Professional Documents
Culture Documents
Alexei Shirov Fire On Board
Alexei Shirov Fire On Board
ALEXEI SHIROV
CADOGAN
C H F S S
CADOGAN CHESS BOOKS
FIRE ON BOARD
AVERBAKH KASPAROV
Chess Middlegames: Essential Knowledge Garry Kasparov's Chess Challenge
DAMSKY PETURSSON
The Heavy Pieces in Action King's Indian Defence, Averbakh
Variation
GUFELD
An Opening Repertoire for the Attacking TAIMANOV
Player Taimanov's Selected Games
LALIC TIMMAN
The Queen's Indian Defence Studies and Games
McDONALD WATSON
Practical Endgame Play Play the French (new edition)
Cadogan Books, 3rd Floor, 27-29 Berwick St, London W1V 3RF
Tel: (0171) 287 6555 Fax: (0171) 734 1733
FIRE ON BOARD
By Alexei Shirov
Foreword By
Jonathan Speelman
2 bxc3 �al+ 3 �d2 (3 'iVbl 'iVxc3+ 4 �c2 'iVxc2 mate) 3... �b2+
4 �c2 �xc2 mate 2...lbe2+ 3 :Xe2 �al+ 4 �d2 �xdl mate 0-1
First published 1997 by Cadogan Books plc, 27-29 Berwick St, London W1V 3RE
Copyright© 1997 Alexei Shirov.
Translations from Russian (games 3, 6, 7, 8, 9, 14, 18, 19, 20, 21, 22, 24, 26,
v Ubilava on p.186, Dokhoyan on p. 211, Magerramov on p. 212, Balashov on p. 213,
Ruzhyale on p. 214, Minasian on p.217, Vyzhmanavin on p. 219) © 1997 Ken Neat.
Translations from German (games 34, 35, 36, 39, 40, 42, 45, 46, v. Nikolic on p. 190,
Kamsky on p.226) © 1997 Graham Hillyard.
All rights reserved. No part of this publication may be reproduced, stored in a re
trieval system or transmitted in any form or by any means, electronic, electrostatic,
magnetic tape, mechanical, photocopying, recording or otherwise, without prior per
mission in writing from the publishers.
Distributed in North America by Simon & Schuster, Paramount Publishing, 200 Old
Tappan Road, Old Tappan, New Jersey 07675, USA.
All other sales enquiries should be directed to Cadogan Books plc, 27-29 Berwick St,
London W1V 3RF
Typeset by ChessSetter
Thanks are due to the following editors, translators, proof-readers and couriers for
their assistance with the preparation of Fire on Board: Andrew Kinsman, John
Nunn, Petra Nunn, Graham Hillyard, Ken Neat, Tim Wall, Mark Huba, Pat Aherne.
Introduction 10
1 Growing Up (1979-1987) 12
2 Winning the World Cadet (1988) 19
3 Professional Chess Life (1989-1996) 24
4 The Botvinnik Variation 184
5 Selected Endgames 211
43. . . �a5 44 �c3 �ab3 45 �f3 �d4 46 60 �c3 �a6 61 �f8 b6 62 �a8+ �b5 63
'ii'g2 l:.h8 47 f5 :If8 48 �f2 �c6 49 ffi l2Je5 'iVeB c4 64 'i¥g8 �-�
50 �h5 �xc4 51 f7 c6 52 �f4 l2Je5 53 'i¥h6
:xf7 54 �xf7 �xf7 55 'iff6 l2Je5 56 'ifxd6
lbcd3 57 'fie7 c5 58 �d2 l2Jc6 59 �g7 �de5 Jonathan Speelman
Introduction
The idea of writing a chess book occurred always been tactical complexity. The
to me a long time ago, but it had always search for the truth in sharp and concrete
seemed very difficult to realize it. So in games with a lot of tactics has always fas
December 1993 when Andrew Kinsman, cinated me, although my busy schedule
who was then the commissioning chess edi didn't permit me to go deep inside every
tor at Cadogan Books, asked me whether game in the book. Therefore they are an
I was planning to write a book, I started notated differently - some have relatively
thinking more seriously about it. In March brief explanations, others very extensive
1994 I accepted his official proposal and analysis. Sometimes I have checked vari
promised to complete this work by the ations with Fritz4 in an attempt to make
end of 1994. my analysis as precise as possible. How
Easier said than done! I had of course ever, I have found it useless to point out
realized that other chessplayers slowed which moves were suggested by Fritz, be
down their chess activities while they cause when a grandmaster works with a
were writing a book, but this was a real chessplaying program, he has to extract
problem for me, as my chess calendar was its variation tree from a lot of rubbish and
always so full of tournaments. Besides, on ultimately his work becomes highly indi
the 27th of November my daughter was vidual anyway. I believe that in the end I
born, so I had to adapt to a new lifestyle. have found a reasonable combination of
To add to the problems, my notebook was the human and electronic brains 'and that
stolen in January 1995 so, with no backup, the computer has not influenced my an
a good deal of my work was lost forever. notations in the wrong way. But that is re
Slowly, step by step, I had to restart my ally for the reader to judge for himself.
work on the book. First I collected some of Now I would like to say a few words
my relatively old annotations from vari about my chess style. When I was just a
ous magazines and tried to make compli boy, I was fascinated by Mikhail Tal's
mentary comments and corrections to games and Alexander Koblents's books,
them (these are marked in the text with both of which were full of tactics. These
the initialsAS, which should be taken to two maestros have had a great influence
mean 'addition') as Kasparov did in his on the Latvian school of chess, of which
The Test of Time. But then I noticed that probably Alexander Shabalov and myself
some of the most important games had are the 'purest' products since we both
only been annotated in lnformator-style, show a lot of creativity in our games .
which in no sense made my task any eas However, I have always tried to be not
ier. Other games that I wanted very much just a tactician - working with a posi
to include had not been annotated at all, tional player such as Bagirov and study
so I had to make completely new notes, ing hard has helped me to develop my
and of course I also wanted to include own strategic understanding, though
games from my most recent tournaments. chess is nowadays so concrete that pure
Fortunately, in the second half of 1995 I strategy practically doesn't exist for me. I
started working with the ChessBase for just try to know typical middlegame ideas
Windows program and my work became that might always be useful, and I am
much easier. quite confident in my endgame ability,
When annotating the games I have at even against the world's leading endgame
tempted to explain their principal strate experts, such as Karpov and Salov. I more
gic themes, but my favourite subject has often gain than lose points at that stage of
Introduction 11
the game (taking into account the posi pleasant memories are the ones that best
tions that have arisen at the start of the demonstrate my approach to chess, and I
ending, of course!) . In fact I believe that hope that the reader will also enjoy them.
the endgame is my strongest area, per If you want to become involved in a world
haps because I am a player with a very of unusual ideas, sacrifices, sharp moves
concrete style. and sometimes strange errors, then I be
In my selection of games for this book lieve that this is the book for you.
I have chosen first of all the most memo
rable, and only then the best games. I Alexei Shirov
think that the games which hold the most Tarragona, October 1996
1 Growing Up (1979-1987)
I was born into a Russian family on 4 July Plata, my sense of fantasy was evoked. I
1972 in Riga, the capital of Latvia. At that dare say that in those early years my
time Latvia was still a Soviet Republic, so prime motivation to improve my chess
Russian-speaking children grew up not was to have the opportunity to travel all
really having to speak Latvian, which was over the world rather than just to achieve
in any case badly taught in school. The success in tournaments.
fact that I never learned Latvian properly I am unable to offer any reasonable
might easily have affected my public rela games that I played between 1979 and
tions in the country but fortunately it 1982 , but the following game, played in
never caused me any problems with Lat October 1983, is probably a deserving one
vian chessplayers. Anyway, now that I live to start this book with.
in, am a citizen of and represent Spain, I
can send that problem to the past. Game1
When I was between four and five my
elder brother tried to teach me the rules Shirov- Zhuravlev
of chess. I learned, tried to play and then Riga 1983
forgot the rules completely. It was hard to
predict that one day I would become a These annotations were made during the
professional player. Some time afterwards preparation of this book.
I iearned to play draughts and began to This was the first game of a mini-train
beat my friends regularly. My brother ing match organized by Alexander Kob
then made another attempt with chess and lents. The late trainer was famous for his
I suddenly realized that it was more fun work with Mikhail Tal leading up to the
when the pieces could move differently. latter'sWorld Championship title victory.
Draughts was soon abandoned and in But very few people appreciated how
March 1979 my chess 'career' started. much of his life he devoted to helping
First my brother and then my father (a young players to develop. I should men
player with a Latvian rating of around tion that Koblents's books were my fa
2 000) grew tired of playing with me and vourite chess books in those years.
in May 1979 I was taken to the chess My opponent in this game, Valery Zhur
school. There I received tuition from Lat avlev, is an experienced international
vian woman master V iya Rozhlapa, who master who was Latvian Champion sev
was quite good at teaching children to eral times and played for Latvia in the
play. (She is still encouraging new talent 1994 Olympiad in Moscow. He made one
and every year I examine the best of them grandmaster norm in the 1970s but had
in a clock simultaneous.) very few opportunities to score another.
I cannot say that I made incredibly His rating is currently around 2 460.
rapid progress at this time, but in my 1 e4 e6
opinion the most important thing was 2 d4 d5
that chess had captured my imagination 3 �c3 dxe4
and my childhood in chess was no less fas 4 �xe4 �d7
cinating than that of any other child. My Zhuravlev obviously believed that the
other hobby at that time was reading easiest way to play against a boy is to
about all the countries of the world and keep things simple, but in fact this strat
when, for example, I heard on the radio egy just gives me fewer chances to go
news that Karpov was playing in Mar del wrong. In the second game Zhuravlev
Shirov- Zhuravlev, Riga 1983 13
21 l':.el !
Game3
Klovans- Shirov
Latvian Championship,
31 Ji.b8!
The point. Riga 1987
31 .:xb8
32 .:xg7 1-0 These annotations were made in March
Black lost on time, but there is no de 1987 and first appeared in Shakhmaty
fence against mate in a few moves. Not a Riga.
perfect game, but still one that gave me a 1 e4 e5 2 �f3 �c6 3 Ji.b5 a6 4 Ji.a4
lot of aesthetic pleasure. 'Congratulations �f6 5 0-0 b5 6 .i.b3 Ji.b7 7 hle1 .i.c5 8 c3
on your nice play, ' said Alvis V ito lin§, d6 9 d4 .i.b6 10 .tg5
who is himself another Latvian master of A plan that Yanis Klovans had success
sacrifice. fully employed previously.
Klovans- Shirov, Latvian Championship, Riga 1987 17
tbxe4
My next major event was a qualifier for 'iVeS 11 0-0 tLih7 12 a3 i..d7 13 b3
the World Junior (Und er-20) Champion Illescas and Kramnik prefer 13 �h1,
ship, held at Borzhomi in April. Kamsky after which the best move is probably
was there as well, so I knew I had to finish 13...h5.
a least half a point in front of him to get 13 f5
•••
into the World Und er-16. If I failed , my Of course 13. ..h5 is also possible here,
chances of making a real chess career in but I wanted to d emonstrate a clear path
the Soviet Union would have been seri to equality that I had prepared at home.
ously d amaged (for example, Kamsky's 14 exf5
father d ecid ed to d efect with his son to 14 f3 ttJf6 is fine for Black.
the USA in 1989 in search of a better d es 14 i..xf5
.•.
17 . . .'iVxe3 +
Also interesting is 17....td7!? 18 .l:[xf8+
�xf8! B
1 8 .tf2 'ifg5 19 �h1
My idea was to answer 19 h4 with
19 ...'iff4 20 gxf5 ..te5 21 �f3 �g4+ 22
� h1 'ii' h3 + leading to a draw by perpet
ual. Khenkin decides to go for more.
19 . . . .td7 20 �de4 'ii'e7
The position is equal.
2 1 'iVd3 ltae8 22 �g3
22 �g2!? was worth consideration.
22 ...ltf6 23 �g2 �g5
40 a4?!
•..
Several leading Latvian players and the was already a new move. 15 flac1 �b7 16
Russian grandmaster Evgeny Sveshnikov flfd1 lic8 had occurred in the game Sem
also participated in that training session kov-Psakhis, Erevan 1988.
and when I left for Romania (with V ladi 15 �b7 1 6 liad1 lieS
•.•
GameS
Shirov- Lautier
World Cadet Championship,
1 7 ltJ2c3 a6?!
Timisoara 1988 Psakhis played 17 ...l2Ja5! against me
two months later at Klaipeda and gained
These annotations were made during the an advantage after 18 "iYg3 l2Jc4 19 �a4 �c6
but it still doesn't seem convincing. Possi This loses by force but 3l...'ti'e8 32 'iVh5
bly Lautier thought that his task (he ltJf6 33 'iVg5 'ti'f8 34 l:!g3 ltJg4 35 f3! tbxh6
needed to win) was already rather easy. If 36 ltxh6 f4 37 .l:g4 .l:c1 38 iVh5! .l:xdl+ 39
so, this probably cost him the game. �f2 also wins for White.
27 ttJd6
..• 32 i..xg7 .l:xg7 33 .tlh3 'ii'xd4
The first inaccuracy. 27...ltJxb2! would Of course 33 ... .l:xg6 34 "VJii h8+ �f7 35
have left Black with the better prospects .tlh7+ also loses.
after 28 ltJxg6 hxg6 29 ltxg6+ i..g7 30 34 "VJiih S + �f7 35 .l:xg7 + iVxg7 36
i.xg7 ltxg7 31 .l:dg3 .l:xc2 32 'iVh6 .tlc7 33 'ifxeS
h4 "VJiie7 34 .l:xg7+ 'iVxg7.
28 i.. dl ltJe4 29 ltJxg6
The only move.
B
29 hxg6 30 ltxg6 + (D)
••.
30 i.. g7?
•••
Both flags were hanging and we had in the Semi-Final of the Soviet Champi
stopped writing down the moves. The rest onship, to be held in Klaipeda in Novem
requires no comment. ber 1988, where I had the opportunity to
36 .td5 37 'ii'd 7 + <it>f6 38 'ii'xg7 +
••• play against such established grandmas
�xg7 39 .tb3 l2Jg5 40 l:Ig3 <it>f6 4 1 i..xd5 ters as Gelfand, Dolmatov and Psakhis,
exd5 42 l:Ic3 <it>e5 43 l:Ic6 d4 44 l:Ixb6 d3 etc.
45 <it>fll2Je4 46 h4 �d4 47 h5 1-0 In the first half of the tournament I
As World Cadet Champion I of course won three games and was close to the
received certain privileges from the So lead, but then I collapsed to finish with 7
viet Sports Committee. The most impor points out of 16. Still, it was excellent ex
tant of these was an invitation to the play perience.
3 Professional Chess Life (1989-1996)
Since 1989 I have played in so many tour Richard Wessman and Norwegian 1M Rune
naments that it is impossible to describe Djurhuus (whom I had met at the World
all of them. Probably the games I played Junior Championship in Colombia, where,
in these years will tell the reader more. by the way, I played poorly) I received in
Here I offer a brief account of my career vitations to play in open tournaments in
during this period. Stockholm at the end of the year and
In March and April 1989 I received an Gausdal in January 1990. When, after a
other privilege from Goskomsport - a two-week nightmare of arranging docu
chance to play in the Budapest open (in mentation, I finally arrived in Stockholm
which I made my first international mas I already felt that I had done the hard
ter norm), followed by a closed category 9 part and the chess would be much easier.
event in Torey, France, where I shared first In fact I scored grandmaster norms in
and second place with grandmaster Ku both tournaments and was awarded the
preichik and achieved my first GM norm. title in May 1990.
The following month I received an The following month I was invited to
other chance to go for a GM norm in the what turned out to be the last Soviet Zo
GMA open in Moscow. However, I started nal tournament and surprised even my
badly and finished half a point short with self by finishing in a tie for the first four
51h out of 9. Compensation was received places (with Leonid Yudasin, Smbat Lpu
in the form of$ 1,000 prize-money (which tian and Alexei Dreyev), which was suffi
was a lot of money in the USSR in those cient for a berth in the Interzonal, a
days) which gave me some security and 13-round Swiss tournament held in Ma
enabled me to make plans for my career. I nila in June and July of 1990. Of course I
had just finished school and had decided had hopes of qualifying for the Candi
that I would play chess full-time for a dates' stage but unfortunately I made a
while. (In fact, after a year of chess I went slow start and had only scored 31h after
to university because there was no other the first 8 rounds. With a last ditch at
legal way to avoid military service. I com tempt I won three games in a row but
pletely quit my studies in the beginning could only manage two draws at the end,
of 1993 but even before that I never spent which was not enough. Still, 71h out of 13
too much time studying.) was a good result for me.
In the summer of 1989 I took part in In August I then went to Santiago for
the Soviet Youth Games (a kind of youth another attempt on the World Junior
team championship), where I scored 6 out Championship. Although I scored 101h/13
of 8 on first board to share first place with the American Ilya Gurevich (whom I had
Boris Gelfand from Byelorussia and the lost to in round 7) made the same score
Moldavian Victor Bologan (who is now a and was awarded the title on some strange
grandmaster). In this event I scored a re tie-break system.
markable victory over Vasily lvanchuk This tournament was also remarkable
(world no. 3 at that time) after he had de because it was there that I encountered
clined a draw in an equal endgame. the Argentinean Veronica Alvarez, who
The Soviet system started to change was participating in the World Girls (Un
dramatically in 1989 and chessplayers der-20) Championship, for the first time.
were able to travel abroad a lot more us We were married in January 1994.
ing our own personal contacts and finan Mter the World Junior I more or less
cial resources. Thanks to the Swedish IM completely broke away from the Sports
Professional Chess Life 25
Committee and was now dependent solely winning the first two games but then
on tournament organizers. In order to ob slipped back to 50% with losses in rounds
tain good invitations I needed to improve 3 and 5. Having won from a dubious posi
my rating as much as possible so I started tion against Gelfand in round 8, I re
trying to beat every weaker opponent I gained some confidence and won two more
played in every competition. Sometimes I games. My final result (8/13) was enough
took many risks, sometimes I had insuffi to compensate for my earlier disasters.
cient energy, but in general my plan Mter Linares I played either good or
worked and in the summer of 1991 I was average tournaments for the rest of the
invited to my first category 15 tourna year. The highlights were first place at
ment in Biel, where I took first place with Munich in May (category 16) and reach
91h/14, a point clear of Evgeny Bareyev. ing the semi-final of the Tilburg tourna
Later that year I began playing for Ham ment, where I was narrowly defeated by
burg in the German Bundesliga, for whom Ivanchuk in a rapid chess tie-break. Al
I made 91h/10 on second board in my first though I didn't qualify for either the
season, which ended in May 1992. F IDE or PCA Candidates' cycles, my re
As a result my rating began to improve. sults in both Interzonals were acceptable.
I was 2655 in January 1992 and this went Returning again to Linares in Febru
up to 2710 in July. I won several tourna ary/March 1994 I reached the peak of my
ments in this period, but my first category career so far. It was the first time that I
17 event, in Dortmund in April, was al had been accompanied to a tournament
most a disaster- 31h points out of 9. by my wife and the start, 1h out of 3, was
In June the Latvian team (Shirov, Ken discouraging. However, in round 4 I beat
gis, Shabalov, Lanka, Bagirov and Klo the Spanish grandmaster Miguel Illescas
vans) went to the Olympiad in Manila. We and I was then joined by my friend,
were naturally very excited about this as grandmaster Victor Bologan (who was my
it was the first time a Latvian team had second at Tilburg 1993, and with whom I
participated since the Buenos Aires event have had several training sessions) who
in 1939. Our final result, fifth place out of had come to assist me. After victories
more than a hundred teams, was quite an against the Bulgarian Veselin Topalov, Va
achievement in my opinion (but one, alas, sily lvanchuk and Judit Polgar, I made
which we couldn't repeat in Moscow in three draws with Kasparov, Karpov and
1994, where we finished 19th). I scored 9 Gelfand. The quality of all these games
out of 13 on top board, but I was disap was quite good, but then Bologan left (he
pointed to lose to Garry Kasparov after is a professional player too) and I again
missing a simple two-move win. started playing badly. However, lady luck
Before the Olympiad in 1992 my pro shone down on me and I won two further
gress had been quite smooth, but then games (against Kamsky and Kramnik), to
things started to get a little shaky. Re finish with 81h out of 13, sharing second
turning to Biel in July/August 1992 for a and third places with Kasparov (Karpov
category 16, 8-player tournament I could was first with 11).
only manage 51h/14 to finish last but one. After Linares my rating was the third
In Moscow in November it was the same highest in the world and I managed to im
story and other results such as 50% at Wijk prove it further with victories in three
aan Zee in January 1993 were equally un- Bundesliga and three F rench League
. .
Impressive. games in March and April. But in May an
My chance to re-establish myself among other crisis began and I had several bitter
the world elite came with the 1993 cate results, sweetened only by sharing second
gory 18 tournament in Linares, Spain in and third places in Horgen, Switzerland
February/March 1993. Here I started well, in September behind Kasparov. My last
26 Fire on Board
event of 1994 was the Moscow Olympiad Madrid in May and I also again made the
where my performance on top board (8/13) best performance in the Bundesliga, but
was nothing special. didn't score so many points as four years
In February 1995 I won the training ago.
