
29 llxdS!
White immediately gains a
bishop and pawn for the rook, and
Karpov - Korchnoi 47
besides, the beleaguered black
f3-pawn is already under attack by
the king, and the queenside pawns
are also becoming objects for at­
tack.
29 exdS
30 llxdS llce6
31 .i.d4 c6
32 lieS (D)
Korchnoi may have underesti­
mated the power of this move. If he
had played 32 lld7, Black would
have had the very strong reply
32 . . . c5 at his disposal, exchanging
off one of White's fundamental
trump cards - the e5-pawn. Now
the white rook is occupying a very
powerful position. White wants to
isolate the queenside pawns by
means of a2-a4, and the f3-pawn is
already under attack.
32
...
llf8
In the game it would have been
hard to decide on the continuation
32 ...lld8 33 �xf3 lld5 34 .:.Xd5 (or
48 Baguio City Wch (14) 1978

his last move gives him no hopes of

saving himself.
33 a4!
In the first place this fixes the
objects for attack on the queenside,
in order to chain the black rooks
down to their defence and divert at­
tention from operations on the op­
posite side of the board.
33 bxa4
34 bxa4 g6
B 35 .:xa5 .:ee8
34 �f4 lhc5 35 .txc5) 34 . . . cxd5. In the event of 35 . . . h5 the move
There have been many arguments 36 .:a6 ties Black down to defend­
about this position. I suggest that ing the c6-pawn and prepares the
the mobility of the white king side advance of the a-pawn.
pawns should in the final analysis 36 .:a7 .:r7
lead him to victory, although it is 37 .:a6
not as simple as it appears at first A cunning little move, forcing
glance. In particular, after 35 �e3 the black rook to occupy a passive
35 ....:h6 36 h5 g6 the tempting 37 position.
�f4 does not work in view of 37 ..• .:c7
37 ...gxh5 38 �g5 h4 ! . 37 ...c5 does not work because of
35 a3 was my intended continu­ 38 .txc5 .:xe5 39 .:a8+ �g7 40
ation, in order to create a fortress .td4.
on the queenside with b3-b4, and 38 .tc5 (D)
only afterwards advance the king- After the bishop has reached d6,
side pawns, for example 35 ...g6 36 the position should be considered
�g3 .:e8 37 f4 and if 37 ....:f8, won.
then 38 e6 with the plan of h5, and 38 .:cc8
if ... gxh5, then f5, creating two 39 .td6 .:as
connected passed pawns. 40 .:xc6 .:Xa4
Nevertheless, trying to open up 41 �xf3 h5
the game on the kingside was 42 gxhS gxh5
Black's only chance of survival, 43 c4 .:a2
since the passive defence to which 44 .:b6 �f7
Korchnoi condemned himself on 45 c5 .:a4
 Karpov - Korchnoi 49

by the fact that he had nothing to

lose, my opponent equalled the
score at 5-5. In four games I lost
everything I had gained in the pre­
vious 27 . I was shattered. All the
same I managed to cast off the bur­
den of painful blunders and made
up my mind that the 32nd game
was to be the decisive one. I played
quietly, confidently, and, having
gained an overwhelming position,
instilled myself with the thought,
46 c6 �e6 'just do not hurry ! ' , as I knew that
47 c7 �d7 the desire to realize my advantage
48 l:b8 l:c8 had already let me down more than
49 �e3 once.
Hoping for 49 e6+? �xe6 50 1 e4 d6
l:xc8 �xd6 and the possibility of a 2 d4 lbf6
rook ending with f- and h-pawns. 3 lbc3 g6
49 ••• l:xh4 4 lbf3 .i.g7
50 e6+! 1-0 5 .i.e2 0-0
Since 50 . . .'it>xe6 (50 ... �xd6 5 1 6 0-0 c5
l:xc8 l:c4 5 2 l:d8+ 'it>xc7 5 3 e7) In the 1 8th game when the same
51 .i.g3 is decisive. opening occurred, play proceeded
along the main variation 6....i.g4 7
Game 10 .i.e3 ltlc6, and here I introduced
Karpov - Korchnoi the novelty 8 'ilfd3 ? ! . Wishing to
Baguio City Wch (32) 1978 avoid another surprise, Korchnoi
Pirc Defence this time chose another plan. In
principle the position after 6 . . . c5
The battle in Baguio lasted for 7 dxc5 dxc5 is thought by theory
more than two months, and things to be slightly better for White, but
started rather successfully. After I thought my challenger might
game 27 the score was 5-2, and it have prepared an improvement,
looked as though the match was so I avoided the early exchange of
coming to an end. But excessive pawns.
confidence weakened me, and aided 7 d5 lba6
50 Baguio City Wch (32) 1978

8 i.f4 liJc7 a) 19 liJxd5 is simple and good:

9 a4 b6 19 ... i.xd5 ( 1 9 ... 'ii'xd5 20 'i'g3 'ifa2
10 l:le1 i.b7 2 1 i.c4 W'xb2 22 lbxt7) 20 c4 i.a8
1 1 i.c4 (or 20 . . .i.e6 2 1 liJc6 'ifxd3 22
The prophylactic 1 1 h3, with the lbxe7+) 21 liJd7.
aim of maintaining the bishop on b) 19 liJd7 liJxc3 20 bxc3 liJxd7
the h2-b8 diagonal, would have led 21 i.xe7 and Black suffers a loss of
to a double-edged game after material.
1 1 ...'i'd7 12 i.c4 l:lad8 1 3 W'd3 e5. In any other competitive situ­
11 ... liJh5? (D) ation I would certainly have played
like that, but in this game I did not
want to take risks.
15 ..• liJd7
Preventing a break in the centre
and hoping for counterplay with
16 ...b5 17 axb5 liJb6.
16 'ife3 i.a8
17 i.h6 b5
18 i.xg7 �xg7
19 i.n ttJr6
20 axb5 axb5 (D)
A pawn attack without a fi­
w anchettoed bishop on g7 is like an
Losing valuable time. 1 1 . . .iid7 infantry attack without a prelimi­
was necessary, trying at any cost to nary bombardment.
provoke tactical complications: 1 2
e5 dxe5 1 3 lbxe5 'i'f5 14 lbxg6
hxg6 1 5 i.xc7 lbg4. However, af­
ter 12 'i'd3 White has the better
chances.
12 i.g5 liJf6
13 'i'd3 a6
14 l:lad1 l:lb8
15 h3
15 e5 ! looks logical. 15 ...dxe5 16
lbxe5 b5 ! 1 7 axb5 axb5 1 8 i.xb5
lbcxd5 and now:
w

21 t:De2 ..tb7
Black's misfortune lies in the
poor communication between his
pieces. If he had managed to carry
out the manoeuvre ...e5, his pieces
would have had an easier passage
from one side of the board to the
other. However, 2 1 ...e5 22 dxe6
t:Dxe6 23 lDg3 would have weak­
ened his pawn structure, and Black
would have been forced to move,
all but crawling, from one flank to
the other along the 8th rank.
22 lDgJ :as
23 c3 %:ta4
24 ..tdJ 'iia8 (D)
White's knife blow has a firm
tactical basis: 25 ...lDfxd5 26 lDh5+
(or 26 lDf5+) 26 ... gxh5 (26 ...�h8
27 'fih6 %:tg8 28 lDg5) 27 'ilfg5+
�h8 28 'ilff5.
26 'ilfxeS t:DcxdS
27 ..txbS %:ta7
Karpov - Korchnoi 51
28 lDb4 (D)
B
Now not only do threats hang
over the black king, but the prosaic
c3-c4 is also unpleasant. Thus, af­
ter 28 ... i.c6 there is the possibility
of 29 i.xc6 'ilfxc6 30 c4 lDb4 3 1
%:td6 exd6 32 lDh5+ gxh5 3 3 'ii'g5+
�h8 34 'ii'xf6+ �g8 35 lDf5,
whilst 28 . . .'ii'b 8 29 c4 'ii'xe5 30
%:txe5 leads to a hopeless ending.
[Editor's note: Black then has
the surprising resource 30 ...l:.a5,
when it is not even clear that White
is better, since 3 1 cxd5? l:.xb5
leaves White with some very weak
pawns, while freeing the b5-bishop
is not possible by other means. Af­
ter 28 . . .'ii'b 8, White might do bet­
ter to try 29 'iig 5.]
28 ••• ..tc8
29 i.e2!
You must agree that there is
something attractive about this
bishop move.
52 Montreal l979

29 �e6 Game 1 1
30 c4 lbb4 Timman - Karpov
31 'ifxcS 'ifbs Montreal l979
32 �n l:r.c8 English, Four Knights

A game from the 'tournament of

stars' which is noteworthy because
the Dutch grandmaster fell into an
opening trap I had stored up for
Korchnoi, but it was not played un­
til after the Baguio match.
1 c4 lbf6
2 lbc3 e5
3 lbf3 lbc6
4 e3
4 g3 is played more often, but
because in the Baguio match I had
managed to achieve a good game
in a variety of lines of the Four
Knights English, Timman chose a
White has an extra pawn and a rarer move.
menacing initiative. 4
..• �e7
35 l:r.g8 Here the more popular continu­
36 lbf3 'iff8 ation is 4 .�b4. At the time this
..

37 'ilfe3 �g7?! modest bishop move to e7 had dis­

37 ...l:r.b7 would have prolonged appeared from the scene. However,
the battle by preventing movement this very game changed its assess­
by the pawns. ment and it again attracted the at­
38 lbg5 �d7 tention of theoreticians.
39 b4 ila8 5 d4
40 b5 lba5 As will become clear, this natu­
41 b6 ral move gives Black a wonderful
In this position Korchnoi sealed game. Evidently, 5 �e2 0-0 6 0-0
the move 4 1 . . .l:r.b7, but the follow­ is a better option for White, and if
ing day resigned the game and the 6 . . d5, then 7 cxd5 lbxd5 8 d3 leads
.

match. to a reversed Scheveningen Sicil­

1-0 ian with an extra tempo for White.
 Timman - Karpov 53

5 ... exd4 1 1 ... dxc4!

6 �xd4
Or 6 exd4 d5 ! 7 cxd5 �xd5 8
.i.b5 0-0 with a good game for
Black.
6 0-0
7 �xc6 bxc6
8 .i.e2 dS
9 0-0 .i.d6
10 b3 'fie7
1 1 .i.b2 (DJ
B
This position had been seen in
practice before our game, and
moreover it was thought to be in
White's favour. Thus, 1 1 . . ..:r.d8 12
cxd5 'fie5 ( 1 2 . . . cxd5 1 3 �b5 .i.a6
14 �d4 ! .i.xe2 1 5 "ii'xe2 'ii'e5 16
g3 l:te8 1 7 l:tac 1 gives White a po­
sitional advantage) 1 3 g3 .i.h3 14
.:r.e1 .i.b4 1 5 'ifc2 .i.f5 1 6 'ilc l
cxd5 17 .i.f3 'fie7 1 8 a3 .i.a5 19 b4
.i.b6 20 �xd5 ! is obviously to
White's advantage; Keene-Jans­
son, Haifa OL 1976.
A move which sharply changes
the assessment of the position. This
idea was also prepared for the Ba­
guio match.
Black solves his fundamental
problem by completely widening
the action of his pieces, which are
directed towards an attack on the
kingside.
12 bxc4
If White had taken with the
bishop all sorts of attacking ideas
would have appeared for B lack,
linked with the weakened control
over g4, for example 1 2...'fie5 1 3
g 3 .i.h3, 1 2. . .�g4 1 3 g 3 �xh2 or
1 2 . . . .i.xh2+ !?. Now the queen's
rook joins the battle with tempo.
12 ... l:tb8!
13 'i!fcl (DJ
Not falling into the trap 13 .:r.b1 ?
.:r.xb2 1 4 .:r.xb2 'ii'e5 ! .
B
13 .. . �g4
54 Montreal l 979

Black's attack is developing

easily and naturally. Not a shadow ltt • • •••
-
remains of White's opening advan­ • �•-� �-
%
Wi
%,·-·�

tage, and on the contrary he is -

�A-
�
� • -
�� �
� -
�
watching with alarm as events de­
velop on the kingside.
• • • •
14 g3 · � · •••
After 14 �xg4 �xg4 1 5 l:r.e 1 . -
� � �
�• u �� �

Black increases the pressure with

� - Bi.� B
1 5... l:r.b4 ! . a NI
%;
: �-���
, "'Z.J �-
�
�� : %=
/. �
w�
1 4 ... l:r.e8
/
" / "

Of course, I could have played B

14 . . . lDxh2 1 5 �xh2 'ili'h4+, imme­ 16 cS
diately forcing a draw. But now the Not, of course, 1 6 �xh2 'ii'h4+
threat of the knight sacrifice on h2 17 �g2 'ii' h 3+ 1 8 �g1 �xg3 1 9
is not so harmless, since the rook fxg3 'ili'xg3+ 20 �h1 l:r.e4 ! (or
would be included in the game 20 . . . l:le6 2 1 �f6) 2 1 l:lf4 �h3 and
along the sixth rank with decisive everything is over for White.
effect: 15 . . .lDxh2 1 6 �xh2 'ifh4+ 16 lDxfl!
1 7 �g2 'ifh3+ 1 8 �g1 �xg3 1 9 17 cxd6 (D)
fxg3 'ifxg3+ 2 0 � h 1 l:r.e6 and the
rook reaches h6.
15 lDd1 (D)
After 1 5 �f3 there follows
15 ...'ili'f6 16 �xg4 ( 1 6 �g2 is bad:
1 6 ... 'ili'h6 1 7 h3 lDe5 ! ) 1 6 . . . �xg4
17 f3 �h3 1 8 l:r.f2 'ii'g6 with a pow­
erful initiative for Black.
15
••• lDxh2!
The sacrifice which helps Black
to increase his advantage. Timman,
of course, saw this blow, but placed
his hopes on the strength of his
replying intervening move. How­
ever, when I began the combina­
tion, I had carefully taken into
account all the tactical nuances.
 Korchnoi - Karpov 55

for White, he has absolutely no lnformator it also won the prize for
choice. the best game ! In the almost thirty­
18 fxg3 'ii'xd6 year history of these prizes, this
19 � 'ii'h6 double has only happened twice.
20 ..td4 'ii'h2+
21 'it>e1 'ii'xg3+ Game 1 2
22 'it>d2 'ii'g2 Korchnoi - Karpov
23 ltlb2 ..ta6 Merano Wch (9) 1981
24 ltld3 ..txd3 QGD, Orthodox
25 'it>xd3 l:.bd8
26 ..tn 'ii'e4+ 1 c4 e6
27 'it>c3 cS! (D) 2 ltlc3 dS
3 d4 ..te7
4 ltlf3 ltlf6
5 ..tgS h6
6 ..th4 0-0
7 :et dxc4! (D)

w
Clearing the final approach to
the white king.
28 ..txcS 'ii'c6
29 'it>b3 l:.b8+
30 'it>a3 :es w
31 ..tb4 'ii'b6 Amazingly, this simple move
0-1 had never been seen in grandmas­
The merit of the novelty used in ter practice before. Chess really
this duel was recognized - it won is inexhaustible - even in such a fa­
the lnformator prize for the most miliar opening as the Queen's
important theoretical game. Re­ Gambit it is still possible to think
markably, in the very same issue of up something new on move seven.
56 Merano Wch (9) 1981
8 e3
After 8 e4 llk6 ! Black has quite
a playable game.
8 cS
9 i.xc4 cxd4
10 exd4
The idea of an early exchange
on c4 was used by Kasparov in
game 23 of our first match. I cap­
tured on d4 with the knight, and it
quickly ended peacefully after 1 0
lL!xd4 i.d7 1 1 0-0 lL!c6 1 2 lL!b3
%lc8 1 3 i.e2 lL!d5 ! with a mass of
exchanges.
10 lL!c6
1 1 0-0 (D)
B w
A typical position with an iso­
lated d-pawn has arisen. But there
is one important nuance which is
favourable for Black. In situations
like this, White's dark-squared
bishop does not usually hurry to
get into the game, but waits for the
best moment. The position of the
bishop on h4 directed me towards
an unusual idea, so I abandoned the
standard idea of playing ...lLlc6-b4-
d5.
11
••• lL!hS!
Exchanging bishops is useful
for Black, but the straightforward
1 1 . . .lL!d5 12 i.g3 would allow
White to avoid it, and in the event
of 1 1 ...lL!e4 1 2 i.xe7 lL!xc3 1 3
bxc3 White i s clearly better.
12 i.xe7
After 12 i.g3 lL!xg3 1 3 hxg3
i.f6 White has immediate difficul­
ties defending the d-pawn.
12
••• lL!xe7 (D)
An important link in Black's
plan; one knight temporarily finds
itself at the edge of the board,
while the second has managed to
take control of d5 . Here White
could rid himself of the weak
pawn: 1 3 d5 exd5 1 4 lL!xd5 lL!xd5
1 5 i.xd5 lL!f4 16 i.e4 'ii'xd 1 1 7

:cxd1 i.e6 with equality. Korch­
noi did not want to give up his ad­
vantage as White so easily, and as a
result he fell into a difficult posi-
tion.
13 i.b3 lbf6
14 lbe5 i.d7
15 'ilfe2 :cs
16 lbe4
White cannot find a clear-cut
plan, and fails to take account of
the fact that every exchange of mi­
nor pieces reduces the dynamic po­
tential of the d4-pawn.
1 6 :fe l looks expedient, inci­
dentally preventing 16 . . .i.c6 be­
cause of 1 7 lbxf7. In that case I
would have preferred to reply with
1 6...:c7 or 1 6...i.e8.
16 lbxe4
17 'ii'xe4 i.c6! (D)
Korchnoi - Karpov 57
an isolated pawn. In that case his
knight would be an able and reli­
able defender of his pawn, while
also attacking the enemy d4-pawn.
At the same time the function of
the white bishop is limited.
18 lbxc6 :xc6!
Very precisely seizing the c-file,
in case White wants to move his rook
to the kingside. After 1 8 ...lbxc6 1 9
d5 exd5 20 i.xd5 the initiative
would have passed into Korchnoi's
hands.
19 :c3 'Mid6
20 g3
It was not worth hurrying to
open an escape hatch. Now White's
prospects on the kingside have dis­
appeared.
20 :ds
21 :d1 (D)
58 Merano Wch (9) 1981
the possibility of attacking on the
kingside, it is important for Black
to preserve his major pieces from
exchange so that he can put pres­
sure on the isolated pawn. At­
tempting to build up on the d-file
does not yet work ( 2 1 . . .'ifd7 22
.i.a4), and he did not want to make
the move . . . a7-a6, as it is not yet
clear what formation the queenside
pawns will need to take on. The
text move chains the bishop to cov­
ering the b2-pawn and leaves the
possibility for an attack by ....:tb4.
22 'ifel 'iid7
Consistently carrying out the
fundamental plan. 22 . . ..:tb4 is par­
ried with 23 l:lc4.
23 .:tcd3 (D)
Attempts to gain counterplay on
the c-file are refuted by tactics:
23 .Uc5 .Ud6 24 .Udcl lLlc6 25 .i.a4
lLlxd4 ! 26 .i.xd7 lLlf3+ 27 �fl
lLlxe1 28 .i.xe6 lLld3.
B not wish to play a2-a3, as he is
23 ••• .:td6
24 'iVe4
In this situation relieving one­
self of the weak pawn does not
work: 24 d5 lLlxd5, and Black un­
ties himself by means of ...'iic 6,
... .:t6d7 and . . .lLlf6. Therefore he
would have to exchange minor
pieces and defend a difficult major
piece ending with a pawn less.
24••• 'ifc6!
25 'iff4
Here the break 25 'ifxc6 lLlxc6
26 d5 just loses the pawn after
26 ...lLlb4.
25 ... lLld5
Forcing the queen to occupy an
unfortunate position. Now 26 'ife4
is impossible because of 26...lLlb4
27 'ifxc6 lLlxc6 28 d5 lLlb4.
26 'iVd2 'ifb6 (D)
w
The unpleasant threat of . . .lLlb4
has arisen. White stubbornly does

afraid of weakening his bishop's
position. Korchnoi's patience was
by now exhausted, and further­
more he had very little time left,
so he exchanged off minor pieces,
depriving himself of any tactical
chances whatsoever and condemn­
ing himself to passive defence.
27 .i.xd5 .:.xd5
28 .:.b3
Beginning an unsuccessful op­
eration, as a result of which the
white rook turns out to be 'offside' .
If Korchnoi had planned to play f2-
f4, he should have done it straight
away, without moving the rook
from the d-file.
28
29 '6'c3
30 f4 (D)
B would probably have led to a win­
The only way of preventing
... e6-e5, but at the price of weaken­
ing his own king.
30 ..• b6!
Korchnoi - Karpov 59
31 .:.b4 b5
32 a4 (D)
32 a5 was threatened, while af­
...
ter 32 .:.b3 Black should continue
32 ....:.cs and 33 ....:.c4.
B
32 ... bxa4
Not for the sake of winning a
pawn, but to distract the white
pieces from the kingside.
33 '6'a3 aS
34 .:.xa4
Or 34 .:.c4 'ii'b5 35 'iWxa4 'iWxb2.
34 ... '6'b5!
Chaining the queen to the de­
fence of the unfortunately placed
white rook, and incidentally threat­
ening to invade on e2. I thought
that White now had to choose 35
b3, in order to free the queen. This
ning rook ending after 35 ....:.bs, al­
though I will not hide the fact that I
wanted to finish the struggle in the
middlegame. I was helped in this
60 Merano Wch (18) 1981

aim by Korchnoi's organic dislike 42 'ifn

of parting with material. Not, of course, 42... Wxa8? be­
35 :d2 eS! cause of 43 'ifxf7+, with perpetual
Here Black had several tempting check.
continuations (35 . . .:cs; 35 ... g5), 43 �h1 Wd5+
but, in making the text move, it was 0-1
necessary to have a sort of feel for The idea of the knight thrust to
the geometry of the position. At hS did not go unnoticed. This duel
first I could not manage this, until I won the next lnformator competi­
found the 37th move, which gave tion for most important theoretical
me personal aesthetic satisfaction game.
and was said not to have been ex­
pected by the majority of those pre­ Game 13
sent in the press centre. Karpov - Korchnoi
36 fxe5 :xe5 Merano Wch (18) 1981
37 ii'a1 'W'e8! ! (D) Spanish, Open Variation

My opponent's choice of opening

somewhat surprised me, as the pre­
vious two games in the Ruy Lopez
had turned out to be rather difficult
for him, especially game 14. Obvi­
ously at the end of the match other
openings seemed even more dan­
gerous to Korchnoi.
1 e4 e5
2 tl:lf3 tl:lc6
3 .tb5 a6
w 4 .ta4 tl:lf6
Making full use of the queen's 5 0-0 tl:lxe4
ability to strike on both the files 6 d4 b5
and the diagonals. 7 .tb3 d5
38 dxeS :xd2 8 dxe5 i..e6
39 l:.xa5 Wc6 9 tl:lbd2 tl:lc5
40 :aS+ <ith7 10 c3 d4
41 'W'b1+ g6 11 .txe6 tl:lxe6
4 t ...:c2 is also good. 12 cxd4 tl:lcxd4 (D)

w 14 tbe4 .i.b6 1 5 tbxd4 .i.xd4 1 6
The Open Spanish underwent
fundamental tests in Baguio and
Merano. In the final stage of the
second of these matches a telling
blow was inflicted on B lack. Now
after 1 3 tbxd4 1Wxd4 14 1Wf3 l:d8
1 5 a4 White achieves only an even
position, as in the old game Capa­
blanca-Lasker, St Petersburg 1914.
In game 14 of this match I had al­
ready used a novelty here, taking
Korchnoi unawares: 13 tbe4 ! . Hav­
ing thought for a record amount
of time, 78 minutes, he answered
1 3 ....i.e7 14 .i.e3 tbxf3+ ( 1 4...tbf5
was the more accurate move my
opponent played in game 1 6, when
he came out of the opening with no
losses), and after 1 5 it'xf3 0-0 1 6
l:fd 1 �e8 17 tbf6+ ! .i.xf6 1 8 exf6
it'c8 1 9 fxg7 l:d8 20 h4 ! he found
himself in a critical position. I then
gained my fifth victory. For this de­
cisive encounter I had prepared yet
another dangerous surprise.
Karpov - Korchnoi 61
13 a4!
The fact that Korchnoi spent 54
minutes thinking testifies to the
strength of this novelty. The his­
tory of game 14 was being re­
peated.
13 ... .i.e7 (D)
The rook move 1 3 . . . l:b8 loses a
tempo, and after 14 axb5 axb5 1 5
tbe4 .i.e7 1 6 .i.e3 White has a seri­
ous initiative; however, 1 3 . . . .i.c5
tbg5 0-0 1 7 axb5 tbxg5 1 8 .i.xg5
.i.xf2+ 1 9 l:xf2 'ii'xg5 20 bxa6
�xe5 deserves attention, as in the
game Tischbierek-Chekhov, Pots­
dam 1 985, when after 2 1 �h 1?,
21 ... �b5 ! would have led to an
even game, although it is true that
after 2 1 'iWfl ! ? White would have
preserved slightly better chances.
w
14 tbxd4 (D)
The manoeuvre 14 tbe4 has no
independent significance, as in this
62 Merano Wch (18) 1981
case play transposes to game 14 of
the match (with 14 a4 instead of 14
�e3). Neither 1 4 h3 nor 14 lle l
gives White many prospects. After
14 axb5 axb5 (a complex game
would have arisen after 14 . . .l£lxb5)
15 llxa8 'iba8 1 6 lbxd4 l£Jxd4 1 7
'ifg4 l£le6 1 8 f4 White has an ap­
preciable advantage.
B black knight should have remained
14 ... l£Jxd4
It must be said that after the
Merano match various defensive
resources after both 13 l£le4 and 1 3
a4 were found for Black. But find­
ing an antidote demands more than
a little time. In particular, here cap­
turing with the queen, 14 . . . 'ir'xd4,
was later tried several times. Here
is one revealing example: 1 5 axb5
'ifxe5 16 bxa6 0-0 1 7 'ir'a4 llfb8 1 8
a7 llb7 1 9 l£lf3 'ir'b5 20 'ikxb5
llxb5 2 1 �e3, which gave White
slightly the better chances in Hjar­
tarson-Smejkal, Bundesliga 1990.
15 l£le4 (D)
Every move has many lines, not
all of which have been exhausted;
here for instance it is also possible
to exchange on b5.
Evidently, this is the primary
cause of B lack's difficulties; the
in its place. 1 5 ... 0-0 is more pre­
cise, for example 1 6 axb5 l£lxb5
17 �e3 'ir'c8 18 'ir'c2 1i'e6 19 f4
llad8 20 l:ta4 l:td7 2 1 l:tfa1 1i'd5
22 h3 f6 23 exf6 �xf6 24 l£lxf6+
:Xf6 25 llxa6 llxa6 26 llxa6 l£Jd4
yielded Black some compensation
for the pawn in Adarns-Yusupov,
Hastings 1989/90.
16 �e3 0-0
17 f4
Threatening to play 1 8 f5, and if
1 7 ... g6, then 1 8 g4. In a difficult
position Black finds the best de­
fence.
 Karpov - Korchnoi 63

17 ••• 'iVxd1
18 lUxd1 l:[tbS
19 l:[d7 (D)

w
24 . . . lbc4 25 ..tc5 ..txc5+ 26
lbxc5 gxf6 27 lbd7 .z:r.cs 28 lbxf6+
is no good.
B 25 .:.act .z:r.xc7
19 ... ..tfS 26 .z:r.xc7 l:[dS (D)
It was necessary to continue
1 9... ..td8, although after 20 a5 lDf8
2 1 l:[d3 Black's position is not the
most pleasant.
20 fS lbd8
21 aS!
Emphasizing the passive nature
of the black pieces; 2 1 .z:r.xc7 bxa4
22 ..td4 l:[b4 23 .z:r.d 1 .z:r.ab8 is not so
clear.
21 ... lbc6 (D)
22 e6! fxe6
23 f6! lbeS w
Defending against the unpleas­ 27 h3!
ant threat of 24 f7+ and 25 tbg5. There is no hurry, and this quiet
Simplifying by 23 ....z:r.d8 24 .z:r.xc7 move deprives Black of the chance
.z:r.ac8 would have led to a hopeless of 27 ....z:r.d1 + 28 � lbg4+.
endgame for Black after 25 f7+ 27 ... h6
�h8 26 .z:r.xc8 .z:r.xc8 27 .z:r.c t . He must not allow the knight to
24 .z:r.xc7 .z:r.cs invade g5.
64 M1rano Wch (18) 1981

28 :a7! 32 b4 :ds
The most clear-cut way of real­ 33 :xa6 �f7
izing the advantage. 28 et:'lcS gxf6 34 :a7+
29 et:'lxe6 :d 1 + 30 'iltf2 i.b4 would Avoiding a trap: after 34 :xe6
have activated the black pieces. et:'lxa5 35 bxa5 :cs the rook ending
28 ... et:'lc4 may turn out not to be winning.
The rook is tied to the back rank: 34 ... �g6
28 ...:d1 + 29 'iPf2 :b1 30 i.d4 lLlc6 35 :d7
3 1 fl+ 'iPh7 32 :as et:'lxd4 33 :xf8 Since exchanging rooks leads to
�g6 34 et:'ld6, and the f-pawn will the loss of a piece - 35 ...:xd7 36
become a queen. et:'lxd7 et:'ld6 37 lbb6 - Black is
29 i.b6 :bs forced to give up the open file and
In the event of29 ...et:'lxb6 30 fl+ resign himself to the loss of his last
�h7 3 1 axb6 Black cannot cope queenside pawn.
with two passed pawns. 35 ... :es
30 i.c5 i.xcS+ 36 a6 :as
31 et:'lxc5 gxf6 (D) 37 :b7 'iPfS
38 :xbS 'iPeS
39 :b7 'iPdS
• • ••• 40 :r7!
a . • • - Precisely here; after 40 :e7 eS
··�
W& �
� -
�·-
� � - � 41 :n f5 ! Black is still holding on.

"'� .
u - - ·�
�
40 ...
41 :r6 (DJ
fS

••• • •
. . . ·� 1. . • • •
" -
u -� ·
� W& � �
� -
�
� � • • • •
W& •
•
� W& -
� ·
�
�• ••a •
w
At this point Black even has an
. �·· · ·
extra pawn. However, White's pos­ �··
u• -
W& •
� �
W& •
•
�
session of the seventh rank, the dif­ . . . ·�
ference in possibilities of the
knights, and the abundance of weak­
. . -�·
nesses in the black camp leave him • • • =
no hopes of saving himself. B
 Karpov - Yusupov 65

In this position Black sealed the 13 lDfd4 .txd4

move 4 l ...e5. 14 cxd4 aS
However, the next day Korchnoi 15 .te3 a4 (D)
resigned without resuming. The Instead 1 5 ... lDb4 1 6 .i.b1 a4 1 7
match, and together with it the era lDd2 a 3 1 8 'it'c 1 ! gives White a big
of our world championship rivalry, advantage.
ended.
1 0
-

Game 14
Karpov - Yusupov
USSR Ch (Moscow) 1983
Spanish, Open Variation

In this game from the 50th anniver­

sary of our national championships
we see again the Open Variation,
so popular in the World Champi­
onship matches at Baguio and w
Merano. Not long before this 16 lDd2
championship, at Linares, Yusupov At Merano (match game 6) I re­
had already chosen this against me, treated the knight to c 1 , and as a
and after a difficult defence the result sustained a defeat. Now I de­
game ended in a draw. This time cided to return to a manoeuvre I
that situation was not repeated. had used in an old game against
1 e4 eS Savon (Moscow 197 1 ).
2 lDf3 lDc6 16 a3
3 .tbS a6 17 lDxe4 axb2
4 .ta4 lDf6 18 l:tb1 .txe4
5 0-0 lDxe4 When I was commentating on
6 d4 bS my encounter with Savon, I esti­
7 .tb3 dS mated the position after 1 8 ...dxe4
8 dxe5 .te6 1 9 l:txb2 �7 20 l:txb5 l:txa2 2 1
9 c3 .tcS 'it'b1 1i'a8 22 l:tc l to be favourable
10 lDbd2 0-0 for White. The text was introduced
11 .tc2 .trs into practice by Yusupov.
12 lDb3 .tg6 19 l:txb2 'ii'd7
66 USSR Ch (Moscow) 1983
20 .id3
In Ivanov-Yusupov, USSR Ch
(First League) 1979, 20 .ixe4 dxe4
2 1 lhb5 lllxd4 22 .l:.c5 .l:.fd8 led to
an equal position. Three years after
this game, Hiibner, in two battles
with Korchnoi, used the new text
move.
20 ... .ixd3
An exchange of bishops hap­
pens all the same. 20 ...b4 does not
work because of 2 1 .ib5 .l:.fb8 22
.l:.xb4 .l:.xa2 23 .ixc6.
21 'i'xd3 (D)
The pawn structure is more
pleasant for White.
B jevo 1 984), my opponent played
21 .•• .l:.fb8
In the second Hiibner-Korchnoi
encounter (Lucerne OL 1 982) an
even position arose after 2 1 ...b4 22
.id2 .l:.fb8 23 .l:.fb 1 'i'g4 24 .ie3
.l:.b6 25 h3 'i'c8, although as a re­
sult of 26 .l:.c2? b3! 27 .l:.xb3 lllb4
Black won the exchange, and in the
end the game. However, returning
material is not compulsory.
22 lUb1
22 .l:.xb5 .l:.xb5 23 'i'xb5 lllxe5
24 'i'b7 'i'c6 25 'i'xc6 lllxc6 26
.l:.c 1 .l:.a6 only leads to an equal po­
sition.
22
••• b4
23 h3
Following 23 a3 bxa3 24 .l:.xb8+
.l:.xb8 25 .l:.xb8+ lllxb8 26 'i'xa3
'i'c6, Hiibner-Korchnoi, Chicago
1982 ended in a draw after 27 'i'e7
'i'd7 28 'i'a3 although according to
Hiibner, by continuing 27 g4 h6 28
f4 llld7 29 f5 lllb6 30 .if2 White
would have retained an advantage.
Interestingly, this move was my
first independent decision. Strictly
speaking, this has nothing to do
with the pawn move, but with
White avoiding the simplification
on the queenside, which occurred
in the above-mentioned games.
23
••• h6 (D)
Not, of course, 23 ....l:.a3 because
of 24 'i'xa3 ! . However, a year later
in a game against Popovic (Sara­
more accurately: 23 ....l:.b6, saving
tempi. White had no great desire to
spend time trying to realize a mi­
croscopic advantage, and a couple
of moves later, after 24 'i'c2 .l:.ab8
25 .l:.c l .l:.8b7 26 'i'c5, he had to ac­
cept a draw.
24 :et .l:.b6
 Karpov - Yusupov 67

with a position which has possi­

· - . . •• bilities for both sides.
� -
�
-
� --· ��·-
� �
-�· . .
.
- �
-�·�
-� u .
-
- u
- � . --� -
��
• .,.m • �
��R, �-� u
� ��- -�
�
-
�:-
� �
�� � � �
-
� �
w
In the event of 24... lDa5 there is
the unpleasant 25 'lfbl , whereupon
25 . . . lDc4 loses: 26 l:txb4 l:txb4 27 w
'iVxb4 l:txa2?? 28 'iVb8+ �h7 29 31 llbc2
'iih l+. 32 l:txc7
25 'ilbl llab8 33 l:txc7
After 25 ...l:ta7 too, B lack will 34 g4!
suffer, as in the game Popovic­ 34 ...'iVe2 was threatened, and if
Timman, Sarajevo 1984. 35 llc2, then 35 ...%lxg2+ ! 36 �xg2
26 l:tc5 lDd8 lDxe3+. The defence 34 'iVc2 is
27 l:tcc2 lDc6 insufficient in view of the reply
If 27 ... lDe6, then 28 f4, with a 34 ... 'ii'f l ! . However, the move by
pawn storm; therefore Black re­ the g-pawn saves White from all
turns the knight to its place. his problems.
28 'ii'c t l:t8b7 34 .•• lDh4
29 l:tc5 li:J.e7 After 34 ... lDxe3 35 'iVxe3 the
30 �h2 lDf5?! (D) black attack has run dry, and a cru­
White has a clear initiative on cial pawn has gone missing.
the queenside, besides which he 35 l:tc8+ �h7
can also send forward the kingside 36 'ii'd l 'ii'a6
pawns, giving his opponent new 37 l:tc2
problems. Black decides to give White's aim is to carry out the
back the c-pawn to stir up some manoeuvre .i.f4-g3, securing his
counterplay. All the same, 30...llb5 king position.
or 30 . . .c6 would have been better, 37 ••• f5 (D)
68 Moscow Wch (9) 1984/5

40 �xh3 "ili'e6+ 4 1 �h2 'ilff5

threatening 42 ..."ili'e4 is dangerous.
Now the rook comes to the aid of
the king.
40 •.. 'ii'e6
41 "ili'hS!
Not allowing 4 l ...h5.
41 'ii'e7+
42 'itxh3 'ii'f7
43 l::thl !
Preventing 43 . . .l::tg 3+ 44 �h4
w lth3+.
In the event of 37 ...1i'a3 38 ...e2 43 ••• 'ii'd7+
b3 39 axb3 "ili'xb3 the position can 44 fS 1-0
be strengthened by means of 40 It is interesting that Yusupov and
l::tc 7. I were jointly handed the prize for
38 �g3! the most beautiful game in the
The very same knight which was championship. The founder of the
meant to bring havoc into White's prize for beauty understood that
camp is cold-bloodedly eaten by every work of art in chess has two
the king. authors ! Then, a little later, this
38 fxg4 victory won the next competition
39 �xh4 gxh3 (D) for the best game in lnformator.

Garne 1 5
Karpov - Kasparov
Moscow Wch (9) 198415
QGD, Tarrasch

In his Candidates matches against

Beliavsky, Korchnoi and Smyslov,
my next opponent for the world
chess crown had used the Tarrasch
Defence successfully. However in
our first duel I managed to find the
w key to Black's position.
40 f4 1 d4 dS
 Karpov - Kasparov 69

2 c4 e6 the knight to f5 straight away than

3 lbf3 cS to play 14 h3 .th5 1 5 .:tad 1 .:tc8 1 6
4 cxd5 exd5 g4 .tg6 1 7 lbf5 first, a s given i n
5 g3 lbf6 earlier opening manuals .
6 .tgl .te7 14 ... .:r.cs
7 0-0 0-0 In our first match the Tarrasch
8 lbc3 ltJc6 Defence was played twice - in this
9 .tg5 cxd4 game and earlier in game 7. It was
10 lbxd4 h6 then that Black first chose to place
11 .teJ .:r.es his rook on c8, which was at that
12 'ili'b3 (D) point considered a novelty.
15 .td4! (D)
A strong move. In game 7 I had
exchanged on e7 - 1 5 lbxe7+ .:txe7
1 6 .:tadl 'ii'e 8, and Black quickly
equalized. He was defeated not be­
cause of his opening formation, but
due to an inaccurate move he made
in the endgame.

B
Kasparov's opponents in the
Candidates matches had played a
variety of moves here - 12 1t'a4, 1 2
'j/fc2 and 12 a3, and Black achieved
a good game every time. I instead
used a rarer continuation, which I
had specially prepared for the en-
counter. B
12 lbaS 15 .tcs
13 'ili'cl .tg4 16 .txcS .:txcS
14 lbfS 17 lbe3!
GM Lajos Portisch thought of Compensation for the isolated
this move. It is better to dispatch d-pawn usually takes the form of
70 Moscow Wch (9) 198415
active piece play, but in this case
the black pieces are quite passive.
Thus, the opening dispute has con­
cluded in White's favour.
17 ••• .ie6
The fork 17 ...d4 is not danger­
ous because of 1 8 :ad 1 .
18 :ad1 Wc8
White was already threatening
1 9 lbexd5 ! �xd5 20 e4.
19 Wa4
1 9 Wb1 :d8 20 :d3 deserved
attention, when 20 ... d4 is bad due
to 21 :Cd l �6 22 .ixc6.
19 :ds
20 :d3 a6
21 :rd1 (DJ
After 2 1 Wd l Black would have
to reply 2 1 . . .11t'c6, as 2 1 . . .lLlc4 22
�exd5 lLlxd5 23 �xd5 .ixd5 24
.txd5 �xb2 25 .ixf7+ �xf7 26
:xd8 lLlxd I 27 :xc8 :xc8 28
lbdl would lead to a rook ending
with an extra pawn for White.
Now the white rook and knight
are very unusually positioned, in a
letter T (it is as if the d3-rook is
suspended between epaulettes on
c3 and e3), and his pieces are
moreover directed at the most vul­
nerable spot in Black's fortress. All
the same, the d-pawn is withstand­
ing the pressure.
21 ... lLlc4
22 lLlxc4
22 lLlexd5 �xd5 23 lLlxd5 .txd5
24 .ixd5 gives nothing in view of
24 . . .:dxd5 ! 25 lbd5 :xd5 26
lbd5 �b6 27 'it'd4 lLlxd5 28 'it'xd5
'it'c 1 + 29 �g2 'it'xb2 with an level
queen ending.
22 ... :xc4 (D)
w
23 WaS
The natural move 23 'ifb3 would
probably have been stronger. After
23 . . . d4, not 24 'ii'xb7 'it'xb7 25
.txb7 :b8 26 .txa6 dxc3 27 .txc4
c2 28 :d8+ lbeS ! , but the quiet 24
 Karpov - Kasparov 71

'ii' b 6! wins a pawn: 24 ...ltld7 25 32 Wg2 .:t7c5 (D)

.:txd4 ! . Black generally has to be
very careful, so as not to lose his
isolani.
23••• .:tcS
24 'ii'b6 .:td7
25 .:td4 (D)
Prophylactic measures such as
25 h3 or 25 a3 would probably
have given Black more trouble.

w
33 .:txc4 .:txc4
34 .:td4 wrs
35 .i.e2 .:txd4
36 exd4 We7 (D)
Insufficiently energetic play by
White has almost allowed Black to
stabilize the position; here 36 .. lbe4
.

37 ltla2 ltld6 38 ltlb4 a5 39 ltlc6

B ltlc4 would have given him more
25 ... 'ii'c7 chances to equalize.
26 'ii'xc7 .:tdxc7
27 b3
Exchanging queens turns out to
be possible because the d5-pawn is
invulnerable: 27 ltlxd5 ltlxd5 28
.i.xd5 .i.xd5 29 .:txd5 .:txd5 30
.:txd5 .:tc2 3 1 .:td8+ Wh7 32 .:td7
.:txb2 33 .:txf7 .:txe2 drawing.
27 ... b5
28 a3 g6
29 e3 Wg7
30 Wh2 .:tc4
31 .i.f3 b5
72 Moscow Wch (9) 1984/5

37 lt::!a2 .tc8
38 li)b4 'itd6
39 f3 lOgS
40 h4 lt::!h6
41 'itf2 lt::!f5
42 lt::Jc2 (D)

• .i.. • •
• • •••
·�
-� -
� � �·�
�
-
� ·
�
w
· · · · ·�· · have been so easy. However, Black
�
-� �·
� u � � �
-
� u has taken the h4-pawn. Exchang­
�- �
�
� -
��8� u ing pawns (if it happened) would
BlLl · ii.= • make achieving a draw simpler. In
fact, the break g3-g4 is harmless,
• • • • and White only has the f4-square
B left through which he can penetrate
Here the game was adjourned, the enemy camp; his pieces, the
and Black sealed the move 42 ... f6. king and knight, cannot both pass
After 42 ...lt::!g7 43 g4 f6 44 .td3 g5 through it at the same time. How­
45 .tg6 ! hxg4 46 h5 White has a ever, White has at his disposal a
dangerous passed h-pawn, but study-like route.
42 ... .td7 would probably have 47 lt::!g2! !
been more reliable. Psychologically i t was impossi­
42 f6 ble for Black to foresee such a
43 .td3 g5 move. It is all too easy to assume
44 .txf5 .txf5 that White will automatically re­
45 li)e3 .tbl capture the pawn on h4. The unex­
46 b4 gxh4 (D) pected knight manoeuvre is based
Black's bad bishop (his a6-, b5- on a pawn sacrifice. By giving up
and d5-pawns are placed on the one of his own pawns White clears
same coloured squares) determines a square which is needed for his
White's stable advantage, but is it king. The material balance is soon
decisive? After 46....tg6, piercing restored, after which the white
a hole in the fortress would not pieces invade.
 Karpov - Kasparov 73

47 ... hxg3+
48 �xg3 �e6
49 lDf4+ �s
50 lbxh5 �e6
51 lbf4+ �d6
52 �g4 .tc2
53 �hS .tdl
54 �g6 �e7 (DJ
After 54... .txf3 55 �xf6 Black
loses the d5-pawn after all.

w
. - - . 66 lbe7 .*.b7
.
- . - .
- - 66...�b3 67 d5 �xa3 68 d6
• • - ·� · i.h3 69 lDd5 will not save Black

-···· - either; the last chance of prolong­

R � � . ing the game lay in 66...i.h l 67
u u � - ltks �d5.
0 • -8· 67 lbf5 .tg2
. . - . 68 lbd6+
69 lbxbS
�b3
�a4
B B.t.B B 70 lbd6 1-0
w
55 lbxd5+ �e6 Game 1 6
56 lbc7+ �d7 Karpov - Kasparov
57 lbxa6 .i.xf3 Moscow Wch (27J 1984/5
58 �xf6 �d6 QGD, Orthodox
59 �fS �dS
60 �f4 .i.hl 1 lbf3 d5
61 �e3 �c4 2 d4 lbf6
62 lbcS .i.c6 3 c4 e6
If 62 ... .tg2, then 63 lbd3 �b3 4 lbc3 .te7
64 lbf4 .*.b7 65 �d3 �xa3 66 �c3 S .tg5 h6
and 67 d5 is decisive. Here White has a choice be­
63 lbd3 .*.g2 tween immediately taking on f6
64 lbeS + �c3 (the Petrosian Variation) or retreat­
65 lbg6 �c4 (DJ ing the bishop to h4, in order to
74 Moscow Wch (27) 198415

carry out the exchange at a more 10 .i.xc4 'iraS

appropriate moment (usually after
...b7-b6). These two plans are seen
alternately in practice, and this is
how it was in my duels against
Kasparov.
6 .i.xf6 .i.xf6
7 e3 0-0
8 'ircl cS
9 dxcS dxc4 (D)
Kasparov and I often borrowed
from each other's opening systems,
which was inevitable, as in those
days our repertoires were in many
ways similar. In particular, this
position was seen in Kasparov­
Timman, USSR v Rest of World,
London 1984, which was played
some months earlier. On that occa­
sion Kasparov gained a convincing
victory after 9 ...'ira5 1 0 cxd5 exd5
1 1 0-0-0 .i.e6 1 2 lL!xdS llc8 1 3
�b1 ! . Now he had to fight against
it himself.
w saw how I should place my pieces.
1 1 0-0 .i.xc3
In the event of 1 1 ...'ii'xc5 1 2
lL!e4 'fke7 1 3 lLlxf6+ 'ii'xf6 14 llfdl
White has tangible pressure.
12 'irxc3 'irxc3
13 bxc3 lL!d7
14 c6 bxc6
15 llab1 lL!b6
16 .i.el cS
17 llfcl! (D)
B
Moving the rook to c 1 is the first
in a long series of precise moves by
White. 17 llfdl suggests itself, but
in the long term this can only lead
to exchanges on the d-file. On c l
the rook is fulfilling a prophylactic
function (defending the c-pawn),
and is not taking the d 1 -square
away from the bishop. The situ­
ation will only become completely
clear in six moves time, and I will
confess that I sensed rather than
 Karpov - Kasparov 75

17 ... .i.b7
After the game this move pro­
voked unanimous criticism from
commentators. In fact, 1 7 ....i.d7,
controlling the b5-square, is more
reliable, for example 1 8 �fl (there
is another possibility linked with
moving the bishop to a6 and the
knight to e5) 1 8 .. .l:tfd8 1 9 l:.b3
l:.ac8 20 l:.a3 l:.c7 2 1 c4 .i.a4 22
l:.bl .i.e8 23 l:.a5 and White had a
minimal advantage in Novikov­ B
Sturua, Lvov 1 985. 19 ..• l2Jd7
But in the game itself White's In the final analysis this loses a
advantage from the start was also pawn; 1 9...l:.ac8 20 l:.a5 l:.c7 21 c4
scarcely noticeable. Therefore, .taB is more resilient.
glancing at this almost symmetri­ 20 l:.aS l:.tb8
cal position, it is difficult to under­ 21 c4 .i.c6
stand immediately why the bishop 22 lbel!
is better placed on d7 than on b7. Paradoxically, before going on
1s �n .i.ds the attack, the white pieces first of
At the cost of a tempo it would all retreat.
have been possible at this point to 22 l:.b4
prevent the rook from landing on 23 .i.d1 ! (D)
b5 : 1 8 ...�c6, but after 1 9 lDe5
.i.a4 20 .i.b5 (20 .i.a6 l:.fd8 2 1
�e2 threatening 22 l2Jd3 i s also un­
pleasant for him) 20 ... �xb5+ 2 1
l:.xb5 l:.fc8 22 l2Jd3 c4 2 3 l2Jb2
Black has not escaped difficulties.
19 l:.bS (D)
The white rook is headed for an
ideal position - the a5-square. The
a2-pawn cannot be taken in view of
the reply 20 c4 and the bishop is
trapped - the cl -rook is already in­
fluencing the course of events.
76 Moscow Wch (27) 198415

. - - ­
If 23 'Lld3 immediately, then
23 ...l:.a4, and the pawn is safe.
Now, however, the march 'Llel­ ••• - -
d3xc5 cannot be prevented. The full R.t.li .t. R R
extent of White's resources is now
revealed, and the hidden strength
a • ••••
of the quiet move 1 7 l:.fc I becomes -�- - -
clear. .ii.. B � R
23 l:.b7
�·
.� -�-
B W B � �'7}.,
- - - -
24 f3 l:.d8
25 'Lld3 g5
26 ..tb3 w
The final preparations; after the 36 l:.xgS ..trs
hasty 26 'Llxc5 'Llxc5 27 l:.xc5 l:.b2 37 �e3
28 l:.xc6 l:.dd2 all White's gains The black h-pawn will in the
would have dissolved. long run cause White distinct
. 26 �f8 trouble, and it would have been
27 'Llxc5 'Llxc5 worth fixing it with 37 h4. For ex­
28 l:.xc5 ample, the following variation is
White still faces a large and la­ possible: 37 . ..l:.f7 38 �e3 ..tg4 39
borious task, but the preliminary c5+ �c6 40 ..ta4+ �c7 41 ..te8
results of the engagement are not l:.h7 42 ..tg6 l:.h6 43 ..te4 with an
difficult to sum up: a pawn is a easy win.
pawn. 37 h4
28 l:.d6 38 �d4 e5+
29 �e2 �e7 39 �c3 ..tb1
30 l:.d1 l:.xd1 40 a3 l:.e7
31 �xd1 �d6 41 l:.g4 (D)
32 :as f5 White's final move was sealed.
33 �e2 h5 (D) In the process of home analysis it
Kasparov is trying with all his became clear that the adjourned
might to gain some counterplay. game was exceptionally sharp, and
With his last move he has weak­ that Black could gain counterplay.
ened the g5-pawn, and I immedi­ Therefore, in order to find a clear­
ately make use of this. cut route to victory, I had to solve
34 e4 fxe4 various problems straight away. It
35 fxe4 ..txe4 is no accident that this ending has

B �g2 llc l 56 �xh2 llxc5 57 llf7
already found its way into mono­
graphs on endgame theory.
41 ••. h3!
In the event of passive defence
with 4 1 ...1lh7, I would have fixed
the h-pawn (but now on h4), after
which Black would have had noth­
ing to hope for. If 4 l ...e4, then 42
�dl .:.n 43 'iit d4 llf2 44 c5+ 'iite7
45 �b3 lld2+ 46 �c3 lld3+ 47
'iitb2 e3 48 �a4 is decisive.
42 g3
Black was counting on my ac­
cepting the pawn sacrifice - 42
gxh3, and then 42 ...1lh7 !, notice­
ably widening the rook's sphere of
influence.
42 .•. lle8
Preparing to transfer the rook
to the second rank via the f-file.
42 ...1lf7 loses immediately to 43
c5+.
43 llg7 .:.rs
44 llxa7 .:.n
45 �b4 llxh2 (D)
Karpov - Kasparov 77
White has a beautiful win after
45 ...1lb2. Here is the main line: 46
c5+ 'iitc6 47 �c4 �c2 48 lla6+ �c7
(48 ... 'iitb7 49 llb6+ �c7 50 'iitc 3)
49 �xc2 llxc2+ 50 �d5 llxh2 5 1
lla7+ �b8 (alternatively, 5 1 ...'iitc8
52 llh7 llh1 53 'iit d6 h2 54 llh8+
'iitb7 55 c6+ 'iitb6 56 c7) 52 llh7
llh 1 53 'iite4 h2 54 �f3 e4+
(54 .. .1la1 55 llxh2 llxa3+ 56 �g4
llc3 57 lle2 llxc5 58 �f5, etc.) 55
.:r.c4 58 g4 e3 59 �g3 winning.
w
46 c5+ 'iitc6
47 �a4+ �d5
48 lld7+ �e4
48 . .. �e6 49 c6 llb2+ 50 �b3+
llxb3+ 5 1 'iit xb3 �e4 52 lld8
�xc6 53 .:r.hs �g2 54 a4 �f5 55
llh4 will not save Black either.
49 c6 llb2+
50 'iita S! llb8
If 50 . . . h2, then 5 1 c7 !, and as
soon as the black queen appears on
78 Moscow Wch (4) 1985

the board, she is lost: S l ...hl'ii 52 S .i.gS h6

.i.c6+. 6 .i.xf6 .i.xf6
51 c7 l:.c8 7 e3 0-0
52 �b6 �e3 8 'ii'c2
53 .i.c6! h2 As the queen move had led me
54 g4! (D) to victory in the game we have just
Taking the f5-square away from seen, there seemed no point in
the bishop. Now it is all over. avoiding this popular position a
year later. However, Kasparov had
prepared an interesting novelty.
8 ••. tba6! (D)

B
54 l:.h8
ss l:.dl .i.a2
56 l:.el+ �f4 w
57 l:.e4+ �g3 A curious manoeuvre, which al-
58 l:.xeS �xg4 lows Black to make use of the
59 l:.e2 1-0 white queen's position on c2 in or-
der to simplify - 9 cxd5 lbb4 ! 10
Game 17 'ii'b3 tbxd5.
Karpov - Kasparov 9 l:.dl cS
Moscow Wch (4) 1985 10 dxcS 'iiaS
QGD, Orthodox 1 1 cxdS tbxcS
12 'ii'd2
1 d4 dS After 1 2 d6 .i.d7 followed by
2 c4 e6 .. .l:.ac8, or 1 2 dxe6 .i.xe6, Black
3 tbc3 .i.e7 has sufficient compensation for the
4 tbf3 l£lf6 pawn.
 Karpov - Kasparov 79

12 .•• lidS The queen retreats and occupies

Not, of course, 1 2...exd5 in view
of 1 3 ltlxd5, while 1 2 ....i.xc3 1 3
1Wxc3 1Wxc3+ 1 4 bxc3 exd5 1 5
l:.xd5 is also favourable for White.
13 ltld4
The only attempt to fight for the
initiative. After 1 3 .i.e2 .i.xc3 1 4
Wxc3 Wxc3+ 1 5 bxc3 l:.xd5 the
time has come to agree a draw.
13 ... exd5
14 .i.e2 'ii'b6
After 14 ...lDe6 1 5 ltlb3 .i.xc3 1 6
bxc3 as well, White i s slightly bet-
ter.
15 0-0 ltle4 (D)
Definite problems remain for
Black after 15 ... lDe6 16 ltlf3 d4 1 7
ltle4 dxe3 1 8 ltlxf6+ gxf6 1 9
Wxe3.
a safe place; when she is needed
she can quickly return to the cen-
tre.
19 ... llc7
20 l:.d2 l:.dc8
After 20....i.xd4 2 1 llxd4 all the
black pieces would be tied up de­
fending the d5-pawn. Possibly in
this case the game would have
ended in a draw, but passive play is
not Kasparov's style. It was diffi­
cult to imagine that changing the
pawn structure would lead in the
future to serious problems for
Black.
21 ltlxe6!
If 2 1 llfd l , Black should con­
tinue 2 1 .. ..i.xd4 and then 22 llxd4
llc2 23 .:.4d2 .i.f5 .
21 ... fxe6 (D)
2 1 . . .Wxe6 22 llfd l is scarcely
any better.

ltlxc3
.i.e6
llac8 w
22 .i.g4
80 Moscow Wch (4) 1985

If 22 llfd 1 , then 22 . . .'iVb4 is pos­ White would maintain a mini­

sible, threatening 23 .. ."tlt'xd2 ! with mal advantage after 27 ...l:lc l 28
complete equality. For the time be­ l:lxc l 'ii'xc 1 + 29 �h2 'ii'c6 30 g3.
ing Black's position looks hopeful, But Kasparov, it seems, underesti­
but while his bishop is firing into mated White's threats.
thin air, White's bishop is capable 28 .i.hS!
of creating dangerous threats - a The bishop bounces off the sides
typical motif in opposite-coloured of the board like a billiard ball,
bishop endings. In the end White's changing direction and ending up
light-squared strategy is trium­ on another, even more dangerous,
phant in this game. As Grandmas­ diagonal.
ter (and pianist !) Taimanov noted, 28 .•• bS
for the next 1 7 moves I was only 28 ...i.d8 29 .i.g6 .i.c7 30 .i.d3
playing on the 'white keys' . 'it'd6 3 1 g3 'it'e5 32 ..g4 'ii'f6 was
22 l:lc4 necessary, erecting a fortress. The
23 h3 'ifc6 text move loses time, and allows
24 'ii'd3 White to improve his position.
After 24 'it'g6, there is the retort 29 .i.g6 .i.d8
24...'it'e8, and White cannot yet in­ 30 i.d3 b4
vade via the light squares. 31 'ii'g4 'ife8
24 �h8 32 e4! .i.gS (D)
25 l:lfdl aS
26 b3 l:lc3
27 'ife2 l:lf8 (D)

w
33 l:lc2 .:.Xc2
After either 33 ...'it'f7 34 l:le2 or
w 3 3 ...'it'c6 34 'ife2 l:lc8 35 l:lxc3
 Karpov - Kasparov 81

bxc3 36 exd5 exd5 37 .i.c2 White's 39 ••. l:td8

initiative has not been dampened 39 . ..d4 40 'ife4 Wg8 4 1 .i.d3 is
down, but B lack should neverthe­ also bad, but 39 ....i.f6 is a more
less feel calmer. dogged defence.
34 .i.xcl 'ii'c6 40 'ii'fS Wg8 (D)
35 'ii'el 'ii'cS
36 .:0 (D)
Defending against the threat of
36 ...l:txf2 37 'ii'xf2 .i.e3. The black
queen's activity will be short-lived,
and clouds are gathering over the
black king.

w
Here the game was adjourned,
and painstaking analysis showed
that White has huge attacking re­
sources.
41 'ii'e6+ �h8
If 41 ...�f8. then 42 .i.g6 'ili'f4 43
B l:tel with the deadly threat 'ife8+.
36 'ii'c3 42 'ii'g6 Wg8
37 exdS exdS 43 'ii'e6+ �h8
38 .i.bl! 'ii'dl 44 .i.fS 1Wc3
39 'ii'e5 45 'ii'g6 �g8
39 'ii'e6 was an alternative, but 46 .i.e6+ �h8
then a white piece has to go onto a 47 .i.fS �g8
dark square at some point! Obvi­ 48 g3 Wf8
ously there is no reason to pursue 49 �gl 'iff6
the pawn - 39 'ii'xd2 .i.xd2 40 l:tdl so 'ifh7 'ili'f7
.i.g5 4 1 :Xd5 l:td8 and now the op­ 5 1 f4 was threatened, winning a
posite-coloured bishops work in piece, and if 50...g6, then 5 1 .i.xg6
B lack's favour. 'ii'g7 52 f4 .i.f6 53 l:tdl .
12 Mo�eow Wcla (4) 1985
!1 h4 .td2
Other bishop retreats are no bet­
ter: 5 1 . . .i.f6 52 l:.e l 'ili'g8 53 'iWg6
'iWf7 54 'ii'g4, or 5 1 ...i.e7 52 l:tel
with the threat of 53 i.e6 or 53
.tg6.
52 l:tdl i.c3
53 l:td3 l:td6 (D)
If 53 ... 'iWg8, then 54 'ii'g 6 'ii'f7
55 'ifb6 is very strong. Now that
the white rook has entered the
game, events develop swiftly.
w If 55 ...i.e5, then 56 i.h3 l:tf6 57
54 l:tf3!
One imprecision - 54 l:te3 -
would be enough to allow Black to
free himself by 54 ...g5 ! .
54 ••• <j;e7
The king is forced to abandon
his refuge. If the f-file is recovered,
the appearance of the rook on e3
will quickly be decisive: 54 ...i.f6
55 l:te3 g5 56 'iWxh6+, and the
rook's action is covered; 54 .. .l:tf6
55 l:te3 g5 (55 ...l:txf5 56 'ii'h 8+
'ii'g 8 57 l:te8+ <j;xe8 58 'ii'xg8+
<j;d7 59 f4) 56 'ii'h 8+ covering the
action of the bishop. Problem mo­
tifs in abundance !
55 'ii'h 8!
55 l:te3+ would have been hasty
in view of 55 ...<j;d8 56 'ifh8+ <j;c7
57 'ii'c8+ <j;b6 58 'ii'b8+ <j;c5 !, and
Black can defend himself with the
help of ...l:tc6 and ...'ii'c7.
55 ... d4 (D)
w
l:te3 ! l:txf2+ 58 <j;gl is the end.
56 'ifc8 l:tf6
57 'ifc5+ <j;eS
58 l:tf4 'iib7+
59 l:te4+ rj}f7
It looks as though Black would
escape without a scratch after
59 ...l:te6, because of 60 i.xe6
'ii'xe4+, but again a problem ma­
noeuvre is decisive: 60 'ifc4! l:txe4
6 1 'ii'g 8+ <j;e7 62 'ifxg7+ and 63
1Wxb7.
 Kasparov - Karpov 83

60 'ii'c4+ �8 12 a4 'ifd7 (D)

61 �h7! lU7 This queen move is relatively
62 'ii'e6 'ii'd 7 rare; the usual continuation here is
63 'ii'e 5! 1-0 1 2 ... h6.
I had prepared 63 . . .:e7 (if
63 . . .'iWd8, then 64 'ii'c 5+ :e7 65
:f4+ �e8 66 'iWc6+ 'ii'd7 67 ..tg6+;
or 63 ...'ife7 64 'iWb8+) 64 'ii'f4+
:n 65 'iWb8+.