(for my opponent) match against Jeroen Now that this book is finally finished
Piket: 51h-21h. An excellent result but not (it really took too long!) my real comeback
in terms of the quality of the games. Then will start! I have nothing else to add and
I, as usual, did well in Linares- this time it is time to let the moves speak for them
scoring 8 out of 13 to share 3rd-4th places selves. I would only like to express the
with Topalov behind Ivanchuk and Kar wish that one day I will have some games
pov. But in April I played so horribly in a worthy of another edition.
category 18 tournament in Dos Her
manas (a suburb of Seville) that I finished Game6
last - for the first time in my life. I still
don't really know what happened because Shirov- Akopian
I often stood well after the opening. Im USSR Young Masters
mediately after Dos Hermanas I went to
Leon where I shared 1st-2nd places with Championship, Tbilisi 1 989
Bareyev in a category 14 tournament (61;2
out of 9). These annotations were made in Febru
After the Leon tournament I had two ary 1989 and first appeared in Shakh
months' rest and this was especially good maty Riga under the title 'Without any
for me since apart from physical prepara prepared analysis'.
tion I needed to think what to do about The 1989 USSR Young Masters Cham
my chess. I think I did the right thing. I pionship went badly for me. Right from
was able to start hard work on chess the start I got stuck in a drawing rut,
again and I believe that one day it will pay then in the middle I gained two wins and
off. This also gave me more confidence reached 'plus one', but I could not main
and made me willing to play and fight. I tain this and finished with a fifty per cent
started playing more creative chess again, score. Many of my drawn games were in
working hard during every game as in my teresting, but my only notable win was
best times. the one over Akopian. In describing this
However it also had one drawback. It encounter, I cannot avoid mentioning
turned out that at the age of twenty-three its thematic predecessor - the game Epi
I was not so full of energy as when I was a shin-Khenkin (56th USSR Championship
teenager. Although I had a solid perform Semi-Final, Barnaul 1988). On certain
ance at Biel in July 1995 (a single round key questions my opinions differed from
robin category 15 tournament), recovering those of Epishin, but here I will concen
well from a slow start, my next tourna trate on my own.
ments in Amsterdam and Belgrade (cate 1 d4 ttJf6 2 c4 g6 3 ttJc3 d5 4 cxd5
gory 16 and 17 respectively) were anything ttJxd5 5 e4 ttJxc3 6 bxc3 i.. g7 7 ttJf3 c5 8
but smooth. A sharp start (31h out of 4 in l:.bl 0-0 9 i..e2
both!), interesting games with plenty of I used to employ the variation with 7
creative ideas but a lack of energy at the i..c4 and 8 ltJe2, but games such as Yusu
end and finally respectable but not espe pov-Kasparov and Belyavsky-Kasparov
cially satisfactory scores. My next tourna from the 55th USSR Championship (Mos
ment, Wijk aan Zee in January 1996, was cow 1988) showed that Black has good
even worse - just fifty per cent. Actually play, and the number of its supporters fell
1996 is not going as well as I had hoped. I rapidly. Instead I chose a variation that
had just one really strong tournament in once experienced a boom, and has now
Shirov-Akopian, USSR Young Masters Championship, Tbilisi 1989 27
Epishin's idea, which he first employed But here is the divergence. Epishin
in the aforementioned game with Khen played 19 l:Ife1 tbf6 (19...f6 20 'ii'e6+! is
kin. It is interesting that, according to better for White) 20 c4 i..f5 21 l:.a1, and
Epishin, this move was devised during after 2l...a6?! 22 h3 h5? 23 i..f4 'ti'd7 24
the game. At that time both Akopian and i..e5 �h7 25 'irff4 l:tg8 26 'ii'g5 he won
I knew of that game only by hearsay, as quickly. But instead of 2l...a6 there was
indicated by our great expenditure of 2l...e5!, when White does not appear to
time. Thus in neither case was there any have sufficient compensation for the two
prepared analysis, but for all that, this pawns. Instead of 21 l:.a1, in Epishin's
game is of no less interest to theory. opinion, it was stronger to play 21 l:.bd1
12 b6
••• i..g4 (otherwise 22 h3) 22 i..xf6 i.. xf3 23
28 Fire on Board
Forintos strives to emphasize the weak There was little time (the control was
ness of White's third move. Unusual play two and a half hours for 50 moves), and so
now arises, since 5 i.g2 cxd4 6 lbxd4 lbc6 I decided to improve the placing of my
or 5 lbbd2 cxd4 6 cxd4 lbc6 is good for pieces while maintaining the tension.
Black. 19 l:Ie3 �b6 20 l::t ea3
5 d5 d6 6 lbbd2 lbf6 7 e4 lbbd7 8 If 20 l:Ib3 fic7 21 lbxb7 l:!xb7 22 l::txb7
i.g2 'tia6! 9 c4 'iixb7 23 �a4 lbef6 24 'tixa7 'V/Iib3 with
White has played c2-c4 in two moves, compensation.
but the black queen is not well placed, 20 ... lbef6 2 1 lbb3
and this balances the chances. Nothing would have been achieved by
9...b5 10 0-0 l:Ib8 21 lbc6 i.xc6 22 l:Ia6 'V/Iic7.
Black could have taken the pawn - 2 1 . :a8
••
10...bxc4 11 'iic2 lbb6, but after 12 a4 i.d7 Not 21...a6?! 22 lba5 with a clear ad
13 a5 i.b5 14 l:Ie1 White would have had vantage.
sufficient compensation. 22 i.d2 a6 23 l:Ila2 l:Ifb8 24 i.a5 Wiia7
1 1 'tic2 0-0 12 l::t e 1 25 i.d3 lbe8 26 'ii'e2 i.c8 27 �g2 l:Ib7
28 i.d2 :b6 29 lba5
12 lbg4?!
.••
An inaccuracy. Mter 12...lbe8 Black has White has make some headway, although
everything in order. after 29 ... i.b7 Black's position would still
13 i.f1 ! have been solid. I could have continued 30
By blocking the position, White gains i.c1 with the idea of i.c2-a4. With time
the initiative. trouble imminent, this would have been
13 b4 14 h3 lbgf6
••• the best decision for Black.
14 ...lbge5 is well answered by 15 lbh2 29 lbb8 30 e5!
•..
tbb6 16 f4 tbed7 17 lbb3, when White en Mter this advance Black's position col
joys a slight advantage. lapses.
15 a3 lbe8 30 ...'iid7 3 1 e6! fxe6 32 dxe6 �xe6?!
The Hungarian grandmaster criticised (D)
this move, but offered no alternative. More tenacious was 32...'iic7 33 i.e4
1 6 lbb3 'ti'b6 1 7 axb4 �xb4 1 8 lba5 l:Ia7 34 i.d5, when White has 'only' a big
i.b7 advantage.
18...lbe5 would have lost to 19 l:.e3, but 33 i.e4! 'V/Iixh3 + ?
now I realized that, even though I had a The third mistake in a row. However,
certain advantage, it would be difficult to after 33...lbc7 34 tbg5 'ti'd7 35 'iVf3 d5
breach Black's defences. (35... :a7? 36 i.xg6! hxg6 37 'V/Iif7+ �h8
30 Fire on Board
-�-M�
� / ' ,/ /uuu/ g
�
7 �ge2
Numerous games have shown that af
ter 7 dxc5 dxc5 8 �xd8 l:Ixd8 9 i..xc5 �c6
22 ...�b3!! 23 gxf6 Black has sufficient compensation for the
If 23 axb3 'ti'a5. pawn. For example, 10 i..a3 b6 11 �ge2 e6
23 ... �xc1 24 VWe3 12 l:[d1 �d7 13 b3 i..b7 14 g3 (White has
Or 24 �xc1 'ti'a5, and White cannot va problems over his development) 14...�de5
cate the d1 square without loss of mate 15 i..g2 tbd3+ 16 �fl a6!? 17 f4 b5 18 .i.c1
rial. bxc4 19 bxc4 (Brenninkmeijer-Gelfand,
24 'ii'a5 25 'ti'xc1 :as
••• Arnhem 1987/88), and here 19...�a5 would
White resigned, in view of 26 a3 'ii'xa3 have led to an equal game. There is no
27 'i'xa3 l:Ixa3 and 28...1:Ifa8. need to comment on 7 d5 - after 7 ...e6 a
well known theoretical position is reached.
Game9 In making the comparatively rare move 7
�ge2 the grandmaster offered a draw. I
Gheorghiu - Shirov declined, not because I wanted to win, but
GMA World Cup Elimination for the reasons given above.
[AS- In fact I just wanted to try a new
Tournament, Moscow 1989 idea. For 7 dxc5, see the game Kramnik
Shirov, Bundesliga 1992/93, Game 40 in
These annotations were made in June this book. ]
1989 and first appeared in Shakhmaty 7 �c6 8 'ii'd2 b6
..•
Riga under the title 'When one is not Black must be wary. Thus 8...a6? is bad
thinking about the result'. on account of 9 dxc5! dxc5 10 'i!Vxd8 l:Ixd8
By the time that this game was played, 11 i..xc5 (Alterman-Shirov, training game
I was not in the best frame of mind - I had in 1988), when the weakness of the b6
only one and a half points out of four, square tells. For example, 1l...�d7 12
32 Fire on Board
i.e3 tba5 13 tbf4, and 13...tbe5? fails to 14 b) 13 cxd5 tbe5 14 tbg1 i.xfl 15 �xfl
i.b6. 'W/d7, with a great advantage to Black.
[AS - Later on I decided that 8...e6 is Correct was 12 dxc5 dxc5 13 i.g2, when
more precise and played it against Car the position is completely level.
sten H,pi (Daugavpils 1990). After 9 0-0-0 12 d5!
..•
(15 'it'xd6 tbc2+ 16 �f2 i.d4! wins for tbcxe5 1 6 tbf4 i.b7!
Black) 15. . .d5! 16 a3 dxe4! 17 axb4 (17 After 16 ... i.xf1 17 �hxfl tbxf3 18 �xf3
'ii xd8 tbc2+! 18 �d2 �axd8+ 19 �xc2 e5 19 i.xe5 tbxe5+ 20 �g2 White's posi
l:.xd1 20 �xd1 exf3 wins) 17...'W/xd2+!, and tion is comparatively well co-ordinated.
Black regains his piece, remaining with a Black does better to retain his bishop.
decisive advantage; 1 7 i.e2
Shirov- Malaniuk, Moscow (GMA open) 1989 33
Game10
Shirov- Malan iuk
Moscow (GMA open) 1989
These annotations were made during the
preparation of this book.
18 4Jf6! 19 4Jf4
••• This was the last-round encounter and
If 19 4Jxf6+ 'ti'xf6 20 'ti'f4, Black wins both players needed to win, as a draw
by 20...'ti'xf4 21 gxf44Jg4+! 22 �g14Je3. wouldn't be worth peanuts.
19 'ti'e7 20 :he1 �adS 2 1 'ifc1
.•. 1 d4 f5 2 lDf3 g6 3 g3 4Jf6 4 i.g2 d6 5
It is already difficult to suggest any 0-0 if.g7 6 c4 0-0 7 4Jc3 'ti'e8 8 b3
thing better for White. All Black's pieces I had prepared this set-up before the
are participating in the attack, which tournament. It looked very natural and
means his queen, two rooks, two bishops had also been played by Karpov against
and two knights! Malaniuk in the 1988 Soviet Champion
21. g5!
•• ship (the most readily available informa
For the clear-cut conclusion of the at tion in those days).
tack, they are joined by a pawn! 8 h6
.••
25 'V/I/e7
••• 37 e4!!
I would seriously consider 25...e5 26 My opponent was visibly surprised but
t2Jd3 i.e6 were I Black. then reacted quickly and ... badly.
26 ttJd3 37 dxe4?
•••
Game11
Shirov- lvanchuk
USSR Youth Games,
Kramatorsk 1989
11 1V b1! with clearly better prospects for The only move to keep the balance. The
White; while after others are too bad, e.g. 18 l':tac1?!lDb6 19
b) 8 ... l2Je7 (played by Dautov, as he re i.xb5 (forced) 19 ... :xc3 20 i.xc3lDc6 21
alized too late that the pawn cannot be l:Ud1 :e7 is clearly better for Black; or 18
taken, against Yurtayev in 1989) White l:Iad1?!lDb6 19 'iWe3 i.xe2 20 'iVxe2lDc6 21
does not need to put his queen on d3 at all i.c3 (or 21 i.c5 l2Ja4) 21...'ii'd6 22 i.b2
and thus can develop properly, maintain lDa5 and again Black is clearly better.
ing the advantage of two bishops and a 18 ...l2Jd6!?
strong centre. Black is forced to exchange some pieces
8 c5 9 e4 lDb6 10 i.e3
••• and does so in the most logical way. An
Finally we reach the line which I had other possibility would have been 18...i.a6
prepared. 19 i.xa5 l2Jxa5 20 iYxc8 'iVxc8 21 l:txc8
10 cxd4 1 1 cxd4l2Jc6 12 i.e2 i.d7!?
••• l:Ixc8 22 i.xa6 bxa6 23 l2Jf4 �f8 also with
But here is another surprise. 12...f5 13 equality.
l:td1 i.d7 had been the accepted line be 19 i.xa5
fore the present game. White can't avoid simplification either,
since the position after 19 'iVe3? i.xe2 20
�xe2 (20 .l:txc8 lDxc8! 21 i.xa5 'iVxa5 22
iYxe2 'iVc3 23 'iVd1 lDb6! is winning for
Black) 20...lDc6 favours Black.
1 9 l:.xc3 20 i.xd8 l:txc 1 + 2 1 l:.xc1
..•
.l:txd8
More complicated but probably still
about equal would have been 21...i.xe2
22 i.a5lDb5 23 d5.
22 lDf4 (D)
Somewhere here I offered a draw but
Ivanchuk refused. It is annoying feeling
that you might be just considered a patzer
(although Ivanchuk could have had purely
chess reasons to decline) and this mobi
13 l2Jh3 l2Ja5! 1 4 i.d2 lized me to the maximum.
Shirov- Ivanchuk, USSR Youth Games, Kramatorsk 1989 37
ltxd4 27 i.e2 l:ta4 28 �b5 and Black doesn't 31. ..:a5 is worthless because of 32
Win a pawn. �c3.
23 <itf2 a6!? 32 g4
Setting up a trap. Hereabouts both players were running
24 ltJd3 rather short of time and stopped writing
24 �e3? g5! would be falling into it, down the moves. Somehow I managed to
leaving Black clearly better. However, keep the situation under control while my
White was not obliged to move his knight opponent seemingly stopped counting the
from a good square on f4. Now I would number of moves made. At some point I
prefer 24 h4!? i.xe2 25 ltJxe2 ltJb5 26.l:[bl! realized that this might be my chance to
and Black doesn't have more than a draw win the game and I eagerly continued
in the line 26...l:td7 27 l:tb3 ltJxd4 28 ltJxd4 blitzing after the 40th move.
.l:txd4 29 l:txb7 l:td3 30 l:tb8+ <ite7 31.l:[b7+ 32 hxg4 33 fxg4 f5?!
.••
�f8 (3l. ..<itf6 32 e5+! <itg6 33 h5 is This gives White a dangerous passed
slightly better for White) 32 l:tb8+ . pawn. There was still a simple way to
24 ...�e7 25 ltJe5 ltJeS! draw with 33 ...ltJc4 34 l:txb5 ltJxe5+ 35
25 ...i.xe2?! was not good due to the in- dxe5 l:txa3+.
termediate 26 l:tc7+! and after 26 ...�e8 34 exf5 exf5 35 g5 <ite7
27 �xe2 White is already slightly better. Moving the king towards the centre is
26 i.xb5 axb5 27 <ite3 l:taS 2S l:tc3 correct as the more concrete 35...f4?! 36
lbd6!? (D) h5 ltJf5 would leave Black in trouble after
Black has managed to create some 37 ltJf3 <ite7 38 <ite4 <ite6 39 <itxf4 ltJxd4 40
pressure due to White's weakness on a3, l:te3+! �f7 41 ltJxd4 l:txd4+ 42 <itf5.
but the position is still equal as the b5 pawn 36 h5 <ite6 37 h6 gxh6 3S gxh6 :as
is not so good either. Besides White's other Black could also have made a draw had
pawns and pieces are in the right places. he been persistent with putting his king
29 <itd3 h5 in the centre, e.g. 38...<itd5 39 ltJf3 and
Still wanting to win. A draw would re only then 39...l:.a8. White has nothing
sult after 29...f6 30 ltJg4 ltJc4 31 ltJe3! better than 40 h7 l:.h8 41 ltJg5, and with
30 h4 citeS!? 31 l:tb3 the precise 41...<itc6! 42 l:tc3+ �b6 43 l:.c5
Now ... f7-f6 was already a dangerous ltJe4 Black equalizes completely.
threat, as can be seen in the variation 31 39 h7
Shirov- Magomedov, Frunze (now Bishkek) 1989 39
8 i.e5?! 13 'ifc3?!
Trying to avoid exchanging the bishops This careless move causes White a lot
might have a point but objectively speak of trouble. It was also bad to play 13 a3?!
ing White should have opted for a quieter in view of 13...i.f5! 14 ltJc3 dxe4 15 fxe4
game with 8 i.xg5 'iVxg5 9 'ii'd2!? (9 e4!?) 0-0-0 (intending ... ltJc5) 16 'ife2 (16 b4
9...'iVxd2+ 10 �xd2 or 8 'iVd2 i.xf4 9 'iVxf4 l:!e8 is clearly better for Black) 16...l:txd4!
'i'f6 10 'ii'xf6 (10 'iVd2!?, intending e2-e4) 17 exf5 i.d2+ 18 �f2 'i'c5! 19 'i'e6+ �d8!
lO ...ttJxf6 11 e4, with a slight advantage 20 \t>e2 (forced) 20... i.xc3 21 bxc3 'iVxc3
in each instance. 22 ttJf3 (again forced) 22...'iVxa1 23 ltJxd4
s . f6!
.. 'ifxd4, when Black is clearly better, but 13
Despite a slight weakening of his posi ltJc3 i.f5!? 14 �f2 with a total mess was
tion Black achieves quick development. probably correct.
Otherwise he would be playing into White's 13 ... dxe4!
hands, for example 8 ... i.f6 9 i.xf6 ltJxf6 I had just overlooked that this natural
10 0-0-0 0-0 11 e4 or 8...ttJf6 9 h4! i.e3 move was so strong.
(9 ...i.h6 10 g4 is clearly better for White) 14 i.xa6
10 iVd3 i.xg1 11 flxg1, with a slight plus There is nothing else. The next moves
for White. are also forced.
9 i.g3 i.e3!? 14 . . .exf3 + 15 \t>f2 fxg2 1 6 \t>xg2 bxa6
Very ambitious, trying to completely dis 1 7 tDf3
rupt the co-ordination of White's pieces.
9 . ttJe7 would have been okay as well, for
. .
costs. It was still possible to turn into a This loss of tempo will be decisive as
slow game by playing something like 12 Black will always be one move short of
e3 ttJa6 13 'iVc3 ltJc7 14 l:.c1, when matters the perfect development! He should have
are far from clear. played 17...'iVd7! 18 :e1+ (18 ltJf2 ltJe7
12 ltJa6
••• followed by castling is just curtains for
Probably the best reply. For 12... dxe4?!, White) 18. . . �f7! (but not here 18 ... ttJe7,
the reader should see the next game. which yields White a strong initiative
40 Fire on Board
after 19 'ti'c4! 'iid5 20 'iVxd5 cxd5 21 i.d6 After the text everything is clear and the
i.f8 22 1Ic1 <it'd7 23 i.f4! g5 24 i.g3 l2Jf5 rest requires little comment.
25 1Ic7+ <itd8 26 1If7 l2Jxg3 27 hxg3) 19 l2Jf2
l2Je7, when it seems that White can only
try something desperate such as 20 l2Jg4,
which works after 20 ...'iixg4? 21 1Ixe7+
�xe7 22 1Ie1+ i.e6 (22 . . . <itf7 23 iVb3+
i.e6 24 'ii b7+ i.d7 25 'ti'b3+ is slightly
better for White) 23 'ti'b4, reaching a draw
by perpetual check, but is completely use
less if Black simply plays 20 ...i.g7 with a
clearly better position for Black. Probably
the only serious chance is to sacrifice the
exchange with 20 1Ixe7+!? 'iixe7 21 1Ie1,
but still after 2l...'ti'd8! (not 21...i.e6? 22
d5! or 2l...'ti'd7 22 l2Je4!, intending ltJffi!) I
don't see anything good for White, as 22
'ii' xc6 1If8 23 i.c7 'ti'd7 24 iVxa8 iVxc7 2 l. i.h6
••
i.h6 12 e4
Both sides have almost symmetrical
(from left to right) pawn structures and
even in the position of some pieces.
B White's only problem is his king on fl ,
but that isn't enough to compensate for
Black's weaknesses and poor develop
ment.
19 0-0-0 20 d5!
•••
Mter the game my opponent didn't give Unfortunately the queen has to change
a clear explanation as to why he decided its route, but it still moves just one point
to diverge from the previous game. Prob on the diagonal.
ably in response to 12...l2Ja6, I would also 27 l2Jb8 28 Wiid5 + Wc8 29 l2Jcb5
•••
have played 13 l2Jc3 but in any case open The game is over.
ing the centre seems a little premature to 29 ...l2Je7 30 l:.c1 + lbbc6 3 1 :xc6 +
me. 'ii'xc6 32 d7 + �b7 33 l2Jd6 + �c7 34
13 fxe4l2Ja6 14l2Jc3 lbb4 l2Je8 + �b7 35 lbxc6 1-0
I would have considered 14...i.f5. This was not a great game but its geome
15 Wile2 try still makes me laugh. Later on Ma
Now White seems better to me. The garam 'promised' to somebody that he
rest of the game was rather easy for me would kill my 7 f3 next time, but we ha
because I could more or less follow the ven't met over the board since then. To
ideas I had in mind when analysing our those readers who know nothing about
previous battle. this player I should mention that later in
15 b6 16 a3 i.a6 1 7 Wild1
••• the same tournament he had some quali
The queen has completed a circle . Amus fying chances for the final of the Soviet
ing, isn't it? Championship and these days he plays
42 Fire on Board
mostly in Asian events. Magomedov played [AS- In fact my intention on 11...b5 was
reasonable chess on top board at the 1994 to continue 12 cxb5 .i.d7 13 h3!? hb5 +
Moscow Olympiad, so perhaps in future 14 cJ;;d2 i.e3 + 15 �c3 �xf1 16 hxg4 bel
Olympiads the 7 f3 story will be contin (all forced) 1 7 'Wixfl �f4 18 'filb5+ t2Jd7 19
ued. tDc5 llbB 20 'Wixd7+ 'VJJixd7 21 'Dxd7 �xd 7
21 l!ad1! with a slightly better endgame.
Game1 4 Now, more than six years after this game,
I can reveal this, as I don't play 4 f3 any
Shirov- D autov more.]
USSR Championship Semi 12 t2Jxd4 exd4 13 'filxd4 0-0
queens White will be a clear piece up. The position after 10 ... l2Jg4 11 i.xa6
l2Jxe3 12 fxe3 bxa6 13 l2Jd5 is worth study
Game1 5 mg.
[AS - According to the game Ulybin
Shirov- Y un ey ev Ankerst, Munich 1992, White is better af
USSR Championship Semi ter 13 . . ."Wid8 14 il'a4 1:Ib8 15 1:Iad1 i..d 7 16
c6 i..e6 1 7 b3!?, as Black has difficulties to
Final, Daugavpils 1989 find a good move.]
1 1 l2Jd2 l2Jxc5
These annotations were made in Novem Mter the game my opponent and I
ber 1989 and first appeared in Shakh came to the conclusion that 1l...i.. b 7 is
maty Riga under the title ' The caged stronger.
queen'. 12 i..xc5 bxc5 13 l2Ja4
Is it advisable to keep queens on the
board in a slightly worse position? Some
times yes, unless the board appears to be
too short of free space. B
1 d4l2Jf6 2 c4 g6 3l2Jc3 i..g7 4 e4 d6 5
i.e2 0-0 6l2Jf3 e5 7 0-0 "ireS!?
This original move has been taken up
by grandmaster V iktor Kupreichik, who
employed it for the first time in 1988 in
two games in Belgrade. However, in two
games with Black against me, Kupreichik
had not ventured the King's Indian De
fence. On the other hand, I was now
granted the opportunity of meeting the
'veteran' of this variation, the Leningrad
master Yuneyev, who, it transpired, has Black's extra pawn falls within a few
been employing 7 ...�e8 since 1977. moves, and his weakness at c7 remains. It
[Note by the Editor of Shakhmaty Riga is true that the bishop pair might give fair
- In the Encyclopaedia of Chess Openings counterplay, but during the game my op
there is a reference to a game Quinteros ponent was unable to demonstrate this.
Bednarski, Wijk aan Zee 1973.] 13 ...�e7 14 �c2 l2Jd7 1 5 l2Jb3 �g5
8 dxe5 It was better to play 15...'ifh4 16 l2Jaxc5
After 8 d5 a5! 9 l2Je1 l2Ja6 10 tbd3 b6 tbf6, although here too White's position is
Black has no problems (Ambartsumian the more promising.
Yuneyev, Frunze 1989). 16 l2Jaxc5 l2Jf6 1 7 I:Iad1 h5
44 Fire on Board
White has the advantage after both Black resigned, in view of 23....i.xa2 24
17 ... ttJh5 18 .i.xh5 'Yi'xh5 19 t2Jd7!, and i..fl 'Yi'c8 25 b3 hld8 26 ttJc6! l:d6 27 t2Jb4
17....i.h3 18 .i.f3 t2Jh5 19 �h1 .i.g4 20 .i.xg4 w1nmng.
ifxg4 21 f3 ifh4 22 �f2!
1 8 'ifd2 ! Vi'h4? Game16
This move loses unexpectedly; neither
does the exchange of queens give Black Shirov- E i n gorn
any joy, for example 18...�xd2 19 l:xd2 Stockholm 1989
l:b8 20 ttJa5 l:b6 21 a3 or 18....i.h6 19
Vi'xg5 .i.xg5 20 ttJa5 .i.g4 21 .i.xg4 t2Jxg4 22 These annotations were made during the
t2Jc6, with better chances for White in preparation of this book, based on my
both variations. notes in lnformator 49.
[Translator's note - Does White have The Rilton cup in Stockholm was the
such a clear advantage after 22...ttJe3!? 23 first of two Scandinavian events where I
fxe3 .i.xe3+ 24 �h1 i..xc5 25 ttJxe5 .i.d6? came to make my final GM norms. I re
AS- It is true that after 22.. . ttJe3! White member having a strict routine in both
has nothing. Correct is 22 ttJa6! with ad tournaments with total concentration to
vantage to White. ] wards achieving the goal - something
which I now lack sometimes.
1 d4 e6 2 c4 i..b4 + 3 ttJc3 ttJf6 4 f3 d5
5 a3 i..e 7 6 e4 c5 7 cxd5
A few months later I tried 7 dxc5 .i.xc5
8 b4 against Yudasin (Lvov 1990) but my
position after 8...i..xg1 9 l:xg1 'V/Iic7 was
just worse.
7 exd5 8 dxc5
•••
20 . .. i.xe6 !? would have been extremely i.xg5 24 gxh7+ �g7 25 'ti'e5+ �xh7 26
complicated but probably not better for 'i¥xg5 ltJc6.
White at the end. He should now play 21 22 i.e7 23 ttJxe6
•••
enough threats after either: The only mistake of the game but one
al) 26 ... ltJf6+ 27 <it>f5! fxe6+ (forced) that loses immediately! There would have
28 \t>g5 'i¥xc3 29 'iVxb7+ ltJd7 30 �g4! (30 been nothing wrong with Black's position
i.b5? 'iVf6+ 31 �g4 'i¥g6+ 32 <it>f3 'i¥h5+ after 23...i.f8! 24 ltJxf8 ltJc6 25 'ii'f6
is only slightly better for White) 30 ...�d6 'ifxa3!! (not 25 ...i.f5+? 26 �e3 l:Ie8+ 27
31 g3! (31 i.b5?! ttJf6+ 32 <it>h4 g5+!! 33 �f3 and White wins) 26 �d2! (forced, as
\t>xg5 'ti'c5+ 34 <it>xf6 i.c3+ 35 <it>g6 'i¥f5+ 26 ltJxg6? i.f5+! 27 'ifxf5 Ild8+ 28 <&t>e4
36 <&t>h6 'ii' xf4+ and 31 'ii'f 3 ltJf6+ 32 <&t>h3 'ii' b4+ wins for Black) 26...'ifb2+ 27 <&tel
'tixf3+ 33 gxf3 ttJd5! are again only slightly 'ii'cl+, and the ending after 28 �f2 'ifd2+
better for White) 31 ... 'tic5 32 i.d3! when 29 ltJe2 (29 i.e2 'ti'd4+) 29...'iVd8! 30 'tixd8
White is clearly on top; even with the ttJxd8 31 ltJxg6! hxg6 32 ttJd4 �g7 is ap
tougher proximately equal.
a2) 26...'ifxc3!, as brute force should fi [AS- This statement was based on my
nally have the last word in the line 27 old analysis and the stronger alternative
'ifxb7 �xe6 28 'ti'd5+ �e7 29 'i:Vc4! ltJc5+ 28 �e2! (D) was missed. In fact it yields
30 �f5! (30 �d5? 'ifd2+ 31 'ti'd4 'iVa2+ 32 White a huge advantage, since Black can
<&t>c6 'tie6+ 33 �b5 'tib6+ 34 �c4 a5! with not regain the piece immediately.
an attack) 30 ...'iff6+ 31 �g4 'ti'g6+ 32
<it>f3 'ti'c6+ 33 <&t>e3 a5! 34 h4! However,
Black does have a way to draw:
b) 24...'ti'a6+! 25 �c2 'ii'xe6 26 i.b5!!
(strong but still insufficient) 26...'ii'g6+
27 �b3 'iVe6+ 28 i.c4 'tib6+ 29 ttJb5 'ti'a5
(29...a6 30 l:Iel+ i.e3 [forced] 31 'tih3! axb5
32 l:Ixe3+ �d8 33 i.xf7 wins) 30 'tih4+
<it>f8 31 'iVh8+ �e7 and I don't see how
White can improve his position.
2 1 ltJg5 g6 22 'iVe5!
Another interesting attempt would have
been 22 f5!?, but it seems that Black can
hold his own in the line 22...i.e7! 23 fxg6
Shirov- King, Gausdal 1990 47
His best chance is 28 . . . V!Iib2 + (White Ernst. I am not especially proud of this
wins after 28 ... i.g4+ 29 �/2 "VJiid2+ 30 <J;;g3) game but the final combination isn't one
29 �{3 V!lib4!, but even then White has to forget.
practically a forced win with 30 0xg6! (30 1 d4 e6 2 c4 i.b4 + 3 ti'Jc3
C£Je6 i.. xe6 31 V!llxe6 + �g7 might give At that time I was so happy with the 4
Black some counter-chances) 30. . . ti'Jd4+ f3 line in the Nimzo-Indian that I even
(30. . . hxg6 31 V!llxg6 + �{8 32 ti'Jb5! ti'Jd4+ provoked my opponents into it from other
33 0xd4 V!lixd4 34 i.e2 is clearly better for openmgs.
White) 31 c3;f2 hxg6 32 V!llxg6+ c3;f8 33 3 c5 4 d5 ti'Jf6 5 f3
.••
"fiif6+ �g8 34 ti'Jd5! "VJiid2 + 35 c3;g1 ti'Jf3 + So I had achieved what I wanted. Nowa
36 gxf3 "fixd5 3 7 c3;f2! 'Wic5 + 38 �g3 i.f5 days everybody knows that 4...c5 is not so
39 fig5+ c3;h8 40 'Wih5+ ! �g 7 41 i.d3 ':f8 strong as 4...d5 in the 4 f3 variation.
42 'f!e1! '!1f6 43 Lf5 :Xf5 44 'Wig4+ �h 7 5 exd5
•••
45 fih4+ �g6 46 '11e 6+ <J;;f7 47 '11h6 and This gives White a clear positional ad
the attack should soon finish the game. vantage. The critical lines are 5... 0-0 6 e4
Thus I must conclude that the main (and d6 7 ti'Jge2 b5 8 ti'Jf4!? e5 9 ti'Jfe2 bxc4 10
not the only!) Black mistake was 20. . . fxe6, ti'Jg3, Shirov-Savon, USSR Championship
since 20 . . . i.xe6 would probably have en 1991 and 5...b5 6 e4 bxc4 7 i.xc4 exd5 8
sured the draw. ] i.xd5 ti'Jxd5 9 'Wixd5 ti'Jc6, with unclear
24 'iVg7+ ! prospects in both cases.
Now it's all over. 6 cxd5 d6 7 e4 0-0 8 ti'Jge2 a6 9 ti'Jg3
24 c3;e8
.•. b5 1 0 a4 bxa4 1 1 'ifxa4 ti'Jbd7 1 2 i.e2
24...�xe6 loses to 25 c3;c2! 'iVxa3 26 ti'Jb6 13 'ific2 a5 14 0-0 c4
i.b5! (26 i.c4+? �d7 27 '11d1+ c3;c6 is not
so clear) 26... ti'Jc6 27 i.c4+ <J;;d7 (27 ... �f5
28 'iif 7+ i.f6 29 V!l/d5+ i.e5 [forced] 30
"fiid3+ c3;f6 31 ti'Jd5+ also wins for White)
28 'f!d1+ V!l/d6 29 '11 xd6+ <J;;xd6 30 ti'Je4+
�c7 31 i.b5.
25 ti'Jc7 + c3;d8 26 "fibS + !
Preventing Black's development.
26 c3;d7
•••
Or 26...�xc7 27 ti'Jd5+.
27 ti'Jxa8 'ti'xa3 28 c3;c2 1-0
This game gave me one of the best feel
ings I have ever had from chess.
Game17
15 ti'Jd1
Shirov- K in g Probably both 15 i.e3 and 15 i.g5 h6
16 i.e3 would have been better than the
Gausda/ 1990
text-move, but it is always tempting to try
These annotations were made during the to create a direct attack against the en
preparation of this book, based on my emy king.
notes in New In Chess Yearbook 1990. 15 ti'Jfd7 16 ti'Je3 i.a6 1 7 ti'Jef5 g6 1 8
•••
20 h4!
20 ltJd2 would also have led to the ex
1 1 �a3!! change of knights, since 20 ...ltJxd2 21
I have attached two exclamation marks :xd2 with the threat of 22 f4! is bad for
to this move for the following reasons: Black. But 20 ...d5! 21 cxd5 �xd5 is much
a) there is nothing better, since if 11 e4 stronger, when Black wrests control of the
simply 1l...f4, the tempting 11 b4 does not d-file, and White has nothing better than
work on account of 11. ..4Jxb4 12 'tib3 a5!, 22 'tig7+ WeB 23 'tixe7+ Wxe7 24 ttJxe4
and otherwise White 'simply' has nothing; fxe4 with a drawn ending. With 20 h4
b) for the exchange White will have White ensures the exchange of knights at
sufficient positional compensation and g5, and Black is unable to win the pawn
excellent attacking chances; advantageously.
c) in a decisive game it is especially un 20 :ad8
•••
pleasant to have to defend, and seizing Against 20...c5 I was intending the sim
the initiative has double psychological ple 21 iHe3 and 22 ltJg5. I had spent most
strength. of my time calculating the line 20 ...We8
1 1 . .. ltJxe5 12 ltJxe5 �xe5 13 ltJf3 �xa1 21 ltJg5 ltJxg5 22 hxg5 �xg5. Now nothing
14 'fixal ltJc5 is achieved by 23 'ii'xd6 'fie7 24 �e5 �g8!
The knight is aiming for e4, from 25 'tixe7+ Wxe7 26 �a3+ �f6 27 �xf8
where it will defend the d6 pawn and con flxf8 28 :d7 :f7 29 :ds l:tg7, but with the
trol f6. But its position in the centre will accurate 23 �a3! White retains the ad
always be insecure. vantage.
50 Fire on Board
21 l2Jg5+ l2Jxg5 22 hxg5 �d7 <itd7 37 'ti'h7+ �f7 38 g6! l:lxh7 39 gxh7,
22...'iYxg5 23 "V/JJg7+ <ite8 24 'ti'xb7 -vJJle7 25 and White wins.
-vJJl xc6+ �f7 26 -vJJl f 3 clearly favours White. 32 .tf6 l:lee8 33 l:lxe6!
23 f4 'iYd8 24 <itf1?! The decisive second exchange sacrifice.
After 24 'ii' xa7 c5 White cannot avoid 33 l:lxe6
•••
the exchange of queens, but in the vari If 33...<itxe6 I was intending 34 .tb7!
ation 25 .tf6 'iYa8 26 'ii' b6! 'ii'a6 27 -vJJl xa6 -vJJla7 35 .td5+ [AS - Simpler is 35 -vJJlb5
bxa6 28 .i.c6 he wins quickly. winning immediately, but I missed it both
[AS- Therefore 24 'ii'xa 7 would simply during the game and in my original anno
win a pawn compared to the game. The tations] 35... �d7 36 'ii' b5+ <itc8 37 a5!,
text is clearly weaker.] and Black is completely helpless.
24 c5
••• 34 .i.d5 l:lfe8 35 a5!
On 24...d5 White would have replied 25 Complete zugzwang.
c5 and then 26 �d3, when Black's posi 35 <itf8
••.
tion is extremely difficult - although this If 35...h6 36 'ti'b6! 'ti'xb6 37 axb6 with
would have been a slightly better choice an easy win.
11 i.c4 0-0 12lt'Jf3 :e8 13 0-0 e5!? 14 21 l:Iad1 'ii' b7 22 'ii h3 'ii'c6
dxe6 fxe6 15lt'Jg5
In the end this move leads to victory
and to the prize 'for brilliancy'. But I also
had available another continuation - 15
f5!? Black is obliged to play 15...gxf5 16
exf5 d5, to which White replies 17 lt'Jxd5!
exd5 18 i.xd5+ <it'h8! (the only move; bad
is 18. ..lt'Jxd5 19 'ii'xd5+ <it'h8 20 'iixa8, or
19...Wf8 20 lt'Jg5! lt'Jf6 21 'ii'xa8) 19 i.xa8
c4!, when Black has some counterplay, al
though it is hard to believe that White's
material advantage will not tell.
[AS- In fact it will tell very easily; I
just wanted to justify my silly search for
sacrifices. ]
15 lt'Jf8 16 f5 h6
••• The critical position. Here I worked out
The only move. If 16 ...gxf5 then 17 i.f4! the lengthy variation 23 l:Ixd7! 'iixd7 24
is quickly decisive. lhf6! i.xf6 25 i.xf6 'iVd4+ 26 <it'h1 l:Ig7 27
i.xg7 �xg7 28 'ii h7+ �f6 29 'iVf7 + Wg5
30 'iVxe8 'ii' xc4, but I overlooked the sim
ple move 31 'iVf7! The continuation cho
sen is rather more complicated, but also
prettier.
23lt'Jd5! exd5 24 l:.xd5!
Of course not 24 i.xd5 + ? 'ii'xd5 25
l':.xd5 lt'Jf8!, when there is nothing clear.
24 �f8!
•••
17 lt'Jxe6 lt'Jxe6 18 fxe6 i.xe6 is less clear. My initial feeling was one of horror,
17 dxe5 18 fxg6! hxg5
••• since I am also obliged to give up a rook.
I was expecting 18 ...lt'Jxg6, against 27 'V/Jlxg7lt'Jxd5
which I had prepared 19 l:Ixf6! i.xf6 20 Fortunately at this point I saw that 28
'l'h5 'ii'g7 21 lt'Jge4 i.g5 (there is nothing i.g5!? lt'J7f6 29 'iixa7 l:Ie7 30 'iVb8! would
else) 22 lt'Jxg5! hxg5 23 i.d3!, and White lead to a win, and I was able to look for a
wins. However, this is not especially prettier alternative. And I found one.
pretty. 28 'iVh8+ �e7 29 g7 l:Ixh6
19 i.xg5 lt'J8d7 20 'ii'f3! l':.a7 On 29. .. .l:.g6 I had planned 30 'i¥f8+ !
The rook may come in useful on the ltJx£8 31 gxf8'ii'+ <it'e6 32 i.g5!, when White
seventh rank. wins.
52 Fire on Board
31 g8'ii' + �e7 32 'ii'd8+ 1-0 The first risky decision. The 'normal'
Black resigned, since he is mated after continuation was 11. ..e6 12 lbxf6+ 'ii' xf6
either 32 . . . rJ;f7 33 'ii' hg8+ or 32 ...�e6/d6 13 'iid2 g5!? with an unclear position,
33 'ii'xh6+ . which nevertheless did not altogether ap
I might add that the alternative win peal to me after 14 .:ac1! With the move
ning continuation on move 23 (23 .:xd7!), in the game Black wishes immediately to
although simpler, is of more or less equal seize the initiative - I was aiming only for
length, so I do not see any aesthetic viola a w1n...
tion here. [AS - It is curious that five years later
[AS- This game was awarded the bril Timman played the same line against me
liancy prize, although I must admit that it in the Donner Memorial (Amsterdam
was kind of a one-sided struggle.] 1995) and I repeated the same moves since
I realized during the game that my old
Game 20 play was entirely correct while 11... e6?!,
for example, would probably have yielded
Lautier- Shirov me a worse position in the variation indi
Manila lnterzona/ 1990 cated above.]
12 4Jxf6+ exf6 13 cxb5
These annotations were made in July 1990 [AS - Here Timman diverged with 13
and have been published in Chess in the 'iid2. After 13...bxc4 14 .i.xh6 .i.xh6 15
USSR and other magazines. "VJ/Jxh61lxb2 an extremely unclear position
This game was my second encounter arose. White obviously has some compen
against Joel Lautier and I was in quite a sation for the sacrificed pawn, but since he
competitive mood for two reasons. First, was in a good tournament situation Tim
because we were both in a must-win situ man went for a forced draw which was
ation fighting for qualification to the Can agreed after 16 4Jh4 lbe7 17 Ji.d5 0.xd5 18
didates', and second because a few months 0.xg6 hxg6 19 fixg6+ rJ;hB. But then in
before this game I had been called 'a sec the Manila Interzonal both youngsters
ond Lautier' in the magazine Europe (Lautier and I) would probably not even
Echecs and I wanted to make it the other have considered variations that might
way around. end up with perpetual check- as I have
Lautier- Shirov, Manila Interzonal1990 53
.
- � �--· ;;: --�
__ � jj�nn:;
20 ...i.e6! 21 �b7 f5! 22 l:xc7 i.d5! 23
'ii'xd5 (but not 23 'ifa7 f4! 24 i.b6 'ii' f6!
B
- •-*--'�
R -
/nm� � with a very strong attack) 23...'ifxc7 the
.t.P4l)�
� -_& -.t.� -_& /nm�
position is roughly equal, although it is
nevertheless White who has to demon
·-· • • strate this.
20 l:e8 21 l:ac1 f5! 22 l:d2
- �_&
�,%; " 1% .%� - (�
.••
14 g5!
...
15 d5!
Lautier is also very consistent. He It seems that White has defended every
could not allow the manoeuvre ...ti:Jc6-e7- thing, and that sooner or later Black's
d5, since then he might even have been in weaknesses on the queenside will tell. So
difficulties. However, there was also an that this should not happen, I decided on
other critical continuation- 15 a4!? In this my next move.
case Black should play 15...ltb3! (15... ltb8 22 he3!!
•••
is weaker on account of 16 d5! ti:Je7 17 The second and principal sacrifice. Once
tLld4, or 16...ti:Je5 17 ltfe1!) 16 d5 ti:Je5 17 and for all Black seizes the initiative, and
ti:Jd4 (17 ltfe1 'i¥b8!) 17...ti:Jc4 18 �c2 Lautier, who is much happier when he is
tLlxe3 19 'ifxb3 ti:Jxf1 20 l:xfl (or any other doing the attacking, is forced to defend.
capture) 20 ... f5!, when he has everything Psychologically he was clearly unprepared
in order. for this.
54 Fire on Board
Game 21
Hauchard- Shirov
World Junior Championship,
Santiago 1990
against him. Finally the bishop came as a 12...'tWa5 or 12...d6 also came into con
nice addition to the rooks and knights of sideration, but the move played enables
that game but the queen was only in the Black to seize the initiative.
mind's eye this time ... 13 .i.xc5
1 d4 ltJf6 2 c4 e6 3 ltJc3 .i.b4 13 cxd5 ltJxd5 14 'tWxc5 .:tc8! is totally
It is a long time since I played the Nimzo bad.