Game 1 8
Kasparov - Karpov
Moscow Wch (5) 1985
Spanish, Zaitsev

1 e4 e5
2 lLlf3 lLlc6 w
3 �b5 a6 13 axb5 axb5
4 �a4 lLlf6 14 :Xa8 i.xa8
5 0-0 �e7 After 14 ... :xa8 1 5 lLlg5 the
6 :et b5 black knight is forced to retreat to
7 �b3 d6 the back rank allowing White to
8 c3 0-0 seize the centre: 1 5 ... lLld8 16 lLldf3
9 h3 �b7 exd4 ( 1 6 . . . h6 1 7 lLlxf7 ! lLlxt7 1 8
10 d4 :e8 dxe5; 1 6 ...c 5 17 dxe5 dxe5 1 8
This rook move was brought 'it'xd7 lLlxd7 1 9 lLlxf7 ! c4 20 lLlxd8
into practice by my long-time sec­ :xd8 2 1 i.a2 :e8 22 �e3 and
ond Igor Zaitsev, and the variation White had a clear advantage in
is therefore named after him. Black lvanchuk-Portisch, Linares 1990)
fortifies the centre, and does not 17 e5 !.
spend time on the prophylactic 15 d5
... h7 -h6. True, there is a danger that Now, as he has taken on aS with
White will repeat moves with 1 1 the bishop, after 1 5 lLlg5 Black can
lLlg5 :f8 1 2 lLlf3 :e8 1 3 lLlg5. reply 15 ... :e7, and then ...h7-h6.
Thus, if he needs a victory, Black 15 ... lLla5
should choose something else. In game 46 of our first match I
11 lLlbd2 i.f8 retreated with 15 . . .lLld8, and after
B4 Mo1cow Wch (5) 1985

1 6 �fl h6 1 7 ltl3h2 ! ltlb7 1 8 i.c2 blocked in by its own b7-knight

ltlc5 1 9 b4 ltla6 20 ltlg4 ltlh7 2 1 and c6-pawn, will quickly free it­
ltlg3 c 6 2 2 dxc6 i.xc6 2 3 i.b3 self and develop energetic activity,
ltlc7 24 W'f3 ltle6 25 h4 W'd8 26 and precisely thanks to his bishop
l:.dl White achieved an advantage. Black will gain the upper hand.
16 i.a2 c6 18 c4
17 b4 ltlb7 (D) The standard 1 8 ltlfl is better,
and after 1 8 ... cxd5 ( 1 8 ... c5 19 i.g5
i.e7 20 ltlg3 g6 21 'ii'd2, preparing
.i.�
-
� �
� -
� .I -
� �·� ·
%%
� ltlh2 and f4) 1 9 exd5 h6 20 ltlg3
···'ii' · · · · l:.c8 21 W'd3 ltld8 ! 22 ltlh2 g6 23
�
-
�
�·� - - -- - �
�
� h4 White maintains the initiative.
-
�
��·· �8-
�
� - •
-
18 ••• l:.c8! (D)
Before retreating the knight to
n
u �8·
�
-
� - � - d8, it makes sense to place the rook
�
-
�%% � -� �
7,� /";"\. . 8
-
�"LJ�
!ii on a more promising file.
i.�-
� �
� � u �8. -
. - - �
.� �-� �
w
After 17 ...ltlc4 1 8 ltlxc4 bxc4
1 9 i.g5 ! cxd5 20 i.xf6 dxe4 2 1
ltlxe5 W'a7 22 ltlg4 'ii'xa2 23 ltlh6+
�h8 24 ifh5 gxf6 25 1li'xf7 i.xh6
26 '1Vxe8+ �g7 27 'ii'e7+ White
has the advantage, since the enemy
bishop has no prospects. However,
now it seems that Black is all but
suffocating from a lack of vital
space. But this impression is de­ 19 dxc6
ceptive. Nimzowitsch once said Handing over the centre allows
that a piece which has been in cap­ Black immediately to gain a play­
tivity for a long time finds incred­ able game. Fascinating events arise
ible strength when it has broken after 19 W'e2 ltld8 ! 20 i.b2 bxc4
free. In fact, in this game my aS­ 2 1 ltlxc4 W'a7 ! (2 l . ..l:.b8 ! also
bishop, hiding in the corner of the gives him good prospects) 22 l:.a1
board, and for the time being cxd5 23 exd5 i.xd5 24 ltlxd6
 Kasparov - Karpov 85

�xd6 25 �xd5 1fxa1 + 26 �xa1 More precise than 24... ..,c2 25

l:tcl + 27 �h2 l:txa1 and Black has 1fa1 ...a4 26 lbd4 lbc5 27 l:te3
enough compensation for his small with a sharp game.
material loss. 25 lbb3?
19 ••. '1Vxc6 After 25 1fb1 White's position
20 c5? would still have been defensible.
Instead of this highly risky ven­ The desire to generate some activ­
ture, 20 �b2 straight away would ity has led him into a difficult posi-
have been better. As it is, the threat tion.
to f7 (20 ... dxc5 2 1 �xf7+) is eas­ 25 ... lbcS!
ily repulsed, and Black seizes the 26 .i.a1
initiative. A forced pawn sacrifice. After
20 ••. lbd8 26 lbxc5 it'xb2 ! 27 l:te2 it'a3 28
21 �b2 dxcS! lbd3 .i.xe4 29 .i.xf7+ lbxf7 30
22 bxc5 l:txe4 b4 Black has a clear advan­
Black also has good chances tage.
after 22 lbxe5 1fa6 ! 23 ..,al c4 24 26 ... .i.xe4
�c3 1fa3 25 l:te3 .i.xb4 26 lbexc4 27 lbfd4 (D)
bxc4 27 �xf6 c3 ! 28 l:tg3 lbe6, or Black would also have won ef­
alternatively 22 .i.xe5 lbd7 23 fectively after 27 lbg5 .i.c2! 28
.i.b2 c4. 'ii'xc2 'ii'xe 1 + 29 �h2 .i.d6+ 30 g3
22 ••. •xc5 lbe4.
23 .i.xeS lbd7
24 .i.b2 �4! (D)
86 Moscow Wch (22) 1985

29 lL:!xcS 'ii'xcS In game 2 1 , when Kasparov had

30 'ii'g4 .:.e8 White, after 5 ... c6 6 e3 ..tf5 7 g4
31 .:.dt ..tg6 .te6 8 h4 lDd7 9 h5, I introduced a
32 'ii'f4 'ii'b 4! novelty, 9 ... lDh6, but did not man­
33 'ii'c l ..te4 age to solve my opening difficul-
34 .:.et 'ii'a5 ties.
35 ..tb3 'ii'a8 6 e3
36 'ii'b2 b4 In game 20 I stopped the bishop
37 .:.eJ ..tg6 coming out to f5 with 6 'i!fc2 but
38 .:.xe8 'ii'xe8 the time I lost in doing this led to
39 'ii'c l lbe4 rapid equality: 6 ...0-0 7 e3 c5 ! 8
40 ..td5 lDc5 dxc5 ..txc5 9 lDf3 lDc6 10 .te2
41 lDb3 lDd3 d4! .
0-1 6 0-0
The final move was sealed. In 1 lDf3 ..trs
the adjourned position Black' s 8 h3
passed pawn guarantees him vic­ Preparing to carry out the stand­
tory and Kasparov considered that ard manoeuvre in this position,
there was no point in resuming. g2-g4. This rook's pawn move had
until that time not been used with
Game 1 9 this move order.
Karpov - Kasparov 8 c6
Moscow Wch (22) 1985 9 g4 .tg6
QGD, Exchange 10 lDe5 lDfd7 (D)

For the third time in a row in the

match a sharp form of the Queen's
Gambit Exchange Variation arose.
In the first two games White had
the initiative, but they ended in
draws. This time, however, I man­
aged to win.
1 d4 d5
2 c4 e6
3 lDc3 ..te7
4 cxd5 exd5
5 ..tr4 lDr6 w
 Karpov - Kasparov 87

After 10 ... lL'lbd7 1 1 h4 Black ( 1 8 ...gxf4 19 tt:'Jxf4) 19 fxg5 ..txg3

feels rather hemmed in. 20 tt:'Jxg3 'iixg5 21 e4.
1 1 tt:'Jxg6 fxg6 18••• 'iixd6
Following the banal 1 1 .. .hxg6 19 f4
Black is deprived of all possible White's kingside pawn advance
counterplay. guarantees him a definite initia-
12 ..tg2 tt:'Jb6 tive.
13 0-0 �h8 19••• gxf4
1 3 . . .tt:'Ja6 is not sufficient for 20 exf4 .:Z.ae8
equality because of 14 'ikb3 ! ; nor is 21 f5 (D)
1 3 . . . g5 satisfactory, due to the re­ 2 1 .:Z.f3 tt:'Jd7 22 g5 followed by
ply 14 ..tg3 i.d6 1 5 ..txd6 1i'xd6 h4 and lL'lg3 is probably more accu­
1 6 e4 ! . rate.
1 4 tt:'Je2 (D)
Both 1 4 'ikb3 and 1 4 ..tg3 fol­
lowed by e3-e4 would have been
promising for White.

B
21 tt:'Jc7
22 .:Z.f2 tt:'Jd7
23 g5 'ike7
B 24 h4 'iie3
14 g5 25 .:Z.d1 tt:'Jb5
15 ..tg3 ..td6 26 'iixe3 .:Z.xe3
16 'iid3 tt:'Ja6 27 �h2! tt:'Jb6
17 b3 'iie7 28 tt:'Jg3 tt:'Jc8
18 ..txd6 29 tt:'Jn .:Z.e7
The immediate 1 8 f4 would have 30 l:td3 tt:'Jcd6
led to a sharp struggle: 1 8 ....:Z.ae8 31 tt:'Jg3
BB Moscow tt 19B6

Here I was preparing to move 36 �xh5

my king to g4 and force back the Black was possibly counting on
knight with a2-a4, maintaining the 36 �f4 llexf5+ 37 t:bxf5 lbxf5
pressure. Obviously, Black should with a probable draw (but not 3 8
also bring his king up to the cen­ �xe4 due to 3 8 ...liJd6+ 39 �e5
tre with 3 l ...�g8, or activate the llxf2).
rook by means of 3 1 ...llel . Instead 36 t:bxf5
of this Kasparov, in severe time 37 :xf5 llfxf5
trouble, made a serious mistake. 38 t:bxf5 llxf5
31 ... t:be4? (D) 39 :Xe4 �h7
40 lle7 b5
In the event of 40...llb5, White
. . . - can achieve his goal with 4 1 a4
.,-
- . -
� - �. llxb3 42 g6+.
··- . . 41 llxa7 b4
42 �g4 1-0
··· · · 8 �-�
� �
The adjourned rook ending is
.
� ���- � easily winning for White, as the
•8•:• � mobility of the h-pawn is decisive.
8 • • ail.=
Game 20
• • • • Karpov - BeUavsky
w Moscow tt 1 9B6
32 �xe4! dxe4 QGD, Orthodox
33 lle3 t:bxd4
34 �h3 1 d4 t:bf6
By 34 f6 ! gxf6 35 gxf6 White 2 c4 e6
achieves a strong position, for ex­ 3 m d5
ample 35 ...lle6 36 llxe4. But be­ 4 t:bc3 �e7
fore moving into decisive action, I 5 i.g5 h6
decided to activate the king. 6 i.xf6 i.xf6
34 •.. lieS 7 e3 0-0
35 �g4 h5+? 8 :et c6
Kasparov gives an impulsive 9 i.d3 t:bd7
time-trouble check; 35 . ..�g8 or 10 0-0 dxc4
35 . . .llfe8 was necessary, defend­ 11 i.xc4 e5
ing obstinately. 12 h3! (D)

B w
Both 12 .i.b3 and 1 2 li)e4 used
to be played, but in both cases con­
vincing routes to equality have
been found for Black. This modest,
but poisonous peripheral pawn
move was first used by Kasparov in
the 23rd game of our second
match. The idea behind it is that the
light-squared bishop for the time
being stays where it is, and so after
1 2 ... l:leS he still has the possibility
of 1 3 'iib 3. Besides this, the g4-
square is taken under control, and
after opening the centre the cS­
bishop will have problems devel­
oping. Thus Black finds himself in
a peculiar zugzwang.
12 exd4
13 exd4 li)b6
14 .i.b3 .i.f5 (D)
In the above-mentioned game
against Kasparov I continued with
14 ... l:leS and brought the bishop
out on the next move. After 1 5 l:le1
.i.f5 16 l:lxeS+ 'fixeS 17 'iid2 'iid7
Karpov - Beliavsky 89
1 S l:le1 l:ldS 19 'iif4 li)d5 20 li)xd5
cxd5 2 1 li)e5 .i.xe5 22 l:lxe5 .i.e6
23 'it'e3 White gained an advan­
tage.
15 l:le1 aS
In the event of 15 ... l:leS 1 6
l:lxeS+ 'fixeS 17 'iid2 the game fol­
lows the variation examined above.
16 a3 'iid7
In game 22 of the return match
with Kasparov I replied 1 6 . . .l:le8,
and after 17 l:lxeS+ ..xe8 1 8 'iid2 I
tried a new idea - 1 S ... li)d7. How­
ever after 1 9 1i'f4 .i.g6 20 h4 'ifdS
2 1 li)a4 I ran into difficulties. An
important novelty was used in
Gurevich-Van der Sterren, Balcu
1 9S6: 1S ...'ifd7 19 l:le 1 l:leS ! 20
l:lxeS+ 'fixeS 2 1 ..f4 .i.e6 ! , and
Black equalized.
17 li)eS .i.xe5
18 l:lxe5 l:lfe8
19 1i'e2 l:lad8
Black has decided that there is
no hurry to push his rook's pawn.
90 Moscow tt 1 986

AxeS
a4 (D)

w
When I played 27 d5, I envis­
aged 29 ll:!e7+ here, but while I
w was checking the variation I turned
It looks as though Black's delay my attention to the possibility of
in pushing his a-pawn makes no 29 .. .llxe7 ! 30 'it'xe7 lbxb2 3 1 'it'e8+
difference, and White must remove �h7 32 'it'xb5 ll:ldl 33 'it'c4 f5 with
the bishop to a2, when the variation a study-like draw: the knight leaps
22 ll:!xa4 ll:!xa4 23 .i.xa4 'ibd4 24 between b2 and d l , and Black
'it'xf5 'it'xa4 cannot worry him. saves himself. In the event of 34 a4
However, White finds an attractive lDb2 35 'it'c6 ll:!dl the queen has to
intermediary manoeuvre. return to c4, as after 36 aS .i..e 4!
22 'it'cS axb3 and then 37 . . . b2 Black actually
23 :e7 'it'd6 takes the upper hand.
Yet another zwischenzug. Before 29 'it'c3 :ds
he gives back the piece, the white 30 lDf4 Adl+
rook penetrates the seventh rank. 31 �b2 ll:!d2
Here 23 . . .'it'xd4 loses because of After 3 1 ...Abl 32 a4 ll:!xb2 33
24 Ae8+! �h7 25 'it'xf5+, while if axb5 the pawn cannot be stopped.
23 ...'it'c8, then 24 'it'xb6 with an 32 b4
obvious advantage for White. Preventing perpetual check by
24 Ae8+ lhe8 the knight.
25 'it'xd6 ll:!c4 32 .i.f5
26 'it'b4 b5 33 f3 g5
27 dS! cxd5 34 bxg5 bxg5
28 ll:!xd5 .i.c2 (D) 35 ll:!e2 (D)
 Karpov - Kasparov 91

Avoiding the final trap - 35 lbb5 8 cxd5 t'Dxc3

t'Dfl+ 36 Wgl t'Dg3+ and White's 9 'ilt'd2 'ilt'xal
knight is lost. 10 bxc3 (D)

B
This variation came into prac­
tice after the famous game Petro­
sian-Fischer, Buenos Aires Ct (2)
1 97 1 , which continued 10 ...'1i'a5
1 1 .i.c4 t'Dd7 12 t'De2 �5 1 3 .i.a2
.i.f5 14 .i.xe5 ! .i.xe5 1 5 .!Dd4
tWxc5 1 6 .!Dxf5 gxf5 17 0-0 with a
dangerous attack for White. Later
various lines were proposed for
Game 2 1 Black, the most effective being
Karpov - Kasparov 1 2 ... .!Dxc5 1 3 0-0 0-0 14 f3 e5 ! 1 5
London/Leningrad Wch (5) 1986 .i.g3 b 5 1 6 .i.a2 'li'b6! 17 Wh 1 a5
Griinfeld Defence with sufficient counterplay.
10 ... 'ilt'xd2+
1 d4 t'Df6 Kasparov prefers to exchange
2 c4 g6 queens. Until now this move had
3 lDc3 d5 resided somewhat in oblivion, and
4 .i.f4 .i.g7 only recently have new ideas been
5 e3 c5 found for Black.
6 dxc5 'it'aS 1 1 Wxd2 lbd7
7 .:.et t'De4 12 .i.b5
t2 LDndol'lllAnlngrad Wch (5) 1986

After 1 2 c6 bxc6 1 3 dxc6 lZ1b6 has completed his development,

(or 1 3 . . .le!f6) White has achieved maintaining powerful pressure.
nothing.
12 .•• 0-0
13 -txd7
If 1 3 c6, then 1 3 ...lZ1c5. Now
Black gains the advantage of the
bishop pair, but as will soon be­
come clear, both will be crippled.
13 ••• -txd7 (D)

z • • •••
-··
- -� i.�
-i�
-i
• • ••• w
17 c6!
• o�• • He does not keep the extra
-
�
� -
� � ·
� -
� - pawn, but in returning it, White
-
• �
u "
�� -
��
• will elicit maximum profit by cre­
• mi � -� A � -� ating a passed pawn and limiting
- � � � L.l � �
�
�
- �
- . � .:
. the movement of Black's rook and
light-squared bishop.
w 17 ••• bxc6
14 e4 f5 18 d6 c5
15 e5 e6! ? Now the light-squared bishop
A n important moment. Why finds some freedom, but the dark­
did Kasparov reject the previously squared one begins to suffocate.
tested 1 5 ...l:l.ac8 1 6 c6 bxc6 17 d6 Perhaps the fate of the bishops
exd6 1 8 exd6 l:l.f6 with an advan­ should have been dealt with differ­
tage to Black (Schmidt-Gross, Nal­ ently - 1 8 .. . g5 1 9 -txg5 -txeS 20
eczow 1 984)? The answer is that cS l:l.cb8, and Black maintains
1 5 . . . l:l.ac8 would be met by 16 c4! hopes of counterplay.
l:l.xcS 17 -te3, as in Seirawan­ 19 h4! h6
Adorjan, New York 1 987. 20 le!h3! (D)
16 c4 l:l.fc8 (D) After the more natural move 20
After 1 6 ... gS 1 7 -txgS -txe5 1 8 lZ1f3, the reply 20 . . .-tc6 ! would
lbf3 -tg7 1 9 l:l.b 1 and l:l.he 1 White have led to a complex game with
 Karpov - Kasparov 93

B B
chances for both sides. However, I 23 lbf2 a2
managed to find a most logical 24 lbd3 lla3
solution to the position. The white 25 llal g5
knight is concentrating on the only 26 hxg5 hxg5
true trajectory towards its optimal 27 .i.xg5 (D)
square, d3 ! As soon as it has
reached its goal, the g7-bishop will
be caught in a trap constructed by
the white e5- and d6-pawns and the
f4-bishop. Incidentally, White pre­
vents the freeing advance ... g6-g5.
20 ••• aS
The passed a-pawn is not dan­
gerous, so Black should have im­
mediately occupied another file
with 20...llcb8.
21 f3 a4
22 llhel ! (D) w
Overprotecting the e5-pawn. 27 ••• rM7
After the hasty 22 t:Df2 Black can After 27 . . . llb8 the duel might
muddy the waters with 22 ... g5 ! 23 conclude with a study-like draw:
hxg5 hxg5 24 .i.h2 f4 25 lbd3 .i.e8 28 .i.f4 llbb3 29 lbxc5 llb2+ 30
26 .i.g1 .i.g6 27 lbxc5 .i.xe5 28 <itc1 llxg2 3 1 .i.d2 .i.h6 ! 32 .i.xh6
lbxe6 .i.xd6 29 lbxg5. llc3+ 33 �d 1 lld3+! 34 lbxd3
22 ••• a3 .i.a4+ 35 �c 1 llc2+ and Black
� I.IMtlll£lrtlrtii'IMI Wch (17) 1986

diUwn perpetual check. But a

ample refutation to this 'study' can
bt found: 28 �e2! l:tbb3 29 lLlxc5
lb2+ 30 �fl and White has the
upper hand.
28 .i.f4 l:tb8
29 l:tecl .i.c6
30 l:tc3 .:taS
31 l:tcl l:tba8
32 lLlcl 1-0

Game 22 w
Karpov - Kasparov The critical position in the
London/Leningrad Wch (1 7) 1986 Smyslov Variation. Almost every­
Grilnfeld Defence thing has been tried here, includ­
ing 12 'it'xd6, 1 2 lLlb5, 1 2 d5, 1 2 h3
This miniature is a classic example and 12 0-0. In the key game Bot­
of a battle which was decided en­ vinnik-Fischer, Vama OL 1962, af­
tirely in the process of home prepa­ ter 12 h3 .i.xf3 1 3 gxf3 .:tfd8 14 d5
ration. White's novelty cannot be lLle5 1 5 lLlb5 'it'f6 1 6 f4 lLled7 17
called unusual - it consists of a e5, Fischer gave Botvinnik an un­
simple one-square advance by the pleasant surprise with 17 ...'ifxf4 ! .
h-pawn, but careful analysis has However, after mistakes b y both
shown that after this move Black's sides this fascinating game eventu­
defence becomes noticeably more ally ended in a draw. However, it
difficult. turns out that White has at his dis­
1 d4 lLlf6 posal one other interesting move,
2 c4 g6 which has not been tested in prac­
3 lLlc3 dS tice until now.
4 lt:lf3 .i.g7 12 eS!
5 'iWb3 dxc4 At first glance this looks para­
6 'iWxc4 0-0 doxical. White not only hopelessly
7 e4 .i.g4 weakens his e5-pawn, but also al­
8 .i.e3 lLlfd7 lows an exchange of queens. How­
9 .:tdl lLlc6 ever, things are not so simple as
10 .i.el lLlb6 that.
11 'iVcS 'iWd6 (D) 12 ... 'iWxcS

13 dxcS �c8 (D)
In reply to the more natural
1 3 ... �d7 White has 14 h3 ! .txf3
1 5 gxf3 ! and the 'doomed' e5-
pawn is invulnerable, as if it is cap­
tured by either knight, 1 6 f4 ! wins
a piece. Otherwise, White will
strengthen the e5-pawn with its
neighbour, and gain a clear advan­
tage.
14 h3!
In game 1 5 , where the novelty
1 2 e5 was used for the first time, I
played 14 �b5, and after 14 ... :b8 !
1 5 �xc7 e6 ! Black managed to
equalize completely. In fact, the
threat to surround the knight by
means of 1 6...a6 forces White to
lose a tempo by playing 1 6 �b5,
and the black c8-knight quickly
transfers to a comfortable position.
The result of that encounter obvi­
ously satisfied my opponent, who
thought that White had no way of
Karpov - Kasparov 95
resurrecting his idea. Meanwhile,
when I again set up the position on
the board, I managed to spot that
the c8-knight's prospects can be
significantly limited.
14 ••• .t:xf3
If 14....te6, then 1 5 �g5 looks
unpleasant: 1 5 ...�xe5 1 6 �xe6
fxe6 1 7 f4.
15 .txf3 .txeS (D)
After the alternative capture
Black also runs into difficulties:
15 ...�xe5 16 .txb7 :bs 17 c6
�4 1 8 :d7.
w
16 .txc6! bxc6
17 .td4 .tf4 (D)
After an exchange of bishops
with 17 ....txd4 1 8 :xd4 :bs 1 9
b 3 a5 the white rook can penetrate
the seventh rank, but after 20 :d7
(20 :a4 is stronger: 20 ...:as 2 1
�e2 :ds 2 2 :d 1 :xd 1 23 �xd 1
followed by the transfer of the
knight to c4) 20... �a7 21 :xc7
96 London/Leningrad Wch (17) 1986
lllb� 22 ltlxb5 .:.xb5 23 .:.xc6 a4!
Black has counterplay.
w deny White's pieces access to e4,
18 0-0 aS?!
It was necessary to move the e­
pawn. My preparation showed that
after 18 ...e5 19 .i.e3 (19 lbe2 lbe7!)
19 . . ..i.xe3 ( 1 9 ... lbe7 is also possi­
ble) 20 fxe3 White has a definite
advantage. However, soon after the
match this position was seen in
Karpov-Timman, Tilburg 1 986,
and the Dutch grandmaster proved
that Black has chances to hold the
position. The game continued thus:
20 ... lbe7 21 .:.d7 lbf5 22 .:.xc7
:res ! 23 .:.d7 :adS 24 .:.fd 1 .:.xd7
25 .:.xd7 lbxe3 26 .:.c7 .:.bs 27 b3
:dB with a swift draw. True, the
feeling remains that somewhere
White could have played some­
thing stronger.
19 :rei
White plans to dominate the
centre. The move g2-g3 will chase
away the bishop, whereupon .i.e5
and .:.d7 will make inroads into
Black's position. In view of the
atrocious position of the knight on
c8, Black has great difficulty op­
posing this plan.
19 ... a4?!
The pawn continues its march,
but this is fiddling while Rome
bums. 1 9 ...e6 would not help in
view of 20 lbe4 when White's
pieces co-operate superbly. It is
hard to see what Black should actu­
ally play here - maybe 1 9 .. .f5 to
but to commit oneself to such an
extreme measure would be diffi­
cult at the board.
What is not in doubt, is that the
pawn advance chosen by Kasparov
carries no real threat, and consti­
tutes a fatal waste of time.
20 .:.e4!
Originally I had intended 20 a3,
to fix the pawn on a4 so as to gang
up on it. However, it occurred to
me that I could ignore the pawn;
especially since if (following 20
a3) the white knight captures on
a4, his black counterpart could
spring out to b5 via a7, with some
chances of real resistance. Clearly
there would be no point complicat­
ing matters in this way, when sim­
ply centralising my pieces is so
strong.
20 ... .i.h6

21 i.e5
Black's planned . . .liJa7 is ren­
dered rather ineffectual by this
move.
21 ••• a3
22 b3
Clearly the best move.
22 ... !lJa7
Although this leads nowhere,
there were already no sensible
moves for Black.
23 l:td7! (D)
23 i.xc7 would regain the pawn
and hit e7, but then 23 ...e6 or the
more precise 23 ... i.g7 would pro­
vide counterchances. Thus I pre­
ferred instead to keep full control
of the long diagonal.
Karpov - Kasparov 97
Black seems to want to justify
his a-pawn's moves by using the
b2-square. However, White can
evict the bishop.
24 l:txc7
The e7-pawn is the next item on
the menu.
24 ... i.b2
Black would do no better with
24 . . .!iJb5 25 l:txc6 or 24 . . .e6 25
l:ta4. Now White must be precise.
25 !lJa4!
This move guarantees decisive
material gains, and is the clearest
way to win. Instead the careless 25
l:txe7 would allow Black counter-
play with 25 ...i.xc3 26 i.xc3 !iJb5
and then . . .l:td8, eyeing the a2-
pawn. Although White is still bet­
ter, Black is in the fight again.
25 ... !iJbS
After 25 ...i.xe5 26 l:txe5 e6
White wins an exchange at least.
26 l:txc6 (D)
98 London/Leningrad Wch (19) 1986

White simultaneously creates a to the system brought into practice

passed pawn and knocks the sup­ by Ragozin, which Najdorf then
port away from the b5-knight. 26 worked on in detail. But here as
lbe7 would be less logical, since a well an opening surprise awaited
knight firmly planted on b5 would Black.
guarantee counterplay for Black. 1 d4 lL!f6
26 ••. l:lfd8 2 c4 g6
Black's moves are forced. 3 lL!c3 d5
27 l:lb6! 4 lL!f3 i.g7
This is the main point of White's 5 ..,3 dxc4
plan. Black's knight has had an un­ 6 Wxc4 0-0
fortunate career in this game, and it 7 e4 lL!a6
is now hounded out of the one good In the two previous odd-num­
square it has found. bered games in this match Black
27 ••• l:ld5 preferred 7 ...i.g4.
28 .i.g3 8 J.e2 cS
I avoided a last, though transpar­ 9 d5 e6
ent, snare here: 28 lL!xb2? :Xe5 29 10 0-0 (D)
l:lxe5 axb2 would even win for 10 i.g5 exd5 1 l lL!xd5 .i.e6 1 2
Black, e.g. 30 l:le 1 l:lxa2 followed 0-0-0 .i.xd5 1 3 l:lxd5 'it'b6 i s not
by 3 1 ...l:la l . dangerous for Black.
Now the advance of the c-pawn
decides the game quickly.
28 lL!c3
29 lL!xc3 .i.xc3
30 c6 .i.d4
31 l:lb7 1-0

Game 23
Karpov - Kasparov
London/Leningrad Wch (19) 1986
Griinfeld Defence

Despite his disappointment in the B

Smyslov Variation (game 17), Kas­ 10 exd5
parov did not abandon the Griin­ 11 exdS .i.£5
feld Defence. This time he turned 12 i.f4 (D)
 Karpov - Kasparov 99

In the Seville match I twice

played 1 2 l:td1 l:te8 1 3 d6 h6 and I. B . I. B*B
then 1 4 h3 (game 1 5), or 1 4 .i.f4 - �
-·-
�
� �-
�
� ··
-·
(game 2 1 ). Both encounters ended �B B B i B
in a draw.
BlD. 8 B.t.B
B•B�fl B
B B BlDB
� .t�
8 � �-
� �8�6
�
-
� -
� �
� .: �
-
�:�
� �
B
use its potential energy, he will
have the initiative, while if Black
manages to set up a reliable block­
ade, his position will become more
promising. Having led the knight
B to b5, I was ready to part with the
12
••• :es b-pawn in order to develop pres­
1 2 . . ....b6 and then 1 3 h3 (or 1 3 sure in the centre after 14 . . . .i.xb2
.i.e5) has independent interest. 15 d6 .i.f6 16 .td3 ! and 17 l:tfe2 .
Here a few practical examples have Kasparov refused the offer.
accumulated, but it is too early to 14 ••• ...f6!
make a definitive judgement. 15 .td3 l£Jb4
13 l:tad1 l£le4 After 15 ....tg4 16 .te5 l:txe5 17
Here 1 3 . . .,..b6 has also been l£lxe5 .txd1 1 8 .txe4 irxe5 19
tried, and then 14 ,..b5 (or 14 l£lh4) l:txd 1 White has gained the initia­
with chances for both sides. tive. Not wishing to condemn him­
14 ltJbS! (D) self to passive defence, Black risks
The leap by the knight to b5 was everything, but the exchange sacri­
prepared at home. Until this point fice is not justified. The capture
only 14 .i.e3 and 14 .i.d3 had been 1 5 ......xb2 is definitely risky, al­
tested, and White had achieved though later practice in the vari­
nothing. ation showed that Black can gain
The basic strategic conflict in counterplay by continuing with
the middlegame is centred around either 15 ...l:tad8 or 15 ....td7.
the d-pawn. If White manages to 16 l£lc7! l£lxd3
lOO LondonllAralragrad Wch ( 19) 1986

17 lllxe8 l:.xe8 'ii'e 3 realizing the material advan­

18 'ii'x d3! (D) tage is not difficult, but in the event
of 1 9 .. .'ii'xa2 20 Wb5 l:.d8 2 1
'ii'xb7 'ii'xd5 Black holds on: 22
'ilfxd5 .:Xd5 23 g4 lbf6!.

18 ... 'ii'xb2
Black's misfortune is that nei­
ther 1 8...lllxf2 nor 1 8 ...lllg3 works w
because of 1 9 'ii'b 5, attacking the 20 lbd2!
undefended rook on e8. Playing for the pin - the main
19 l:.de1 motif of White's idea.
The straightforward 19 d6 (inci­ 20 ••. 'ii'a4
dentally threatening to play 20 g4) 21 'ilfc4
19 ...lld8 ! 20 We3 h5 2 1 llb 1 'ilfxa2 Forcing an exchange of queens;
22 l:.xb7 'it'd5 23 l:.xa7 lbxd6 24 21 lbxe4 is not so clear: 2 1 . . .l:.xe4
.i.xd6 'ilfxd6 25 lbg5 is unpleasant 22 l:.xe4 .i.xe4 23 'i!fd2 c4 with
for Black, and it is therefore prob­ counterplay.
ably worth playing. However, I de­ 21 ••• 'ilfxc4
cided to pin the knight, limiting the 22 lbxc4 .i.c3
enemy pieces' activity. Of course 22 ...b5 23 lbd2 lbf6 24 l:.xe8+
19 l:.fe1 does not work in view of lLlxe8 25 l:e1 , etc., loses immedi­
1 9 . . .'ilfxf2+ 20 �h1 'ilfxel + ! 21 ately.
l:.xel lbf2+ 22 �g1 l:.xe1+ and 23 liJd2 .i.xd2
23 ... lbxd3. 24 .i.xd2 .i.d7 (D)
19
..• 'i!fb4? (D) During the game many people
This is the decisive mistake. Af­ thought that Black had got himself
ter 1 9 . . . lbf6 20 l:.xe8+ lbxe8 2 1 out of his mess: the d2-bishop is
 Karpov - Kasparov 101

under attack, and . . . .i.b5 is threat­

ened.

- - �. -·­
· · --*- - · - ·
- - -·-
- ·8- -
- -·- -
-
A �� -%m·;; -"'B
. "" �-
A � �
LJ � %

�
0+
�
LJ
, y B��
/�
'""'
!.;.;
% -""'
%- , !/;::"
37 �e2
w
�f7
.
. . � .: �
. � � 38 ..td3 �e6
w 39 �c4 lLle5+ (D)
25 .i.f4! .i.b5
26 f3! g5
After 26 . . . .i.xfl 27 'itxfl lLlf6 - - - -
28 l:lxe8+ lLlxe8 29 .i.e5 ! the d­ -
- . -
- . -••
pawn finally shows what it is capa­ � •.•
- - - •
�
ble of: 29 . . . f6 30 d6 ! and he has to
give back the piece.
.
. .
. - �
-

27 .i.xg5 .i.xn -<&it- - -

27 . . . lLlxg5 is no better after 28 • - .8.
.:xe8+ .i.xe8 29 h4 ! , and White 8. - •
- • -8u"
wins the knight.
28 �xn lLJd6 - - - .
29 .i.e7 lLlc8 (D) w
29 . . . lLlc4 30 d6 lLlb6 3 1 .:bl 40 �d4 lLlc6+
lLld7 32 .:xb7 is also bad. 41 �c4
30 .i.xc5 .:d8 Here the game was adjourned,
31 .:eS f6 and Black resigned without resum­
32 .:r5 b6 ing. After 4 1 . ..lLle5+, both 42 'itb5
33 .i.d4 lLle7 and 42 ..td4 lLlc6+ 43 ..te4 lead to
34 .i.xf6 .:xd5 victory, as the white pawns are ir­
35 .:g5+ .:xg5 repressible.
36 .i.xg5 lLlc6 1-0
102 DNIHJI OL (14) 1986

Oame 24 therefore in reply to h2-h4, the

Karpov - Sznapik symmetrical . . . h7-h5 has now
Dubai OL (14) 1986 gained popularity as a means of
Sicilian Dragon hindering the movement of the
white g- and h-pawns.
This encounter took place in the 1 3 ..tgS (D)
last round of the Olympiad; our
team could not afford to drop even
half a point, so this game was deci­
sive. By winning the match 4-0, we
won the Olympiad, ahead of Eng­
land by only half a point.
1 e4 c5
2 lDf3 d6
3 d4 cxd4
4 lDxd4 lDf6
5 lDc3 g6
6 ..te3 ..tg7
7 f3 0-0 B
8 Wd2 lDc6 Attempting to blow up the en­
9 ..tc4 ..td7 emy fortress immediately with the
10 0-0-0 lDeS help of 1 3 g4 is unsuccessful, and
1 1 ..tb3 l:tc8 the bishop move to g5 looks more
12 h4 solid than ideas such as 13 ..th6, 1 3
Sometimes 12 ..tg5 is played im­ l:tdgl and 1 3 �bl .
mediately, but obviously, including 13 ••• l:tc5
the moves h2-h4 and . . .h7-h5 is in 1 3 ... lDc4 and 1 3 . ..lDh7 are less
White's favour for a future attack convincing.
on the kingside. 14 �b1
12 ..• h5 This waiting move, which has a
Black previously did not prevent definite idea that will become clear
the h-pawn from moving. Its sac­ later, was suggested to me by Gel­
rifice - 1 2 ... lDc4 1 3 ..txc4 l:txc4 14 ler while I was preparing for my
h5 ! ? - has been seen countless match in Merano. I had to wait five
times, but a definitive conclusion whole years to use the idea in a
has not yet been made. However, it game !
most often favours White, and 14 ••• b5 (D)
 Karpov - Sznapik 103

analogous position has arisen in

practice several times, but without
the moves �c 1 -bl and ...b7-b5. In
that case the manoeuvre 1 6 l£Jd5
(now it is made a move earlier) af­
ter 16...l:lxd5 ! 17 i.xdS 'irb6 is not
dangerous for Black in view of his
pressure on b2 ( 1 8 ...l£lxf3 is threat­
ened). Including the moves �b1
and ... bS seriously changes the
situation - the exchange sacrifice
w on d5 is now insufficient, as the b­
15 g4! file is covered by his own pawn,
Until that time 1 5 l:lhe1 was and after 17 ...l:lxd5 1 8 i.xd5 gxf3
played here, simply centralizing. 19 lbr5 it is only one step to victory.
15••• hxg4 17 ••• l:le8 (D)
A fundamental reply, which we
looked at in 1 98 1 . Some rounds
earlier in the USSR-Bulgaria match,
��- � -�
�- ·ri@;· �
� ·
�
Kiril Georgiev chose another prob­
-
- � �- �'-
ri@; -
... -
lematic continuation against me, • �
. - � - -
ri@; ' �
�
�
1 5 ... a5, which we had not had time
to prepare for. The correct reaction
B
• l. ��- m•
-"l.J- -
was 1 6 gxh5 ! , but I went for the . ��-··
more modest 1 6 i.xf6 i.xf6 1 7 a3 Bi.. B � B
and the game soon ended in a draw.
16 h5! l£lxh5
�"
u�- pm � ri@; �
�- -
�
Obviously, it was worth Black B<t>B : • • :
giving up the exchange immedi­ w
ately - 16 ...l:lxc3 ! ? 17 bxc3 l£lxh5 1 8 l:lxhS!
1 8 l:lxh5 gxh5 1 9 'ifh2 lbc4 20 White sacrifices the exchange
'ifxh5. White has a certain initia­ himself, hopelessly weakening
tive, but the whole game still lies the guard around the enemy king.
ahead. 18 •••
gxhS
17 l£Jd5! 19 'ifh2! (D)
Now the idea behind White's This precise position was the
1 4th move becomes obvious. An subject of our old analysis. Then
104 Dubai OL (14) 1986
we established that White's small
material outlay is repaid with inter­
est when the black king is attacked.
Because the defence cannot man­
age without ... e7-e6, Black should
move his queen from d8, and we
considered it necessary here to re­
turn the exchange - 1 9 ... llxd5 20
.i.xd5 (20 exd5 l£lxf3 21 l£lxf3
gxf3 22 'ibh5 .i.f5 is not so clear)
20 . . . 1ib6. Now if 2 1 'ifxh5, then
2 l . . . e6 is possible, with a resilient
defence, but Black's fortress can be
assaulted by 21 l£lf5 ! .i.xf5 22 exf5
ltJxf3 23 'ii'xh5 (the pawn must be
taken at precisely this moment; if
23 .i.xf3 gxf3 24 'ifxh5, then
24...'ifc5 !) 23 ...ltJxg5 24 'ifxg5, and
pressure by White on the g- and h­
files threatens the most unpleasant
consequences. 19 ... a5 is terrible
for Black, for example 20 'ifxh5 a4
2 1 llh 1 axb3 22 'ifh7+ �f8 23
.i.h6 .i.xh6 24 'iVh8#.
19 ... llc4
Black prefers to give the rook
back for the bishop rather than the
knight, but it does not ease his
problems.
20 .i.xc4 bxc4
21 'ii'xhS f6
22 f4! liJf7 (D)
He will be saved by neither
22 . . . e6 23 l£lxf6+ .i.xf6 24 fxe5
.i.xg5 25 1i'g6+ nor 22 ...fxg5 23
fxe5 dxe5 24 llh1 (24 l£lf5 ! is even
stronger for White) 24 ... exd4 25
'i'h7+ �f7 (25 . . .<it>f8 26 'ii'g6 ! ,
threatening both 27 l:l f l+ and 27
l:lh7, gives White a decisive attack)
26 llfl + <it>e6 27 'ii'xg7 .i.c6 28
'ifxd4.
23 .i.b4!
There is no sense in rushing, as
Black is totally helpless.
23 ..• 'ifb8
Again, if 23 ...e6, then 24 l£lxf6+
.i.xf6 25 1i'g6+ is decisive; after
 Karpov - A. Sokolov 105

23 ...lbh6 there is 24 f5 l:tf8 25 lbf4 8 cxd5 lbxd5 9 .tg2 0-0 10 lbxd5

.te8 26 lbg6 .txg6 27 fxg6 fol­ exd5 1 1 0-0 lbd7 a position arose
lowed by lbf5. which was well known from the
24 l:th1 c3 first Karpov-Kasparov match. In
25 b3 'ifb7 situations like this, White's altered
26 r5 lbeS move order (.tg2 and lbc3) does
27 lbe6 1-0 not matter. But now, after Black
If 27 . . ..txe6, then 28 'iWxe8+, has already castled kingside, the
and otherwise there is no defence position of the knight on c3 is not
against 28 .txf6 .txf6 29 'iWh7#. that harmless for him, since White
can seize the centre by means of
Game 25 e2-e4.
Karpov - A. Sokolov 8 e4 d5
Linares Ct (10) 1987 9 cxd5 .txn
Queen's Indian 10 �xn exd5
l l eS lbe4
1 d4 lbf6 12 'iie2 (D)
2 c4 e6 In game 8 I played 1 2 l:tc l , pre­
3 lbf3 b6 venting immediate regrouping by
4 g3 .ta6 means of ...'iid7 and ...lbc6-d8-e6.
5 b3 .tb4+ However, this loss of tempo makes
6 .td2 .te7 itself felt, and Black could gain a
7 lbc3 satisfactory position by playing
It is curious that the Queen's In­ ...c5 and ...lbc6.
dian was played in all five games
where I had White in this match
against Sokolov, and this position
arose automatically. It must be said
that the results exceeded expecta­
tions - I won three games, and the
other two were drawn. However,
my success was due not so much to
good opening play, but to victories
in the ending.
7 ••. 0-0
Other games in the match against
Sokolov continued 7 . . . d5, and after B
IM UMNI a (10) 1�1'1

Thll time I managed to find a has great attacking potential con­

more precise arrangement for my nected with preparing f4-f5.
pieces. The position of the queen 18 -txd4 'W'f5 (D)
on e2 besides everything else has
the important virtue of preventing
... f7-f5; after 12 ...f5 1 3 exf6, the
e4-knight is forced to abandon the
centre, and White has the e-file and
the e5-square at his disposal.
12 �xc3
13 -txc3 'W'd7
14 �g2 �6
The idea behind this somewhat
fanciful knight manoeuvre is to
bring it to e6 to blockade White's
pawn centre. w
15 l:the1 19 l:tad1 -tb4
Introducing a natural plan: 20 :n �e6
White will throw the f-pawn for­ It looks as though Black is on
ward and simultaneously control the point of seizing back the initia­
the e-file. The move 1 5 l:tac l was tive, but, as so often happens, he is
played here previously, allowing a tempo short. If the rook were on
Black to construct a defensive for­ c8, White would not have the pos­
tress: 15 ... �d8 1 6 -tb2 �e6 17 sibility of 21 'ii'd3 due to 2 1 ...l:tc2+
l:thd l l:tae8. 22 l:tf2 'ii'xd3 23 l:txd3 �xd4 24
15
••• �d8 l:txd4 :Xf2+ and 25 ...-tc5 .
16 �g1 c5 21 'ii'd 3!
1 6 ...�e6 is still impossible: 17 Destroying Black's blockade on
f4 f5 18 exf6 l:txf6 1 9 f5 ! wins a the light squares.
piece. In the event of 16 ... f5 17 21 'ifxd3
exf6 -txf6 1 8 f4 followed by �f3- 22 l:txd3 l:tac8
e5, White has the advantage. 23 �f3
17 f4 cxd4 23 f5? is too hasty: 23 ...�xd4 24
Black cannot manage without l:txd4 l:tc2+, when 25 l:tf2? fails to
opening the c-file, and he decides 25 ...l:txf2+ 26 �xf2 -tc5.
to do so immediately. 1 7 . . �c6
. 23 l:tc2+
1 8 �f3 g6 is risky, since White 24 :n :res (D)
 Karpov - A. Sokolov 107

33 lbf3 .:.cl (D)

Obviously, now Black is simply
placing his hopes on a counterat­
tack against the white queenside
pawns. On other parts of the board
he has lost the fight.

25 f5 !
The phalanx of white pawns fi-
nally begins its forward march.
25 lbxd4
26 lbxd4 lbfl+
27 �xfl .:.et
28 g4 �8 w
29 �3 .:.n+ 34 a4 .:.bl
30 �g3 .:.et 35 lbd4 i..e7
Checking from behind achieves 36 h5
nothing: 30....:.g1 + 3 1 �f4 .:.n + 32 I spent a lot of time on this
lbf3 , while Black already has to move. Of course, the continuation
bear in mind the direct threat of 3 1 36 g5 hxg5+ 37 hxg5 would have
lbc6. given me chances for victory, but it
3t �f4 is more useful to fix and 'seal up'
After 3 1 g5, in the event of Black's kingside and his pawns. I
3 I .. ..:.g1+ (or 3 l ...�e8 32 lbf3) 32 am not exaggerating when I con­
�f4 .:.n+ 33 lbf3 things are not firm that at this point I had already
looking good for Black, but after calculated a beautiful combination
exchanging rooks with 3 I . . . .:.c3 32 a dozen moves long. But at the
.:.xc3 .i.xc3 33 lbc6 aS 34 �f4 same time I will not hide the fact
�e8 his resistance could have been that I foresaw the possibility of the
prolonged. pawn break.
3t h6 36 a6
32 h4 �e8 37 � i..cS
/OB LlnaNs Ct ( 10) 1987

38 l0e2 d4 43 ...�f8 caused more trouble in

39 l0f4 �d7 analysis. At first I examined the
40 e6+ �e8 move 44 lbc7, but I stopped every­
Both now and later, taking on e6 thing when I found that after
is impossible since the g7-pawn 44 ... .Ue 1 + 45 �d3 .Ue5 ! Black can
would be lost. hold the position. Only the next
41 �e4 aS morning did I manage to find the
Here the game was adjourned. correct route: 44 e7+ ! i..xe7 45
Analysis showed that the game �xd4 .Ugl 46 �c4 .Uxg4+ 47 �b5
could be won by study-like means. i..d8 48 .Uc3 ! .Uh4 49 �c6 ! .Uxh5
42 .Uf3! (D) 50 <i'd7 .Uxf5 and here 5 1 lbxb6! is
the end. 5 1 . . . i..xb6 allows a rather
funny mate: 52 .Uc8+ i..d 8 53
.Uxd8#. Therefore Black has to give
back the exchange with 5 1 .. ..Uf3,
and after 52 .Uxf3 i..xb6 the inevi­
table break b3-b4 is decisive. A
unique case, in which one king is
the main director and executor of
an attack on the other!
44 <i'd3! .Uxg4
45 f6! (D)

B
The sealed move, which seemed
the strongest. The rook lies in an
ambush, anticipating the black
king wishing to cross the f-file.
42 ••• .Ubl
Preparing for 43 �d5, in which
case after 43 ....Ugl 44 l£Jd3 .Uxg4
45 lbxc5 bxc5 46 �xc5 fxe6 47
fxe6 �e7 Black would have held
the position. However, White's
knight goes off in quite a different B
direction. This is White's quintessential
43 l£Jd5! .Ugl ! idea. He is threatening 46 l£lc7+
 Karpov - A. Sokolov 109

and 47 e7+, and none of his pawns

can be taken: 45 ...fxe6 46 f7+ �d7 - - . ­
(46 . . . �f8 47 li:Jc7 and 48 li:Jxe6+; • • •••
46 . . .�d8 47 f8'ii'+ �xf8 48 llxf8+
'iti>d7 49 �xb6+) 47 li:Jxb6+ 'iti>c7 48
. . - .
li:Jc4, threatening li:Je5 and li:Jg6.
- . . -�
45 . . . llg5 46 li:Jc7+ 'iti>d8 47 e7+ � �ttJ�
- - d • d
•
�xe7 48 fxe7+ 'iti>xe7 would have · � ·�· .
been more stubborn, with hopes
for the kingside pawns.
• • • •
45 ••• �d6 . . - .
46 li:Jxb6 llgS w
Again 46 . . .fxe6 will not do: 47 53 li:Jxb4 was threatened, while
f7+ 'iti>f8 48 li:Jd7+. Black is not after 52 ...�c5 53 li:Jxa5 �g5 54
helped by 46 . . . gxf6 either: 47 li:Jc6 �xh5 55 b4 the a-pawn will
llxf6 ! fxe6 48 llxe6+ �e7 49 li:Jd5 queen.
leads to a winning pawn endgame 53 li:Jxd4 �b4
for White. Black would not be saved by
47 fxg7 llxg7 53 . . .�f2 54 li:Jc6 �g5 55 �xa5
The final hope for salvation lay �xh5 56 �e2 ! , followed by 57
in 47 . . .fxe6 48 li:Jc4 �e7 (if Black 'iti>f3 and, if necessary, 58 �g2. The
plays 48 ... �b4, then 49 �xa5 ! is queenside pawns can easily move
decisive) 49 li:Jxa5 llxg7 50 li:Jc6+ forwards. Surprisingly, the knight
'iti>d7 5 1 li:Jxd4, although I suggest turns out to be stronger than the
that I would have coped with the bishop with play on both flanks.
diffkulties. 54 li:Jc6 �el
48 li:Jc4 �b4 SS �e2 �c3
49 exf7+ llxf7 56 'iti>d3 �el
so llxf7 'iti>xf7 (D) 57 'iti>c4 ..tgs
The storm of the combination 58 li:JxaS!
has abated, and the game has Both the most effective and the
moved into its technical phase. most striking line. Black cannot re­
This gives the impression that the fuse to take the piece: 58 ...�xh5
worst is over for Black, but there is 59 li:Jc6 'iti>g4 (after 59 ... �g5 60 b4
more to come. h5 White wins a decisive tempo by
51 li:JeS+! �6 means of 6 1 li:Jd4, threatening 62
52 li:Jc6 �el li:Jf3+) 60 b4 h5 6 1 aS h4 62 a6 h3
1 10 Stvllll Wch (2) 1987

(or 62 .tf2 63 lbd4) 63 a7 h2 64

... 9 f3 e3! ? (D)
aB'ii' h 1'ii 65 lbe5+.
58 .txaS
59 b4 .tdS
60 aS 'itxhS
61 '1tb5 .tgS
62 a6 .te3
63 'itc6 1-0

Game 26
Kasparov - Karpov
Seville Wch (2) 1987
English, Four Knights
w
1 c4
I must confess that I was not in­ It was precisely this move, sug­
itially expecting Kasparov's con­ gested in its time by Igor Zaitsev,
version to the English Opening in which plunged my opponent deep
this match. However, this surprise into thought (until then Black had
at the start strangely turned into a always taken automatically on f3).
gift for me straight away - at least, 10 d3!
Kasparov spent almost an hour and The correct reply, so White has
half on his first ten moves ! It is in­ not wasted his thinking time. Let
teresting that the novelty I used us look at two other possibilities:
was prepared at the beginning of a) 10 dxe3 b6 ( 10 ... h6 1 1 lbh3
the 1 980s for the Merano match. d5 12 lbf4! dxc4 1 3 e4 is worse) 1 1
However, instead of 7 lbg5, Korch­ e4 and then:
noi retreated the knight to e 1 , and a l ) 1 l .. . .ta6 immediately is
so it remained unused. dangerous : 12 f4 h6 13 lbxf7 !
1 eS cl;xf7 14 e5 .txc4 (14 ... lbg8? 1 5
2 lbc3 lbf6 .td5+ '1tf8 1 6 .ta3+ lbge7 17 'iid3
3 lbf3 lbc6 d6 18 'iih7 ! winning) 15 exf6 gives
4 g3 .tb4 White an advantage.
5 .tg2 0-0 a2) 1 1 ... h6 12 lt:Jh3 .ta6. The
6 0-0 e4 c4-pawn is doomed, and Black can
7 lbgS .txc3 be quite satisfied at the results of
8 bxc3 :es the opening.
 Kasparov - Karpov Ill

b) 10 d4 iLia5 1 1 'ifd3 b6 12 c5
.i.b7 ! and Black has the initiative, · ·i.� · -··
as in H.Olafsson-Naumkin, Bel­ �··
• �a • ••
d .•
grade 1988. .
. .
. -
- .
.
10..• dS
The only move, because other­
�
- ··�
. -
·�
� �
�
- - .
wise the e3-pawn will quickly be
• �"' '/�!iJ-
Ao .
% oA B
"�' '�"'
surrounded. /,- !'?
' !1,::: �

1 1 'ii'b 3! 8 Bl · 8 -�B
Yet another precise move; after
1 1 cxd5 iLixd5 two lines are possi­
�
� �
- •
.:*
�
ble: w
a) 1 2 iLie4 f5 1 3 c4 fxe4 14 d4, by directing the knight to e5,
cxd5 exf3 15 l:txf3 ilJd4 1 6 l:txe3 White is better.
.tg4 17 .te4 'iff6 1 8 .tb2 'ifb6 ! 19 16 .tb2 .tg4
.txd4 'ii'xd4 20 'iic l .txe2 21 1itg2 17 c4 (D)
.txd3 with a winning position for Of course, it is very tempting to
Black. threaten the enemy king, but this
b) 1 2 'ii' b3 ilJa5 ( 1 2 ...'iixg5 1 3 gives Black the initiative. 17 iLif3
f4 iLixf4 1 4 l:txf4 leads to a sharp 'ifd7 1 8 .tal .l:te7 ! ? 1 9 l:tfc l .i.h3
game) 1 3 'iia3 'iixg5 14 'iixa5 'ii'e5 is not dangerous either - Black has
( l4 . . . b6 15 f4) 1 5 d4 ( 1 5 f4? iLixf4 sufficient counterplay. However,
16 'iixe5 iLixe2+ 17 1ith l l:txe5 1 8 by continuing with the quiet 1 7
.i.b2 .:r.b5) 1 5 ...'ifd6, and the game l:tfe l White would have retained
is roughly even: 1 6 .ta3 'ifc6 17 f4 good chances.
'iixc3 1 8 'ifa4 iLif6. 17 dxc4
11 ilJaS 18 .txf6 gxf6
12 'ii'a3 c6 19 ilJe4 litg7
13 cxdS cxdS (D) 1 9....:Xe4 ! ? 20 .i.xe4 f5 ! 2 1 .tf3
14 f4 ilJc6 ilJd4 22 dxc4 i.xf3 23 exf3 e2 24
15 l:tb1 'ii'c7 l:tfel 'ifxc4 is also interesting.
After 15 ....i.g4 16 .l:tel (the most 20 dxc4
precise; after 16 l:txb7 .txe2 17 Now the advantage fully trans­
lbxf7 'ifa5 ! 1 8 'ifd6 .i.xfl 1 9 fers to Black. The continuation 20
iLih6+ lith8 20 ilJf7+ White has to iLixf6 ! ? litxf6 21 .:r.b5 l:tad8 22
settle for perpetual check) 1 6... h6 'ifc3+ lite? 23 'ifxc4 would have
( 1 6 . . .'iid7 1 7 h3) 1 7 iLif3 'ifd7 1 8 led to uncertain complications.
1 12 S1vlll� Wch (2) 1987

20 ... l:.ad8
More reliable than 20....txe2 2 1
lbxf6! .
2 1 l:.b3 (D)
Apparently the decisive error.
After the continuation 2 1 lbc3
li:)d4 22 li:)d5 there is still a stub­
born fight to face. Its direction de­
pends on the queen's choice of
whether to go to d6 or c4:
a) 22... 'ii'd6 23 'irxd6 l:.xd6 24
B l:.xb7 .txe2 25 l:.e1 .i.xc4 26 lbxe3
Here many commentators have ex­ l:.de6 27 l:.b4 lbe2+ 28 �f2 .i.d3
amined the more natural 20 'ii'c 3, 29 lbd5 lbxf4 30 lbxf4 :Xe1 3 1
and in this case after the correct re­ lbxd3 l:.8e2+ 32 �f3 l:.e3+ 3 3 �f2
ply 20 . . .'ird8! (20 .. .'1Ve7 2 1 l:.xb7 ! l:.1e2+ 34 �1 :Xg2 35 �xg2 lhd3
'ii'xb7 22 lbxf6 �f8 23 lbxg4 ! and 36 l:.a4, and as a result chances are
White has the upper hand) 2 1 dxc4 equal.
'ii'd4 22 :Xb7 l:.xe4 23 'ii'xd4 l:.xd4 b) 22 ...'ii'xc4 23 li:)xe3 'ii'xe2 24
24 .txc6 the chances are roughly li:)xg4 'ii'xg4, with a tense situation
even. on the board.
Another dangerous manoeuvre
is 20 lbd6, and Black has to play
carefully: 20 . . .l:.e6 2 1 lbxc4 l:.d8
22 f5 l:.ee8 23 l:.b2 lbd4 24 l:.xb7
lbxe2+ 25 �h1 lbxg3+! 26 �g1
lbe2+ (neither 26 ...'ii'xb7 27 .txb7
lbxfl 28 �xfl .txf5 29 lbd6, nor
26 . . . lbxfl 27 l:.xc7 e2 28 'ii'xa7
e1 'iV 29 l:.xf7+ �h6 30 l:.xf6+ !
�g5 3 1 'ii'g7+ �f4 32 'ii'h6# will
do) 27 �h l 'ii'g3 with perpetual
check. However, in the last vari­
ation Black could take a risk:
27 ...'ii'c 8 ! ? 28 lbd6 l:.xd6 29 'ii'xd6
.txf5 with enough compensation
for the exchange.
 Karpov - Farago 113

24 l:td3 �xe2 S a3 �xc3+

25 hd8 hd8 (D) 6 bxc3 cS
7 1i'g4
I had prepared this variation for
my match against Korchnoi at Ba­
guio. Ten years later, I finally got a
chance to try it in practice.
7 1i'c7
8 1i'xg7 l:tg8
9 1i'xh7 cxd4
One of the key positions in this
opening.
10 ltle2
A versatile manoeuvre: White's
w knight is covering the queenside
26 l:te1 l:te8 and simultaneously supporting the
The game is decided. Black has pawn advance on the kingside.
prepared a simple mating idea: 27 10
..• ltlbc6
lDd6 ltlxd6 28 1i'xd6 �f3 29 :Xe8 The pawn on e5 is invulnerable:
1i'fl#. 10 ...1i'xe5 1 1 �4 1i'f6 12 cxd4 with
27 1i'a5 bS a big advantage to White.
28 ltld2 1i'd3 1 1 f4
29 ltlb3 �f3 White's basic idea is to prepare a
30 �xf3 1i':d'3+ pawn advance on the kingside, and
31 �g1 l:txe1+ 1 1 �f4 runs contrary to this. Inci­
32 1Wxe1 lLle3 dentally, l l cxd4 is no good due to
0-1 1 1 .. .ltlxd4 ! .
11 �d7
Game 27 12 1i'd3 dxc3
Karpov - Farago 13 1Wxc3 (D)
Wijk aan Zee 1988 White takes the pawn immedi­
French, Winawer ately, and with it opens up both the
c-file and the entire queenside, so
1 e4 e6 he has to be extremely careful. 1 3
2 d4 dS lDd4, 1 3 ltlg3, 1 3 �e3, and 13 l:tb1
3 lbc3 �b4 have all been played several times,
4 eS ltle7 but with material equality Black
1 14 W(Jl Cltlll bt 1988
can easily count on a successful
outcome. There is no difference in
value between taking with the
queen or the knight on c3, and in
any case I have tried both routes. If
the queen takes on c3, the black c6-
knight turns out to be pinned, while
the queen herself does not feel too
settled (the threat is the advance
...d5-d4 ); if the knight takes it then
the black knight on c6 is free to
manoeuvre.
B
13 ••• ti::lfS
14 l:tb1!
This move has been given an ex­
clamation mark not because it is so
cool or particularly strong, but due
to other considerations: this game
is unique in the number of short
moves made by the white rooks -
out of the 27 moves remaining in
the game, 1 5 are made by the
rooks!
14 ••• l:tc8
15 ..td2 b6
In his game against Tal in the
same tournament, Farago used the
successful novelty 1 5 ... a6 ! and
equalized. I still think that all
White's resources have not yet
been exhausted here.
16 g3 (D)
The exchange 16 ti::lg3 ti::lxg3 1 7
hxg3 ti::le7 1 8 "iixc7 l:txc7 19 ..td3
l:txg3 20 l:th8+ l:tg8 2 1 l:txg8+
ti::lx g8 22 ..tb4 led to a better end­
game for White in Martic-Dras­
kovic, corr. 1 987, but I prefer to
maintain the tension on the board.
Incidentally, only now does this
game acquire independent signifi­
cance. Of course, I would have
liked to play g2-g4 in one go, but if
16 l:tgl , then 1 6 ...1fd8 ! with the
threat of ...ti)cd4 and ...1fh4+.
16 .b7
17 .d3 ti::lce7
18 l:tg1 l:tc4! (D)
 Karpov - Farago 1 15