Indian Defence. The choice of opening is 13 d4 14 'iid2 .:te8
•••
explained by the fact that the French player I spent a lot of time calculating the rook
was seconded by the Soviet grandmaster sacrifice 14...ltJa5! ?? 15 ..txf8 ltJxc4 16 'tWd3
Iosif Dorfman, and, knowing his experi .:tc8. If now 17 .i.b4 then after 17 ...ltJe3 18
ence (four Kasparov-Karpov matches!), I <itf2 e5! Black's attack is indeed danger
did not want to use my main repertoire. ous. Unfortunately, Black's idea is de
[AS- I still don't really play the Nimzo, stroyed by 17 :c1!, when after 17 ...ltJe5
so I can't really say much about the open 18 'ifd2 <it'xf8 (18...l:.xc1+ 19 'tWxc1 'tWxf8
ing of this game. ] 20 'tWc7! wins for White) 19 l:.xc8 .i.xc8 20
4 'ifc2 0-0 5 a3 .i.xc3+ 6 'ifxc3 b6 7 e3 it only remains for him to resign. But
J..g5 .i.b7 8 ltJh3 after 14.. J�e8 I no longer see a way to
As yet this move has not become as equality for White.
popular as 8 f3. In my opinion, it deserves 15 e4
closer consideration. 15 e3 was no less dangerous for White-
·
Not 11 e3 g5! 12 .i.g3 ltJe4 13 'it'c2 'it'a5+ It is surprising that, when making this
when Black is better. move, I had already seen the possibility of
11. ..ltJc6 12 .i.f2?! a queen sacrifice on move 28.
This seems to be a mistake. The correct 17 b4
12 e3 g5! (12. ..d5? 13 .i.xf6 'tWxf6 14 'iixf6 17 ltJd3 ltJd8 would have been no better
gxf6 15 cxd5 exd5 16 ltJf4 is favourable for for White.
White) 13 .i.g3 d5! leads to terribly com 17 ltJd8! 18 ltJd3 ltJe6 19 .i.e2
•••
plicated play.
���.i} �--A,·
?;�Ciriii
0/_..///: fA�
$/.:: /:,, ' '"' ////1?�
• • • •
RLSa R R
"
�;;; �
�� -�·tij
. '�
%tiP!
�d ��0�+�0��A -..- A� ��;
�a"', � %�
/on;� / /' /'un.-�
;;;
/
�
////:- ,/i� m�£if%��
� ��� F?
-'///�;'('� ;;jo �
// ;/�
19 .i.xe4!
•••
56 Fire on Board
Game 2 2
Lutz- Shirov
World Junior Championship,
Santiago 1990
on d5, White's queenside pawns will be Black at last reminds White of his main
more of a weakness than a strength. weaknesses - his queenside pawns.
16 i..d3 cxd5 17 ltJxd5 26 i..g5!?
The position arising after 17 cxd5 i..d 7 White launches a desperate attack, but
18 a4!? is also worth studying. this proves unsuccessful, since the black
17 ltJxd5 18 cxd5 i..d7
••• pieces are excellently co-ordinated. How
It is clear that once Black has placed ever, I no longer know what to suggest
his queen at b6 and his rook at c8, he will instead. 26 a4 is rather strongly met by
have everything in order. 26...l:Ic4, while if 26 lle2, then 26...i..b5!
19 l:.c1 'iVb6 20 �h1 l:.ac8 21.l:Ixc8?! 27 l:lf2 iVxa2 and Black is simply a pawn
White shouldn't have simply conceded up. Finally, after 26 f4 exf4 27 i..xf4 iVxb4
the c-file. 21 ltJh3 was preferable, and if, White again has no play.
as in the game, 21...'ii'd4 22 i..e4 �g8 23 26... ltJf5 27 g4 hxg4 28 fxg4 ltJd4 29
'i'e1! (weaker is 23 l:le1 l:lxc1 241Vxc1l:!c8 i..xg6 'ifxa2 30 i..e4 l:lf8! 31 'fid3
25 'fid1, leading to a position from the Other moves would also have lost: 31
game) he has reasonable play in view of l:Ie3 l:If2 32 l:Ih3 i..a4 33 �c1 i..c2! or 31
the threat of 24 g4. ltJh3 i..a4 32 'ii'd2 iVxd2 33 i..xd2 i..d7! 34
21. l:Ixc8 22 ltJh3 iVd4 23 i..e4 �g8!
•• l:Ig1 ltJe2 and Black wins in both cases.
(D) 31. i..xg4 32 iVg3 i..f5 33 i..g2 l:lf7 34
.•
An essential move, otherwise 24 ltJg5 i..h6 i..d7 35 iVg5 ltJf5 36 'fid8+ �h7
with the idea of 25 g4 would have been (D)
very unpleasant. 37 i..c1?
24 l:Ie1 White could have set his opponent
24 'fie1!? was interesting. much more difficult problems by 37 i..xg7
24 ltJe7!? 25 ltJg1?
••• l:Ixg7 38 i..e4 'ii'f 2 39 l:ldl. At any event, it
Up till now Lutz had played quite well. was only with difficulty that in analysis I
But now he should have continued 25 i..g5! found a way to win: 39...iVe3! 40 iVh4 +
'i'xd1 26 l:Ixd1 �f7 27 l:Ic1!, and although �g6 41 ltJf3 (or 41 l:Ie1 1Vg5 and wins)
after 27...l:.xc1 + 28 i..xc1 ltJf5! Black has 41..J:th7 (in fact 41...iVxf3+ is a lot sim
good prospects, White's position is by no pler than 41...l:Ih7; White may try 41 l:ld3
means hopeless. I should mention that, instead of 41 ltJf3 but the endgame after
instead of 27 l:.c1, 27 i..xe7? �xe7 28 .i.xg6 41...iVxg1+ 42 �xg1 �f7+ 43 'ti'g3 ltJxg3
58 Fire on Board
Now it is all very simple. There fol This long castling was introduced into
lowed: serious tournament practice by Kasparov
38 'ii b8 .:tg7 39 i.e4 'iif2 40 .:td1 'iVh4 in his game against Campora at the Dubai
0-1 Olympiad in 1986 and I myself had a lot
With the time-trouble over, White re of success with it before I got a lost posi
signed. tion against Milos in Buenos Aires 1993.
Although I won that game too, I still un
Game 23 dertook not to play 10 0-0-0 anymore.
10 .'i¥a5
••
27 l2Jd5 wins) 27 lDxb5 "ii'b6 28 lDd6 is cur White has achieved his best possible
tains. set-up and has various winning ideas such
25 "ii'd2 dxe4 26 fxe4 :Ih4! (D) as 3 7 ltcfl and 38 g4 !? In time-trouble
Black throws in the towel after just one
move.
36 l:.h5? 37 �xd4! 1-0
•••
Game 24
Shirov- Piket
Groningen 1990/9 1
These annotations were made in January
199 1 and first appeared in Shakhmaty
Riga.
This game is memorable for me be
Black sets up some play against the h2 cause it was very much liked by Alexan
and d4 pawns and White needs to play ac der Koblents, who called me soon after he
curately. saw it published. That telephone conver
27 g3 :Ih6 28 e5! :idS 29 "ii'e3 sation was one of my last contacts with
Threatening 28 l2Je4. the old maestro as he soon moved to Ber
29 l2Je6
••• lin where he lived his last years. He
29 .. . "ii'b6 would not be a great improve passed away in 1993.
ment due to 30 l':.hfl ! l2Je6 (30 . . . :Ixh2 3 1 1 d4 l2Jf6 2 c4 g6 3 l2Jc3 �g7 4 e4 d6 5
e 6 wins) 3 1 l2Jxe6 fxe6 (3 1 . . . :Ixe6 3 2 .l:tf4! �e2 0-0 6 l2Jf3 e5 7 0-0 l2Jc6 8 d5 l2Je7 9
wins) 32 "ii'f3 �g6 3 3 �xg6 hxg6 3 4 h4! l2Jd2 a5 1 0 a3 l2Jd7 11 l':.b1 f5 12 b4
with a huge advantage for White. axb4 13 axb4 �h8 14 �c2 lDf6 15 f3 c6
30 �xb5! ltJxd4 3 1 �c4 According to Piket, this was a new
White's advantage in pawn structure is move, which he had prepared specifically
smoothly transformed into piece supe for this game. But I think that 15 . . . c6 had
riority. Black does his utmost to co-ordi already occurred somewhere before, with
nate his position but his task proves to be White replying 16 �hl.
impossible.
3 1 . .. �d7!
3 1 . . . l2Jf5 can be answered by 32 "ii'e4
l2Je7 33 h4! 'arresting' the rook on h6.
32 a3!
White is in no hurry, as he only needs to
improve the position of his pieces.
32 �f5 + 33 �a2 �e6
•••
16 bxc6
••• I preferred this move to the tempting
On 16 . . . ttJxc6 I would have replied 1 7 26 'iVd3 !?, since I could not find anything
liJb5!, with the idea of 1 8 'i¥d3, when White in particular in reply to 26 . . . ltJg4 27 'iVf3
has the advantage. h5 28 ttJxb7 'ifxb7 (28 . . . l:.f8? 29 ttJd8! 'iVb6
17 b5 cxb5 30 lLlf7 + l:lxf7 3 1 'ifxf7 tLlf2 + 32 'iVxf2
The immediate 1 7 . . . d5 doesn't work 'iixf2 33 �xe7 wins for White) , and I did
on account of 18 b6 ! l:lb8 19 �a3, with a not want to have to be satisfied with a
great advantage to White. slight advantage after 29 �c5 !
18 ttJxb5! 26 h6 27 'ifb3 ! tiJfg8
•••
18 cxb5 could now have been answered Now 2 7 . . . ltJg4 is no longer so strong,
by 18 . . . d5 19 b6 �b7, when all Black's since White replies 28 l:ld6 'iVc8 29 �g1 ! ,
problems are solved. But now it will not when there is no way for Black to co-ordi
be easy for him to defend his d6 pawn. nate his forces, whereas the white pieces
18 ttJh5 19 ttJb3!
••• are placed on dominant squares (29 . . . 'iff8
But not 1 9 g3 tLlc6 ! 20 'ifd3 ttJd4, when 30 'iff3 ! �c6 3 1 :Xc6!, or 29 . . .i.c6 30 ltJe6!).
the position becomes unclear. [Translator's note - After 31 l:lxc6 !
19 fxe4 20 fxe4 'ifb6 + 2 1 �hl
••• Black can play 3 1 . . . 'ifxf3 32 gxf3 tLlxc6
.l:xfl + 33 fxg4 :xa3 34 ttJxa3 �f8 with equal
The only move. Totally bad was 2 1 . .. lLlf4 chances. ]
22 �xf4 exf4 23 l:.bd1 �e5 24 ttJxd6! �xd6 2 8 �c4!
25 'ifc3 + ! �g8 26 c5 �xc5 2 7 �c4 + �e6 At this point I could not believe that
28 ttJxc5, when White wins. the battle could drag on for long. All my
22 �xfl ttJf6 23 �a3 �b7 pieces are on the attack, and are operat
Understandably, 23 . . . ttJg4 24 c5 dxc5 25 ing so harmoniously. But . . .
i.xc5 'iVf6 26 h3 did not appeal to Black at 28 'iVb6!
..•
all.
62 Fire on Board
It turns out that Black's forces are also there would be new threats to parry. I had
well co-ordinated, and are ready to parry to compose myself, quickly evaluate all
White's onslaught. Here I was obliged to the possibilities, and land a decisive blow.
think. I failed to find a forced win, al 35 �xg6!
though I managed to calculate some vari As soon as this move is made on the
ations. As always, there was little time board, the smoke disperses, and it becomes
left, but I realized that the preceding play clear that the black king' s defences have
had to be brought to a logical completion. been destroyed.
The remainder of the game resembles a 35 ...'iVh5
genuine melee, but by no means a finish Forced. Black loses after both 35 . . . �xg6
ing blow. 36 i.xg8, and 35 . . . 'iVel + 36 i.gl i.d4 37
29 �d7 'iVf2 30 �d6 i.c6! �f8 + �h8 38 �f7+ �g7 39 �e6 + .
After 30 . . . i.a6 3 1 �xe5 ! i.xc4 32 �exc4 36 �xe7 �xe7 37 �f5
it would all have been immeasurably eas Had I had more time, I would have con
ier for White. cluded the game differently, by 37 e5 !
3 1 i.c5 'ii'h4 i.xe5 (37 . ..'ii'xe5 38 'iVg8 + ! ! �xg8 39 i.d3 +
In time-trouble I was afraid of the �h8 40 �f7 mate) 38 i.d3 + �g7 (38 . . . �g6
queen sacrifice 3 1 .. . 'ii'f4 32 l:Ifl i.xd7 33 39 'i!Vb7 + �h8 40 �f7 + �g8 4 1 �xe5
.:xf4 exf4, but now I realize that after 34 'iVxe5 42 i.xg6 and 39 . . . i.g7 40 �f5 are
h3 the absence of Black's commander also winning for White) 39 �e8 + �f7 40
condemns him to a rapid demise. i.c4 + �g6 4 1 'iVa6 + �h7 42 �f6 + i.xf6
32 �xe5! 43 'ti'xf6 'ti'xc5 44 'ti'f7 + �h8 45 'ti'f8 +
The first blow against the opposing �h7 46 i.d3+ and White wins.
army (32 . . . i.xe5 33 �f7 + ) and a feeling 37 i.e5 38 i.g1 i.f3!
•••
of contentment . . . but an instant reply - This is not yet the last gasp, but a final
32 i.a4!
••• trick with the opponent's flag about to
. . . and a feeling of horror, since I had fall.
overlooked this move. I had to respond 39 'ti'b7!
likewise, by also attacking a rook. Of course not 39 �xe7 i.xg2 + 40 �xg2
33 'ifb7 i.xd1 34 'iVxa8 �h7 'i!Vg4 + 4 1 �fl 'ti'd l + , with a draw by per
petual check.
39 .i.xg2 +
.••
Game 25
Adams - Shirov
Bie/ 199 1
These annotations were made during the
preparation of this book, and based on my
notes in Informator 52.
Piket played so quickly, that he effec In my first 15th category tournament I
tively gave me no respite at all, whereas managed to perform well. Not only the re
my clock was inexorably advancing. At sult (91h out of 14 and a clear first place)
first I wanted, without thinking, to reply but also the games, full of fight, gave me a
35 �d3, but I saw that after 35 . . . �f6 ! feeling of satisfaction. One may say I was
Adams - Shirov, Biel1991 63
lucky in several cases and in a way it's months before this game. 10 . . . 11xf7 had
true because somehow the situations of been automatic before.
mutual tension worked out well for me 1 1 ltJxe4
and gave me some extra points. l l lDxc6 bxc6 1 2 liJxe4 �g8 is not what
I put a lot of energy into the games and White normally wants.
I didn't get tired. One week later I went to 1 1 . ltJxd4 12 i.xd4 e5
••
the Lloyds Bank open in London and won Now Black's idea is clear - he will have
it as well. Unfortunately nowadays I am two powerful central pawns, since White
less consistent . . . doesn't have 13 i.c5 (as would have been
My game against Adams is a typical ex the case with 10 . . . lhf7) .
ample of my play in Biel: aggressive play 13 i.e3
from the opening, putting on the pres 13 'iff3 + ? �e8 is clearly better for
sure, a little risk in complications and get Black.
ting the upper hand. 13 d5 14 liJg3
•••
Threatening 22 . . . l:Ih5 . During and af White 's kingside pawns look terrific
ter the game I thought that my attack by but probably Black is not lost because of
this point was already nearly decisive. the pin and the aforementioned a-pawn.
Closer analysis proves that this wasn't He should now continue 32 . . . 'iVc2 ! 33 f6
true, but it is certainly psychologically i.. f8 34 'ii'g5 + �f7 35 'ii'h 5 + 'ii'g6 36 �f3
unpleasant to be White in this position. 'ii'c2! and since the line 37 'ii'xb7+ �g6 38
22 f4 'iff3 i.. c5 + 39 �fl �f5 40 �g2 'iVxf3 + 4 1
22 l2Jg3 l:If4 23 i.. e 7 �f7, followed by Wxf3 a5 leads to a draw, I don't see any
24 . . . lth8, looks extremely dangerous for real winning chances for White. This
White. proves that had Adams played 23 i.. g3 he
22 �g4!
••• would have had a clear practical advan
Now 22 . . . 1':.h5? fails to 23 i.. g5 ! tage, because the line indicated above
23 .i.g5? would have been very difficult to find over
But here this move is a big mistake. 23 the board.
i.. e 7 (which I thought was the White 's 23 l:Iaf8!
•..
only move! ) is also extremely dangerous Now the threat of 24 . . . d4! is lethal. I al
due to 23 . . . l:Ie8 24 i.. d 6 d4! with the idea ways enjoy it when all my pieces are in
of 25 .i.xe5? i.. d5 ! and Black wins after 26 the attack.
Shirov - Kozul, Biel 1991 65
24 .:f2 13 h4 .:fc8
24 .:tf3 .l:txg5 25 .l:tg3 would not help be The game develops to White's advan
cause of 25 . . . 'i¥xg3, while 24 l:Iae1 d4 tage after 13 . . . l:Iac8 14 h5 e5 15 hxg6 hxg6
should be more or less the same story as 16 d5 liJd4 1 7 lbxd4 exd4 18 i.d2, as in
in the game. Shirov-Pieshina, Daugavpils 1990.
24 ... d4! 25 cxd4 exf4 26 i.. xf4 [AS - That game continued 18 . . . 'ii'b 6 19
26 l2Jxf4 .l:txg5 27 l2Jxe6 l:txf2 28 'ii'xf2 �gl l:IfeB 20 I:tb1 i.a4! ? 21 "VIIifl 'ii'c5 22
'ii'xe6 also loses the house. i.d3 (White stands slightly better here)
26 i.. c4
••. 22. . . �a3 ?! 23 '!!xb7 l:Ic3 ?! 24 i.xc3 dxc3 25
Now Black wins the piece. There fol "Wibl i.d4 26 l:Ih3 'ii'c5 27 "Wib4 i.x{2+ 28
lowed r:tifl c2 29 'iVxc5 i.xc5 30 i.xc2 i.xc2 31
27 i.e3 l:txf2 28 i.xf2 i.xe2 0 -1 .l:tc3 (now White is winning) 31 . . . i.b6 32
'f!xc2 'f!xe4 33 d6 'f:.d4 34 d7 �g7 1 -0 (35
Game 26 :IkB and wins).]
14 h5 lbd8
Shirov- Kozul
Bie/ 1 99 1
This game was annotated in August 1991 w
advantage, since the advance of the pawn Preferable was 15 . . . i.b5 16 i.xb5 'iVxb5
to f3 proved to be in his favour. Interest 17 �f2 .:Xc1 18 'iVxc1 lbc6 19 'iVb1 'iVa6 20
ing positions also arise after 12 . . . l;Id8 1 3 hxg6 hxg6 2 1 e5 l:td8 22 'ife4 'iVxa2 23
h4 h 5 o r 13 . . . h6 . �h4 �f8, as in Shirov-1. Gurevich, World
[AS - The last word in this variation is Junior Championship, Santiago 1990) .
12.. . "VIIia3! ? The games Ftacnik-1. Gurevich [AS - This note is just a trick. See the next
(Biel Interzonal 1 993) and Kamsky-Anand game, played a few weeks later in Lon
(Las Palmas 1 995) are good examples of don.]
this.] 16 i.b3 .:txcl 1 7 i.xcl fib6?!
66 Fire on Board
Game 27
Shirov- Ernst
London (Lioyds Bank) 199 1
These annotations were made during the
preparation of this book, based on my
notes in lnformator 52.
Before starting to annotate this game, I
would like to delve a little into its history.
In the 1990 World Junior Championship
in Santiago, I lost a decisive game against
Ilya Gurevich in the same variation of the
20 ltJe6 2 1 i.xe7 g5
••• Griinfeld Defence. After that game I tried
The only move. to find an improvement for White, but did
22 d5 ltJd4 23 �xg5 b4 24 i.e3! 'iia6 not succeed for almost a year. In June
On 24 . . . �b5 there follows 25 'it>f2, and 199 1, after a break of a couple of weeks
White is close to a win. from chess, I decided to undertake some
25 �xd4 .txd4 26 'ii'd2 �g7 preparation for the Biel tournament. I set
Black appears to have compensation up the position after Black's 2 1st move
for the pawn. What is to be done against and in the same second the move a2-a4
27 . . . �b5? occurred to me. Although it doesn't win
27 d6! �b5 immediately, as I had originally thought,
it certainly promises White the better
chances. The Polugayevsky variation (i. e.
the variation characterized by 10 l':.c1)
may already be part of chess history, but I
still hope to have that kind of inspiration
over the chessboard sometimes.
1 d4 ltJf6 2 c4 g6 3 ltJc3 d5 4 cxd5
ltJxd5 5 e4 ltJxc3 6 bxc3 �g7 7 �c4 c5 8
ltJe2 ltJc6 9 �e3 0-0 10 .:tel (D)
lO cxd4 1 1 cxd4 'il/a5 + 12 'it>fl �d7
••.
13 h41Uc8 14 h5 ltJd8 15 f3
When I played this move against Ilya
Gurevich in the World Junior, it was a
novelty.
15 �b5
•••
Shirov - Ernst, London (Lloyds Bank) 1991 67
19 �a6
•••
B 22 a4!
This is my improvement over the Shi
rov-I. Gurevich game, which went 22 'tWe4?
'tWxa2 23 'ti'h4 �f8 24 l:td 1 (24 'ti'h7 ttJxd4
wins for Black) 24 . . . ttJxe5 25 'tWh7 g5 ! 26
�d2 (26 i.xg5 ltJg6 is also better for Black
- I. Gurevich) 26 . . . 'ti'a5 2 7 ltJg3 �b6 ! 28
ttJf5 'tWg6 ! 29 'tWxg7 + 'tWxg7 30 ltJxg7 ltJc4
and Black won the endgame.
22 ttJa5?
•••
21 iYh2 ttJxd4 22 'tWh 7 + �f8 23 �xg7 + 24 . . . �c6 would have prolonged the game
�xg7 24 i.xd4 + with a winning position) without changing the result, because the
2 1 'ti'h2 ttJd3 + 22 �g3 f5 ! 23 iYh7 + �f7 ending after 25 �h4 g5 26 'ti'xg5 'ti'g6 2 7
24 .:h6? f4 + ! 25 i.xf4 ttJxf4 26 �xf4 iYa6 ! 'tWxg6 fxg6 2 8 ltJf4 is hopeless for Black.
and Black wins. 25 i.g5 'tWe6
68 Fire on Board
Game 28
Shirov- Speelman
London (Lioyds Bank) 199 1
These notes were made in 1991, with the
assistance of, and translated by, Jon Levitt.
1 d4 d6 2 e4 t2Jf6 3 �d3
Normally I play 3 l2Jc3, but the game
Anand-Timman, Linares 199 1 , had made
a big impression on me. 1 1 t2Jxe5
3 e5 4 c3 l2Jc6 5 t2Jf3 �e7
... Another plan would have been 1 1 l2Je2,
In the aforementioned game Timman with the idea of g2-g4 and l2Jg3.
continued 5 . . . �g4, but after 6 d5 l2Je7 7 c4 1 1 t2Jxe5 12 'ii'xh5 t2Jxc4 13 0-0 l:te8
•••
l2Jg6 8 g3 �e7 9 l2Jc3 White had an edge. Probably this is the correct move, but
6 d5 lbb8 7 c4 only if played in conjunction with a sub-
Black has lost two tempi with his sequent . . . �f8. Speelman was worried
queen's knight, whereas White has lost about 1 3 . . . �f6 14 f4, but after 14 . . . �xc3
only one with the c-pawn, so White has a 15 bxc3 g6 16 'tWe2 cxd5 1 7 exd5 b5 both
clear advantage. sides have chances. Stronger is 14 'ir'e2
7 0 -0 8 l2Jc3 t2Jbd7 9 �c2
.•. l2Jb6 (but not 14 . . . �xc3? 15 'tWxc4 �f6 16
Usually such a move is only played af dxc6 �e6 1 7 'tWd3 and White wins since
ter . . . l2Jc5, but I was afraid that after 9 1 7 . . .bxc6 is met by 18 e5) 15 dxc6 bxc6 16
0-0, 9 . . . l2Jh5 is possible (with the plan 'i¥d3 g6 1 7 �h6 :e8 18 :ad 1 d5 19 �b3
. . . g7-g6, . . . l2Jg7, . . . f7-f5). I thought that 9 �e6 20 exd5 . In this variation neither
�c2 prevented 9 . . . l2Jh5 in view of 10 l2Jxe5 20 . . . l2Jxd5 2 1 'tWf3 �xc3 22 bxc3 'tWh4 23
l2Jxe5 1 1 'i¥xh5 (when White's bishop on �c 1 (or 23 �d2) , nor 20 . . . cxd5 2 1 'i¥f3 d4
d3 is no longer en prise). It was only after 22 l2Je4 �g7 23 �g5 offer Black much
Black's reply that I saw 1 l . . . �g4. How hope.
ever, 9 i.c2 is quite a good move anyway, 14 l1d1
Shirov - Speelman, London (Lloyds Bank) 1 991 69
20 f5
••• White would be winning after 34 . . . l:tadS
Mter 20 . . . .i.e6 I would have played 2 1 35 l:te7 �d3 + 36 �gl, but now 35 �gl
i.b3, since if Black wins his pawn back fails to 35 . ..�d4 + .
with 2 1 . . . .i.xb3 22 axb3 �xc3 23 bxc3 35 �f2 l;tad8 3 6 ttJb6 �f6
�xe4, White would have a won endgame 36 .. . �f5 might have caused me serious
after 24 c4! problems in time-trouble. Mter 37 ltJc4
21 �b3 + �e6 �f6 3S :xeS + l:txeS White does not have
2 1 . . . �g7 22 exf5 �xf5 23 �e3 is hope 39 .:taS (as in the game), and 3 7 l:te2 �d3
less for Black. 3S .:td2 is precarious: 3S . . . �f6 ! 39 �gl
22 exf5 gxf5 23 �e3 ltJc4 24 �xc4 I:tel + 40 �h2 �c3 ! It was necessary to
i.xc4 find 3 7 l:taS! when Black can choose be
Black has the two bishops, but no less tween 37 . . . %:txaS 3S ltJxaS l:.xaS 39 l1xe5
than four pawn islands as well! White I:txa2 + 40 l:te2 .:tal ( 40 . . . I:txe2 + 4 1 �xe2
should be able to win comfortably. How and the f4 pawn will go) 4 1 .:td2 �e6 42
ever . . . I:tdS + �f7 43 lifS+ �g6 44 b4, winning
70 Fire on Board
for White, and 37 . . J!d2 + ! 38 �g1 .:xa8 39 pressed too hard against Stohl and Mokry
4Jxa8 .i.c3 ! with drawing chances. and lost both games after having refused
37 :xeS + :xeS 38 :as .i.d8 39 �gl draws.
Not 39 g3? �f7. As a good consolation, two of my three
39 �f7 40 :a7+ �g6 41 .i.f2
••• won games were of a high quality.
Eliminating any counterplay. 1 d4 d5 2 c4 dxc4 3 e4 4Jc6 4 �e3
41. .. :el + e5?!
The opposite-coloured bishops ending is In this position this move looks some
lost but Black saw no reasonable option. what suspicious. More common is 4 . . . 4Jf6
42 .i.xel �xb6+ 43 �f2 ..txa7 44 5 4Jc3 e5 6 d5 4Je7, although White has
.i.xa7 also been doing well here recently.
[AS I should add that the amazing
-
B
Game 29
Shirov- Murshed
Brno 199 1
These annotations were made during the
preparation of this book, based on my
notes in lnformator 52.
Following the Biel and London tour
naments, I was less successful at Brno. I
Shirov- Murshed, Brno 1991 71
At this point I was feeling very optimis dangerous for Black, for example 1 7...'ifxe4
tic. 1 2 . . . l:xb2 1 3 g3 l2Jg6 14 l2Jd2 is clearly 18 f3 'i'e3 19 �el l l:b7 (19 . . . 'i'xc3 20 l:xc3
in White 's favour, so what is the point of l2Jh4 2 1 l:Ixc7 l2Jxf3 22 �f2! l2Jd2 23 l:dl
Black's play? But . . . l2Je4 + 24 <it>e3 is obviously better for
12 i.c5!
••• White) 20 l:Ig1 'ti'xc3 2 1 llxc3 l2Jh4 22 l:Ixg7
The next two moves took me nearly an l2Jg6 23 l2Jg3 ! with a clear advantage. Now
hour, so that from now on I had barely 23 . . . l2Jf6? is met by 24 d6 cxd6 25 %Ixg6 ! ,
more than half an hour to reach the time winning.
control. 18 l2Jxc3 l2Jf4 19 l:Ig1 g6 20 l2Je2! <it>d8
13 i.xc5! For the time being Black defends well.
Otherwise White loses, for example 13 Much weaker would have been 20 . . . l2Jd3
g3? l:tb6 14 'ifa4 i.xe3 1 5 fxe3 l:xb2! (but 2 1 l:g3 ! l2Jxb2 22 l:c1 �dB 23 %Igc3 with a
not 15 . . . 'iif6 + ? 16 <it>e2 l:xb2 + 1 7 l2Jd2 won position.
with an unclear position) and Black's at 21%Ic1
tack looks decisive.
13 'iig5 14 l2Jd2!
•••
21 ...l2Jf6?
But this is a mistake. After 2 1 . .. l2Jd3 22
%Ic2 f5! 23 exf5 gxf5 White can only claim
a slight advantage. Now Black gets into
trouble.
Now 14 . . . 'ifxg2 + doesn't work in view 22 l2Jxf4 exf4 23 d6!
of 15 <it>e2 l:txb2 (15 ... 'ifxh1 16 l2Jgf3! 'ifxa1 This is stronger than 23 f3 l2Jd7, when
17 l2Jxe5 l:d8 18 l2Jxf7! wins for White) 16 White only holds a slight edge.
�d3 ! ! 'ti'xh 1 (or 16 . . . l:xd2 + 1 7 <it>xd2 win 23 .:cs
.••
ning) 1 7 l2Jc4 (the other knight heads for 23 . . . cxd6 24 i.xd6 l':.xb2? 25 i.e5 l:b6
e5 this time! ) 1 7 . . . l:b8 18 l2Jxe5 lidS 1 9 26 i.c7 + loses immediately, but Black could
lbxf7 and White wins. have tried 23 . . . l:Ib7 24 f3 l:Ie8, though 25
15 i.a3 �d3 + 16 l2Je2 l2Jxg2!? .:g2 still promises White an edge.
16 . . . 'i'xe4 17 l':.gl l2Jf5 18 'iixc7 'iixd5 19 24 f3 lle8 25 l:Ig5! cxd6
'i'c2! yields White the advantage. The text After 25 . . . l':.e6 26 l:gc5 ! (not 26 l':.a5?!
looks logical. cxd6 27 l':.xc8 + <it>xc8 28 l':.xa7 l2Jd7! when
17 'ifc3! 'ifxc3 Black has some counterplay) 26 . . . l2Je8
Probably Murshed thought that he (forced) 27 dxc7 + l:xc7 28 i.b4! White has
would be able to hold the endgame, but it a clear advantage.
is not so easy. In any case he made the cor 26 llxc8 + �xeS 27 i.xd6 l2Jh5 28
rect decision because the position is very llc5 + <it>d7 29 i.e5
72 Fire on Board
20 0 -0
..•
Despite his pawn minus, White can 20 ... �d7 2 1 4Jc5 + <it>d6 22 4Jxa6 :b6 23
quietly increase his initiative. His pieces 4Jc5 4Jd5 24 4Jb3! is also bad for Black.
are much more active than Black's and he 2 1 4Jc5 :res
has an excellent pawn structure. The exchange sacrifice 2 l. . . a5 22 4Jd7
15 l:td8
••• axb4 fails to 23 4Jxb8 l:txb8 (23 ... 4Jxb8 24
15 ... 0-0 is no great improvement on the axb4 also wins) 24 i.. xc6 ! 4Jxc6 25 .:tacl
game due to 16 a3 'ti'g6 1 7 f4! , when White 4Ja7 26 :Xc7 4Jb5 27 :b7! and White wins.
stands clearly better, but 15 ... 0-0-0!? is in 22 l:tacl
teresting. Still, it seems that White is bet Now White's advantage is decisive.
ter after 16 4Jc5 l:td5 1 7 'ti'g4 + <it>b8 18 22 4Ja7 23 4Jxa6 l:tb6 24 4Jxc7 g6 25
•••
is unclear. 4Jd5?!
16 'ti'e6?!
••• Zsuzsa starts to err in time-trouble and
An unfortunate decision in an already loses quickly, but objectively her position
difficult position. It would have been bet was hopeless in any case.
ter to play 16 . . . 'ii'g6, although after 1 7 28 l:tcl lbf4?! 29 4Jxd4 l:td8 30 4Jxc6
4::\c5 ! (not 1 7 f4?! 'ti'xe4 1 8 i.xe4 �d7 1 9 l:.xd3 3 1 :tal lbe2 + 32 <it>h2 4Jc3 33 a4
4::\c5 + <it>c8 when matters are not clear at 1 -0
74 Fire on Board
Game 31 13 exf5
A novelty. Mter 13 c5 Black would play
Shirov- Nunn 1 3 . . . f4, transposing to well-known vari
Bundesliga 199 1 ations.
13 gxf5
•••
These annotations were made during the During the game I expected 13 . . . ttJxf5
preparation of this book, based on my as the most logical choice. Compared to
notes in lnformator 53. the aforementioned Lutz-Shirov game,
This was my debut game in my first White hasn't yet played b2-b4 and I hoped
Bundesliga season. The German Team to exploit this somehow, but now I am not
Championship consists of seven weekends sure whether White has the slightest ad
of two games and one of one game every vantage after 14 ltJf2.
year. A tough routine (the games start at 14 f4! e4 15 ttJf2 t2Jg4
2 p.m. on Saturdays and 9 a.m. on Sun On 15 . . . h4 I was planning 16 ttJh3 with
days) and the importance of the matches a possible ltJg5 one day, and I quite like
demands a high level of concentration. this for White. On my database I found
In my first season (I am now nearly a some games in which Black played 15 . . . c6,
veteran, having competed for four years) I but I presume that after 16 dxc6 bxc6 17
was especially successful, probably be i.e3 White stands excellently.
cause I felt very relaxed - such was the at 16 tbxg4 fxg4?
mosphere in the Hamburg team. It seems to me that this is a serious
1 d4 ttJf6 2 c4 g6 3 ttJc3 i.g7 4 e4 d6 5 mistake. Mter !6.,..hxg4 1 7 i.e3 White
ltJf3 0 -0 6 i.e2 e5 7 0 -0 ltJc6 8 d5 ltJe7 9 has a pleasant game, but I don't believe
ttJe1 ttJd7 that his advantage is anything special, be
At that time 9 . . . ltJe8!? was still not very cause Black is very solid.
well known. Today it seems to be Black's 1 7 ltJxe4 i.xb2 1 8 l:b1 ! i.d4 + 19
main weapon (with it, John Nunn gained <it>h1 t2Jf5
his revenge against me at Amsterdam Black's counterplay on the kingside is
1995) . in fact not dangerous at all. However, he
10 liJd3 f5 1 1 i.d2 t2Jf6 had little choice as opening the centre
When I played this variation as Black, I would only favour White, for example
used to employ 1 1 . . . <it>h8 and only after 12 19 . . .c5 20 dxc6 bxc6 (20 . . . ttJxc6 2 1 l:Ib5 !)
b4, 1 2 . . . ttJf6 13 f3 h5 (see Lutz-Shirov, 2 1 i.a5 !? 'ifxa5 22 'ifxd4 t2Jf5 23 'i¥d3 with
Santiago 1990 - Game 22 in this book) . a clear plus. Neither was 19 . . . i.f5 advis
12 f3 h5!? able, in view of 20 i.d3 b6 2 1 'ifc2 with the
Avoiding the line 12 ... f4 13 g4. idea of 22 tbg3 or 22 ltJg5 .
Shirov - Nikolenko, USSR Championship, Moscow 1 991 75
22 . . . l';Ie8 23 ltJe6 i.. xe6 24 l:txe6 ! l:xe6 26 . . . i.. xe6 2 7 ltJxe6 l:txe6 28 'ifxg4 +
(24 . . . ltJg7 25 l:h6 also wins) 25 dxe6 'iff6 ltJg7 29 dxe6 is absolutely hopeless.
26 'ife2 ! , intending 27 'ife4, when White is 27 i..h7+ �g7 28 'ifd3! i.. xe6 29 'ifxc3!
clearly on top. Actually 29 dxe6 l';Ixf4 ! 30 'ifg6 + �h8
23 i..b 4!! (or 30 . . . �f8 3 1 'ifxh6 + �e7 32 liJf7) 3 1
White has discovered the most vulner i.. g8 ! ( I had overlooked this during the
able point in Black's position - the bishop game, seeing only 3 1 'ifxh6? 'iff8 ! , when
on d4 (in the very centre!) which cannot everything is unclear due to the threat
be protected and lacks useful squares. 32 . . . l:fl + ) 3 1 . . .'ife7 32 i.. f7! would also
The rest of Black's play is agony, but White win nicely but from an aesthetic point of
had to play precisely to the end. view I still like the text more.
23 a5
••• 29 ...i.. g8
Forced. 29 . . . i.. d 7 is also met by 30 i..b 2 with the
24 i.. a3 i.. c 3 25 :e2 idea of 3 1 ltJe4.
Now the threat is 26 l:cl i.. b4 27 i..b 2, 30 i..b 2! 1 -0
occupying the long diagonal and winning The threat of 3 1 i.. xg8 is irresistible
the game. (30 . . . i.. xh7 loses to 3 1 ltJe6 + ), so Nunn re
25 h4
••• signed.
A desperate attempt to create tactical
counter-chances. Game 32
Shirov- Nikolenko
USSR Championship,
Moscow 199 1
These annotations were made during the
·
preparation of this book, based on my
notes from lnformator 53.
1 e4
Since 1987 I had played 1 d4 almost ex
clusively and this game marked my return
to 1 e4, which I started playing more and
more often from that moment. My open
ing choice in this game was influenced
26 :e6! by the fact that Nikolenko always used to
The most powerful way. I rejected the play the same system with Black and I de
obvious 26 .:tel ltJg3 + (the only chance) cided to try a new idea of Lanka's.
76 Fire on Board
16 t2Jxf8 .:.xf8 17 l2Jg5 ! is clearly better for Nikolenko is also trying to be as active
White. as possible and he doesn't mind sacrific
14 e6! ing back. Besides, 17 . . . l2Jxd5? seems to lose
As usual, having sacrificed a pawn one by force to 18 l2Jg5 + rt;g6 (or 18 . . . rt;g8 19
must be very energetic. 14 'ii'd3 g6 15 e6 'ii'h 5 h6 20 'ii'f 7 + rt;h8 2 1 'ii'xb 7 hxg5 22
f6! leads to nothing. i.. xc6 l!b8 23 'ii'f7 i.. xh2 + 24 rt;h 1) 19
l!xf5 ! ! 'it>xf5 20 i.. d3 + rt;f6 2 1 'ii'f3 + l2Jf4
22 i.xf4 �e7 23 i.xd6 + 'ii'xd6 24 'ti'f7 +
�d8 25 'ii'xb7, whereas the text keeps ten
sion.
18 rt;xh2 iVxd5 19 c4!
Another pawn sac, this time to get
Black's knight misplaced and unpro
tected. This move demanded very precise
calculation.
19 t2Jxc4 20 t2Jg5 + rt;g6
.•.
Now White has to continue 22 iYh5 ! h6 23 for the piece) 28 . . . 'tWg3 + 29 <it>g1 iYe1 + ; so
dxe5 hxg5 ! (only so, both 23 . . . �xe5? 24 White's best try is
�f3 ! and 23 .. .'jWxb5 24 'tWf7 + <it>h8 25 'iVg6 b) 23 'ii'g3 ! 'iVxfl 24 tt:le4 + <it>f7 25 �h6!
hxg5 26 iYh5 + <&t>g8 2 7 'iif7 + <it>h8 28
1:xg5 just lose) 24 b3 !
'i'e8 + <it>h7 30 iYh5 +) 27 . . Jhe8 28 iYxe8 + By now there was no defence, for exam
�h7, as White has nothing better than 29 ple 25 . . . <it>e8 26 'ii'xg7 :f7 27 'iVg8 + <&t>e7 28
'flh5 + with perpetual check. �g5 + <it>d6 29 iYxf7 iYxa1 30 'tWxb 7 is cur
2 1 �xc4 iYxc4 22 iYf3! :fs? tains.
Finally B lack makes a big mistake. It 26 iYc7 + <&t>g6 27 dxc6 �c8 28 �f4 +
was also bad to play 22 . . . �xd4? 23 iYxb7 <it>f6
'i'xfl 24 iYf7 + <it>h6 25 �e6 + with mate to If 28 . . . �g5 then 29 iYe7 + �g4 30 �d5
follow, but the alternative knight move, f4 31 �xf4 ends the game.
22 . . . �e7! , would surprisingly promise 29 'ti'd6 + �f7 30 iYd5 + �f8 31 �e3
Black good chances of resistance: 1-0
a) 23 iYxb7?! 'ii'xfl 24 'tWb6 + �h5 25 Black resigned as he is mated after
lLle6 :a6 ! 26 4Jxg7 + �g4 27 'ii'b 7 (or 2 7 3 1 . . . 'tWxa1 32 �c5 + .
'i'c7 iYf2 ! 2 8 tt:lxe8 iYh4 + ) 2 7 . . .'it'd3 ! 2 8 A very complex game which actually co
4Jxe8 (28 'iVf3 + ?! 'iVxf3 2 9 gxf3 + <it>xf3 30 incided with Mikhail Tal 's last birthday
tt:lxe8 f4 gives Black excellent compensation (he was fifty-five) and he was also playing
78 Fire on Board
in the tournament. Shortly before Niko by Glek) doesn't seem to equalize either,
lenko resigned I said to Tal that the game in view of 1 7 i.. xa6 bxa6 ( 1 7 . . . ltJxg2? 18
had been my modest present to him. It �xg2 'iixh3 + 19 �g1 bxa6 20 lDh2 wins;
seemed to me that he was happy with 17 . . . 'ifxa6 18 i.. xf4 exf4 19 liJd5 l!b8 20
this. lDxc7 �c6 21 'iid6 ! i.. xb2 22 l!ab1 'iixd6
[forced] 23 cxd6 i.g7 24 e5! wins for
Game 33 White) 18 i.. xf4 exf4 19 t2Jd5 i.. e 5 ! 20 'iia4
(intending l!ad1 and ifa5) 20 . . . l!b8 2 1
Shirov- Kovalev l!ad1 l!xb2?! 2 2 liJd4! 'ii'e8 2 3 'iia 3, when
Bundesliga 199 1 White is clearly better. Mter 16 . . . c6 17
i.. xa6 bxa6, Kovalev's suggestion of 18
These annotations were made during the 'iid6!? is interesting since 18 ..txf4 exf4 19
preparation of this book, based on my ltJd4 'ifc4! looks unclear. It is probably
notes in Informator 53. best for Black is to stick with wait and see
As I have mentioned, in my first season tactics such as 16 . . . �h8 or 16 . . . �h7, both
everything went my way. From these of which were tested in tournament prac
games, I rate this one, with its mutual tice after this game.
tension, hard calculation and creativity, 17 �xg2 'iixh3 + 18 �g1 i.. g4 19 i.f1 !
the highest. It is very important to force Black's
1 d4 liJf6 2 c4 g6 3lDc3 i.. g7 4 e4 d6 5 queen back a little, since after 19 i.. e 2 f5
liJf3 0 -0 6 i.. e2 e5 7 0 -0 ltJa6 8 i.e3 ltJg4 his initiative could have become very dan
9 i.. g5 'iie8 10 dxe5 dxe5 1 1 h3 h6 gerous.