• • •• • •
....... . . .
. -·- -
. -
- ··� ��· ·
-- ��
�
-
�-
- --- � �}, . -
"
u �
-
� -·
�\Wr �
rff// � -�
� �- "'

· � mltj•
- - . � u
•
,Wit : �
-
·� �
� i.. "
�
w B
Black's rook is ideally placed 24 h3 ii'd8
on c4: it is working both along 25 'ii'g3!
White's fourth rank and the c-file, Following Black's example of
and furthermore it is covering the rearranging his pieces on the c-file,
fl -a6 diagonal from the possibility I also decided to swap round my
of transferring the queen to a6 or queen and rook.
the bishop to a4 (or b5). 25 'ilfe7
19 g4 ll'lh4 26 .:o ii'cS
20 l:.g3 ,.c8 27 'itd1
21 l:.b2 ll'lhg6 White is planning to begin ma­
22 l:f3! (D) noeuvring the knight, therefore it
Overprotection of the f4-pawn. is best to remove the king from the
The white rook on g3 was beauti­ possibility of a check from c4 be­
fully fulfilling its defensive func­ forehand, simultaneously defend­
tion on the third rank and the g-file, ing the pawn on c2. Incidentally,
but now it is time to get down to ac­ transferring the king to dl now is
tive operations, for which it is nec­ better grounded than it was on
essary to regroup. move 10.
22 ••• ll'lc6 27 ••• l:.a4 (D)
Black cannot repeat the position In a complex position Black is
with 22 ... ll'lh4 because of 23 l:.f2 ! , losing the thread of the game. It
and then 2 3. . .l:.xg4 i s impossible in was necessary to play 27 ... d4 28
view of 24 'ilfh3 ! , when White is ll'lc l l:.c3 29 .i.xc3 dxc3 30 l:.b5
better. 'ilfd4+ 3 1 l:d3 'ilfxf4 32 'ilfxf4 ll'lxf4
23 m .:hs 33 l:.xc3 ll'ld5, and Black holds on,
1 16 Brussels World Cup 1988

although 3 1 i.d3 is stronger for 33 .Ub4

White. 34 .!i)b3
Matters are coming to a head;
34...i.xb3 is impossible because of
35 i.b5.
34 d3
35 'ili'xd3 'ilt'xd3
36 i.xd3 i.xb3
37 .Uxb3 .!i)cS
Not 37 . ..l:xh3 because of 38
i.b5+.
38 .Uc3 </;e7
39 i.n .Ud8
40 .Uc4! 1-0
w
28 .Uc3! Game 28
Giving Black no respite. 28 lbc 1 Karpov - Timman
lba5 29 l:c3 � 30 l:b4 is weaker. Brussels World Cup 1 988
Now Black should return the rook Queen's Gambit Accepted
to c4, as retreating the queen leads
to an almost forced loss. 1 d4 d5
28 ... 'ile7 2 c4 dxc4
29 .!i)ct 3 e4
The white pieces have unex­ The classical continuation is 3
pectedly developed stormy activity .!i)f3, 4 e3 and 5 i.xc4, immediately
exactly where Black usually domi­ winning back the pawn. However,
nates. White is threatening 30 i.b5, sometimes I like playing in pure
and if29... lba.S, then 30 .Ub4 is rea­ gambit style, acting in accordance
sonable. with the name of the opening!
29 ••• d4 3 .!i)f6
Somewhat spoiling White's in­ 4 e5 .!i)d5
tentions, but having torn itself 5 i.xc4 .!i)b6
loose, the central pawn will need 6 i.d3
constant defence. It is still not known which re­
30 .Ucb3 .!i)a5 treat is better for the bishop, d3 or
31 .Ub4! .!i)b7 b3. Before we continue I would
32 .Uxa4 i.xa4 like to pay a little attention to the

second of these: 6 .i.b3 lDc6 7 .i.e3
.i.f5 and now:
a) 8 lbe2 e6 9 0-0 lDa5 I 0 .i.a4+!
c6 1 1 .i.c2 .i.g6 12 lbbc3 lbac4 1 3
'iVc 1 .i.e7 1 4 .i.xg6 hxg6 1 5 lbe4
l:h4 1 6 lD2g3 'iVd5 1 7 'iVc2 0-0-0
1 8 .i.g5 .i.xg5 1 9 lbxg5 l:d7 20
lDf3 l:tf4 21 l:ad1 g5 22 1fc1 ! gave
White an appreciable advantage in
Karpov-Speelman, Brussels 1 988
and the game ended in my favour.
b) 8 lDc3 e6 9 lbge2 .i.e7 10 a3
0-0 1 1 0-0 lba5 I 2 .i.c2 .i.xc2 1 3
'it'xc2 lDac4 14 l:adi lDxe3 1 5 fxe3
and I quickly managed to seize the
initiative in Karpov-Short, Linares
Ct (9) 1 992.
6 ... lbc6
7 lDe2 (D)
White can also play 7 .i.e3 :
a) 7 . .. lbb4 looks good: 8 .i.e4
f5 ! 9 exf6 exf6 1 0 lbc3 f5 1 1 .i.f3
lD4d5 1 2 .i.d2 .i.e6 1 3 lDge2 'iVd7
1 4 0-0 0-0-0 1 5 l:e 1 ( 1 5 a4 ! ? de­
served attention) 1 5 ...l:tg8 ! , and
Black had solved all his opening
problems in Karpov-Short, Linares
Ct (3) 1 992.
b) 7 ... .i.e6 8 lbc3 'iVd7 9 lDf3
0-0-0 10 h3 lbb4 1 I .i.e2 f5 12 0-0
h6 1 3 a3 lD4d5 14 lDe1 lDxc3 1 5
bxc3 .i.c4 1 6 lbd3 e6 I 7 a4 g5 I 8
1fc2 'iVc6 1 9 :tfc I gave White the
better chances in the game Karpov­
Ivanchuk, Reggio Emilia I99112.
7 .i.g4!?
8 .i.e3
Karpov - Timman 1 17
B
After 8 f3 .i.h5 and 9 ... .i.g6
(White must avoid continuing 9
lDf4 owing to 9 ...'iVxd4 10 lbxh5
'iVxe5+) Black has a comfortable
game (the position resembles the
Alekhine Defence, Four Pawns At­
tack).
8 ... .i.xe2
9 .i.xe2 'iVd7
Black does not use up a tempo
playing ...e7 -e6, preferring instead
to increase the pressure on the d4-
pawn. In a situation like this one
immediately has to resort to tac-
tics.
10 lbc3 0-0-0
11 a4! ?
With the intention of distracting
White's attention from the d4-
square.
1 1 ... a6!?
If I l .. .lbxd4, then 1 2 a5 lbxe2
( 1 2 ... lba8 is no better: I 3 a6 b6 I4
.i.f3, and the black knight is
trapped in the corner of the board)
I 18 Brussels World Cup 1988
1 3 axb6 �xc3 14 iic2 cxb6 1 5
'it'xc3+ with a very strong attack
for White: 1 5 ... �b8 16 e6 iixe6 17
0-0, etc.
12 aS �dS
13 i.f3 �db4 (D)
1 3 ... e6? loses immediately in
view of 14 �xd5 exd5 1 5 i.g4.
Exchanging on c3 or e3 favours
White, as he gains a strong pawn
centre. However, 1 3 . . .�cb4 looks
stronger.
w rook and knight, but in a slightly
14 e6!?
The prelude to all further play
with an unusual correlation of
forces. After the quiet 14 0-0, tak­
ing the central pawn is risky:
14 ...�xd4 (Black's best chance is
probably 14 ... e6, with a complex
game) 1 5 i.xd4 'it'xd4, and now
alongside 1 6 Wb3 e6 17 llfd 1
iixe5 1 8 i.xb7+ �xb7 1 9 llxd8
i.d6 20 llxd6 iixd6 21 �a2 c5 22
�xb4 cxb4 23 'it'f3+, which gives
White a minimal advantage, he
could immediately take the bull by
the horns with 1 6 i.xb7+ ! �xb7
17 Wf3+, and then both 17 ...c6 1 8
llfd I �d3 19 lla4 Wd7 20 � e6
2 1 �d6+ followed by 22 llxd3,
and 17 ...�b8 1 8 llfd1 �3 1 9 lla4
'it'd7 20 lla3 ! are unfortunate for
Black.
14
.•• 'ili'xe6
After 14 .. .fxe6 1 5 �a4 Black
does not become disentangled:
15 ... �xd4 16 i.xb7+!.
15 dS 'ili'eS
There is another interesting at­
tempt, 1 5 ...Wg6. Then the continu­
ation 1 6 i.h5 'ifxg2 17 'it'g4+!
looks tempting, but if instead
Black replies 1 6 ...�c2+ ! 17 �fl
�xe3+ 18 fxe3 Wf6+ 19 �g1 e6,
the complications are not at all in
White's favour. Probably here as
well everything would have been
reduced to a queen sacrifice for
different way: 1 6 0-0 e6 17 dxc6
llxd 1 1 8 cxb7+ �b8 19 llfxd l
i.d6. White's initiative compen­
sates him for his small material
loss, although the chances can be
considered roughly equal.
16 0-0 e6 (D)
17 dxc6!
The logical continuation of the
plan White began on move 1 1 .
17 llxd1
18 cxb7+ �b8
 Karpov - Timman ll9

24 ... lDe5 loses: 25 l:tdxc7 lDxf3+

26 gxf3 Wxa5 27 lDc5 Wxc7 28
lDxa6+ �xb7 29 l:txc7+! �xa6 30
l:txg7 and White has a technically
won ending.

w
19 l:tfxd1 i.cS
19 . . . i.d6 is more stubborn, pre­
venting the white rook from pene­
trating the seventh rank.
20 i.xcS 'iWxcS
21 l:td7 fS w
22 l:tad 1 lDc6 25 .:Xc6 'iWxaS
White was threatening l:td8+, 26 l:txe6 �a7
and in the event of 22 . . . c6 23 h4 ! 27 g3 (D)
(but not 23 l:txg7 lDd3 ! 24 l:txd3 27 h4 ! is even stronger, com­
'iWe5 ! ) White has an obvious posi­ pletely depriving Black of counter­
tional advantage. play on the kingside.
23 lDa4
23 lDe2 !? is also good: 23 ...'ir'b5
24 l:tcl lDe5 (the most obstinate
continuation is 24 ... lDa7 ! , though
here as well Black experiences se­
rious problems) 25 l:tdxc7 lDxf3+
26 gxf3 Wxa5 (Black cannot con­
tinue 26 . . . 'iWxe2 due to 27 l:tc8+
�xb7 28 l:t l c7#) 27 l:tc8+ �xb7
28 l:txh8 'ir'd2 29 �fl is insuffi-
cient for Black.
23
24 l:tcl
120 Btlfort World Cup 1988

27 g5! 39 .i.g2
28 ltxh7! ltb8 Co-ordination between the rooks
29 hJ g4 and the bishop creates decisive
30 hxg4 fxg4 threats.
31 .i.g2 'it'a1+ 39 'it'd7
32 ..th2 'it'xb2 40 lth8 c4
33 lthh6 'it'a2 41 .i.e4 1-0
34 ltef6 c5
34...ltxb7 would have lost: 35 Game 29
.i.d5 ! 'it'a5 (35 ...'it'xd5 36 ltxa6+ Karpov - Kasparov
..tb8 37 lth8+) 36 .i.xb7 ..txb7 37 Belfort World Cup 1 988
ltxa6 'ifxa6 38 ltxa6 ..txa6 39 f4!. Griinfeld Defence
35 ltf4 (D)
1 d4 l£lf6
2 c4 g6
3 lbcJ d5
4 cxd5 lbxd5
5 e4 lbxc3
6 bxcJ .i.g7
7 .i.c4
The plan of 7 lbf3 c5 8 ltb1 0-0
9 .i.e2 has enjoyed great popularity
in recent years. Another plan of
development for White involves
supporting his mighty pawn centre
B with an early .i.e3, followed by
35••• 'it'd2? ! bringing the light-squared bishop
35 . . .c4 was necessary, although out to e2.
it is true that after 36 .i.d5 White 7 c5
has taken three black pawns for his 8 lbe2 lbc6
own on b7, and finally an endgame 9 .i.e3 0-0
would arise with a rook, bishop 10 0-0 .i.g4
and pawns on f2 and g3, against the It is well known that in the event
queen, which is good for White. of 10 . .. cxd4 1 1 cxd4 lba5 1 2 .i.d3
36 .i.n ! Lb7 .i.e6 1 3 d5 White can sacrifice the
37 ltxa6+ ..tb8 exchange, gaining a dangerous at­
38 ltf8+ ..tc7 tack.

1 1 f3 lLlaS (D)
Now in reply to 12 i.d3 there is
the possibility of 12 ...cxd4 13 cxd4
i.e6, when sacrificing the exchange
with 14 d5 is not so effective, as in
a lot of variations Black has a
queen check on b6. However, in
this case White is not at all forced
to give up the rook for the bishop.
w
12 i.xf7+
I specially prepared this cap­
ture on f7 for the Seville match.
Immediately after it had ended
Kasparov remarked that the vari­
ation is hopeless for White. But if
you judge from what happened in
the five games with it in Seville,
you cannot say that it is all that
easy for Black to solve the prob­
lems of the opening. Moreover the
proposed post-match duel in gen­
eral ended miserably for him. Only
later was it established that in the
event of i.xf7+ and the exchange
Karpov - Kasparov 121
of light-squared bishops, Black
gains a favourable game.
12 l:txf7
13 fxg4 l:txfi+
14 Wxfl
The point of White's idea is not
to take the pawn, as his activity
compensates Black for his small
loss, but to block in the enemy g7-
bishop with the help of the pawn
chain c3, d4, e5, g5 and h4.
14 ... 'iVd6
This queen thrust occurred in
four out of five games in Seville,
and it has also been used in many
other encounters. The other con­
tinuations, 14 ...cxd4, 14 ...'ili'd7 and
14 .. .'ii'e8 , are not really justified.
15 eS 'iVdS
16 i.f2 l:td8 (D)
w
The threat of a capture on e5
forces the white queen to vacate
d 1 . In this game I moved her to a4,
while later against Timman I
122 B1(fo11 World Cup 1988

played 17 Wc2, a game which also provoke such a useful pawn move
proved to be very engaging. as . . . b6? The fact is that after 1 7
17 . . . Wc4 1 8 'ifb2 .i.h6 19 h4, and 'ifa4 b6 1 8 'ifc2 in some variations,
now: for example in reply to 1 8 ...'ifc4,
a) 19 ...'iff7 20 �g1 (20 g5 ! ltX4 White switches the queen to e4
2 1 e6 ! is correct; thanks to this and, by attacking the knight, wins
zwischenzug, White diverts the an important tempo. Thus, moving
queen, and does not allow Black to the queen out to a4 has its advan­
double on the f-file: 2 1 . . .'iff5 22 tages. But I will not undertake to
�g3 Wxe6 23 Wxb7 llf8 24 lle 1 ! decide definitively where the
�e3+ 25 �g1 l:txf2 26 gxh6 cxd4 queen is more comfortable, al­
27 cxd4 llf8 28 'ifb2 �4 29 llxe6 though I have tried both possibili­
1 -0 Vyzhmanavin-Emst, Stock­ ties in practice.
holm 1 99 1 ) 20 ...l:tf8 2 1 �g3 �c4
22 We2 Wxf2+ ! 23 'ifxf2 .i.e3 24
'ifxe3 �xe3 and a roughly even • • •••
endgame arose in Karpov-Kaspar­ � -
-·�
� - -·
� -
ov, Amsterdam 1988. • • •••
- �- N �
b) 19 ... llf8 20 g5 ! (a universal
move which guarantees White an
- - u �
-
�
advantage in this case) 20...'ifd3 2 1 •• D ·�·
'ii'b 1 ! 'ife3 2 2 1We 1 .i.g7 2 3 �g1 -
�
�
� �
u �
-
� �
� -
�
�
We4 24 �g3 ! (White returns the
pawn but seizes all the critical
� · ·tb· � �p�
squares) 24...'irxh4 25 �4 llxf2 a • •�•
(a desperate sacrifice, similar to B
what happened in the Belfort game 17 ... b6
I am commentating on) 26 �xf2 White gets a better game after
cxd4 27 lld 1 d3 28 'ire3 ! �c6 29 17 ...�6 1 8 'irb3 c4 19 'ifxb7.
�xd3 'ira4 30 'irf3 'ifa5 3 1 e6 �d8 If Black prefers 17 . . .�c4, then
32 �4 .i.e5 33 llJd5 'ifc5+ 34 �h1 after 1 8 �f4 1Wf7 1 9 g3 �d2+ 20
1 -0 Karpov-Timman, Rotterdam �g2, White has the advantage in
1 989. the event of 20. . .lbe4 (Gutman) 2 1
17 'ifa4 (D) 'ifc2 �xf2 22 'irxf2, but Black has
Now after . . .b6, the queen will the strong reply 20 ...g5 ! as demon­
(as we see) retreat to c2. The ques­ strated by Gutman. He further sug­
tion is, why does White choose to gested that White should play 1 8

g5 ! himself, followed by 18 . . .:fs
1 9 �g1 Wf7 20 ..tg3 �3 2 1 ..tf4
l'bd5 22 :n ; Black's initiative has
been arrested, and his bishop is un­
der lock and key.
After the continuation 17 . . .:fs
1 8 �g1 1lf7 19 ..th4 ! l'bc4 20 'ifb3
the bishop again does not manage
to break free: 20 . . . ..th6 2 1 g5 !
.i.xg5 22 .i.xg5 11f2+ 23 �h1
Wxe2 24 h3 and White is apprecia­
bly superior.
18 Wc2
The queen has nothing left to do
on a4.
18 ••• :rs
Now the black pieces fall into a
clamp. 1 8 . . .:cs is more precise,
with counterplay on the c-file.
19 �g1 11c4 (D)
Black's bishop still cannot break
out to freedom: 19 .....th6 20 h4 'iff7
2 1 l'Dg3, alternatively 1 9 . . . l'Dc4 20
h4.
w 23 exf6 exf6 24 l'Df4.
Karpov - Kasparov 123
20 Wd.2!
White continues to play to limit
the mobility of the bishop. In the
event of 20 'ife4 l'Dc6 ! ? the threat
is to take on e5, and 2 1 Wxc6 Wxe2
is natural, but not a cause for de­
light.
20••• We6
20 . . .11f7 is no good: after 2 1
l'Dg3 everything turns out happily
for White - the knight moves to e4,
and the queen to e2. Black is not
saved from difficulties by 20.....th6
2 1 11xh6 Wxe2 either, in view of
22 We3 'iVxg4 23 dxc5 bxc5 24
Wxc5.
21 h3 l'Dc4
22 1lg5! (D)
An important moment. Besides
l'Df4, now ..th4 is also threatened
when the time comes.
B
22 h6
22 . . ...tf6 does not work due to
124 Be/fort World Cup 1988

23 'ifcl 'ii'f7 Forced, as if 27 ... b4, then 28

In order to gain counterplay liJf5 and after 28 . . .l:tf7, White has
B lack should have chosen 23 . . .b5 29 e6 'ii'xe6 30 l:te1 Wd7 3 1 cxb4.
with the idea, for instance, of 24 28 l:te1 b4
tbf4 iif7 25 tbd3 (after 25 -'.g3 29 'ii'g6 wrs
White would preserve his advan­ 29 . .. bxc3 loses immediately: 30
tage) 25 ... b4 !?. On the other hand liJf5 Wf8 31 e6 l:txf5 32 gxf5 liJd6
23 . . . 'ii'd5 does not achieve Black's 33 dxc5.
aim because of 24 'ii'c2! , and after 30 liJe4 (D)
23 ... h5 the queen returns to g5.
24 �g3 g5
This move was condemned by
the commentators, but alas, the
recommended 24 ...iid5 25 tbf4
'ii'e4 26 tbe6 would also have led to
a difficult position for Black. Here
are three key variations:
a) 26 ...l:tc8 27 'ii'b l ! 'i!Ve3+ 28
-'.f2 Wxc3 29 'i!Vxg6 'i!Vxal + 30
'it>h2 with inevitable mate.
b) 26 ... liJe3 27 'ii'd2 cxd4 28
cxd4 l:tc8 29 l:tel l:tc2 30 l:txe3 B
'ii'c6 3 1 d5 and Black's position is 30 ••. l:txf2
bad. He could try driving the queen This exchange sacrifice hardly
to e2 by playing 27 ... ltJc4 28 'ii'e 1 improves the situation, and White
lLle3 29 'ii'e2 cxd4, but 30 lLlxf8 only needs elementary accuracy.
-'.xf8 3 1 'ii'f3 ! 'i!fd3 32 cxd4 'ii'xd4 31 Wxf2 bxc3
33 l:te1 brightens up White's situ­ 32 'ii'f5+ �g8
ation completely: 33 ...liJc2+ 34 -'.t2 33 'ii'c8+ �h7
ltJxe1 does not work because of the 34 iixc5 iif7+
intervening 35 Wb3+. 35 �g1 c2
c) 26 ...cxd4 27 lLlxf8 liJe3 28 36 liJg3 -'.f8
'ii'd2 dxc3 29 'ii'e 2 -'.xf8 30 'ii'f3 37 liJf5 ..ti>g8
with a big advantage. 38 .:tcl 1-0
25 'it'c2 'it'd5 Black's pieces on the kingside
26 -'.f2 b5 could not break free. This was de­
27 liJg3 .:tf7 clared the best game in the next
 Karpov - M. Gurevich I 25

edition of Jnformator, and as it was In game 1 8 of the above-men­

played in the French town of Bel­ tioned match, at a point when my
fort, someone as a joke called the opponent was not particularly try­
variation with 1 7 iVa4 the 'Belfort ing for an active game, he played 9
Variation' . li:Jbd2, and after 9 ... li:Jbd7 10 0-0
0-0 1 1 l:tel c5 1 2 e4 dxe4 1 3 li:Jxe4
Game 30 J.b7 ! the game was totally equal.
Karpov M.Gurevich
- 9 li:Jfd7
USSR Ch (Moscow) /988 10 li:Jxd7 li:Jxd7
Queen's Indian 1 1 li:Jd2 0-0
12 0-0 l:tc8
1 c4 lbf6 13 e4 (D)
2 lbf3 e6
3 d4 b6
4 g3 J.a6
5 b3 J.b4+
6 J.d2 J.e7
In the game Karpov-Short, Am­
sterdam 1 988, after 6 . ..J.xd2+ 7
'ifxd2 0-0 8 J.g2 c6 9 0-0 d5 10
'ifb4 lbe4 1 1 l:tc l lbd7 1 2 'ir'a3
J.b7 1 3 cxd5 exd5 1 4 lbc3 f5 1 5
e 3 the dark-squared strategy led
White to victory; after 1 5 ...'ii'f6 16
l:tc2 a5 17 :td 1 l:tac8 1 8 lbe2 g5 1 9 B
lbc 1 g4 20 li:Jh4 m y opponent had 13 bS
got nowhere. After 1 3 . . . dxe4 14 J.xe4 b5 1 5
7 J.g2 c6 1Wc2 h6 1 6 l:tfd l bxc4 1 7 bxc4
The gradual preparation of ... d5 White had the initiative in Fta�nik­
has recently supplanted immediate Adorjan, Szirak 1986. Obviously,
action in the centre by means of 1 3 ...c5 is also insufficient: 14 exd5
7 ... d5 or 7 ...J.b7 8 li:Jc3 d5. These exd5 1 5 dxc5 dxc4 1 6 c6! cxb3 1 7
move-orders were seen a few times l:tel b2 1 8 J.xb2 li:Jc5 1 9 J.a3 !
in my first match against Kaspa- gave White a clear advantage in
rov. Gheorghiu-Csema, West Berlin
8 J.c3 dS 1 986.
9 li:JeS 14 :tel bxc4
126 USSR Ch (Moscow) 1988

In the sixth game of my first

match against Kasparov, I captured
with the other pawn, 14 ... dxc4, and
after 15 bxc4 lbb6 ( 1 5 ...bxc4 would
have led to a position from the cur­
rent game against Gurevich) 1 6
cxb5 cxb5 17 l:l.c l .i.a3 1 8 l:l.c2
lba4 1 9 .i.a1 l:l.xc2 20 •xc2 •as
2 1 'iWd 1 White gained an apprecia­
ble advantage thanks to his danger­
ous passed d-pawn.
15 bxc4 dxc4 B
After 1 5 ....i.xc4 16 lbxc4 dxc4 20 lbd2 e5?!
1 7 a4 'iWb6 1 8 .i.fl ..,a6 1 9 a5 c5 20 20 ... c5 ! is more precise: 21 d5
d5 l:l.fd8 2 1 e5 White had more lbes with a double-edged game.
than enough compensation for the 21 l:l.ad1 ..,e7
pawn in Dol.matov-Ehlvest, USSR 22 .i.n •e6
Ch (Minsk) 1987. 23 dxeS .i.xeS
16 ••4! ? 23 ... lbxe5 24 .i.xe5 .i.xe5 25
Until now this encounter had .i.xc4 is no better.
followed game 2 1 of my return 24 f4 i.c7
match against Kasparov. Then I 25 e5 lbb6 (D)
continued 1 6 'ifc2, and after After 25 ... .i.b6+ 26 �h1 lbcs
1 6... ..,c7 1 7 lbfl e5 ! 1 8 lbe3 exd4 the whole game would still lie
1 9 .i.xd4 .i.c5 chances were even. ahead, but now White gains a seri­
However, White's resources in this ous advantage.
variation have not yet been ex­ 26 f5! ...h6
hausted. 27 lbe4! lbd5
16 .i.b5 28 .i.d2 .i.b6+
17 ..,c2 .i.a3 29 �h1 •h5
18 lbb1 (D) 30 .i.e2 ...h3
White could have maintained a 31 lbgS ...h6
minimal advantage after 1 8 lbxc4 32 .i.xc4
.i.xc4 1 9 ..,a4 c5 20 'iha3 cxd4 2 1 Taking on f7 would have led
.i.xd4. White to his goal more quickly: 32
18 .i.d6 lbxf7 ! 'ifh3 33 lbd6, etc.
19 a4 .i.a6 32 ... .i.xc4
 Karpov - M.Gurevich 127

43 ... fxe6
44 l:xe6 �h8
45 .:eel ..d7 (D)

w
33 'ibc4 •hs
34 •e4 .i.fl
35 g4 ..h4
36 w ..d8 (D) w
46 aS! .i.xa5
46 ... .i.c7 also loses; 47 .i.xb4
cxb4 48 ..xc7 !.
4 7 •xc5! lbd3
48 •xas lbxel +
49 l:xel .:ed8 (D)

B
37 .:n .i.b6
38 l:del l:e8
39 l:e2 •c7
40 .:et �7
41 •c4 c5
42 �g2 lbb4 w
43 e6! 50 �g3! �7
White makes the decisive break. 51 ..a3 l:d7
128 USSR Ch (Moscow) 1988

52 �f4 'tWbS Transferring the game from a

53 l:teS 'iWc6 closed opening to an open type of
54 'ife3 l:tb7 game. Incidentally, the move e2-e4
ss l:te6 'iWa4 in this situation was given new life
56 l:te4 'ifbS by Timman in his match against
57 �eS 'iVc6 Short, Belgrade 1 987.
58 �d4 1Wd6+ 8
••• lLlxe4
59 �h3 l:tf8?! Short's preference was 8 ... dxe4
An oversight caused by time 9 �xf6 �xf6 10 1Wxe4+ c;li1f8 1 1
trouble in a difficult position. �c4 �g7 1 2 0-0 l:te8 13 'iff4 �e6
60 �cS 1-0 1 4 �xe6 l:txe6 1 5 l:tfe1 l:txel + 1 6
l:txe1 lLld7 17 lbe4 �e7 1 8 h4 'iVb8
Game 3 1 19 lLle5 f6 20 lLlg5 ! lLlf8 21 hS
Karpov - Yusupov gxhS 22 'iff5 'ii'c8 23 lLld7 ! 'i'xd7
USSR Ch (Moscow) 1988 24 'ii'xd7 lLlxd7 25 l:txe7+ �g6 26
QGD, Exchange Variation lLlf3. White had the better end­
game, and eventually came out on
1 c4 e6 top. Taking on e4 with the knight
2 lLlc3 dS did not seem to have been tried be­
3 d4 �e7 fore, but the novelty did not take
4 lLlf3 lLlf6 me unawares.
5 cxdS exdS 9 �xe7! �xe7 (D)
6 �gS c6 Strangely enough, the alterna­
7 'ii'c2 tive 9 ... 'it'xe7 loses straight away:
Instead 7 e3 (and then 'iVc2 and 10 lLlxdS ! cxd5 (if lO ...'i'e6, then 1 1
�d3 in one order or another) is lLlc7+) 1 1 'iVxc8+ 1i'd8 12 �bS+
most frequently played, or first 7 �e7 1 3 'i'xb7+ with a wipe-out.
'ifc2 and then 8 e3. In any case, at 10 lLlxe4 dxe4
this point the e-pawn is not moved 1 1 'ifxe4+ �e6
more than one square. 12 �c4 'iWaS+
7 ••• g6 1 2 . ..l:te8 is probably more pre­
A year later at Rotterdam, Yusu­ cise, because then 1 3 lLlgS is im­
pov chose 7 ... lLla6 against me, but possible on account of 1 3 ...'ii'a5+
again did not solve his opening and 14... 1t'xg5.
problems, and the game ended in 13 � !
my favour. Yusupov very likely underesti­
8 e4 mated this retreat by the king. After
 Karpov - Yusupov 129

Thanks to this mighty break in

the centre, White gains an impor­
tant square on d4 and freedom on
the dark diagonals for his queen, in
return for the sacrificed pawn. I
spent almost an hour thinking
about it, but the time is worth it for
a move like that!
16 ... cxdS
17 ..ibS!
The key to my idea. Without this
w attacking move White would find
1 3 lLld2 lLld7 14 0-0-0 .:l.ae8 Black himself at a dead end. If 1 7 lLld4,
would have overcome his opening then 1 7 . . .'iWe5 ! is strong, and 1 8
problems. 1t'a3+ leads to an equal position af­
13••• 'iWfS ter 1 8 ...'ird6, while if 1 8 Wxe5
14 'it'e3 lLld7 lLlxeS 1 9 ..ib5, then the simple
1 4 ...�f6 loses: 1 5 d5 ..ixd5 1 6 1 9 . . .lLld7 gives Black an excellent
..id3 'irg4 17 We5#. After 1 4...�f8, game. Of course, before throwing
White has the good continuation forward the d-pawn, I had to fore­
1 5 ..ixe6 'ilfxe6 1 6 'irh6+ �g8 17 see this fine bishop manoeuvre to
g3 lLld7 1 8 �g2. b5.
15 .:let .:l.ae8 (D) 17 ... a6
After 17 ...�f8 there is the possi-
bility of 1 8 'ilfc3, followed by
lLld4xe6+ and ..ixd7. Black should
have thought about returning the
pawn by means of ... d5-d4, and
transferring his queen to c5.
18 ..Wa3+ �d8
Right through the game the
black king does not find a safe ha­
ven, and now 1 8 ...�f6 loses to 19
..ixd7 ..ixd7 20 'iWc3+.
19 'ii'aS+
But not 1 9 lLld4 because of
16 dS! ! 1 9... 'ii'f4.
JJO USSif Ch (Moscow) 1988