Nowadays 1 1 . . . ltJf6!? is often played, 19 'iih5 20 i.e2!
•••
keeping the h6 square open for the bishop Just here! Mter 20 i.g2? �h8, intend
and not weakening the kingside pawns. ing . . . f7-f5 , Black would have a great
game. Now White is aiming to exchange
some pieces.
dered horribly against Igor Glek (Moscow Although objectively this might be the
1991) with 16 �a4??, and after 16 . . . ltJxg2 strongest move, it allows White to get the
I could have resigned. In spite of a long better game without real effort. From the
fight I couldn't save that game. practical point of view 20 . . . f5 was inter
16 ... ltJxg2!? esting: 2 1 lDxe5! (21 lDh2? simply allows
Now this is just a dubious though inter Black to get a third pawn for the piece af
esting sacrifice. 16 . . . g5 (recommended ter 2 1 . . . i.. xe2 22 'iVxe2 'iVxe2 23 l!xe2 f4)
Shirov - Kovalev, Bundesliga 1 991 79
21 ... i.xe2 22 lbxe2 f4! (22 .. .i.xe5 23 'iVd5 + Kovalev does his utmost to keep the
�h7 24 iVxe5 l:Ife8 25 'iVh2 wins for White), fires burning. 24 ...fxe4 25 �hl ! would not
and now White has to find the precise vari have given him any serious chances, for
ation 23 lbxf4! (23 i.xf4 :ad8! 24 'iib3 + example:
�h7 is unclear) 23 ...iVxe5 (23 ... iVg5 + 24 a) 25 . ..l:.xd2 26 i.. xd2 lbxc5 (26 ... :xf2
'i'g4 ! i.xe5 25 'iixg5 hxg5 26 lbxg6 i.. xb2 27 i..e 3 wins) 27 i.. e3 lbd3 28 :gl! lbxf2 +
27 :abl i.. c 3 28 l;Iec l is winning for 29 i.. xf2 :xf2 30 'iVd7! ! �h8 (30 . . .�h7 3 1
White) 24 lbxg6 'iix e4! 25 lbxf8 l:Ixf8 26 · i.. f5 wins) 3 1 l:Iafl with a won position; or
'ii'h 5 ! (intending i.. h 6) 26 . . . :f6 (26 ... l:If5 b) 25 ... lbxc5 26 i.. xc5 :xd2 2 7 'iVxd2
2 7 �hl ! is clearly better for White) 2 7 iVxg4 28 l:Ie3! l:If3 (28 ...l:If5 29 'iVd8 + �h7
l:ad l ! :g6 + 2 8 �fl and Black's position 30 l:Ig3 wins) 29 �d5 + �h8 30 'iVxe4! and
is in disarray. again White is winning.
21 lbd2 f5 22 exf5! 25 i.. g5!
This seems to open files for Black, but The decisive move; the rest is a matter
in fact White now gets control of the very of technique. It was not too late to err, e.g.
important e4 point. 22 i.. xg4? fxg4, with 25 lbg5? f4! or 25 �fl !? hxg4 (forced) 26
the idea of 22 . . .:f3, is perfectly acceptable 'iib 3 + 'iif7 2 7 'iixf7 + l:Ixf7 28 lbg5 :rn
for Black. (intending ...:g6, . . . f5-f4) , when in both
22 gxf5
••• cases matters would have been far from
Forced. clear.
23 i.. xg4 iYg6! 25 fxe4
•.•
After long thought Kovalev finds a 25 ...l:Ixd2 loses to 26 i.. xf5 ! iYxf5 2 7
move which greatly complicates White's 'ifxd2 'iVg4 + 2 8 lbg3 h 4 29 :e4.
task. 23 .. .fxg4 24 lbce4 would have been a 26 i..xd8 :xd8
lot easier, for example 24 . ..l;If3 25 lbg3 ! Black cannot create any dangerous
'ii' h3 (25 ... 'iVh4? 26 lbxf3 wins) 26 'iVb3 + threats as he is too short of material, for
�h7 2 7 lbde4! (intending 28 l:Iadl, win instance 26 ...l:.f4 27 l:.xe4! or 26 . ..lbxc5 27
ning) 27 ...lbxc5 28 i.xc5 ! :xb3 29 axb3 b6 lbxe4! l:.f4 28 'iVd5 + ! �h8 29 iYxc5 l:Ixg4+
30 i.. e 3 a5 3 1 :adl and the game is over 30 lbg3 and White wins.
because Black's queen is completely use 27 'iib 3 + �h8 28 lbxe4! 'iixg4 +
less. Of course, 28 . . .hxg4 29 �g2 is even
24 lbce4! worse.
White has two extra pieces so he should 29 'iVg3 'iVf5 30 'iVg5!
aim to eliminate Black's counterplay. Finally forcing Black to exchange the
This could not be achieved by 24 'iVb3 + ? queens.
�h7 2 5 'iVxb7 owing to 2 4. . . e4! 30 'iVxg5 + 3 1 lbxg5
.••
24 h5!
•••
80 Fire on Board
Game 34
Bareyev- Shirov
Hastings 1 99 1/92
These annotations were made in January
1992 and first appeared in Schack.
Before the tournament I thought that
the fight for first place would mainly be
between myself and Evgeny Bareyev, as
happened at Biel 199 1 . However, at the
time of this game both Bareyev and
Simen Agdestein were far ahead while I
was on fifty per cent and had lost all hope With the bishop pair, a lead in develop
of first place. ment and open central lines, Black has
Having won this game I made an at more than adequate compensation for the
tempt to catch the leaders but only fin pawn.
ished in third place. 20 <it>h1?
1 d4 d5 2 c4 c6 3 ttJc3 ttJf6 4 ttJf3 e6 5 A serious mistake. After 20 e5 i.c5+ 2 1
e3 ttJbd7 6 i.d3 dxc4 7 .i.xc4 b5 8 i.d3 �h1 (Sherbakov-Kaidanov, USSR 1988)
i.b7 9 0-0 a6 2l . .. ltJd5 22 ttJxd5 i.xd5 23 b3 :adS 24
9 . . . b4 10 ltJe4 i.e7 is currently more 'iVe2 'iVf7! Black has an excellent position,
popular, but the older text move still has but White is by no means lost.
some life left in it. 20 :Iad8!
..•
[AS - Nowadays everybody plays 9 ... a6 Stronger than 20 . . . i.xf4 2 1 i.xf4 'iVxf4
and one has to wonder why 9 . . .b4 is al 22 ttJd4 'iVe5 23 ttJf3 or 22 . . . 'iig4 23 1Vf3 ! ,
most forgotten.] when White has chances to save himself.
Shirov - Smejkal, Bundesliga 1991/92 81
White could go two pawns ahead with 22 The most accurate. 3 1 . . . .i.c6 32 l:lcl
�xe6 l:lxd3 23 ltJxc7 ltJxe4 24 ltJxe4 .i.xe4 i.e8 33 b4! is not totally clear.
25 ltJxa6, but after 25 .. . l:lc8! he would face 32 l:le1 .i.g6 33 l:lc1 �f8 34 b4 .i.e4!
an irresistible attack. Only after this move was I sure of win
22 ltJg4 23 ltJce2
••• ning.
No better is 23 .i.gl i.xd4 24 i.xd4 35 l:lc8 + We7 36 l:lc7 + <itxe6 37 l:la7
1:txf4, or 23 ltJd5 .i.xd5 24 exd5 l:lxd5. .i.c2
23 ltJxe3 24 'iixe3
••• A last finesse.
3M :xa6 + Wd5 39 l:la7 i.e5 40 l:la8
i.c3 41 l:lc8 Wd4 42 l:.c7 �d3 43 h4 .i.a4
44 g4 .i.f6!? 45 h5
B 45 g5 hxg5 46 hxg5 i.xg5 4 7 l:lxg7 .i.d2
48 l:lg3 + 'iii>c 2 49 l:lg2 �c3 50 l:lg4 .i.cl 5 1
Wg2 .i.a3 change nothing.
45 �e4 0- 1
•••
Game 35
24 :Xd4I 25 ltJxd4 �b6 26 l:lad1 l:ld8
Shirov- Smejkal
••.
lbxe5 20 dxe5 .i.f5 2 1 l:tad 1 'iVg5 22 l:td7 Now White's attack crashes through.
�e7 (22 . . .�xd7 23 lbxg6 + 'iVxg6 24 'iVxg6) 15 . . . �e6 also fails to solve the problems
23 g3 'iVxf4 24 'iVh8 + �f7 25 'iVf6 + �g8 on account of 16 lbxf5 (but not 16 .i.xe6?
26 :xe7 :xe7 27 gxf4; and .l::txe6 1 7 exf5 l:th6 ! 18 �c1 dxc3 ! when
Shirov - Thorhallsson, Reykjavik 1 992 83
Black is clearly better) 16 . . . �xf5 1 7 l::t xf5 16 . . . i.e6 because of 1 7 �xf5 �xb3 18 �g4
White's attack is very powerful, for exam 'ii'f6 19 tLie7 + �xe7 20 �xe7 :Xe7 2 1 axb3.
ple: But it is precisely here that 2 1 . . . l:ae8!
a) 1 7 . . J�xe4 18 l'hf7 �h8 19 �f3 ! iVh4 gives Black good chances of saving him
20 l':.fl and either self. However, 18 axb3! is stronger than
al) 20 . . . l::t e 3 2 1 'ii'f5 l:e5 22 �f8 ! ! l:xf5 18 'ii'g4, for example 18 . . .l::t e 5 (if 18 . . . �f6
23 �xg7 + �g8 24 l: lxf5 �d5 25 �xd4 then 19 l::tf4! 'ifg6 20 iVh3 ! h5 2 1 l::t afl
'iVh6 (25 . . . 'ii'g4 26 l::txd5 ! cxd5 27 �xd5 with a decisive attack) 19 l::t adl ! iVe8 20
wins for White) 26 1':.7f6 'ii'c l + 27 l:fl fol 'it'g4 l:.xf5 2 1 l::txf5 and White has a clear
lowed by l:6f3-g3; or advantage. I might also mention that
a2) 20 . . . l:el 2 1 l:xg7 ! �xg7 22 �f7 + 16 . . .�f6 1 7 �xf5 yields White a strong at
<it>h8 23 �c5 ! ! dxc3 24 �f2 l':.xfl + 25 �xfl tack as well.
'i'g4 26 �f6 + �g7 2 7 �d4 is just crush 1 7 .:ad1 'Wh6 + 18 �h1 �b5
ing. The only chance for Black is The alternatives 18 . . . �e6 19 exf5 �xa3
b) 1 7 . . . �h8, but still after 18 'ifg4! (or 19 . . . �d5 20 �xc4 �xc4 2 1 f6 winning)
dxc3 19 .:dl �c8 20 �xf7 l::td8 2 1 l':.dfl he 20 fxe6 fxe6 2 1 .:d 7 .:f8 22 l::t xg7 + �xg7
should feel sad. 23 �e5 + and 18 . . . �a6 19 exf5 ! �xa3 20
However, Black had the strong move �xc4 �e7 (or 20 . . . �f8 2 1 f6 �h6 22 l::t d4)
15 . . . �h4! at his disposal. During the game 2 1 :de l ! �xel 22 f6 ! do not alter the re
I could not see a fully satisfactory reply to sult.
this, but later I found that 16 cxd4 ( 1 6 19 tLI:xf5 �:xf5
lLlxf5 �xf5 1 7 .:xf5 �xe4 i s unclear) On 19 . . . �e6 the simplest is 20 �xc4!
16 . . . �e6 1 7 e5 ! ( 1 7 �xf5 �xf5 18 l::txf5 iVxc4 (or 20 . . . �xc4 2 1 �g5 ! 'ii'e 5 22 �d6
'i'xe4 is again not promising for White) 'ii'f6 23 �e7 + I;lxe7 24 l::txf6) 2 1 tLixg7
17 . . . f4 ( 1 7 . . . l:ad8 18 ttJxf5 �xf5 19 l:xf5 �xg7 22 �g5 + �h8 23 �d6 and White
:xd4 20 �f3 ! is clearly better for White; WillS .
An idea of Alexander Shabalov' s . Only by 18 i.c4 is fatal for Black, the knight on
time will give an objective answer to the c3 must be eliminated before capturing on
assessment of this move. It is clear, though, g5 . However, the strong text move cost my
that the it leads to completely different opponent too much time on the clock. He
types of position to the traditional lines of now had only half an hour left for the re
the Meran. maining 27 moves to the time control.
[AS - The theory of7 g4 has developed a 14 i.d3 i.xc3
great deal since 1992, but there is still Again the best decision. During the game
room for investigation.] I was afraid of 14 . . . b6, but White obtains
7 0-0
••• a big advantage with 1 5 cxd5 ! , for exam
Mter this White obtains a strong initia ple 15 . . . cxd5 16 �b5 i.xd2 + 1 7 l:.xd2, or
tive without making any material conces 15 . . . exd5 16 i.h7 + <ifi>h8 1 7 �e2 ! i.d6 (or
sions. 7 . . ..�)xg4 8 l:.g1 is critical, and 7 . . . h6 17 . . . i.xd2 + 18 l:.xd2 and 19 lt:Jf4) 18 'iVxc6,
should also be considered. or finally 15 ...i.xc3 16 i.xc3 exd5 17 i.h7 +
[AS - Two weeks later, in Oakham, Aka �h8 18 i.f5 .
pian played 7. .ti:Jxg4 against me and lost.
. 15 i.xc3 hxg5 16 hxg5 �e4 1 7 i.xe4
Then in Dortmund (April 1 992) Kasparov dxe4 18 �xe4 l:tf5
came up (after thinking for about twenty
minutes) with 7. . . dxc4 ! ? against Adams
and won quickly. Kasparov said that chess
is an interesting game if moves like 7 g4
are possible!]
8 g5 �h5 9 i.d2 f5!?
On 9 . . . a6 (intending . . . b7-b5) White
forces the play with 10 c5 i.c7 1 1 �e2! fol
lowed by 12 �g3.
As the immediate 13 . . . hxg5 14 hxg5 �e4 Now Black has problems with his
15 �xe4 dxe4 16 'iVxe4 l:.f5 1 7 c5 followed queen. 25 f5 is threatened.
Shirov - Plaskett, Reykjavik 1992 85
24 <itd8
.••
Game 37
Shirov- Plaskett
Reykjavik 1 992
1 d4 e6 2 e4
It's strange that I chose this move since
at that time I didn't employ 1 e4 openings
very much. Perhaps the reason was that I
was not very familiar with Plaskett' s pet
line 1 d4 e6 2 c4 b6.
2 ... d5 3 l2Jc3 �b4 4 e5 c5 5 a3 i.. xc3 +
6 bxc3 'iVa5
Normally 6 . . . l2Je7 is played here.
7 �d2 l2Je7 8 l2Jf3
Possibly I should have tried 8 'iVg4 in
25 d5! order to exploit the early advance of the
Unleashing the second wave of the at black queen.
tack on the black king. Despite serious 8 ...l2Jbc6 9 h4!? f6!
time-trouble my opponent defends him An interesting reaction to White 's
self very ingeniously for a while. pawn advance. Since he is well developed
25 cxd5 26 cxd5 �d7
••• Black immediately tries to open the cen
If 26 . . . exd5 then 2 7 l:xd5 + i.d7 28 f5 tre. Mterwards I was so impressed by this
wmmng. move that I thought that Black was al
27 f5! exf5 28 e5 f4!? ready almost better. Of course, this is an
This gives White more practical prob exaggeration. 9 . . . 'iVa4 10 'i¥b1 c4 1 1 h5 h6
lems than 28 . . Jlc8 29 e6 l:e7 30 'ii'b 4! b6 was what I had expected and 9 . . . �d7 had
3 1 'ii'd6 ! 'ii'xg5 + 32 �b1 , when the threat been played before.
of 33 �b4 l2Jg6 34 l:c1 is irresistible.
29 e6 �a4! 30 .:d2 l:f5 31 'ii'f2 !
3 1 e 7 + ? <it>d7 i s unclear.
3 1 . f3!? 32 .:e4?!
••
1 1. 'iYa4!
•• 15 l2Jxf6 exd4
1 1 . .. e5? 12 c4 is clearly better for White. Now 15 . . . 'ti'xc2 can be met by 16 :c1
12 :b1? 'iVf5 17 'iYxf5 i.xf5 1B dxe5 l2Jxe5 19 i.h6 ! ,
A very weak move which gets White when White is clearly better.
into severe trouble. Instead he should have 16 i.e2 'ifxc2
played 12 i.d3 with good attacking pros 16 . . . dxc3? is quite bad in view of 1 7
pects. i.g5 'ifxc2 1 B 0-0 with a tremendous at
12 c4!
••• tack.
Very strong. I had only counted on 17 :b5 dxc3?
12 . . . a6 13 i.d3 ! c4 ( 1 3 . . . cxd4 14 'ifh5 + Mter this Black gets mated virtually by
�dB 15 'iYf7 and 13 . . . h5 14 'iYf3 are clearly force, but White's attack was already very
better for White) 14 'ifh5 + �dB 1 5 i.xh7 strong. Here are some sample variations:
'iYxc2 ( 1 5 . . . 'ti'xa3 16 'iYf7!) 16 i.xc2 l:.xh5 17 . . . i.e6 1B i.g4! dxc3 19 i.g5 ! a6 (or
1 7 l2Jf3 with a clear advantage or 12 . . . cxd4 19 . . . l2Jd4 20 i.xe6 ! lbxb5 21 0-0 'iYg6 22
1 3 i.b5 'ti'xa3 14 cxd4 'iYd6 15 'ii'h 5 + �dB 'iYxg6 lbxg6 23 l2Jxd5 + �eB 24 :e1) 20
16 l2Jf3 with good compensation. :xb7! (20 :xd5 + !? i.xd5 21 0-0 'iYg6 22
13 'iYh5 + �d8 14 l2Jg4 'iixg6 hxg6 23 lbxd5 is also promising)
20 . . . �cB 2 1 :xe7! 'iYb1 + 22 i.d1 l2Jxe7 23
0-0 h6 24 'iie 2 'iib 6 25 i.e3 'iYc6 26 i.d4
l2Jf5 (26 . . . l2Jg6 2 7 :e1 is clearly better for
B White) 2 7 i.xc3 and the threat of 2B i.a4
is very unpleasant. Of course this is just a
brief analysis of the position after 1 7 :b5
but it illustrates Black's difficulties. Still,
there was no reason for 17 . . . dxc3.
14 e5?
•••
19 . . . b6 20 h6 i.b7 2 1 'iYg7 + 'iYxg7 22 hxg7 19 . . . <it>c7 would have come to the same
l2Jge7 23 gxhB'iV l:.xhB) 20 i.e2 'iixg2 2 1 after 20 i.f4+ <it>b6 2 1 i.e3 + �c7 22 'iff7 +
�d2 ; but 1 6. . .'iYf5 ! 1 7 'iYxf5 tbxf5 1B tbxf6 �dB (or 22 . . . <it>d6 23 i.f4+ l2Je5 24 'iYd5 + )
lbxe3 19 fxe3 �e7 20 l2Jg4 b5 with a slight 23 l2Jd5 ! 'iYb1 + 24 i.d1 i.d7 25 'iYf6 + .
endgame advantage (20 . . . h5 2 1 l2Jh2 e5 is 20 'ti'c5 + �xf6 2 1 'iYg5 + ! <it>f7 22
also a shade better for Black). i.h5 + �e6 23 i.g4+ <it>f7
Tiviakov - Shirov, Oakham 1992 87
26 .1g5 + . ltJce3
25 .1g5 + �f8 26 'ti'd6 + �g8 For the moment White prevents the ad
26 . . . Wf7 27 'ti'f6 + . vance 19 . . .f5 but Black doesn't give up on
27 .1e6 + �g7 28 'iVc7+ 1 -0 this idea.
18 �h8
•••
with 27 . . . l2Jf4 (27 .. .'ti'g5 28 l::tdxd5 .:xf5 29 favourite games. I really enjoy this sort of
'i'd1 yields White sufficient compensation fight.
for the exchange) 28 .:Xd6 'ti'g5 (28 .. .'iVxd6?
is just bad in view of 29 'iVxf4 ! ; while Game 39
28 . . . �xf3 29 l::t xd8 l::txd8 30 gxf3 l::td 5 ! 3 1
l:xa6 l::txa6 3 2 �xa6 l::td 1 + 3 3 �fl l2Je2 + Zapata- Shirov
34 �g2 l2Jf4+ ends up with a repetition of Manila Olympiad 1 992
moves) 29 "ii'e3 'ti'xf5 30 f3 l2Jh3 + 3 1 �h 1
(31 gxh3? �xf3 32 "ii'xe5 + 'ifxe5 33 l:Ixe5 These annotations were made in June
l:g7 + 34 �f2 �g2 + is clearly better for 1992 and first appeared in Schack .
Black) 3 1 . . . �xf3 32 'ti'xf3 l2Jf2 + 33 �g1 Although I played several interesting
ti:Jh3 + , etc. games in Manila, somehow this one was
27 l2Jf4 28 'ti'h6 l:If6 29 'ti'g5
•.• the most memorable for me. I would also
like to mention that this Olympiad was
one of the most pleasant events I have
ever played in, since it was excellently or
B ganized and there was very little dirty
chess politics around - in complete con
trast to the 1994 Moscow Olympiad. If the
Olympiad is well organized it provides a
unique opportunity to see the chess world
as a whole since you meet so many people
there.