19 ••• �e7
The path to the queenside is
fraught with danger, e.g. 1 9 ... �c8
20 l:.c l + �b8 2 1 'ii'c7+ �as 22
li)d4 'ii'f6 23 .txa6 ! l:.b8 24 'it'a5
'it'd8 25 l:.c7 winning.
20 'ii'b4+ �6
If 20 ...�d8, then 2 1 li)d4 'iVf6
22 .txa6 bxa6 looks reasonable.
Then after 23 l:.c l ! (23 li)c6+ �c7
24 l:.c 1 l:.a8 25 lLla5+ �d8 26 'ii'b7
�e7 27 tLlc6+ �d6 28 'ir'b4+ �c7 B
only leads to a repetition of the 24 tLlgS
position) 23 ...l:.eg8 24 l:.c6 there is Threatening 25 g4 ..Wh6 26 l:.xe6
no defence against the threat of 25 and 27 lLlxf7+.
l:.xa6 and 26 l:.a8+: 24 l:.hf8
a) 24...'it'e7 25 'iia5+. 25 .te2! ..Wh6
b) 24 ... 'ii'e5 25 l:.xe6! . 26 .tf3 l:.e7
2 1 'ii'd4+ Retreating the king does not
A pure geometric pattern. The help: 26 ...�c7 27 ..Wf4+ and 28
white queen is moving like a pen­ tLlxe6; 26 ... �c8 27 l:.c 1 + �d8 28
dulum. 2 1 .txd7 does not work in .txd5 .txd5 29 ..Wxd5.
view of 21 ...'ir'd3+. 27 'ii'b4 lLlf6
21 .•• �e7 28 'ii'd6+
22 .td3! 'ii' bS 28 'ii'f4! would have been imme­
22 ...'ir'f6 cuts off the path of the diately decisive, as it attacks the
king's retreat after 23 'iVb4+ [edi­ knight and simultaneously threat­
tor's note: for example 23 .. . �d8 ens 29 tLlxe6+ and 29 Wfb8+.
24 'ii'xb7 'ifxf3 !? 25 'ir'a8+ (25 28 ••• l:.d7
gxf3?? .th3+ 26 �g1 l:.xe 1 + 27 29 'ii'f4 tLlg8 (D)
.tfl l:.xfl#) 25 ... �e7 26 'ifxe8+ If 29 ... 'ir'g7, then 30 l:.xe6!.
l:.xe8 27 gxf3]. 3 0 .tg4! �c8
23 h4! (D) Taking on g4 leads to an elegant
It is almost as if White is con­ mate: 30 ....txg4 3 1 'ifb8#.
structing a study on the theme of 31 .txe6 fxe6
domination. 32 l:.cl+ �d8
23 •.. �d8 33 tLlxe6+ �e7
 Karpov - Malaniuk 131

3 i.g2 g6
4 c4 i.g7
5 .!Df3 d6
6 0-0 0-0
7 .!Dc3 'ife8 (D)

w
34 'ii'xf8+ 'ii'xf8
35 .!Dxf8 �xf8
36 .Uh3 .!De7
37 h5 �g7
38 h6+ �6
39 llf3+ �e6 w
40 .Ue1+ �d6 One of the most topical posi­
41 .Uf6+ �c7 tions in the Leningrad Variation of
42 g4 .!Dc6 the Dutch Defence. Black is in­
43 .Ue8 d4 tending ...e5, creating a flexible
1-0 pawn centre. e8 is the most suitable
square for the queen, as from there
Game 32 she is influencing the centre, and
Karpov - Malaniuk might come in useful on the king­
USSR Ch (Moscow) 1988 side.
Dutch Defence 8 b3!
The appearance of the bishop on
Malaniuk is one of the outstanding a3 could essentially hinder Black's
experts in this opening, and there­ plans. Other lines are less danger­
fore I prepared particularly thor­ ous for him.
oughly for our game in this event. 8 ... .!Da6
As a result I managed to come up An immediate march by the e­
with an interesting novelty. pawn, 8 ... e5, has not yet been fully
1 d4 f5 investigated, but it is evidently not
2 g3 .!Df6 without danger for Black.
132 USSR Ch (Moscow) 1988

9 .i.a3! 12 llad1 �h8

White's idea consists of actively 13 e4
positioning his pieces, with his Having finished mobilizing his
bishop on a3, his queen on d3, and forces, White generates activity in
rooks on d 1 and e 1 , and then mov­ the centre of the board.
ing the pawn to e4. In this game I 13 fxe4
managed to carry out this plan in 14 lbxe4 �fS (D)
ideal fashion. Black is forced to set
aside his standard advance ... e5 in­
definitely.
.•.9 c6
10 'iVd3
It is interesting that in this na­
tional championship Malaniuk
chose the Dutch Defence seven
times ! Two rounds before our game
Gavrikov had played the more
timid 10 llc 1 against him. After
10 . . .h6 1 1 e3 �e6?! 1 2 We2 'i!fd7
1 3 lbd2 lbc7 14 'ji'd3 llab8 15 �b2
�h8 1 6 d5 cxd5 17 lbxd5 lbcxd5
1 8 cxd5 �f7 Black built a fortress
which was not easy to penetrate,
and in the end he even won. How­
ever, this time matters did not turn
out so successfully for him.
10 ... �d7?!
This bishop move has very un­ The extremely unpleasant 2 1
pleasant consequences. 10 ...llb8 Wa5 was threatened.
deserves attention, as a series of 21 lbxe6 �xe6
games has testified. The fate of the 22 l:.de1 �d7 (D)
whole variation possibly depends He should have brought his
on the assessment of this position. bishop back to c8. There now fol­
11 llfe1 l:.d8 lows a decisive exchange sacrifice,
1 l . . .d5 is more stubborn, al­ which, incidentally, has been ma­
though after 1 2 lbe5 White main­ turing for a long time.
tains the initiative. 23 l:.xe7! �xe7
 Karpov - Hjartarson 133

29 l:l.xe5 dxe5
30 'iVxeS rM7
31 d6 j_f5
32 c5
Black is in total zugzwang.
32 h5
33 g4 hxg4
34 hxg4 j_d3
If 34 ...j_xg4, then 35 'iVf6+ 'it>e8
36 'ii'xg6+ and 37 'ifxg4 is deci-
sive.
35 j_d5+! 1-0
An effective final blow ended
the game; taking the bishop is im­
possible in view of 35 . . . cxd5 36
'iVxd5+ 'it>e8 37 'it'e6+ mating.

Game 33
Karpov - Hjartarson
Tilburg 1988
QGD, Slav Defence

In this tournament there was a rare

occurrence, at least in my experi­
ence - I played three straight
games in the same variation, and
moreover it was not in a classical
variation, but a very sharp one, in
which the balance of forces on the
board is disturbed in the first ten
moves. My results were quite fa­
vourable, 211213 . Detailed analysis
of these rather complex duels only
took place much later.
1 d4 d5
2 c4 c6
w 3 lt:lf3 lt:lf6
/34 Tilburg 1988

4 tt'lc3 dxc4 12 'ii'e2 i..xd2+

5 a4 i..f5 Here in the same tournament
6 lDe5 Timman introduced the novelty
This is the so-called Central 1 2 . . .'ifh4+ against me, but the in­
Variation (White is preparing f3 tervening check did not achieve its
and e4). aim: 1 3 g3 i..xd2+ 1 4 'itxd2 'fle7
6
••• e6 1 5 'ii'e 3 ! with a patent advantage.
This move has recently com­ 13 'itxd2 t!fd5+
pletely replaced two other theoreti­ In the ending after 1 3 ... 'flxe2+
cal continuations, 6 ... lDa6, and 14 i..xe2 the black pawns do not
6... lDbd7. compensate for the missing piece.
7 f3 It is another matter if the queen ex­
Logical; after 7 e3 or 7 i..g5 it is change takes place on d5, because
easier for Black to equalize. then Black strengthens his pawn
7
••• i..b4 chain.
8 e4 14 'itc2 lDa6
Leading to immense complica­ 15 tt'lxc4 (D)
tions, as Black is forced to sacrifice
a piece. As a rule after the cautious
• • ••• -
8 tt'lxc4 or 8 i..g 5, positions arise
with even chances.
-
-
,�.
� �
�
�'-
-
�
�'
8 ... i..xe4 �. , . , . .
9 fxe4 tt'lxe4 . . .. . .
Black has three pawns for the
piece, so there is approximate �·ttJ· • •
material equality on the board. • • • •
10 i..d2 'ii'xd4 Rc;t>�
u
-
�-�
� -
��u
� ft
10 .. .'ii'h4+ does not work: 1 1 g3
lDxg3 12 hxg3 'ifxh1 13 'ii'g4 with a • •i.• :
a strong attack for White. B
1 1 tt'lxe4 'i!Vxe4+ A peculiar, though standard pat­
1 1 ...i..xd2+ is no good: 12 lDxd2! tern from the Slav. Black has quite
'ifxe5+ 1 3 i..e2 b5 14 axb5 cxb5 1 5 a wide choice here, but recently
l:la5 a6 1 6 tt'lxc4 'ii'c7 17 lDd6+ castling one side or the other has
<1iie7 1 8 0-0 gave White an irresist­ been seen most frequently. Theory
ible attack in Khalifman-Ehlvest, is sceptical about 1 5 ...l:r.d8, 1 5 ...b5
USSR Ch (Moscow) 1 988. and 1 5 ...lDb4+; 1 5 ...'iff5+ deprives
 Karpov - Hjartarson 135

Black of the possibility of swap­ More accurate than 21 :g1 e5 22

ping queens on d5, while exchang­ .td3 b5 ! 23 axb5 lbxb5+ 24 �a4,
ing on other squares leads to an when the game is roughly even.
unpleasant endgame. 21 ••• eS (D)
15
•.• 0-0-0 2 l . ..�b8 is no good because of
With castling long, as opposed 22 1i'g3+.
to short, the rook immediately
lands on d8.
16 'ii'e5
White is aiming at the g7 -square
and is prepared to swap queens,
just not on d5.
16
•.• f6
Defending the pawn and forcing
the queen to clarify her intentions.
Neither 1 6 ... lbb4+ nor 1 6 . ..:he8
has vindicated itself in practice.
17 'ii'eJ! c5
Black has also tried 17 ...'ii'f5+, w
1 7 . . . lbc5 and 1 7 .. .'�b8, but with­ 22 'ii'c3! b6
out any particular success. 23 lbbJ 'W'xb3+
18 �b3! 23 . ..'W'e6 is more obstinate: 24
In my game against HUhner at .tc4 'ii'e7 25 lbxd4 :xd4 26 :hd1
Tilburg I preferred 1 8 .te2, and af­ :hd8, and Black holds on.
ter the continuation 1 8 ...lbb4+ 1 9 24 'W'xb3 lbxb3
�b3 lbc6 2 0 �c3 lbd4 21 .tf3 25 �b3 :d4
lbxf3 22 gxf3 'ii'd4+ 23 'ii'xd4 26 h4
:xd4 24 b4 :ds 25 bxc5 :xc5 a Black has material equality and
level endgame arose. This unex­ a flexible pawn chain, but I quickly
pected raid by the king looks rather managed to disturb its harmony.
strange, but it saves White an im­ 26 :hd8
portant tempo by immediately 27 .tc4 �c7
freeing the c-file for the rook. 28 h5 :g4 (D)
18 lbb4 Instead of pursuing the g-pawn,
19 :et lbc6 Black ought to generate some ac­
20 �a3 lbd4 tivity on the queenside: 28 ... a6! 29
21 lba5! .txa6 (Black was threatening ... b5)
J 36 Seanle Ct (3) 1989

29 . . ..:tb4+ 30 �c3 (in the event of 32 e4

30 �a3 White could even lose - 33 .Uhxf6 h5
30 . . . .:ta8 !) 30 ....:txa4 with chances 34 li6f4 lid4
for a draw. 35 .Uf7+ .Ud7
36 lixg7 lixg7
37 lif4 lig3+
38 �c2 lig2+
39 �c3 .Ug3+
40 �d2 lig4
41 .Uf7+ �d6
42 �e3 a6
43 �xa6 1-0

Game 34
Hjartarson - Karpov
Seattle Ct (3) 1989
Spanish, Zaitsev
29 h6!
By stealing along the edge of the 1 e4 e5
board, this pawn undermines the 2 lLlf3 llJc6
black kingside. 3 �b5 a6
29 ••• lbg2 4 �a4 lLlf6
Attempting to preserve a dy­ 5 0-0 �e7
namic pawn chain by means of 6 :et b5
29 . . . g6 is unsuccessful on account 7 �b3 d6
of the weakness on h7: 30 .:thd 1 ! 8 c3 0-0
.Ugd4 (30 ....:txd 1 3 1 .Uxd 1 .:th4 32 9 h3 lieS
�b5 .Uxh6 33 l:td7+ �b8 34 �a6 10 d4 �b7
with decisive threats) 3 1 lixd4 1 1 lLlbd2 �f8
lixd4 32 �g8 ! lid? 33 �c2 and 12 a3
Black will be put in zugzwang. As well as 1 2 a4, this restrained
30 hxg7 l:txg7 pawn advance is also seen fre-
31 lien .Ud6 quently. White guards b4 from in-
32 lih6 vasion by the black knight, and
Black's fortress has been de­ then prepares to develop the initia-
stroyed, and this sums up the result tive on the queenside.
of the opening. 12 ... h6
 Hjartarson - Karpov 137

13 ..tc2 liJb8
14 b4 liJbd7
15 ..tb2 aS! ? (D)
A move I had prepared specifi­
cally for this match; 1 5 ...g6 is a
more widely used continuation.

- ... -
. jilli
... • ... d
-
·
�
� - -
... �· ·
� ­
� .t. �
.
. � - � - � ­
� ·d
- · � - -� . 20 bxc5
B

�-/� �
� fj -
�- · d
�
• 20 lbxb5 immediately is much
� ·� � u . 1":\ . fj
.t.z.J. more precise: 20 ...cxb4 21 axb4
� ... /�
'. . �
� --'<% � -�� 0A ­
�
- : � l:lxal 22 ..txal d5 23 exd5 (23 e5
••
�
R �. m
·· · 7• • 1'1
..txb4 leads to an approximately
� . � � equal position) 23 ...l:lxel+ 24 1Wxel
w liJxd5 25 'lie4 liJ7f6 26 ..txf6
16 ..td3 c6 liJxf6 27 "ilxb7 "ilxd3 with a sharp
17 liJb3 axb4 game, in which Black has enough
18 cxb4 compensation for the pawn.
1 8 axb4 liJb6 1 9 liJa5 "ilc7 20 dxc5
would have led to equality. 21 liJxb5 ltJxe4!
18 ••. exd4 22 'lic2?!
19 liJfxd4 (D) This is a serious error. White had
Now Black seizes the initiative. to play 22 "ilf3, when the following
Things are also fine for him after variations are possible:
1 9 ..txd4 c5 20 ..txf6 liJxf6 2 1 a) 22 ...liJd6? fails to 23 l:lxe8
..txb5 l:be4. The correct continu­ 'tWxe8 24 liJxd6 ..txd6 25 "ilxb7
ation is 1 9 liJbxd4 c5 20 bxc5 dxc5 l:lb8 26 '1id5.
2 1 liJxb5 ltJxe4 22 liJe5 liJxe5 23 b) 22 ...'lib6 23 ..txe4 ..txe4 24
..txe4 with a roughly even game. l:lxe4 'lixb5 25 a4 is unclear.
19 ••• c5! c) 22...l:lb8 should be answered
A very promising pawn sacri­ by 23 ..tc4, with a complicated po­
fice, with which Black takes aim at sition. Instead after 23 ..txe4?!
the enemy centre. ..txe4 24 l:lxe4 l:lxe4 25 'lixe4
138 Slcelleftea World Cup 1989

.:txb5 26 'ili'd5 'ilt'b8 27 'i't'xd7 .:txb3 Black threatens the deadly

28 �cl .:tb1 29 .:.Xbl 'i't'xb1 30 'ilt'd2 ...tZ::lf3+ or ...tZ::lxh3+.
c4 3 1 �h2 1t'b3 Black is slightly 27 .:te3 �d6
better. 28 h4 tZ::le6
d) 22 ... �df6 !? 23 �xf6 �xf6 29 li)d1?
24 .:txe8 �xf3 25 .:txd8 .:txd8 re­ White would fare no better with
sults in an equal position. 29 .:txe6? fxe6 30 'it'g6 'it'e8 . His
22 ••• �df6 best chance was 29 li)b5 !? �f4 30
23 �c3 (D) .:th3 li)e4 when Black would still
White must certainly avoid the have some obstacles to overcome.
continuation 23 .:tad 1 'ilt'b6 24 tDc3 After his actual choice I was able to
�xf2! . launch a winning attack.
29 tZ::lg4
30 .:txe6 �h2+!
31 �h1 1t'xe6
32 f3 ile1 ! !
0-1
White's position collapses: 33
fxg4 'it'xh4 wins on the spot, while
33 'ii'c4 gives Black a choice be­
tween 33 ...�f4 and 33 ....:ta4.

Game 35
Karpov P. Nikolic
-

B Skelleftea World Cup 1989

23 ••• �g5! Bogo-Indian Defence
24 �bS
More resistance was possible by 1 d4 tZ::lf6
24 tZ::ld 2, although after 24 ...'ilt'b6 2 c4 e6
Black threatens 25 ...'ilt'c6. 3 tZ::lr3 �b4+
24 llxe1+ 4 �d2
25 .:txe1 ikc7 A more popular continuation
26 �n than 4 li)bd2.
Now, however, 26 tZ::ld2 can be 4 ••• ile7
met by 26 ...1t'f4!. Still, 26 .:te3 was Black has several alternatives,
worth considering. including exchanging bishops on
26 ••• 'ii'c6! d2.
 Karpov - P. Nikolic 139

s g3 0-0 17 'ird1
6 .tg2 i.xd2+ First and foremost in order to
7 Wxd2 d6 control the h5- and g4-squares.
8 ltlc3 e5 After the immediate 17 f4 Black
9 0-0 l:te8 could have chosen the exchange
10 e4 .tg4 1 7 . ..exf4 in order to gain counter­
11 dS .txf3 play by means of ...ltlg4 (or ..ltlhs)
.

12 .txf3 cS and ...'ifh4.

13 l:tael a6 17 l:te7
14 b3 ltlbd7 18 l:te3 Was
15 .tg2 l:.ab8 19 l:tfe1
16 a4 (D) White has reinforced the e4-
By freezing play on the queen­ pawn, and it remains for him to
side, I am decisively moving the transfer the bishop to h3 and ex­
battle's centre of gravity to the change it for the d7-knight, which
kingside. Now if 1 6 ... a5, then 1 7 would give him a significant spa­
ltlb5 b6 1 8 ltla7 and ltlc6, followed tial advantage.
by f4. 19 ... �h8
20 .th3 l:tg8
Nikolic prevents the advance 21
f4, after which the game would
now continue 2 l ...exf4 22 gxf4 g5.
In that case White would not have
enough time to rush the pawn
through to e5, and Black would
thus gain a comfortable post for his
knight on that square.
21 �h1 Wd8
22 'ird2 ltlhS
23 ltle2 g6
24 aS ltlg7
16 'ird8 White needs time to regroup his
Preventing further movement by forces on the queenside, but Black
the white pawn. Otherwise after a5 has apparently prepared for activ­
and ltla4, I would have advanced ity on the kingside. However...
the pawn to b4 and put pressure on 25 .txd7 Wxd7 (D)
the weak b7-pawn. 26 f4! l:tge8
140 Skelleftea World Cup 1989

31 lbxe4 ikf5 (D)

.
- .
- ·
---�
- 3 l . . .'iWg4 32 c5 .:txe4 33 .:txe4
. . .... . . . . 'ii'f3+ 34 �gl 'iWg4+ 35 �f2 'ii'h4+
•• • ••• 36 <ite2, etc., is worse, but 3 l ...lbh5 !
�'� -
>
�8-
!'! � • d is interesting, and in the event of
32 c5 c;l;>gS 33 cxd6 .:txe4 34 .:txe4
·8·8· • lLlg3+! 35 hxg3 'iWh3+ the compli­
� � -�
. A . Ml
do• " . !'! cations end with a perpetual check.
•
B B pm o.z.J
�U � �
�
• • a •�
w
27 b4! cxb4
In the event of 27 ...exf4 28
lbxf4 the essential superiority of
the white knight over his opposite
number becomes clear: by landing
on d3 it supports the attack in the
centre.
28 'ii'xb4 f5
Black is prepared to exchange w
the e- and f-pawns, then move the It looks as though everything is
knight over to f5, achieving a posi­ in order for Black now. 32 lbxd6
tion of dynamic equality. There­ does not work: 32 .. .l:be3 33 l:lxe3
fore I have to play very carefully, ifxf4 threatens mate on fl , which
but also quite decisively. would cost White his knight if he
29 lbc3 plays 34 .:txe8+ lLlxe8. After 32
The immediate 29 c5 ! deserved lLlg3 Black should again continue
attention: 29 ... dxc5 30 'ii'xc5 exf4 32 ....:txe3 33 .:txe3 'ili'xf4 34 .:txe8+
3 1 lbxf4 fxe4 32 l:Xe4 l:be4 33 lLlxe8 and White has nothing better
.:txe4 .:tc8 (33 ....:txe4? 34 'ii'f8#) than a draw: 35 'iWxb7 'tie l + 36
34 .:te7 'ii'g4 35 'ii'f2 , and White's c;li>g2 1Wxc4.
chances are obviously better. De­ But not all White's resources
laying the break c4-c5 for three have been exhausted...
moves complicates the situation. 32 c5!
29 exf4 The queen joins in to defend the
30 gxf4 fxe4 knight, and White makes a passed

pawn. Events on the board take on
quite a stormy character.
32 dxcS
33 'it'c4 'ifxf4
34 d6 :Z.eS? (D)
Black is still under the impres­
sion that his doubled rooks are
strong. 34 ... :Z.d7 would have been
more forceful, preventing the pawn
from advancing. After 35 'ii'c3 'ii'e5
Black holds on, while 35 'ii'xc5 is
not dangerous for him either:
35 . . .'ii'e5 36 'ii'xe5 :Z.xe5 37 :Z.c l
lbe6 38 :Z.c8+ �g7 39 :Z.e8 �f7 40
liJg5+ [editor's note: White may
try 40 :Z.xe6 �xe6 4 1 liJc5+ �xd6
42 :Z.xe5 :Z.c7 43 liJd3] 40 ...:Z.xg5
4 1 :Z.3xe6 :Z.d5 42 :Z.8e7+ :Z.xe7 43
:Z.xe7+ �f6 44 :Z.xh7 :Z.xd6 45
:Z.xb7 with a draw.
w away on d7, but, as it perishes, the
35 d7! :Z.d8
If 35 ...:Z.f8 36 'ii'c3 ! :Z.d5, then
37 liJxc5 is decisive.
36 'iVcJ! hS?!
Karpov - P. Nikolic 141
Allowing White to create a
small study. After 36 ...h6 the posi­
tion is still unclear, for example 37
:Z.f3 'ii'g4 38 :Z.f7 ...e6 39 :Z.efl
liJfS ! .
37 :Z.f3 'it'h4
38 :Z.f7
Not, of course, 38 •xe5? due to
38 ......xel + 39 �g2 ,..e2+.
38
••• Ilg8
39 'it'cl
39 :Z.xg7 :Z.xg7 40 d81i'+ •xd8
41 'ii'xe5 is also good, and after
4 l . ..'�i'e7 42 liJxc5 •xe5 43 :Z.xe5
White achieves a technically win­
ning endgame. But I managed to
find a more elegant line.
39
••• gS
39 . ..liJf5 does not work because
of 40 :Z.xf5, therefore Black is
forced to advance yet another
pawn from his king's shelter.
40 liJg3 :Z.dS
41 :Z.f6 :Z.d8 (D)
The game ends with immediate
mate after 4 1 . .. �h7 42 ...bl+.
42 'iib 1 !
An elegant finale.
42
•.• �g8
It looks as if Black has extri­
cated himself, and with his next
move he will seize the daring run­
pawn has the last word.
43 :Z.e8+! 1-0
Let us see the end of the study:
43 ...:Z.xe8 44 dxe8...+ liJxe8 45
142 London Ct (8) 1989

managed to glean the least advan­

tage out of the opening.
8 i.xe7 'iVxe7
9 llcl
In game 4 I had continued 9
it'c2, and in game 6, 9 cxd5, and in
both cases Black quickly equal­
ized. This time I was better pre­
pared for the game.
9 c6
10 i.d3 li)xc3
1 1 llxc3 dxc4
12 i.xc4
Recapturing on c4 with the rook
also deserves attention.
Game 36 12 .•• li)d7
Karpov - Yusupov 13 0-0 eS
London Ct (8) 1989 Neither 1 3 ... b6 nor l 3 ...lld8 is
QGD, Lasker very promising for Black. It must
be said that from the Lasker De­
This decisive duel was very strik­ fence, the game has transposed to
ing. lines of the less weighty Capa­
1 d4 li)f6 blanca System, and moreover the
2 c4 e6 position of the pawn on h6 (instead
3 lf)fJ d5 of h7) is favourable for White, as it
4 li)c3 i.e7 gives him an additional object for
5 i.g5 0-0 attack.
6 e3 h6 14 i.b3 (D)
7 i.h4 li)e4 This waiting move with the
Not waiting for the white bishop bishop was prepared specially for
to take on f6. This move constitutes this game. Here White would nor­
the Lasker Defence, where Black mally play 14 h3, 14 jfc2, 14 'iib l ,
in many respects has no preten­ or 1 4 dxe5 li)xe5 1 5 li)xe5 it'xe5
sions, but relies on creating an im­ 1 6 f4.
penetrable fortress. In games 4 and 14 exd4
6 Yusupov had already turned to 15 exd4 lDf6
this defence, and in fact I had not 16 lle1 'iVd6
 Karpov - Yusupov 143

• •.t.• -··
m B mu · -
- · ··- m
••• • •
. .
- - -� - .
. � . .
-
d£• R� ­
� � U"'l.J.
8� • B 8_rJ
•
. .\lllr
- �� . �
.. �
-�
B w
17 ltleS! lDdS 2 1 ... f5 is no good because of 22
18 l:.g3 '6'f4 ! (or 22 'ii' h5 g4 23 ltlxg4); in
White does not hide his inten­ the event of 21 ... ltlf6 White has the
tions. strong 22 'itf3, or 22 '6'f4! ltlh5 23
18 ••• .i.fS 'ii'f3 ltlxg3 24 ltlxf7 'iixd4 25 l:.e7,
This move is a novelty. Theory etc.
gives the continuation 1 8 ....i.e6 19 22 hxgS!
'iid2 �h8 20 l:.e4 and 21 l:.h4, 22 Wh5 is insufficient: 22... fxe5
which is somewhat preferable for 23 hxg5 .i.g6 24 gxh6 �h7 and the
White. black king is safely covered.
19 'tWhS .i.h7 22 ••• hxgS
The paradoxical move 19 ......e6!?, In the event of 22 ... fxg5 the re­
suggested by Taimanov, is interest­ ply 23 f4 simply becomes more
·

ing.
20 'tWg4! gS (D)
After 20 . . . g6 there is the very
strong reply 21 h4. 20 . ..'it'f6 does
not work because to 21 ltld7. After
20 ....i.g6, the following variation
is possible: 2 1 ltlxg6 'tWxg6 22
'iixg6 fxg6 23 l:.xg6 l:.ae8 24 l:.xe8
l:.xe8 25 �1 l:.e4 26 l:.d6 l:.xd4 27
l:.d8+ with a big advantage to
White.
21 h4 (6
valid.
23 f4
After the continuation 23 l:.h3
fxe5 24 l:.xh7, White would have
won beautifully in the event of
24...�xh7: 25 .i.c2+ �g7 (25 ...�h6
26 'iWh3+ �g7 27 'iWh7+ �f6 28
dxe5+ Wxe5 29 'iWh6+) 26 'fi'xg5+
cltf7 27 dxe5 We6 (27 ...'6'b4 28
e6+ �e8 29 .i.g6+; 27 .. .'ile7 28
'iWg6#; 27 ....:tae8 28 'iWh5+) 28
.i.f5 . However, Black can hold on
144 London Ct (8) 1989
thanks to 24 ... 'iff6. Besides mobi­
lizing the f-pawn, White would
have gained a dangerous initiative
after either 23 'iih5 or 23 ltlf3.
23 ..• .:tae8 (D)
The critical moment. This rook
manoeuvre forces Black's demise.
He should have sheltered the king
in the corner with 23 ... �h8, al­
though in this case as well I could
have reached a winning ending
with four rooks with 24 fxg5 fxe5
25 g6 'iixg6 26 'ii'xg6 i.xg6 27
.:txg6 exd4 28 .:te4 .:tf7 29 i.xd5
cxd5 30 .:th4+ .:th7 3 1 l:xd4 l:d8
32 .:tg5 .:thd7 33 a3 a6 34 a4 b5 35
axb5 axb5 36 b4 - Black is in zug­
zwang.
queen sacrifice - 25 gxf6+! ! i.xg4
26 .:Xg4+, and then:
a) 26 ... �h8 27 ltlf7+ .:txf7 (if
27 ...�h7, 28 .:tg7#) 28 .:Xe8+ .:tf8
(28 ... �h7 29 .tc2+ �h6 30 .:th8+
and 3 1 .:txh7#) 29 f7 ! ltlf6 (or
29 . . .�h7 30 i.c2+ �h6 3 1 .:tg6+
'iixg6 32 i.xg6, etc.) 30 .:txf8+
'ili'xf8 3 1 .:tg8+ ltlxg8 32 fxg8'ii+
'ifxg8 33 i.xg8 �xg8 34 �f2 and
it's all over.
b) 26 . .. �h7 27 i.c2+ �h6 (al­
ternatively 27 ...�h8 28 .:th4+ �g8
29 .th7+ �h8 30 ltlg6#) 28 .:tg6+
�h7 (28 . . . �h5 29 i.d 1 + �h4 30
l:e4+ mates) 29 f7 and White wins.
25 g6! i.xg6
26 dxeS!
More accurate than 26 .:txe5
�h7 27 'iWh5+ �g7 (27 ... i.xh5 28
.:txh5+ 'ii'h6 29 i.c2+ �h8 30
l:xh6#) 28 .:txg6+ 'ifxg6 29 l:g5
.:te 1 + 30 �h2 .:te6 with an unclear
game.
26 ••• 'ili'e6
Things are no better for Black
after the continuation 26 ...'iic5+
27 �h2! (but not 27 �h1 .:tf6 !), or
26 . .. �g7 27 exd6 .:txe 1 + 28 �h2
.:th8+ 29 .:th3.
27 i.xdS cxdS
28 'iWxg6+ 'iixg6
29 .:txg6+ �h7 (D)
If 29 ...�t7, then 30 .:td6 .:td8 3 1
.:r.n + �e7 32 .:txf8 .:txf8 3 3 .:txd5
winning.
30 .:td6 .:cs
 Karpov - Timman 145

.
- - ... �
. - .
.
• • • • ••
. . •:•
. · · �'� .
• • • •
• • • •
80 • .8•
.
. . �
�� �
. �
w
After 30....:d8 White has the de-
cisive 3 1 .:d 1 .
31 .:e3 l:.c2
32 .:d7+ �g6
33 .:xb7 .:e8
The rook ending is hopeless for
Black, but resigning the game was
tantamount to resigning the match
too, and so Yusupov played on,
hoping for a miracle. If 33 ...1:.ff2,
then 34 .:g3+ �h6 35 .:bs ..th7 36 Game 37
.:b4 is decisive. Karpov - Timman
34 a3 d4 Kuala Lumpur Ct (4) 1990
35 .:d3 .:xe5 Griinfeld Defence
36 .:xd4 .:g5 (D)
If 36 ....:ee2, then 37 .:g4+ �f5 1 d4 lLlf6
38 .:gg7. 2 c4 g6
37 .:d6+ �h5 3 lLlf3 j.g7
38 .:h7+ �g4 4 g3 c6
39 .:d4+ �5 s j.gl d5
40 .:d5+ �g6 6 cxd5 cxd5
41 .:g7+ �xg7 7 lLlc3 0-0
42 .:Xg5+ �f6 8 lLleS e6
43 .:b5 a6 9 0-0 lLlfd7
44 .:b6+ �e7 10 f4 lLlc6
146 Kuala Lumpur Ct (4) 1990

1 1 .i..e3 f6 After the text move the threat of

12 ltld3 ltlb6 the knight appearing on f5 is more
13 b3 'ile7 dangerous for Black: 25 ... .if7 26
14 a4 .i..d7 fxe6 'ilxe6 27 lLlfS. To defend
15 .i..c t .:tfd8 against this, Timman sacrifices a
16 e3 .i..e8 pawn.
17 .i..a 3 'ilf7 25 ••. e5 (D)
18 :et .i..rs
19 .ixrs Wxrs (D)
As a result of complex manoeu­
··- -*-•• •
vres, an advantage has crystallized �· - . .•
·-�-V
�.• V �

for White on the queenside.

.
�� �
M·- -�- �
I . _ ... . . . - � r...
� •�- r/� • � •. �
-·· . •• •�- z.w�"l.JU �
-
•• �
··�·• �·� �·
m .
wa.
�- - � � -� i. � �
-�
• ••• � • - • -:=
�

A �
Q � " �� -
�- U �
�
� w
�
" V

�- �
�-
r.... � rt�
��� -
�"l.JU
- � i. " 26 Wet
�
"� V

- -� m
' �\lllr � m �
Obviously, the continuation 26
�
� �- � � �-: � .i..xd5+ �h8 27 l:tfcl exd4 28 e4 !
is stronger, returning the pawn;
w the inevitable break e4-e5 would
20 g4 'ile7 have guaranteed White a serious
21 'ild2 .:tac8 advantage. Killing two birds with
22 ltle2 l:tc7 one stone (attempting to control
23 l:tc5 ltlc8 the c- and f-files simultaneously)
24 f5 g5 does not work, and Black, by liqui­
White is very slightly better af­ dating the danger in the centre,
ter the continuation 24.. . ltld6 25 achieves a playable game, which
lLldf4 gxf5 26 gxf5 ltlxf5 27 ltlxd5 for a long time will be typified by
exd5 28 l:txf5. manoeuvring. However, thanks to
25 ltlg3 such tardy action the game quickly
25 fxe6 'i!he6 26 ltlc3 ltl8e7 27 moved into an ending filled with
e4 is not so clear. fascinating events.
 Karpov - Timman 147

26 b6 33 liJdl l:.c6
27 l:.c2 e4 34 l:.xc6 l:.xc6
28 liJf2 liJd6 35 l:.xc6 'ifxc6
29 'ii'd2 l:.dc8 36 l2Jc3 �f8
30 l:Ucl aS 37 �f2 �e7
31 �n 38 �el �f8
If it did not deprive the white 39 �dl 'ii'c8
queen of mobility, this move would 40 �el �g7
have been very perilous for Black. 41 l2Ja2 l2Jxa2
Attempts to rebuild with 3 1 'ii'c3 42 'ifxa2 'ii'c7
are repulsed by means of 3 1 . .. �d7 43 �f2 �f8
and 32 ...l2Je8. 44 'W'b2 �e7
31••. li)b4 (D) 45 �e2 �d8
3 l . . .li)xd4 !? suggested itself, 46 �el �c8
but after 32 exd4 e3 33 'ii'd3 exf2+ 47 �d2 �b7
34 �xf2 l:.xc2+ 35 l:.xc2 l2Je4+ 36 48 'ii'c l 'W'e7
�f3 a draw is not far away (swap­ 49 �el �d7
ping the good c6-knight for the bad so � l2Je8 (D)
one on f2 is not favourable for
Black). Timman's position in the
match obliged him to maintain the
• • •• •
pressure on the board in the hope ···i.- • .
of seizing the initiative from me. -� . . � - .�
�
- ·
-·· -8--
8.�� Ri-
u ·8-·
.8.
d d � u �
�
•
d •.li.m
d � � u
.
. �
- .
. .
.
w
51 'W'hl !
I did not want to agree to a draw,
so I decided to bring about some
confusing complications, by leav­
ing the queenside to its fate, while
actively occupying the kingside.
148 Kuala Lumpur Ct (4) 1990

Calculating variations to the very papers such as 'Will Karpov Hold

end was obviously out of the ques­ Out?' . In fact, my preliminary
tion. analysis showed that only White
51 'Wib4 has chances to win, and his pawns
52 h4! 'ii'xb3 are more dangerous than his oppo­
53 hxg5 fxg5 nent's.
54 'Wxh7 'ii'xa4 61 ..• <j;c7
55 'Wie7 'ii'c6 A good sealed move, but, of
56 'ii'xg5 a4 course, we had also analysed oth­
57 'ii'e7 'ii'd6 ers. In the end it was clear that
58 'Wixd6 lbxd6 6 1 . .. i.d7, 6 l . ..i.e8 and 6 l .. .'.tc6
59 i.dl i.b5?! would have left Black with less
It was barely worth obstructing chances to save himself.
the path of the pawn. After 59 ...b5 62 'iti>g3 lbc4
the game would most likely have 62...<iti>d7? loses: 63 <ifi>f4 i.c4 64
ended in a draw. g5 a2 65 lbxa2 i.xa2 66 g6 ri;e7
60 lbe2 a3 (66 ...lbe8 67 i.a4+) 67 g7 <j;f7 68
61 lbcl (D) f6 i.c4 69 i.g4 <itg8 70 i.e6+ lbf7
7 1 <j;f5 i.b5 72 ..ti>g6 i.e8 73 i.xd5
b5 74 i.a2 b4 75 i.b3 .
• • • • 63 i.e2
••• • • But not 63 <j;f4? lbb2 ! with the
- - . �� '.
� threat of ...lbd3+.
•-*-• · · � ­ 63 ••• i.e8!
This is the strongest defence.
.
- " u·· -�· � 63 ...lbxe3 is insufficient: 64 i.xb5
- . "
- � ;Q/ � . lbc2 65 g5 lbxd4 66 f6 <j;d6 67 g6
.
�� .
� �
� .
. <j;e6 68 g7 ..ti>f7 69 i.e8+ <j;g8 70

� i. •
..ti>f4 winning.
�
-
� " •
� . 64 ..ti>f4 (D)
B Here is a striking variation after
Here Timman sealed a move he 64 g5, ending in a problem-like
had been thinking about for almost mate: 64 ...lbxe3 65 <itf4 lbc2 66
half an hour. Black's activity on the <j;e5 lbb4 67 <ifi>f6 <iti>d6 68 g6 lbc6
queenside puzzled many commen­ 69 i.b5 e3 70 g7 l1Je7 7 1 i.xe8 ! e2
tators, and the following day there 72 lbxe2 a2 73 lbc3 ! lbg8+ 74 <j;f7
were even headlines in the local lbh6+ 75 <ifi>f8 a1 'iW 76 lbb5#.
 Karpov - Timman 149

It would not have been too bad 67 ltJal b5

ending like this, but unfortunately,
Black has a defence: 64 ...�d6 65
.i.xc4 dxc4 66 ltJa2 .i.a4 67 ltJc3
.i.b3 68 tiJb5+ �e7 69 ltJxa3 c3 70
�f4 b5 ! 7 1 �xe4 b4 72 �d3 bxa3
73 �xc3 .i.d5 drawing.
64 ••• tiJb2
68 .i.dl b4 (D)
Bringing the king slightly closer
to the pawns does not work:
68 .. .<iPd6? 69 .i.b3 tiJc l 70 cJ;h6 b4
71 f6 �e6 72 Wg7 liJxb3 73 liJxb3
.i.a4 74 ltJc5+! cJ;d6 75 f7 al 'iW 76
f8'iW+ cJ;c6 77 'il'c8+ cJ;d6 78 'il'd8+
Wc6 79 'i!i'd7+ 'it>b6 80 'ii'b7+ cJ;a5
8 1 'iWa6#.

. .... . .
�
-
�� - � ·
� -
� .
• • • •
�
-
/'·�
� - ·
�·-
� ���
� �
�
- � � ... -
��-
�
" -� ·
�
�
• • •o •
B
.... . . .
65 �g5 � .JL. •
After 65 ltJa2, the long variation w
65 . . . �d6 66 �g5 �e7 67 �h6 69 .i.b3 tiJcl
�f6 68 tiJb4 .i.f7 69 .i.fl ! tiJd I 70 After 69 ...'it>d6 70 cJ;f6 Black's
g5+ �e7 7 1 �g7 ltJxe3 72 f6+ king is kept away and (he white g­
�e6 73 .i.h3+ tiJf5+ 74 .i.xf5+ pawn is irrepressible.
�xf5 75 rJifxf7 e3 76 'iii>g7 ! e2 77 70 .i.xd5 'iii>d6
ltJc2 a2 78 f7 e l 'ii' 79 liJxel al 'i!l 71 .i.c4 .i.b5
80 f8._,+ �xg5 8 1 'iWf2 promises There is also the interesting pos­
White good prospects, but after the sibility of 7 l ...b3 72 .i.xb3 liJxb3
correct 65 ... b5 ! nothing appropri­ 73 liJxb3 'it>d5 74 tiJal 'it>c4 75 cJ;f6
ate can be found. 'it>c3 76 cJ;e7 cJ;b2 77 'it>xe8 'it>xal
65••• tiJd3 78 f6 cJ;b2 79 f7 al 'iW 80 f8'iW
It is too late for 65 ...ltJc4 66 �f6 'ifa4+ 8 1 'it>f7 'i!i'b3+. This position
ltJxe3 67 cJ;e7. arose in our analysis, when we
66 tiJb3!? a2 considered that 82 'it>f6 Wxe3 83
150 Brussels Ct (4) 1991

�e5 �c2 84 'iff4 �d3 85 d5 73 000

would have given chances for Only now does Timman make
victory, but in our post-mortem, the decisive error. 73 ...b3! 74 .i.xb3
Timman demonstrated the more lbxb3 75 lbxb3 .i.c4 76 lbal �f6 !
striking continuation 82 d5 ! 'ii'xe3 would have led to a draw.
83 'ifb4+ �c2 84 g5 'iff3+ 85 �g7 74 .i.e6 i.d7
e3 86 'ifc4+ �b2 87 'ifb5+ �c 1 88 75 g5 b3
d6 e2 89 'ifc5+ �b2 90 'ffb6+ �c 1 76 g6 1-0
9 1 d7 el'if 92 d8'ff 'ffec3+ 93 If 76 ...i.xe6, then 77 fxe6 b2 78
'ifdf6, etc. g7+ �g8 (78 ...�e7 79 g811f bxal'ff
72 i.g8 �e7 80 'it'f7+ �d6 8 1 'ifd7#) 79 e7 �
73 �h6 (D) 80 g8'it'+ �xe7 8 1 'it'g5+ �f7 82
In the event of 73 f6+ �f8 74 'ifd5+ and after two more checks
.i.e6 b3 (74 ... .i.a4 leads to the White takes the b2-pawn.
same position as in the game, but
74 . . . .i.d7 ! will save Black, and if Game 38
75 i.xd7? then 75 ...b3 76 �f5 b2 Karpov - Anand
77 g5 lbe2! 78 g6 lbg3+ 79 �g5 Brussels Ct (4) 1991
lbh5 ! 80 �xh5 bxal'ff 8 1 g7+ � QGD, Semi-Slav
82 .i.e8+ �xf6 and Black unex­
pectedly takes the initiative) 75 1 d4 d5
.i.xb3 lbxb3 76 lbxb3 i.c4 77 lba1 2 c4 c6
�f7 78 �f5 White has the upper 3 lbf3 lbf6
hand. 4 lbc3 e6
5 e3 lbbd7
6 'ifc2
After 6 .i.d3 dxc4 7 i.xc4 b5 8
i.d3 the classical variation of the
Meran arises. Moving the queen
forward to c2 is a standard Anti­
Meran possibility.
6 000 .i.d6
7 .i.e2
Besides this move, White has a
choice between 7 i.d3, 7 g4 and 7
b3.
7 ••• 0-0

8 0-0
Black's counterplay is linked
with the advance . . .c5 or . . .e5 (or
both), usually after a preliminary
exchange on c4. All my games as
White in this match against Anand
developed in this fashion.
8 .•• dxc4
9 .ixc4 'fle7 (D)
In the second game of this match
after 9 ... a6 10 l:td1 'fle7 1 1 h3 b5 I
did not place my bishop too suc­
cessfully: 12 .id3 c5 1 3 lDe4 c4 14
lDxd6 'flxd6 15 .ie2 .ib7 1 6 .id2
l:tfc8, and Black solved his open­
ing problems.
w is quite tangible.
10 h3 a6
In game 6 Anand improved his
play by means of 10 ...c5 1 1 dxc5
.ixc5 1 2 e4 .id6 1 3 lDd4 (incred­
ibly, at that very moment on the
same stage in the quarter-final
match between Korchnoi-Timman,
1 3 lDb5 lDe5 14 �e5 .ixe5 15 f4
Karpov - Anand 151
'ifc5+ 16 �h2 .id7 ! was being
played, and Black had a favourable
game) 1 3 . . . lDe5 14 .ib3 .id7 15
.ie3 lDg6 and everything was in
order for Black.
My dispute with Anand contin­
ued to the 8th, and decisive, game
of the match. On that occasion I
changed my move order and
played 10 a3. Then in the event of
10 ...c5 1 1 dxc5 .ixc5 Black has to
bear in mind 1 2 b4 .id6 1 3 lDb5,
but Anand instead advanced his e­
pawn - 1 0...e5.
Thus, in this opening variation
White can play a3 or h3, while Black
can play ...c5, or ...e5, and any of
these moves could essentially in­
fluence the future course of the bat­
tle. Modem opening theory is built
on precisely such nuances. I am
convinced that assessments here
are still to be redefined more than
once.
1 1 e4! e5
12 l:tdl ! b5
After 12 ... exd4 13 lDxd4 �5 14
.ifl White's pressure in the centre
13 .tn c5
14 d5 c4?!
A serious inaccuracy; the mod­
est 14 ...l:tb8 was correct, and then
in the event of 15 a4 b4 1 6 lDb1 b3
17 'iie2 l:tb4 Black would have
gained counterplay.
15 a4 l:tb8
152 Brussels Ct (4) 1991

Now 1 5 ...b4 does not work be­ 22 .!Dxc5 i.xc5 23 .:tal b2 24

cause of 16 lbe2 .!Dc5 17 .!Dd2 aS i.xb2 .!Dxe4 25 'ii'xe4 .:l.xb2 26
1 8 lL!g3 and the advanced black .:l.d2 either.
pawns become objects for attack. 20 'ii'e2
16 axb5 axb5 The c4-pawn for the time being
17 .:l.a5! (D) is invulnerable: 20 'ii'xc4 .:l.b4 2 1
�c6 lL!b8.
20 .!DeS
_... . -·· 21 .!Dxc5 i.xc5
� �
�
-
� -
��ma
� - l. ­
� l. 22 .:tal c3!
�· -
�
-
� - �- . .·� Postponing this pawn's march

a••�- • until a later date is unsuccessful,

viz. 22 .. .'ii'd6 23 i.d2 i.a6 24
··-�- - i.c3, and White's superiority is be­
�
-
� �
� � · - ltj�
-
�
�� yond doubt.
,.•·
�-� \lllr ��d
• �
!;,' . 23 .!Dxe5!

• -
d � .: d
·.i.�
�
8 24 i.xb2 .:l.xb2 (24....!Dxe4 25
I had this position on the board
during my preparation for the
game. The attack by the black
queenside pawns, which is inci­
dentally forced, seems dangerous,
but in fact these foot soldiers will
soon be stopped.
17 ••• b4
18 .!Da4
1 8 lL!b5 •d8 ! 1 9 .!Dxd6 'ii'xa5
20 .!Dxc4 'ii'c7 !? is not so clear.
18 ... 'ii'd8
19 .:l.a7 b3!?
After 1 9 . . . .!Db6 20 i.e3 b3 2 1
'ii'b l it i s more difficult for Black to
create counterplay, and he does not
have full compensation in the event
of 1 9 ... c3 20 bxc3 b3 2 1 'ii'e 2 lL!c5
Anand had probably only reck­
oned on 23 bxc3, and then 23 ...b2!
1i'xe4 .:l.xb2 26 'ii'x e5) 25 'ii'xb2
.!Dxe4 26 .:l.d2 .!Dxd2 27 'ii'xd2 e4
28 lbel e3 ! would have led to a po­
sition with chances for both sides.
After this bold knight move the en­
emy pawn reaches White's second
rank, but fortunately its career
comes to an end there.
23 ••• c2
24 .:l.d3 (D)
I chose this continuation, al­
though I also examined 24 .:l.el
.:l.e8 25 .!Dc6 'ii'b6 26 'ii'f3 which
has roughly the same value.
24 ••• 1t'e8
Now 24 . .. .:1.e8 loses: 25 .!Dc6
'ii'b6 26 e5 .!Dxd5 27 .:l.xd5 'ii'xc6
28 \i'c4.
 Karpov - Short 153

35 lbd5 li:)xcJ
-.t.- -·- 36 l:td3 li:)a2
� -
- � -�··· 37 lbb3 .i.c8
- - - -�
' !:'-'

38 .i.c4! g6

- ·8� - 39 l:ta3 cHi'

- - 8 - �· •
40
41
l:txcl
l:tcJ
li:)xcl
l:te8
- ·-:- - 8 42 lbcl l:te7
B • -'iVD
� 8.
'�
!:'-' .- !i!
43 .i.n -
10

a m -i.=
� + if? ' !i!
Game 39
B -
Karpov Short
25 li:)c6 l:tb6 Linares Ct (7) 1992
26 .i.e3 li:)xe4 QGD, Tartakower
27 .i.xcS li:)xc5
28 l:te3 'iVd7 1 d4 d5
29 'iVc4 l:txc6 2 c4 e6
29 ...'iVd6 30 l:tc3 li:)d7 3 1 l:txb3 3 ll)cJ li:)f6
is no better. 4 .i.g5 .i.e7
30 dxc6 'iVd1 5 e3 0-0
31 l:te1 'iVd6 6 ll)f3 h6
32 'iVcJ! 7 .i.h4 b6
White has a clear-cut plan to win The Tartakower Variation of the
a pawn: the queen makes way for Queen's Gambit, also known to
the bishop on c4, after which the Russian players as the Bondarev­
threat of l:ta5 arises. sky-Makogonov Variation, is very
32 ••. 'iVd5 widely used in modem practice,
32 ... .tf5 33 'it'e5! 'iVxe5 34l:txe5, and moreover many grandmasters
etc., does not help. willingly play it as both Black and
33 c7! White. When the h-pawn attacks
33 l:ta5 would have been too the bishop, you always have some
hasty in view of3 3 ... li)e4! 34 l:txd5 hesitation: should you exchange on
li:)xc3 35 l:tc5 li:)a4 ! 36 l:tb5 li:)xb2 f6, or wait until ... b6 has been
37 l:txb3 li:)dl and Black unexpect­ played, and temporarily retreat the
edly wins. bishop to h4? In my encounters
33 .i.b7 with Kasparov, this bishop di­
34 l:taS li:)e4 lemma arose 24 times (equivalent
154 Linares Ct (7) 1992

to a whole World Championship this looks rather passive. After 1 2

match !), and most frequently the 0-0 consider the following exam­
bishop avoided immediate ex­ ples:
change. a) 12 .. .'ifd6 1 3 'ifb3 lLld7 I4
8 .i.e2 .i.b7 :rei .:r.ad8 I 5 .:r.ab1 .:.res I6 .i.d3
9 .i.xf6 .i.xf6 .i.a8 (Speelman-Short, London Ct
10 cxdS exdS (D) ( I ) I 988) I7 h3 and then e4 gives
At first glance the position White a small but stable advantage.
seems fairly simple, but this vari­ b) I 2 . .. .:r.es 1 3 'iib 3 and then
ation is quite poisonous, and de­ (D):
mands careful play by Black, as
we shall see. Curiously enough,
this position arose eight times in
my matches against Kasparov. In
spite of the stormy nature of our
battles, these disputes strangely al­
ways ended in peace.
B lack is already preparing to
carry out ... c5, and the natural reac­
tion is b4, and after an exchange on
c5, l:[bl .

B
b i ) 1 3 ...'ifd6 I 4 a4 lLld7 1 5 a5
l:tad8 I 6 axb6 axb6 17 l:ta7 'iib 8
IS .:r.a2 b5 I9 lLle1 .i.e7 20 lLld3
.i.d6 2I g3 lLlb6 22 .i.f3 .i.c8 23
l:tfal .i.f5 24 lLlc5 lLlc4 25 lLle2
gave White a secure plus in P.Nik­
olic-Short, Manila IZ 1 990.
b2) Black tried to improve with
1 3 . . . a5 in Karpov-Short, Amster­
dam 1 99 1 . The game continued I4
w a3 lLld7 1 5 b5 c5 1 6 lLlxd5 .i.xd4
1 1 b4 cS ( 1 6 ... .i.xd5 I7 'ifxd5 lLle5 1 8 'iib3
By playing I l ...c6 Black could ltlxf3+ I9 .i.xf3 .:r.cs 20 .i.c6 is no
have avoided hanging pawns, but good). After 17 .:r.ad 1 lLle5 ! 1 8

lt:ixe5 -'.xd5 1 9 lt:ic4 'ifg5 20 g3
'iff5 2 1 1lfe1 'ife4 22 f3 'ifxe3+ 23
'ifxe3 llxe3 24 lt:ixe3 -'.xe3+ 25
�fl -'.d4 26 llxd4 cxd4 Black
managed to stand firm, but 17 exd4
llxe2 1 8 1lfe 1 ! ? would have given
Black certain difficulties. What­
ever, at our next meeting in the
Candidates match, Short pushed
his c-pawn two squares.
12 bxcS bxcS
13 llb1 'ii'aS
As a whole series of games
against Kasparov showed, 1 3 .. .-'.c6
is more passive.
14 'ifd2 cxd4
15 lt:ixd4 -'.xd4
16 exd4 (D)
Not, of course, 16 'iVxd4 lt:ic6 1 7
'iVd2 d 4 1 8 exd4 -'.a6 an d Black i s
already better [editor's note: but in
The Queen 's Gambit for the At­
tacking Player 1 9 -'.xa6 'ifxa6 20
lt:ie2 llfe8 2 1 llb2 is analysed to
an ending with marginal winning
. chances for White].
16
••• -'.a6
1 6 . . . -'.c6, as seen in the 40th
game of my first match against
Kasparov, is weaker. I was playing
B lack, and I only managed to hold
the position with difficulty.
17 lt:ibS 'ii'd8
18 0-0
Now the black knight has two
possibilities - it can develop to d7
or c6. After 1 8 . . . lt:id7 a rapid rout
Karpov - Short 155
B
ended the game Vaganian-Geller,
New York 1 990: 1 9 1lfc 1 lt:if6 20
f3 ! lle8 2 1 a4 1le7 22 -'.d3 -'.c8?
23 'iVf4 ! a6 24 lt:ic7 g5 25 ...xf6
lle1 + 26 1lxe1 'ifxf6 27 �hl 'ilfxd4
28 1lbd1 ! , etc.
18 ••• lt:ic6
19 .:.Cd1
Soon after our Candidates match
the Olympiad took place (Manila
1 992), and Short again chose this
variation against Azmaiparashvili.
The Georgian team captain obvi­
ously supposed that after our duel
Short would have prepared some
sort of improvement, so he played
something different: 1 9 a4. I have
also played this pawn move, if
slightly later, but here White de­
cided there was no time to lose. Af­
ter 1 9 . ..'iff6 20 llfd1 1lfd8 2 1 1lb3
llac8 22 h3 ...g5 23 'ilfxg5 hxg5 24
llg3 ! f6 25 -'.g4 1lb8 26 1lc3 Black
had a difficult endgame.
19 ••• 'iVf6
156 Linares Ct (7) 1992

20 �n (DJ
A modest novelty, which I had
prepared specially for the match. It
is more precise than 20 �f3, as in
the game Yusupov-Beliavsky, Lin­
ares 1 988, where after 20 ... .:tab8
2 1 a4 .:tfd8 22 'iic 3 �c8 23 .:tbc 1
a6 24 'iixc6 'iixc6 25 .:txc6 axb5
26 axb5 the players agreed a draw.

w
26 .:te3! lbe6
26 .. ..:tc2? loses directly: 27 .:te8+
'iii>h7 28 'iixc2 1fxc2 29 �d3+.
27 .:te5 'ii'g4
28 �e2! 'ii'g6
29 g3 .:tbc8 (D)

. �. . . • •
B • • •••
20 .:tabS • ·41\·�-
:res
21 a4
22 .:tb3! �xb5
·�•·a •
23 axb5 lbd8 (D) • � -� IJ
. ·
� -
� ·
�
�
• • • B
� �

24 'ifa2!
Now the white rook on the third
�� �-
'ii' IJ
· � ��
/� u u �
rank prepares to move to one of
three important squares, a3, e3 or -
·
� ·
� -
� .:. ·
� -
� �
� �
f3. w
24 ••• .:tc7 30 �h5!
If 24 ....:tb7, then 25 .:te3, threat­ The d5-pawn is going nowhere,
ening 26 'ifxd5 and 26 .:te8+. The and so there is no point hurrying to
weak d5-pawn is causing Black a take it: 30 .:txd5? .:tc2; 30 jfxd5
lot of trouble. .:td8 3 1 'ji'e4 'ji'xe4 32 .:txe4 .:tcd7.
25 .:ta3! 'ji'fS 30 ••• 1i'f6
 Karpov - Kamsky 157

31 'ii'xd5 g6 38 . .. 'ilt'xd4 39 l:txd4 l:taa2, though

32 .te2 l:tc3 it is true that here the precise 40
33 h4 l:ta3 (D) �f l ! would have maintained an
appreciable advantage for White.
39 l:tg8+ �xg8
40 'ii'xf6 l:txe2
41 :at lbe4
42 l:txa2 l:txa2
43 'ii'd4 l:txf2+
44 �gl l:te2
45 d6 1-0

Game 40
Karpov - Kamsky
Moscow Alekhine mem 1992
w Grtinfeld Defence
34 'ii'e4 l:tcc3
35 �g2 �g7 This game, played at the Moscow
36 d5 lbc5 super-tournament celebrating the
37 'ilt'd4 l:tc2 l OOth anniversary of Alekhine's
38 l:te8 (D) birth, went, as they say, wonder­
fully well. A bitter struggle was
conducted over the whole board,
• . •:• • first on one flank, then the other,
� -
- �
�� ·.·�
- and in the end the black pieces be­
•
mu -
. ."" -
�
� -
- ... came totally disorientated.
· � - � - •:n
-
1 d4 lbf6
2 c4 g6
-
�� - � . � -� � 3 lbfJ .tg7
- �
.

- �
-
�
- � � u 4 g3 c6
B .I Bi.B�· First White (by playing 4 g3 in­

. •:• . stead of 4 lbc3) and then Black

(by 4 ... c6 instead of 4 ...d5) avoids
B the sharpest continuations in this
38 ••• l:taa2? opening, giving the game a posi­
Black could have shown more tional feel.
dogged resistance by continuing 5 .tg2 d5
1 58 Moscow Alekhine mem 1992

6 cxdS cxdS 1 2 .. . ltle7, so that after 1 3 e4 he

7 ltlc3 0-0 could exchange pawns and seize
8 ltleS e6 the d5-square. Although Black
9 0-0 managed to equalize, it was felt
9 .i.g5 has been played several that his development was defi­
times, but it is not compulsory to nitely awkward.
fix the position of the dark-squared 13 e4 ltle7
bishop at this early stage. With the idea of placing the
9 ltlfd7 bishop on c6 and strengthening the
10 f4 (D) centre.
14 ltlxd7 'ifxd7
15 eS (D)

B
Supporting the knight with the
pawn is more logical than retreat­ B
ing it to f3. 15 .•• l:.ac8
10 •.• ltlc6 1 5 ...l:.fc8 was Kasparov's pref­
1 0 ...f6 is also played quite fre­ erence in Karpov-Kasparov, Se­
quently, but practice has shown ville Weh ( 1 ) 1987. This game went
that it has more disadvantages than 1 6 l:.cl .i.f8, when White could
advantages. have sacrificed a pawn with 17 g4
1 1 ..t.e3 ltlb6 l:.c4 18 f5 !?, achieving a serious in­
Here as well 1 1 . . .f6 is possible itiative. Instead the game contin­
(see game 37). ued 17 .i.f3 l:.c7 1 8 b3 l:.ac8 1 9
12 .i.f2 .i.d7 'ilfd2 ltlc6 20 'ilfb2 a6 2 1 .i.e2 'ilfe7
In Karpov-Kasparov, Seville 22 ltlb1 l£Jb4 23 ltlc3 ltlc6, ending
Wch (3) 1 987, Kasparov preferred soon in a draw by repetition.
 Karpov - Kamsky 159

16 l:lcl a6 (D) The first in a series of fine

Now Black has a ' hanging' moves by the queen, which charac­
knight on b6, which will soon put terized this game. In the variation
him on the spot. 1 6. ..l:lc7 would 20 l'Lle2 l:lxc 1 2 1 l:lxcl l:lxc l + 22
have been safer, and if 17 'ifb3, tbxc 1 ltlc6 23 lDd3 White main­
then 17 . . .ltlc4 1 8 l'Llb5 l:lc6 1 9 tains his spatial advantage, but the
ltlxa7 l:la6 20 l'Llb5 l:lb6 with coun­ position has lost its internal energy,
terplay. Generally, while Black's and his chances on the kingside are
position on the queenside looks not all that great.
quite reliable, the less said about
the kingside the better.
•
.••
. �
- ••
-
· · -·- · · ·
·- •••••
B BiD B
•• u � u
�8· .
• 8 !!U
· " mu- •·
8•8 8 • M
� .JL
�U R
.
- � •
- -:�*!
B
20 ••• lbc6
21 fS! �a3
17 b3 All the black pieces have trans­
Emphasizing the awkward posi- ferred to the queen side, but for the
tion of the knight on b6. time being this is not dangerous,
17 ltc7 because the white pieces on the
1 8 'ifd2 :Cc8 other half of the board are not yet
19 g4! co-ordinated. 2 1 .. .exf5 22 gxf5
The starting gun is fired for 'ifxf5 is unsuccessful in view of 23
White's attack. The constrained �h3 'ifh5 24 �xc8 �h6 25 'ifh3
nature of the black pieces makes �xcl 26 'ifxh5 gxh5 27 �xb7
counterplay significantly more dif­ �b2 28 �xc6. 2 l ...�b4 would not
ficult. have led Black to his aim either,
19 �f8 due to 22 �hl intending 23 �el .
20 '1Ve3! (D) 22 l:lcd1 l'Llb4
160 Moscow Alekhine mem 1992

Now the black pieces are getting
under each other's feet. The knight
is blockading his own bishop, a
fact of which I instantly made use.
However, 22 ...lbxd4 23 W'xd4 is
no better, since the b6-knight is
hanging (23 ....tc5 24 Wf4 .txf2+
25 1fxf2).
23 'ii'h6 (D)
Stronger than 23 lbbl lbc2! 24
W'h6 ..tf8.
when White simply has an extra
pawn.
24 lbbl!
Now the bishop has to desert the
important a3-f8 diagonal.
24•.. .tb2
25 'ii'd 2!
The white queen's shuttling is
causing Black more than a little
bother.
25•.. lbc2 (D)
The startling 25 ... a5 ?! can be re­
futed by means of 26 a3 (the trap is
26 ti'xb2 l:r.c2 27 'ii'a3 l:r.xa2)
26 . ..l:r.c2 27 W'e l ii'bS (27 ...lbc6
28 l:r.d2 !) 28 axb4 .:.e2 29 it'xe2
"it'xe2 30 bxa5 (or 30 .:.d2 'ii'xg4 3 1
.:.xb2) 30...lbd7 3 1 .:.d2 .