1 e4 c5 2 l2Jf3 e6 3 d4 cxd4 4 4Jxd4
4Jf6 5 l2Jc3 l2Jc6 6 l2Jdb5 d6 7 �f4 e5 8
�g5 a6 9 4Ja3 b5 10 4Jd5 i.e7 1 1 i.xf6
�xf6 12 c3 �b7 13 lbc2 l2Jb8!? (D)
Against Tiviakov at Oakham 1992 I
29 i.xg2 !
••• played 13 . . . 0-0 14 �e2 4Jb8, but White
Here it comes ! See the note to White's could have exploited this delay with 14
20th move. a4! bxa4 15 l:.xa4. I was inspired to try the
30 �xg2 l:Ig7 3 1 'ifxg7 + move in the game by Krasenkov' s idea of
White's position is already hopeless, for continuing, after 13 . . . l2Jb8 !? 14 l2Jce3 l2Jd7
instance 3 1 'ti'xe5 l2Jh3 + 32 �fl (or 32 15 4Jf5 , with 15 . . . 0-0 ! , which promises
�h1 l2Jxf2 + 33 �g1 l2Jxd 1) 32 . . . %:.xf2 + 33 Black good play after both 16 l2Jxd6 �xd5
<it>e1 'ti'xh4 ; and 3 1 l:.xe5 l:.xg5 32 l:.xg5 1 7 exd5 l2Jb6 18 l2Je4 4Jxd5 and 1 7 'ti'xd5
lLle2+ 33 �fl lbxc3 34 l::td3 l::txf2 + . l2Jb6 18 'ti'd3 g6! But the Colombian's next
3 1 ... �xg7 32 .l:.xe5 'ticS! moves cast doubt on the text.
A very precise move. Soon White will [AS - The theory of 1 3. . . l2Jb8 is still de
have to drop another piece. veloping. Nowadays 1 4 c4 and 1 4 g3 are
33 l::te 7+ critical and quite popular.]
Also losing are 33 l:Ig5 + �h6 34 lLlf3 14 a4 bxa4 15 l2Jce3!
'i'xc3 and 33 l::txd6 'ti'g4. Very strong! White plans the set-up
33 l:.f7 34 l:.xf7+ �xf7 35 i.d5 +
••• 'iVa4 and l:.d1, after which his pieces co
Neither 35 l:Ixd6 'ti'g4 36 l:.xa6 l2Jh3 + operate harmoniously.
nor 35 h3 'ti'xc3 36 l::txd6 l2Je2 + 37 �fl [AS - In a later game Kasparov played
'i'a1 + 38 �xe2 'ti'e5 + can save White. 1 5 �a4 against me (Horgen 1 994) and
35 ...�f8 36 l2Jg2 'iig4 37 l:Id2 'iig5 0-1 although not everything was clear in that
White resigned as it's all over. From game, this does look more promising than
an artistic point of view this is one of my the text.]
90 Fire on Board
27 'iVc8!
.••
b4!, with a clear advantage. Black must and the forthcoming manoeuvre (3l .. "'J.c4, .
Kramnik - Shirov, Bundesliga 1992/93 91
42 h4 43 b6 h3 44 b7 h2 45 b8'tW
••• A new idea, whereby Black sacrifices a
h1iY 46 'tWe8 'iVe1 + 47 �b2 'tWxe4 0-1 second pawn in order to turn his lead in
development to account. I owe special
Game 40 thanks here to my team-mate Karsten
Muller, with whom I made a thorough
Kramnik- Shirov check of the critical variations.
Bundesliga 1992/93 13 lbxe7+
The only plausible move. 13 �xb4 axb4
These annotations were made in Decem 14 lbxb4 lbd7! is not playable.
ber 1992 and first appeared in Schack . 13 �h8 14 hd8 +
•..
Although in 1992 Kramnik was just 1 7 The only viable alternative, 14 lbd5 , is
years of age, his play, especially with of interest from the theoretical stand
White, was already terrifying. No wonder point. Black has two replies:
that I was preparing for this game for sev a) 14 . . . lbc2 + 15 �f2 lbxa3 16 bxa3 b5
eral days and fortunately it wasn't in 1 7 lbh3 bxc4 18 �xc4 l:Iac8, and now nei
vain. The novelty I invented was recog ther 19 �b3 a4 20 �xa4 tbxd5 2 1 exd5
nized to be the best one in Informator 55. �xd5 nor 19 lbb6 l:Ixd1 20 l:Ixd1 l:Ixc4 2 1
1 d4 tbf6 2 c4 g6 3 tbc3 �g7 4 e4 d6 5 l:.d8+ lbg8 22 lbxc4 �xc4 promises White
f3 0 -0 6 i.e3 c5 7 dxc5 dxc5 8 'tWxd8 + much, although the second of these two
�xd8 9 i.xc5 tbc6 1 0 �a3 a5 1 1 l:.d1 variations is fairly complicated and needs
�e6 12 tbd5 further testing. Also interesting here is
I was of course very familiar with the Kramnik's suggestion of 16 . . . �f8; after
game Kramnik-Nunn, Manila 1992. Mter 1 7 l:.b1 lbd7 he considers that Black has
92 Fire on Board
sufficient compensation for the two sacri In view of the obvious sequel 18 . . .bxc4
ficed pawns. 19 i.. xc4 �xd5 20 exd5 i.. xd5 2 1 i.. xd5
b) 1 4 . . . b5 !? (this is what I intended to .:Xd5 + 22 �e2 we agreed peace terms.
play) 1 5 i.. xb4 axb4 16 �xb4 .:xd 1 + 1 7
�xd 1 i.. f8 1 8 �d5 .:xa2 ! 1 9 �c2! i.. g7 Game 41
( 19 . . . .:a1 ! ?) 20 �b1 ! , and now Black can,
as he prefers, either play for a draw with Gelfand- Shirov
20 . . . I:.xb2 + 2 1 �xb2 �xe4 + 22 �c2 �f2, Linares 1 993
or maintain the tension by continuing
20 . . . l:.a7!? These annotations were made during the
[AS - There is more theory on 1 4 �5 preparation of this book, based on my
now, but I th ink it's all available on data notes in Informator 57.
bases and I suggest that the reader draws My first Linares tournament started
his own conclusions about recent develop successfully enough, but then I lost two
ments.] games and dropped to fifty per cent. I also
14 .:xd8 15 �d5
••• drew a completely winning game against
Analogous to 13 i.. xb4 above, 15 i.. xb4 Anand in round 6 and before the present
axb4 16 �d5 .:a8! 1 7 �xb4 �d7! is also game I was no longer thinking about a
bad for White. good performance. The game appeared to
be a breakthrough. I won it (as the reader
will see, not without luck) , jumped to plus
one and then made 3� points from my
B last five games. I was also satisfied with
the high quality of my play.
At the end of the tournament I realized
that, when on form, I needn't be scared of
anybody.
1 c4 e6 2 �f3 d5 3 g3 c6 4 b3!? a5!?
Formally a novelty but the idea of this
pawn advance is quite well known. How
ever, my opponent's reply came as a sur
prise.
5 i..b2 a4 6 i. g2
6 bxa4 �d7 would have yielded Black
15 ... �c2 + 16 �d2 good play for the pawn. The text is rather
I had expected 16 �f2, when I was also provocative . . .
planning to continue 16 . . . �xa3 1 7 bxa3 6 a3!? 7 i.. c3 b5?
•.•
b5. Then after 18 �h3 there is a choice So many pawn moves! However, Gel
between 18 . . . bxc4 19 i.. xc4 .:c8 20 i..b 3 a4 fand's reply immediately sobered me up.
2 1 i.. xa4 �xd5 22 exd5 i.. xd5 23 i..b 3 Correct was 7 . . . �f6, with the idea of a
i.xb3 24 axb3 I:.c2 + 25 �g3 i.e5 + 26 f4 later ... b7-b5.
:c3 + 2 7 �g4 h5 + 28 �h4 i.. f6 + 29 �g5 8 c5!
l:.xb3 with equality, and 18 . . . l:.c8!?, which, Now White has a clear advantage as
though complicated, should not change the a3 pawn should drop off one day. But
the assessment of the position after 1 9 accuracy is always necessary . . .
�hf4 ( 1 9 �xf6? i.. xf6 20 cxb5 i.. d4 + 2 1 8 ...�f6 (D)
�e2 l:.c1 ! ) 19 . . .bxc4 2 0 �xe6 fxe6 2 1 �b6 9 b4
�xe4 + 22 �e3 :c6 23 �xe4 I:.xb6 24 9 d4 �e4 10 i..b4 would have been
i.xc4 l:.c6. slightly more precise, when White clearly
16 ...�xa3 1 7 bxa3 b5 18 �h3 V2-V2 stands better.
Gelfand - Shirov, Linares 1993 93
for, when again White is better. Now White cannot get his pieces into
10 l2Jxc3
••• play in time. Also unsatisfactory was 16
Three moves in a row with the knight: 'ii'd 3? ! h6 1 7 ltJh7 (both 1 7 ltJh3 e5 and 1 7
a 'logical' sequel to the preceding pawn ltJel 'ifd7 1 8 ltJgf3 l:ta4! 1 9 ltJc2 e5 are
play! good for Black) 1 7 . . . i.e7 ! (not 1 7 . . . :xh7?
1 1 ltJxc3? 18 l2Jg5 hxg5 19 .i.xd5 exd5 20 hxg5 .i.xg5
Now Black gets an excellent game. Even 2 1 �g2) 18 lbxd4 'Yi'd7, when Black wins a
here, White could have obtained a slight piece, but 16 'ii'c 2! h6 1 7 tbh3 e5 18 d3 of
edge by 1 1 dxc3 g5 !? 12 e4! (12 'Yi'b3 g4 13 fered better chances of survival.
tbfd2 h5! is unclear) 12 . . . �g7 13 exd5 exd5 16 h6 1 7 tbh3 e5 18 e3
•••
The knight has to give up living. impression in the opening. This time,
though, I was better prepared!
1 d4 l2Jf6 2 c4 g6 3 lDc3 d5 4 cxd5
lDxd5 5 e4 l2Jxc3 6 bxc3 i.. g7 7 i.. c4 c5 8
W ltJe2 lDc6 9 i.. e3 0-0 10 I:tc1 !?
In my view this sharp move is the only
way to fight for an advantage. The main
variations after 10 0-0 promise White no
advantage.
[AS - Nowadays the Polugayevsky vari
ation is not very popular. I think Black has
a couple of ways to reach comfortable
equality.]
1 0 cxd4 1 1 cxd4 'ti'a5 + 12 �f1 i.. d7
•••
22 l2Je7?
•••
Black' s apparent activity is only short The strongest reply. After 13 . . . 4Jc7 14
term. The text move makes it clear that f4 i..d4+ 15 <it>h2 e5 16 f5 4Jf6 17 g4 White
sooner or later the hammer will fall with stands better. Worth considering is 13 ... e5!?
4Jf6 + . 14 iVe2 'ti'a6!? 15 'ti'xa6 bxa6 with unclear
3 7 4Jg3 + 38 <it>dl 4Je4 39 <it>e2 4Jg3 +
••• play, although deactivating the g7 bishop
40 <it>dl 4Je4 4 1 i.. b 6 l:tf3 42 i.. d 8! llfl + like this would not suit everyone.
43 �c2 lU2 + 44 �b3 i..c5 45 4Jf6+ ttJxf6 14 f4 4Jf7!
46 l:tg7 + <it>t"8 47 gxf6 .l:U'3 + 48 <it>c2 l:.f2 + I had not really examined this move
49 <it>dl l:Ul + 50 �e2 :t'2 + 5 1 <it>el lha2 with Kotronias. It was becoming clear
52 i.. e 7+ i.. xe7 53 fxe7+ �e8 54 d6 1-0 that Kasparov was on the right track, but
[AS - This was one of the best endings as a result of my opening preparation he
of my career so far.] was using up a lot of time searching for
the best continuation at every move, and
Game 43 was gradually getting into time-trouble.
Mozny's opponents chose alternative pos
Shirov- Kasparov sibilities:
Linares 1993 a) 14 . . . 4Jd3, and now after 1 5 e5 (not
15 'ti'xd3? c4 + 16 i.. e3 cxd3 17 i.. xb6 axb6
These annotations were made in March with advantage to Black) 15 . . . c4+ 16 �h2
1 993 and first appeared in the German 4Jc7 1 7 4Jce4 White stands well;
magazine Schack 64. b) 14 . . . h6 15 fxe5 l:txfl + 16 'ti'xfl hxg5
In Linares I played my third game 1 7 i.. xg5 iVxb2 18 lld1 ! with a big advan
against the World Champion. Kasparov had tage for White, Mozny-Ankerst, Munich
won the first two, albeit with some luck in 1992.
the second, but this time I no longer felt
as if I was squaring up for an unequal
struggle.
[AS - Hopefully I will return to such
confidence playing against people like
Kasparov as soon as possible . . .]
1 d4 4Jf6 2 c4 g6 3 g3
I normally play 3 4Jc3, but it is usually
advisable to side-step Kasparov's prepa
ration. I also wanted to try out a new idea.
3 i.. g7 4 i.. g2 0-0 5 4Jc3 d6 6 4Jf3
•••
.t• i B i B i B
30 l:txd6
•••
- • �
"-----� �---·: .•
""----�
� - B�Di¥8
- �MB w� 8
�� �
iL.�
D �s �
Ynm� • • /nun/ -
• • �� .
33 'ii'f6?!
•••
On 36 :hl Kasparov (with his flag equalize. The 4 . . . a6 variation has been
hanging) intended 36 ... i.xe4 37 :h7+ 'V/Jig7 seriously investigated and developed by
38 :xg7 + �xg7 39 'V/Jixe6 �f5 + 40 <it>h3 the Moldavian trainer Viacheslav Cheba
%:.d3 + 4 1 <it>h2 :d2 + , when the white king nenko and his pupils, among whom I
cannot escape perpetual check, since for should particularly mention Gavrikov and
example 42 <it>gl? l:.g2 + 43 <it>fl �g3 + 44 Bologan. It is through them that I know
�el :e2 + 45 <it>d l i.c2 + loses the queen. this line, although nowadays it is also
I also looked at 36 :hl, finally rejecting it practised by many other players.
because of 36 . . . 'V/Jig7! 37 i.xd3 l':.xd3 + 38 [AS - The Chebanenko system usually
�g2 .:d2 + 39 �fl (after 39 <it>f3? l:.xb2 40 arises after 1 d4 d5 2 c4 c6 3 �3 �{6 4
l:[xb2 'V/Jic3 + ! ! 4 1 �e4 'ti'xb2 Black is prob �{3 a6. The text move-order allows White
ably winning) 39 . . . J:.d4 ! , when the only to play 5 'V/Jic2!?, which came into practice
problems I can see are White's. in 1 994.]
36 i.xe4 37 'ti'h7 + 'V/Jig7 38 'V/Jixg7 +
••• 5 �f3
�xg7 39 l:.xe4 �f5 + 40 <ifi>f3 <it>f6 Here 5 a4 is interesting.
The time-trouble was over and at first I 5 b5 6 cxd5
•••
thought I had some winning chances, but This doesn't seem very ambitious. Gen
after a few moves it became clear to me erally White plays 6 b3 here.
that the position is a draw. 6 ... cxd5 7 �e5 �bd7 8 f4 e6 9 i.d3
4 1 J:.c4 :d5 42 llf8 + <it>e7 43 liaS i.b7 10 0-0 i.d6 1 1 i.d2 0-0 1 2 i.e1
ttld6 44 llc3 �e4!?
44 :c2 c4 45 llh2 �f5 46 :h7+ �d6 Black has developed comfortably and is
doesn't lead anywhere, since Black has now ready to start fighting for the initia
the emergency exit on c5. tive. Also possible was 1 2 . . . �b6 13 i.h4
44 c4 45 l':.a6 llc5 46 b3 <it>f6 47
••• i.e7 with equal chances.
bxc4 lf2-lf2
In view of the variation 47 . . . �xc4 48
�e4 �d6 + 49 <it>d3 lld5 + 50 �c2 �f5 5 1
%:.d3 J:.c5 + 52 J:.c3 I offered a draw, which
Kasparov accepted.
Game 44
Seirawan- Shirov
Buenos Aires 1993
These annotations were made in May
1993 and first appeared in Revista Inter
nacional de Ajedrez.
During the Buenos Aires tournament I
met my future wife who was demonstrat 13 'ti'b1 !? �df6
ing the games for the public. Possibly it Continuing with my aggressive strat
prevented me from showing my best egy. After 13 . . . f5 Seirawan was planning
chess there but at least I had some excite 14 a4! , which seems to lead to equality af
ments such as this game. ter 14 . . . b4 [AS - 14 . . . bxa4 seems to be
1 d4 d5 2 c4 c6 3 �c3 �f6 4 e3 a6!? equal as well.] 15 �xe4 fxe4 (but not
Normally I play the Semi-Slav Defence 15 . . . dxe4? 16 i.c4 i.d5 1 7 i.xd5 exd5 18
(4 . . . e6, etc . ) , but sometimes I get bored 'ii'a2 with the advantage - Seirawan) 16
employing the same openings over and i.e2 �xe5 1 7 fxe5 l:.xfl + 18 i.xfl i.e7.
over, even if they are good enough to 14 i.h4!? l:.c8?!
100 Fire on Board
After making this move I began calling Mter the game Seirawan suggested 16
myself an idiot, because I saw 15 i.xe4 a4!?, but I don't believe that Black has any
dxe4 16 �g4. Better, perhaps, is 14 .. .'�ia5 problems after 16 . . .bxa4 17 i.xffi lbxf6 18
immediately, and if 15 %Ic1, then 15 .. J:lfc8. l:txa4 'iVb6 19 'W/a2 l':.a8, or even 16 . . .b4 1 7
I must point out that 14 . . . �d2 was bad, lbxe4 l':.xc1 + 1 8 'ii'x c1 lbxe4 1 9 i..xe4 dxe4
because of 15 'ii'c2 (but not 15 i..xh 7 + ?? 20 �c4 'iVd5 21 b3 l::tc8 22 l::ta2 l::tc6 23 l::tc2
�xh7) 15 . . . �xf1 16 i..xh7 �xh7+ (I don't ffi.
see a good defence after 16 . . . �h8 1 7 %Ixfl, 16...�xe4 1 7 �xe4 dxe4 18 �d7
for example 17 . . . g6 18 i.xg6! fxg6 19 'iVxg6
'iVe7 20 'iVh6 + ! �g8 2 1 %If3 or 1 7 . . . b4 18
%If3 ! bxc3 19 %Ih3) 17 i..xd8 �xe3 18 'iVf2 ! ,
with a small advantage ( 18 . . . �xg2? fails
to 19 i.b6 and White is winning) .
[AS - Of course, instead of 1 B ... 'Dxg2 ?
Black can simply play 18. . . i.xe5 1 9 fxe5
(19 dxe5 d4!) CUg4 20 VJI/g3 0xe5 21 i..e 7
%!feB and his chances are just a little
worse.]
15 %Ic 1?!
When I saw this I sighed with relief.
However I should point out that 15 .i.xe4
dxe4 16 �g4 doesn't win, as the post-mor
tem analysis later showed. Mter 16 . . . i..e 7
1 7 i.. xf6 i.. xf6 18 �xf6 + (18 �xe4 i..e 7) Here I spent about 20 or 25 minutes,
18 .. .'iWxffi 19 lbxe4 'iVg6 20 lbg3, 20 . . .'iVxb1 leaving less than half an hour for the fol
2 1 l':.axb1 l':.c2 22 .::. f2 l:tfc8 doesn't equal lowing 22 moves. But Seirawan had only
ize completely because of 23 l':.e1, but 15 minutes and this fact greatly influ
Black can play a curious queen sacrifice: enced my decision. I saw that the position
20 . . . %Ic2 2 1 l':.f2 l':.fc8 ! 22 f5 ! %Ixf2! 23 fxg6 after 18 . . . l:fe8 19 lbc5 i.d5 20 b4 'W/b6
l:xg2 + 24 �fl fxg6 25 e4 l':.cc2 26 'ii'e 1 h5! would be about equal and then started
27 %Ic1 %Ixb2 and it doesn't seem that thinking about a tempting rook sacrifice
White has anything better than 28 l::tb 1 ( 18 . . . 'ii'd2 !?) . At first I found nothing to
l::txa2 29 l':.a1 with equal chances, since af counter 19 �xf8 'W/xe3 + 20 i.. f2 'iVxf4 2 1
ter 28 h4, the continuation 28 . . . g5 ! is very g3, because 2 l . . . 'iVf3 2 2 l:txc8 e3 runs into
strong. the unfortunate 23 'iVxh7 mate, but I con
[AS - Here I made a big mistake in my tinued calculating and finally made up
old analysis. After 28... g5 White wins eas my mind, convinced that in any case I
ily with 29 'Wie3 gxh4 30 'Dxh5 l:g4 31 d5, would not be worse after the sacrifice.
which means that instead of the incorrect Probably I was wrong, since had Seira
queen sacrifice 21 ... 1::tfc8 ?!, Black should wan made the right moves, I would have
try to survive in the worse endgame aris had to find the only way to survive.
ing after 20 ... 'VJ/ixb1 21 1::t axb 1 i::tc2 22 1::tf2 [AS I still like this sacrifice but I am
-
1::tfc8 23 i::te1 :Xf2 24 �xf2 i::tc2+ 25 i::te2 not sure I would have done it, had I seen
i::tc 1 ! ?] more during the game.]
15 'iVa5
••• 1 8 'W/d2?! 19 lbxf8 'W/xe3 + 20 i..f2
•••
'i'd4 + and White cannot win. Simple and decisive. White resigned.
[AS - Another commentary error. In
stead of 29 :XeS, 29 g3! wins since Black Game45
has no perpetual check and ultimately
loses his bishop. The correct continuation S h irov - Lutz
is 21 ...i..xc8! (not 21 . .. 'fixh2+) 22 g3 'Vi!if5! Munich 1993
23 'Dxe6 i..xe6 with sufficient compensa
tion for the exchange]. These annotations were made in May
2 1 . .:xc 1 + 22 'iVxc1 'i'f5!!