........,
• • - • • a"i
·- ····­
23 'iWe8
• ··8�·
�
�-��
u -
·�J � �
� -
�
•••

In the new situation 23 . ..lbc2
does not guarantee equality, as af­
ter 24 lbe2 the black rooks have no
entry square for their invasion.
23 .. .'ii'e7 is no better: 24 lbbl .tb2
25 ..th4 'ii'f8 26 'ii'd2 .:.c2 27 1Wel ,
and the black pieces are soundly
tied up on the queenside, despite
their superficial activity. 23 ...lbd3?!
does not work either, in view of 24
lbxd5 ! lbxd5 25 .:.xd3 .tf8 26
'ii'g 5 h6 27 fxe6 ! 'ii'xe6 28 .txd5,
-�· . -
� ··· fli.B
attJa : a : =
w
Black has been demonstrating
the maximum activity for the last
five moves. However, the threaten- ,
ing-looking tandem of the bishop
on b2 and knight on c2 has to be
defended, and only reduces the
 Karpov - Kamsky 161

activity of his own rooks. Having

siopped his opponent's aggression,
White begins to develop his attack.
26 �hl !
Freeing g l for the bishop, which
in its turn uncovers the fl -rook.
26 ••• "ile7
27 i.gl lbd7
28 ·.Uf3
With every move White's attack
is gaining energy, while Black is
still at a dead end.
28 ••• 'ifb4
29 'ifh6!
Black has no time to take the d4- The black queen cannot yet re­
pawn due to the threat of 30 .Uh3 turn to f8 in view of the reply fxe6.
lbf8 3 1 f6. 29 'iff4 would have 31 ..• b6
been slipshod: 29 ...i.xd4! 30 i.xd4 There is no other way of freeing
lbxd4 3 1 .:Xd4 .:c 1 + 32 .:.n .:Xfl + the 'hostages' on b2 and c2. Black
33 i.xfl 'ifel 34 lbd2 lbxe5 with is preparing 3 1 . ..a5 in order to se­
sufficient counterplay. cure the b4-square for his knight to
29 ••• 'iff8 retreat.
30 'ifgS (D) 32 .Udn aS
In the event of 30 'ifh4 the cap­ 33 h4 lbb4
ture 30 ... lbxd4 does not work due 34 a3
to 3 1 i.xd4 i.xd4 (or 3 1 .. ..l:lc l 32 Obviously not 34 'ifxb2 owing
i.xb2 ! .l:lxdl + 33 .l:lfl ) 32 .Uxd4 to 34 ... .Uc2. After the text the rook
.Uc 1 + 33 .Ufl .Uxfl + 34 i.xfl .l:lc 1 invasion on c2 is not dangerous for
35 lbd2 and White maintains a White.
material advantage. However, both 34 .Ucl
30 ... 'ifd8 and 30 ...'ifg7 are quite 35 'iff4 lbc6
reliable for Black. Now after either 36 .th3!
of these moves White should con­ Threatening 37 fxe6 fxe6 38 g5
tinue 3 1 'ifd2, and in comparison lbf8 39 'ifxf8+ ! ! 'ifxf8 40 i.xe6+
with the position on move 28, the �g7 4 1 .l:lf7+! , winning.
situation has altered appreciably in 36 lbd8
White's favour. 37 i.e3!
1 62 Moscow Alekhine mem 1992

The c l -square is under control, li)fS 47 'ife7 ! 'ifd7 48 .i.h6 ! - an

therefore the bishop scrutinises the amusing finale!
h6-square. 43 'ili'c2 li)xe5
37 ..• b5 Despair. 43 . ..'ili'f8 44 'ifc7 'ife8
38 ft3fl! (D) 45 f6 is also bad, and if 43 ... li)b8,
Having appreciably strength­ then 44 li)d2 ltJdc6 45 lbf3 is suffi­
ened his position on the kingside, cient, as White is effectively play­
White suddenly changes tactics. ing with an extra piece. But now,
Paradoxically, because of the awk­ having accepted the sacrifice, I
ward position of his pieces, espe­ managed to find a beautiful forced
cially the b2-bishop, Black, after win.
exchanging rooks, will perish pre­ 44 dxeS 'iWxeS
cisely where he once dominated. In 45 'ili'c8! 'ili'e4+
a few moves the white queen be­ 45 ...'ibe3 allows mate in three.
comes mistress of the position on 46 i.g2 'ili'xbl +
the c-file. 47 �h2 i.b2
48 'ili'xd8+ �g7 (D)

· · - •••
•l •
� ···• • · mit
-·
• •••••
�
-·· �
�z . �-� � -
�
�� �u
�
- - u
�
� �tffi � "'
" �

o � • m •.t
-
--� �- �
� r• �
-
�
.l[). • .:. •�
B
38 b4 w
39 axb4 axb4 49 f6+! i.xf6
40 llxc2 llxc2 50 i.h6+! �h6
41 :n ftxfl 51 'ili'xf6 'ifcl
42 'ili'xfl .i.a3 52 g5+ �h5
Alas, after 42 ....i.c3 there is the 53 �gJ!
immediately decisive 43 f6 'iff8 44 Strangely, 53 �h3? would have
'ili'c2 i.el 45 'ifc7 'ili'e8 46 i.fl let victory slip from my grasp:
 Kamsky - Karpov 163

53 ...1Wf5+ 54 1Wxf5 gxf5 55 .i.f3+ 7 ltJ1f3 .i.d6 (D)

�g6. One of the key positions in the
53 ..• 'ii'c7+ modern treatment of the Caro­
54 �h3 1-0 Kann. Besides 7 ....i.d6 Black may
For this victory in the Alekhine choose between 7 ... 'fic7, 7 ... i.e7
memorial, I was awarded a prize and 7 ...h6.
for 'playing in Alekhine's style' .
Strictly speaking, I do not feel this
was not a very precise decision. In
fact, in the work of the fourth
World Chess Champion there were
many examples of the battle being
transferred from the queenside to
the kingside. But here we were ac­
tually dealing with a rarer theme,
as on the contrary, I managed to
transfer the battle from the king­
side to the queenside with decisive
effect! w
Then this beautiful victory also 8 1We2
soon won the best game prize in In­ More active than 8 0-0, which
formator. Kasparov once played against me
(Amsterdam 1 988); then after the
Game 4 1 continuation 8 ...h6 9 lbe4 ltJxe4 10
Kamsky - Karpov .i.xe4, by playing 10...ltJf6 !? Black
Dortmund 1993 could have equalized. In fact I
Caro-Kann Defence chose another plan: 10 ...0-0 1 1 c3
e5 1 2 .i.c2 l:le8 1 3 lle1 exd4 1 4
1 e4 c6 l:lxe8+ ..,xe8 1 5 'fixd4 'fie7 16
2 d4 d5 .tf4 .txf4 17 'fixf4, and White
3 ltJd2 dxe4 gained an advantage.
4 ltJxe4 ltJd7 8 ... h6
5 ltJg5 ltJgf6 9 ltJe4
6 .i.d3 The standard sacrifice on e6 is
6 .i.c4 is another popular con­ not dangerous for Black in this
tinuation. situation: 9 ltJxe6 fxe6 10 .i.g6+
6 ... e6 �e7 1 1 0-0 lDf8 1 2 .td3 .td7 1 3
164 Dortmund 1993
lbe5 •e8 14 f4 �d8 1 5 c4 �c7 1 6
�d2 .:.cs 1 7 c5 �e7 1 8 f5 �b8 1 9
.:.ael �d8, and he had a clear ad­
vantage in the game Amason-0st­
enstad, Torshavn 1987.
9 •.• tDxe4
10 1i'xe4
In the event of 1 0 �xe4 0-0
Black quickly plays ...c5 or ...e5
with a good game.
10 lDf6
1 1 ... h4 (D)
B 12 lDeS
1 1 'ii'e2 is played more fre­
quently, but in fact on h4 the queen
looks quite threatening. 1 1 . .. 0-0
will not do because of 1 2 �xh6 !
gxh6 1 3 'ii'xh6, and Black cannot
defend himself from the threat of
14 lLlg5 . After l l ...'iia5+ 1 2 �d2
'ifh5 1 3 'ifxh5 lDxh5 1 4 lDe5 0-0
1 5 0-0-0 he has a prospectless posi­
tion.
1 1 . . .lDd5 is recommended by
theory, but after 1 2 'ifxd8+ �xd8
1 3 c3 White preserves a small but
stable positional advantage. What
should Black do?
11
••• �e7!?
I thought of this paradoxical ma­
noeuvre, keeping the king in the
centre, in 1 988, when I was prepar­
ing for the above-mentioned game
against Kasparov in Amsterdam. I
was afraid of the white queen being
transferred to the king side, but ex­
changing queens struck me as be­
ing rather insipid. I sat for some
time over this position before I
found the correct solution. So this
important novelty had to bide its
time for five whole years !
With this odd king move Black
unexpectedly brings harmony to
his forces. Now the threat of ...g5-
g4 has been created, winnilfg a
piece, and if White is to count on
keeping his opening advantage, he
will have to act decisively.
The most ambitious solution.
Bearing in mind that the enemy
king is stuck in the centre, White
decides to sacrifice a pawn. Of
course, he will hardly be happy af­
ter 1 2 �f4 �b4+ 1 3 �d2 (else
1 3 ... g5) 1 3 ...�xd2+ 14 'itxd2.
12 �xeS
13 dxeS 'ila5+
14 cJ (D)
1 4 �d2 'ifxe5+ 1 5 �e3 is also
possible, and it is very dangerous
 Kamsky - Karpov 165

for B lack to take on b2, altho1:1gh 20 f4 (in order to close the c l-h6
the position of the c-pawn is hardly diagonal from checks) 20...gxf4 2 1
better for White than in the game i.d4 seems dangerous, for exam­
itself. ple (D):

.. . . . .
�
- •.t. . - ··� �
-
�
-
- � �·- -
-
� � -
� -
� - - � - mu ·
�
�
-
�
� - �
� - � -
�
Wl1
- u � i.�
·
� � ­
�
� � -� -
, �
� �
�,\&��-�"u
. � �:- � �- �
�
w
. B
14 'ii'xe5+ a) 21 ...ifd5 loses on account of
15 i.e3 b6 22 l:.e5 Wxg2 23 l:.xc5! bxc5 24
16 0-0-0 g5 ifxc5+ �e8 25 ..txf6.
17 'ii'a4 cS b) 2 l ...'it'c7 is also in White's
18 libel ..td7 favour: 22 i.eS ifc6 23 i.xf6+ (a
19 'ii'a3 l:.hd8 (D) rare combination arises after 23
..txf4 ! ? 'ilt'a4?? 24 ..td6+! �e8 25
l:.xe6+! fxe6 26 i.g6#) 23 ...�xf6
24 i.e4 Wa4.
c) 2 1 . . .'it'g5 ! 22 l:.e5 ifh4 ! res­
cues Black, and 23 g3 does not
lead White to his goal in view of
23 . .. fxg3 24 hxg3 ifxg3 25 l:.xcS
bxcS 26 ifxcS+ �e8 27 i.xf6
iff4+.
20 ••• 'ii'c7
The black queen abandons the
centre of the board, and restrains
all approaches to the king.
20 g3 21 ..td4
166 Dortmund 1993

Now after 2 1 f4 there is the

strong retort 2 1 . .. ttlg4 22 .td4
gxf4 23 h3 ttle3.
21 ... .te8!
22 �b1 l:td5
22 . . . .tc6 ! 23 .te5 'ifd7 24 .tc2
'ifb7 is perhaps even better.
23 f4 .:.Sd8
24 .tc2
24 .te5 1i'c6 ! 25 c4 l:t5d7 26
'ifc3 ttlg4 27 fxg5 hxg5 28 l:tfl
l:td4 ! favours Black.
24-..• l:t5d6
25 .txf6+ �xf6
26 fxg5+ hxg5
27 l:txd6 l:txd6
28 c4 have been stronger.
The only move, as otherwise the 34 W'h7+ �d8
white queen turns out to be ex­ 34 .. ..tf7 35 .tg6 l:tdl + would
cluded from the game. have led to a draw (35 ...'ifxc4 36
28 �e7 b3 ii'e6 37 l:tgl ).
29 ii'e3 f6 35 h6 l:td2
30 h4 (D) 36 W'f5 'ifxf5
Here in mutual time trouble, 37 .txf5 .td7? (D)
Kamsky offered me a draw, but I
turned it down, as Black's chances
are evidently better: White has
been deprived of the advantage of
the bishop pair, and he is still a
pawn down.
30 gxh4
31 gxh4 'ifd7
32 W'h6 e5?!
In time trouble I decided to de­
fend the f6-pawn, and missed a
chance to do something much
more effective: 32 ...l:td2 ! 33 l:tfl
 Karpov - Kramnik 1 67

38 i.g6? 49 l:txb6 l:txa2

A mistake; Kamsky failed to 0-1
make use of my blunder in time If you exclude Black's mistake
trouble. After 38 �c 1 ! Black would in time trouble, the game was quite
even have to fight for a draw: logical, and its result was fully in
38 ...l:td4 (the rook cannot abandon order.
the d-file because 39 l:td1 would
win the bishop) 39 i.xd7 ! l:txc4+ Game 42
(after either capture on d7 there Karpov - Kramnik
follows 40 l:td1 , and the h-pawn Linares 1994
will inevitably promote) 40 �bl QGD, Semi-Slav
l:th4 (40 ... �xd7? 4 1 l:th l ) 4 1 l:td l
l:txh6. 1 d4 dS
38 l:th2 2 c4 c6
39 h7 �e7 3 �f3 �f6
40 i.d3 i.e6 4 �c3 e6
41 l:tg1 fS 5 e3 �bd7
42 l:tg7+ �f6 6 i.d3
43 l:txa7 e4 (D) Around this time I was avoiding
the Meran System, preferring the
Anti-Meran 6 'ii'c2 (see, for exam­
ple, game 38), but this did not
mean that I had excluded it from
my opening repertoire for ever.
6 dxc4
7 i.xc4 bS
8 i.d3
After the move 8 i.e2 it would
be more difficult for White to count
on an advantage, as the game Kar­
pov-Kasparov, Moscow Web (29)
1 984/5, showed: 8 ... i.b7 (8 ...b4 9
�a4 i.e7 10 0-0 i.b7 1 1 a3 a5 1 2
'ii'c2 0-0 1 3 l:td 1 c5 ! 1 4 �xc5 lill.c5
15 dxc5 'ii'c7 1 6 axb4 axb4 17
l:txa8 l:txa8 1 8 i.d2 i.e4 ! 1 9 'ifc4
'ifxc5 20 jfxc5 i.xc5 gave Black
1 68 Linares 1994
an equal position in Polugaevsky­
M.Gurevich, Antwerp 1 993) 9 a3
b4 1 0 lLla4 bxa3 1 1 bxa3 !l...e7 1 2
0-0 0-0 1 3 !l...b2 c5 112-112.
8 ... a6
After 8...!1...b7 9 a3 b4 10 lLle4 a5
1 1 lLlxf6+ lt:Jxf6 1 2 e4 !l...e7 1 3 'ii'e2
c5 14 !l...b 5+ �8 15 dxc5 !l...xc5 1 6
!l...d 3 h 6 17 0-0 g5?! 1 8 !l...e 3 !l...xe3
1 9 'iixe3 rlig7 20 li:Jd2 lDd7 2 1
lt:Jc4 'iie7 2 2 'ifd4+ e 5 2 3 'ifd6
'iWxd6 24 lDxd6 !l...c6 25 !l...c4 I
gained a big advantage in Karpov­
Shirov, Linares 1994.
9 e4 c5
10 d5 c4
My theoretical duel with the
young chess star soon continued in
Monaco, albeit in a rapid-play. Af­
ter the continuation 1 0 . . . 'ilfc7 1 1
0-0 !l...b7 1 2 dxe6 fxe6 1 3 !l...c 2,
Kramnik could have returned to
the game we are examining with
1 3 . .c4, but of course it ended sadly
.
for him, so he continued instead
with 1 3 ...!1...e7 !? ( 1 3 ...!1...d6 14 lDg5
gives White a big advantage) 14
lt:Jg5 W'c6 1 5 'ii'f3 h6 16 'ii'h 3. Here
1 6...lLlf8 !? would have led to a
rather complicated game, but in­
stead Black preferred a very dubi­
ous exchange sacrifice: 1 6 . .. hxg5?
17 ...xh8+ rlif7 1 8 ...h3 g4 19 'ilfh4
lt:Je5 20 f3 gxf3 2 1 gxf3. Black has
no compensation for his material
loss, and later I could have taken
the upper hand by force more than
once, but I was caught out by the
time control.
1 1 dxe6 fxe6
After 1 1 .. .cxd3 1 2 exd7+ 'ilfxd7
1 3 0-0 !l...b7 14 .:.et , it is well­
known that Black faces an uphill
struggle for a draw.
12 !l...c2 !l...b7
13 0-0 'ilfc7
14 lDg5 lLlcS (D)
Black played 1 4 ...'ii'c6? ! in the
game Gligoric-Ljubojevic, Linares
1 99 1 , but the stormy complications
of 1 5 'irf3 ! !l...c5 1 6 'irh3 rlie7 17
lDf3 b4 1 8 lLle2 lDxe4 19 lDed4
'iib6 20 !l...xe4 !l...xe4 2 1 'ifg4 !l...xd4
22 'irxe4 !l...f6 23 'ii'xc4 were
clearly not in his favour.
Moving the knight to c5 looks
reasonable, but...
w
15 e5!
A clever break in the centre.
There are other possibilities, 1 5
'irf3, 1 5 'ire2, and 1 5 f4, but as I

have the two bishops, I want to
open up the position ! Generally,
the idea of sacrificing a pawn with
e4-e5 is not new, but I am not
aware of it being seen in precisely
this position before.
15
••.
'ii'xeS
The pawn must be taken, as after
1 5 . . .'ifc6 1 6 f3 Black's battery on
the long diagonal is liquidated, and
things are looking bad for him.
16 :et 'ii'd6
17 'i!kxd6!
This turn of events is completely
unexpected; White is a pawn down,
but he nevertheless voluntarily ex­
changes queens. Rather paradoxi­
cal !
17 ... .txd6 (D)
w
18 .te3!
Quietly finishing his develop­
ment, as the e6-pawn is going no­
where.
18 ... 0-0
Karpov - Kramnik 169
After 1 8 ... �d3, White should
play the simple 19 .txd3 cxd3 20
.:r.ad l , and Black loses both pawns
in the centre. 1 8 . . b4 is no better:
.
1 9 .txc5 .txc5 20 �a4, and either
the knight or rook can take on e6.
Finally, 18 ...�g4 is no good be­
cause of 19 .txc5 .txc5 20 .:r.xe6+
'ifild7 2 1 .:r.dl+ r/;c7 22 �. etc.
19 .:r.ad1 .te7
Black has to lose a tempo - after
19 ...:ads 20 .:r.xd6 .:r.xd6 2 1 .txc5
.:r.d2 22 :c l Black does not have
sufficient compensation for the
material he has given up because of
the threat of 23 .te3.
20 .txcS .txcS
21 �xe6 :res
22 h3! (D)
Not so much to open an escape
hatch as to advance the g-pawn.
The raid 22 �xc5 .:r.xc5 23 .:r.e7
.:r.bs allows Black to hold the posi­
tion.
1 70 Linares 1994

The critical moment. The black

• • • • •• •
• • • •
pieces are poorly co-ordinated, and
the question is whether he can im­
prove their interaction. i.R . RltJ•
- �
-
22 •.• i.f8?!
••• • •
22... b4 loses: 23 lba4 i.a7 (oth­
erwise 24 lbb6) 24 lbxg7 'it>xg7 25 ••• 08&
l:te7+. If 22 ... i.b4, then 23 a3, •
""� �� •� .t. •
-8
forcing Black to give up his bishop, 8 "i.�
u � • •
• •
as after 23 ...i.a5 there is the cap­
ture 24 lbxg7 'it>xg7 25 l:te7+. A
. . �
. . � �
�
later game, Topalov-Lautier, Dos B
Hermanas 1994, featured 22...l:tab8, 28 .l:.f2, with the idea of i.g6-
and after 23 g4 i.f3 24 l:td2 b4 25 f7+ and l:tf4, is tempting, but af­
lba4 i.a7 26 g5 lbd5 27 l:te5 l:tb5 ter 28 ... .tc5 (28 ...lbc5? 29 lhf8+)
Black eventually managed to stand 29 lbxc5 lbxc5 30 l:te5 White has
firm. However, the impression re­ only a minimal advantage: 30...lbd3
mained that he was on the edge of 3 1 i.xd3 cxd3.
the abyss for a long time. 28 ... �xf8
23 g4 h6 In the event of 28 ...l:r.xf8 the
Or 23 ... .tf3 24 l:td4 ! b4 25 lba4 rook manoeuvre to d6 becomes
lbd5 26 l:te5 ! , etc. If 23 ... g6, then more valid.
24 lbxf8 l:txf8 25 g5 followed by 29 l:td6
26 l:td7 is very unpleasant. 29 lbd5 also looks reasonable,
24 f4 i.f3 aiming to exchange off Black's ac­
25 l:td2 (D) tive bishop.
25 l:td4 looks more energetic, 29 ... b4
and if 25 ...i.c6, then 26 g5 hxg5 27 30 lbe4 .i.e8?! (D)
fxg5 lbd7 28 l:th4 and the black 30 ...i.d7 was necessary: 3 1 h4
king falls into a mating net. How­ (3 1 lbg3 l:te8 !) 3 1 ...l:tc6 and Black
ever, by continuing 25 ...l:te8 Black can still breathe.
could threaten 26 ... l:r.xe6 27 l:r.xe6 31 lbg3!
i.c5 or 26...i.c5 immediately. I was placing great hopes on this
25 i.c6 move - the appearance of the
26 g5 hxg5 knight on f5 or h5 is extremely un­
27 fxg5 lbd7 pleasant.
28 lbxf8 31 ... lidS
 Karpov - Kramnik 1 71

Evidently 37 a4 is more logical,

rounding up the black a-pawn later.
37 ... l:td3
38 �gl l:txb3
39 h4 (D)

w
Now 3 l . . ..i.d7 does not work:
32 l:te7 .i.xh3 33 ll)hS, etc.
32 ll:)fS l:txd6
33 ll)xd6 .i.g6
34 .i.xg6 ll)xg6
35 ll)xc4 l:td8 (D) B
35 ...l:tf8 36 l:te4 ll)f4 37 h4 �h7 39 ••• ll)f8?
38 ll)eS is suicide. Of course, the defence 39 ... �f8
40 h5 ll)e7 4 1 l:te5 �e8 would

- . .•• have been more stubborn, although

�
-
� �
�
-
� . �
� - -­
after 42 l:ta5 (or 42 l:tc5) White,
with his extra pawn, must emerge
•• . -
� victorious .
40 l:te8! 1-0
• • • D
.ttJ. • • It is becoming clear that the
black king cannot break out to free­
. . . -� .
dom; if 40 . .t�f7. then 4 1 ll)d6+ !
�B • • • and the rook is defended. While
�
-
%z -
� �
� �
�
�
�
Kramnik was grieving over the fact
that moving his king from g8
w would not free his pieces from their
36 l:te4 b3 paralysis, his flag fell.
If 36 ...l:td3, then 37 �g2 and the It remains only to say that while
h-pawn marches forward. my game against Topalov from the
37 axb3 same tournament won the best
1 72 Linares 1994
game prize in lnformator, this
game in the very same edition won
the prize for the most important
theoretical game.
Game@
Karpov - Beliavsky
Linares 1994
Catalan Opening
We have come to the end of the Li­
nares tournament, and with it the
end of this book. Victory over Be­
liavsky in the final round would
guarantee me a lead of 2 points
over the rest of the field ! But this
game is not only remarkable for
the context of the competition, as
in it I also managed to use a piece
of opening preparation which had
been awaiting its hour of glory for
about twenty years !
So, let us return to the chess
board for the last time.
1 d4 ltlf6
2 ltlfJ d5
3 c4 e6
4 g3 i..e7
5 i..g2 0-0
6 0-0 dxc4
- 7 ..,cl a6
8 a4
So far this is all very well
known. For example, I once had
the position after move 7 against
the very same opponent, although
on that occasion we were playing
with opposite colours! After 8
'iVxc4 b5 9 'ifc2 1&..b7 10 i..d 2 lDc6
1 1 e3 ltlb4 12 i..xb4 1&..xb4 1 3 a3
i..d6 14 ltlbd2 .:c8 15 b4 a5 a com­
plex position with chances for both
sides arose in Beliavsky-Karpov,
Brussels 1 988.
8 i..d7
9 Wxc4 i..c6 (D)
w
10 i..g5 i..d5
Matters developed differently in
Kasparov-Karpov, Moscow Wch
(22) 1 984/5 : 10 ... a5 1 1 ltlc3 ltla6
1 2 .:act 'ilfd6 1 3 ltle5 i..x g2 1 4
'iPxg2 c 6 1 5 i..xf6 gxf6 1 6 ltlf3,
and again White has· preserved
only a small advantage out of the
opening.
1 1 'ii'd3 c5
12 ltlc3 cxd4
13 ltlxd5 'ii'xd5 (D)
It is necessary to capture on d5
with the knight. In the old game
Htibner-Karpov, Tilburg 1 979, the

w 14 ...
exchanges 1 3 ... lLlxd5 14 i.xe7
'flxe7 15 lLlxd4 lLlc6 16 lLlxc6 bxc6
left White only a symbolic advan­
tage.
Taking on d5 with the queen is a
new move ...
1 4 h4! !
. . . new for Beliavsky, but not for
the author of these pages ! I had
faced this position on the board in
1 974, as I was preparing for my
match against Korchnoi in Mos­
cow - the match which was soon to
give me the title of 1 2th World
Champion ! It goes without saying
that it is very rare for an opening
novelty to be used two decades af­
ter it was prepared. Furthermore, I
must admit that this novelty is shat­
tering: Black's position collapses
like a house of cards.
Capturing on d4 with the queen
is known to lead to a better end­
game for Black, and taking with
the knight is impossible, since the
Karpov - Beliavsky 1 73
g5-bishop loses its support. Now,
with the bishop defended by the h­
pawn, the recapture in the centre
can be made by the knight. Besides
this, the light-squared bishop ac­
quires incredible strength, which is
extremely important. This is a very
rare example of an advance by a
rook's pawn practically deciding
the outcome of the game from the
very start.
lLlbd7
14 ... lLlc6 suggests itself, but
White has at his disposal the fol­
lowing cunning variation: 15 i.xf6
i.xf6 16 lLlg5 !, and again the pawn
on the edge of the board helps, this
time by defending the knight:
16 . ..iif5 17 i.e4 it'a5 18 i.xc6
i.xg5 19 i.xb7 .:.a7 20 i.e4. The
pawn has been won back, and
White's positional advantage is ex­
tremely tangible.
15 lLlxd4 'ii'd6 (D)
w
1 74 Linares 1994
If 15 ...Wa5, then 1 6 lt:Jb3 ! .
1 6 l:lfd1 lt:JcS
17 ...c4
This is stronger than 17 lt:Jf5? !
lt:Jxd3 1 8 lt:Jxd6 lt:Jxb2 19 l:ld2 .i.xd6
20 l:lxd6 lt:Jc4 or 1 7 'i'c2 ...b6 1 8
a5 'ii'b4, when White's advantage
is not that large.
17 ... l:lfd8 (D)
w rounds ! With four draws (I man­
18 b4! lt:Jxa4
Beliavsky is confused, and this
is another consequence of White's
1 4th move. He should have played
1 8 . . . lt:Jce4 1 9 .i.xf6 lt:Jxf6 (but not
1 9 . . ..i.xf6? 20 .i.xe4 .i.xd4 2 1 e3
.i.xal 22 l:lxd6 l:lxd6 23 .i.xb7
with a big advantage) 20 .i.xb7
l:lab8 2 1 'i'xa6 'ilt'xb4 22 lt:Jb5 .
White would have taken the upper
hand, but the game would have
continued.
19 ...b3! (D)
Now it's all over: Black loses a
piece.
etc.
20 e3 1-0
This victory in the final round
was my ninth at Linares; it was
even more remarkable that I won
six straight games in the first six
aged to avoid being beaten) I
reached an inconceivable score,
1 11 1 3 . And, incidentally, the sec­
ond and third placed players, Kas­
parov and Shirov, were 2112 points
behind me.
Without false modesty I can say
that in the whole of chess history,
tournaments where on the one
hand, all the stars of the chess
world are gathered together, and on
the other, the winner has demon­
strated such notable superiority
over the remaining contestants, can
be counted on your fingers.
 Karpov - Beliavsky 1 75

Kasparov once called the tradi­ official, because it would have

tional battle at Linares the World been a wonderful crowning addi­
Championship Tournament. It is a tion to my collection of champion­
pity that this title was not made ship titles !