•• 1993 and first appeared in the German
This is the key move of the combina magazine Schack 64.
tion which began with 18 . . . 'iVd2. The tournament in Munich which was
run by the late Heinrich Jellissen was one
of the best organized in the world. When
he died it was a shock to learn that he had
severe financial problems. Many chess
players and friends of his who had in
vested big sums in his suspicious business
never saw their money again. It doesn't
feel great that in a way I played for the
money of my colleagues (I didn't invest
anything with Jellissen myself) but this
can happen to anybody.
My best game from the tournament
was the present one. It was also impor
tant for determining the winner of the
tournament. Munich appeared to be the
23 i.e3 strongest event I have ever been clear
23 tLlxe6 would have lost to 23 . . . e3 ! ! 24 first in.
'ii' xe3 'iVd5 25 <ifilfl 'iVh1 + 26 i.g1 (26 �e2 1 e4 c5 2 ttJf3 d6 3 d4 cxd4 4 tLlxd4
'iVxa1) 26 . . . 'iVg2 + 2 7 �e1 i.b4 + 28 <ifild1 ttJf6 5 ttJc3 tLlc6 6 i.g5 e6 7 'iVd2 a6 8
i.f3 + 29 �c 1 i.d2 + ! 30 'iVxd2 'iVfl + 3 1 0-0-0 h6 9 i.e3 i.e7
�c2 .i.e4 + 3 2 Wb3 'iVc4 + 33 �a3 'i'a4 In this well-known position I thought
mate. for a long time. I had frequently played
23 .i. xf8
••• the black side of this variation, but this
Now Black is a little bit better, but my was the first time I had had it on the
best chance was still his time-trouble. board as White. I thought I remembered
[AS - The text is probably stronger than having had most problems against 10 f3,
23.. . <ifilx{B 24 .i.f4!} so this was my choice, although in retro
24 'iVc7 'iVf3 25 'i'f4 'i'e2 26 'iVf2 'iVd3! spect it was not such a good one.
Not 26 . . . 'iVc4? 27 l:.c1 ! 'i'xa2? 28 l:.c7. 10 f3?! tLlxd4!
27 'ii'd2 I had simply forgotten that this move
After 27 l:c1 i.d5 28 'i'd2 ffi! Black keeps leads to equality.
a small advantage. [AS - The mark ' ?!' is too emotive. Of
27 'i'c4 28 b3 'iVd5 29 .:c1 i.d6 30
••• course, 10 f3 is as common as 10 f4.]
'ii' a5 h6 3 1 'i'd8 + <ifilh7 32 'iVd7?? 1 1 i.xd4
A terrible but understandable mistake [AS - Against Alon Greenfeld (Pardu
when his flag was almost ready to fall. bice 1 994) I played 1 1 'iVxd4 and won. The
[AS - After 32 <ifilg2 White would have game Bologan-Lutz, Bundesliga 1994/95, is
retained good drawing chances.] probably the critical example of this line.]
102 Fire on Board
1 1 . e5 12 .i.e3 .i.e6! 13 g4
•• but 22.. . l:txh6 winning back the piece un
Starting the kingside advance with 13 der favourable circumstances. If 23 'ifxh6,
h4 also leads to equality after 13 ...'ifa5 14 then 23 . . . dxc3 24 'ifh8 + .i.f8 25 l:Ih2 .i.xb1
�b1 l:c8 15 ttJd5 �xd2 16 tLlxf6 + gxf6 1 7 26 .i.xb1 cxb2 + 27 Wd2 l:Id8 + 28 We3 l:Ixd1
l:xd2 f5. and Black wins.
13 .. .'�ia5 1 4 a3 b) 2 1 .i.f2 prevents the motif of the
On 14 �b1 , 14 ... l:.c8 is good, for exam above-mentioned variation ( ...l:.xh6 is not
ple 1 5 h4? ! l:Ixc3 ! 16 'ifxc3 �xa2 + 17 Wc 1 possible) , but Black has the initiative, for
d5 ! with an attack for Black. example 21. ...i.a2 22 c3 dxc3 and now:
14 b5!?
••• b1) 23 i.b5 + doesn't work here, because
This tempting advance is at the same there is no white bishop on h6; Black wins
time very committal. The simple 14 ...l:Ic8 with 23 . ..�f8.
is good for Black: 15 .i.d3! (this is better b2) 23 'ife2 (or any other move with
than the two alternatives 15 h4 l:.xc3 1 6 the queen) loses after 23 ....i.xb4!.
�xc3 �xc3 1 7 bxc3 d5 and 15 ttJd5 'iixd2 + b3) 23 bxc3 .i.xb1 24 �b2 (or 24 .i.xb1
16 ltxd2 ttJxd5 1 7 exd5 .i.d7, intending l:Ixc3 + ) 24.. . l:Ixc3 + 25 �xc3 �xc3 + 26
. ...i.g5, when Black is slightly better) �xb1 .i.xb4 and Black should win .
15 . ..d5 ( 1 5 ... b5 16 �b1 ! is good for White 16 .i.xd5 1 7 exd5 l:Ib8 18 �b1
.••
but 1 5 . . . ltxc3 is quite playable: 1 6 'iixc3 Again the best move. 18 .i.c4 would have
'ifxc3 1 7 bxc3 d5 18 g5 ! .i.xa3 + 19 �b1 been bad because of 18 . . . �c7 with the
with an unclear position) 16 ttJxd5 'ifxd2 + double threat of ...'ii'xc4 and . ..bxa3.
1 7 ltxd2 ttJxd5 18 exd5 .i.xd5 19 .i.b5 + !
i.c6 20 .i.xc6 + l:.xc6 with equality.
15 h4 b4
'iixb4 24 c3, and here it is White who is at 'iVxa5? i..xa5 23 b3 %:tb8!) 22... 'iYxc3 23 i..xc3
tacking; %:tf4 24 i..e 2 a5 with counterplay (pointed
c) 19 . . .'iVa4 20 b3 (not 20 axb4? %:txb4 out by Lutz).
21 c3 %:tb7 22 'iYc2 'iVa5 followed by . . . 0-0 22 'iVc7
•••
and . . . %:tfb8, when Black develops an at Very bad for Black would have been
tack on the king) 20 . . . 'iVxa3 2 1 g5 with 22 . . . 'iVxd5 23 'iYb3 'iVc6 24 b5 and White
good play for the pawn. WillS.
18 i.d8?
••• 23 �b2!
To be frank, I hadn't even considered The best move, whereas 23 'iVb3 'iVd7
this move. When I started pondering my 24 i.. h3 'iVe8 is much less promising. And
reply, . . . i.. d8 at first seemed good to me - after 23 'iVxc7 i.. xc7 24 �b2 l':.b8 25 �b3
after all, the queen is protected on a5 and J:ta5 ! 26 c4 4Jd7! Black performs a balanc
Black threatens simply to take on a3. But ing act and manages to hold everything
then I discovered the flaw ... together.
19 axb4 %:txb4 23 l2Jxd5?
•••
Mter 19 . . .'iVxb4 20 iixb4 %:txb4 2 1 i..xa6 The decisive error, after which White's
Black does not have enough for the pawn. victory is not in doubt. Also losing were
20 'iVc3! 23 . . . iid7 24 'ifc6 ! and 23 . . . a5 24 'iVxc7
This move poses Black great problems. i.xc7 25 �b3 l:xb4+ 26 i.. xb4 axb4 2 7
20 0-0
••• i..b 5. But 2 3 . . . 'ifb7 would have made fur
Bad is 20 . . . %:ta4 2 1 'iVxa5 %:txa5 22 i.. c 4, ther resistance possible, even if the posi
and the rook is caught with the moves b2- tion after 24 'ii'b 3 iid7 25 i.. h 3 iie8 26
b4, c2-c3 and i..b 3. And 20 . . . %:tb8 loses a l:a1 l:xa1 27 l:xa1 is obviously advanta
pawn after 2 1 'iVxa5 i.xa5 22 i..xa6. geous for White.
2 1 i.. d2 ! 24 'iVxc7 i..xc7 25 �b3lDb6 26 i..e3 a5
White must turn his attention to the There is nothing else.
queenside. There is nothing to be obtained 27 c3!
on the opposite flank, for example 2 1 g5? A last refinement. The threat of i.xb6
4Jxd5 ! 22 %:txd5 'iVxd5 23 'iVxb4 'iVxf3 and followed by �xa4 decides.
Black is winning. 27 axb4 28 i.xb6 1-0
•••
21. %:ta4
•• And since after 28 . . . %:ta3 + 29 �xb4 %:tfa8
Here 2 1 . . . 4Jxd5 fails to 22 'iVd3. 30 i..xc7 Black is short of both a perpetual
check and two bishops, he resigned.
[AS - A strange way to trap a rook. I
have never had any other experiences like
this.]
Game46
Kotro nias - S h i rov
Chalkidiki 1993
These annotations were made in Septem
ber 1993 and first appeared in the Ger
man magazine Schack 64.
The Chalkidiki tournament was played
practically on the beach and I had a feel
22 b4! ing that my chess was sometimes more
Now Black has problems with his rook. relaxed than it should have been. But I
Weaker would have been 22 b3 (or 22 really like two of my games from that
104 Fire on Board
event and have included them both in this 24 i.. xe4 ltJxe4 25 'iih6 + �g8 26 :g1 + ) 23
book. Perhaps 'beach chess' isn't so bad fxg4, etc.
after all! 13 e5 d5!
1 e4 c5 2 tbf3 d6 3 d4 cxd4 4 tbxd4 This is better than 13 . . . dxe5 14 ..txe5
tbf6 5 tbc3 tbc6 6 i.. g5 e6 7 �d2 a6 8 �a5 15 :h3, when again White is slightly
0-0-0 h6 9 ..tf4 i.. d7 1 0 tbxc6 i.. xc6 1 1 better - Kotronias.
'iVe1 'ii' c7?! 14 :h3 tbd7 15 �b1 tbc5 16 h5 0-0-0!?
Curiously enough I had made the same 16 . . . 0-0 is also possible, but after 17 'iVe3
mistake once before, against Topalov at �h7 Black's king position is none too
Oviedo 1 992. Apparently I didn't learn safe.
anything from this experience. 17 tbe2 i..b 5!?
[AS - At Buenos Aires in 1 994 (the Lyev I wanted to get rid of my 'bad' bishop,
Polugayevsky theme tournament) Iplayed but there was also nothing wrong with
the correct 1 1 . . . i.e7 against Karpov and 17 ... �b8 18 tbd4 i..d 7.
equalized after 12 e5 tbh5 1 3 i..e3 Wic7 14 18 g4!
..te2 g6 (this is what Kotronias showed According to Kotronias, 18 l:Ic3 fails to
me after our game) 1 5 i..xh5 gxh5 1 6 i..f4 18 . . . i.. c4! (18 . . . �b8 is also good) 19 b3 tbe4
0-0-0 1 7 f3 Jl.g5! ? 18 i.xg5 hxg5, etc. The 20 bxc4 i..b4 ! with a very strong attack.
game was drawn.] White could have tried 18 tbd4 i.. xfl 19
'iixfl , but after 19... �b8 Black has no prob
lems.
18 �b8 19 i.. g2 i..xe2 20 '1Wxe2 l:td7!
..•
12 h4?!
Now the mistake can be corrected.
Stronger in my opinion is 12 tbd5 ! ..txd5
13 exd5 l:.c8 14 c3 e5 15 �b1 with a slight At first I had a generous respect for the
advantage. white bishop pair, but then I realized that
12 ... i.. e 7 Black has enough counterplay along the
In the above-mentioned game against c-file, and as a result perhaps even has
Topalov I hit upon the incorrect idea of the better prospects.
12 . . . b5 1 3 tbd5 i.. xd5 14 exd5 e5? ! (better 2 1 :d4
is 14 . . . :c8, but White is slightly better) 15 White plans c3-c4, which is not so good
�b1 i.. e 7 16 i.. c 1, and Topalov defeated immediately on account of . . . d5-d4, but he
me spectacularly after the further moves will not achieve this aim. Preferable was
1 6 . . . 0-0?! 1 7 g4 ! tbxg4?! 18 :g1 h5 19 f3 21 b3, intending i..c 1-b2.
tbffi 20 i.. d3 �h8 2 1 :xg7! �xg7 22 'ti'g3 + 2 1 . l:tc8 22 l:tc3
••
22 .'iWd8 23 a4!?
•• 31 c4? dxc4 32 'ife4 l:tc7 33 l:cl
Weakening the king's position. It was There is nothing better.
better to play 23 i..g3 b5 24 f4, when Black 33 'ifd7!?
•.•
bxa4 l:xc2 31 'ifxc2 l:Ixc2 32 �xc2 'ifxa4 + Or 3 7 bxc4 'ii' xf4 with a won game.
winning) 30 . . . l:xc4 3 1 bxc4 l:Ixc4, and 37 i.. xd2 38 bxc4 i.. xf4 39 i.. e 4
.•.
i.e3 54 i.e4 i.g5 55 i.c6 i.e7 56 i.b5 i.d6 Viktor Korchnoi has often played this
57 i.c6 i.c7 (the first zugzwang) 58 i.b5 (or way, but it was the first time that this
58 i.e4 b5 59 axb5 a4 winning) 58 . . . e4 + move had occurred in my own tournament
59 �e2 i.d6 60 i.c6 �e5 6 1 �e3 i.c5 + 62 practice.
�e2 �f4 63 i.d7 e3 (the second zugzwang) 10 f5 1 1 f3 h6
..•
64 i.e6 (on 64 �e1 there follows 64 . . . �f3, This plan was first tried in Korchnoi
and the threat o£. .i.c3 + forces White to Yurtayev, Manila Olympiad 1992 . Yurtay
play i.c6 + , after which he loses the g4 ev's idea appears dubious, but I would not
pawn) 64 . . .b5 65 axb5 a4 66 i.f5 a3 6 7 like to pass final judgement on it.
i.e6 i.b6 (the third zugzwang) 6 8 �e1 12 lLld3 b6
�f3 (the g4 pawn is again lost) . 12 . . . g5 13 c5 lLlf6 14 :c 1 ! lbg6 15 cxd6
5 1 . .. �b4 cxd6 16 lbb5 ! is clearly better for White,
Now Kotronias sealed, but the rest can as 17lLlc7 llb8 18 i.xa7 is threatened.
be understood without much commentary. 13 b4 g5 14 c5 lbf6
52 i.xg6 �xa4 53 i.f7 b5 54 i.xe6
�a3 55 �d5 a4 56 �c6 b4 57 �b5 b3
58 i.f5 b2 59 i.c2 i.f2 60 �a5 i.e1 + 61
�b5 i.d2 62 i.f5 �b3 63 i.e6+ �c3 64
i.a2 a3 65 �a4 b1 'iV 0- 1
66 i.xb1 is met by 66 . . . �b2.
Game47
S h i rov - G elfa nd
Chalkidiki 1993
The game was annotated in September
1993 and was published in the German
magazine Schack 64.
The system of the Chalkidiki tourna 15 1Ic1!
ment was so strange that I find it hard to An important improvement. Against
explain to readers. I can only mention Yurtayev, Korchnoi played 15 cxd6 cxd6
that finally I had to play Boris in four 16 b5 l2Jg6 1 7lLlb4 lbf4 18 lLlc6 'iYe8, when
games, but since I lost both games to him Black stood well. However, worth consid
in the quadrangular tournament, Boris ering is 15 lLlf2 f4 16 i.d2, reaching a
was bound to be the tournament winner well-known type of position, but with the
independent of the result of the last two moves b2-b4 and . . . b7-b6 inserted. This
games. I won one game and drew the favours White, as it is now easier for him
other, so I could claim a moral victory be to open up the queenside.
cause my total score (there were also 15 lbg6
•••
Adams and Kotronias in the event) was Also good for White are 1 5 . . . i.d7 16
half a point more than Boris'. l2Jf2 and 15 ... g4 16 fxg4!; but 15 ...:f7 comes
I should mention that apart from the into consideration.
strange system, the tournament was well 16 exf5!
organized and for me it was a unique ex Only so! Both 16 lbf2 lbf4 and 16 cxd6
perience of combining tournament chess cxd6 17 exf5 lbe7 18 g4 lbexd5 lead to un
with beach relaxation. clear play.
1 d4 lbf6 2 c4 g6 3 lbc3 i.g7 4 e4 d6 5 16 i.xf5
•••
i.e2 0-0 6 lbf3 e5 7 0-0 lbc6 8 d5 lbe7 9 16 . . . l2Je7 fails to 1 7 g4 l2Jexd5 1 8 l2Jxd5
lbe1 lbd7 10 i.e3 l2Jxd5 19 'iYb3 c6 ( 19 . . . i.b7 20 c6 and
Murey - Shirov, Luxembourg (European Club Cup) 1 993 107
A new move. White's idea can be seen i.xc4 wins for Black) 16 . . . i.f7 17 'iVxe4
in the line 6 . . . i.e7 7 i.h3! g6 8 0-0, intend fxe4 18 t2Jxa8 t2Ja6 with an excellent in
ing 9 f3, with a slight advantage, as Murey itiative for the sacrificed rook. After I
played a very long time ago. replied to 14 t2Jf3 my situation became es
7 i.g5! i.g7 8 'iVd2 0-0 pecially dangerous since I already had
Every opening move was taking me a less than half an hour to reach move 40
lot of time because I could not (and still and in such a complicated position this is
cannot) understand what was going on. not enough. Besides, my opponent contin
The natural 8 . . . h6? is wrong because of 9 ued playing impressively quickly.
i.xf6 'iVxf6 10 t2Jd5 "VJiif7 1 1 t2Jxg6. 14 cxd4 15 t2Jc7 t2Ja6
..•
and played the rest of the game without Black is now a pawn up, but unfortu
inspiration. It's a pity that there were still nately his king is too passive to hope for a
no practical winning chances for myself. Will.
If 24 l2Jd2 then 24 . . . .i.c6, with the idea 40 l:e6 would have been simpler.
of 25 . . . b6, and 24 l2Jxd6 can also be an 40 ...l:g2 + 41 �h3 .:td2 42 lbxd6 .:txd5
swered by 24 . . . .i.c6, intending 25 . . ..:td8! 43 lieS + �g7 44 lbe4 a5 45 �g3 4Jg6
24 .:teS!
••• 46 .:as l2Je7 4 7 �f3 l:b5 4S l2Jd6 .:tb3 +
Very precise. 24 . . . .i.c6 25 l2Jxe2 fxe2 26 49 �e4 l:b4+ 50 �e5 .:tb1 5 1 l2Jc4! .:tb4
l:txf8 + 'it>xf8 27 �f2 b6 28 lbxb6 axb6 29 1/2-%
<&t>xe2 would have been really unpleasant Draw and ... Hamburg lost the knock-out
for Black. match by half a point (had I won it would
25 l::t e 1 .i.c6 26 l2Jd5 have been the opposite result) .
During the game I feared 26 a3 !? b6 27 This was the last game I played for
4Jxb6 axb6 28 b4, but then I found that af Hamburg before switching to Berlin; I
ter 28 . . . .:te3! 29 b5 (29 l2Jxe2 b5 !) 29 ... .i.a8 ! ! would have relished a farewell win. How
Black i s winning! ever, the game itself was very enjoyable.
26 .i.xd5 27 cxd5 %:te3! 2S a3 b5
•••
of 4. Then I recovered completely and These two moves took Kramnik more
won three interesting games. The present than an hour. Later he told me that he
game was important to prove that at that had been calculating 14 . . J:txc3 15 l!xc3
time I was one of the elite players. I was (in my opinion 15 bxc3 is not bad either)
on my way to the 3rd position in the rat 15 . . . h6, but had decided that after 16
ing list which I obtained clearly half a fxe5! l2Jg4 17 i.xe7 l2Jxe3 18 l!xe3 'ii'xe5 19
year later . . . just for a while. i.xf8ltxf8 Black stands worse. Yet is not
1 e4 c5 2 l£lf3 t2Jc6 3 d4 cxd4 4 t2Jxd4 easy for White to prove that he has seri
t2Jf6 5 t2Jc3 d6 ous winning chances in this line.
Like yours truly, Kramnik always plays 15 i.xe6 exf4
the Rauzer with Black. However, the line The only move, since 15 . . . fxe6 16 �h3
he chose in this game was not one that I ltf7 17 f5 is just bad for Black.
had expected. 16 i.xf4
6 i.g5 e6 7 "ii'd2 i.e7 8 0-0-0 0-0 9 f4 For a while I considered 16 �h3?! l!xc3
t2Jxd4 1 0 iVxd4 �a5 1 1 i.c4 i.d7 12 1 7 :xc3 �xg5 , but then Black has no
l!d3! problems.
This move, first played by Shabalov 16 ...fxe6 17 'ifh3 <3;f7
against Inkiov in Gausdal in January 1991, Again the only move. Now it seems that
has not proved popular. White has the advantage, but he has to be
12 ... e5 very energetic to prevent Black from con
[AS - Two months later, in Belgrade solidating.
1 993, Kramnik came up with a new move 18 ltb1
12...':ad8! His games in that event against Not, of course, 18 i.xd6?? l!xc3 19 bxc3
Kotronias and Hodgson, as well as the i.xd6 20 :xd6 �a3+.
game Shirov-Kramnik in Groningen (De 18...�c5?!
cember 1 993) sent White back to the draw Now White's edge is more or less clear.
ing board.] After 18 . . . :c6 I was going to play 19 l2Jd5 !?
exd5 20 exd5 l2Jxd5 21 .i.d2 ! , but I was not
sure about anything in this position.
13 �e3!
A novelty which Vasilios Kotronias had
analysed deeply and shown me at Linares 19 g4!
in 1993, where he was my second. 13 'ii'g 1, I could have won a pawn with 19 i.e3
which was played in Lanka-Kotronias, 'ii'e 5 20 .i.d4 'ii'g5 2 1 i.xf6 i.xf6 22 :xd6,
Gausdal, July 199 1 , is nothing like as but after 22 . . . l!fe8 23 l!fl �e5 Black has
good. sufficient compensation.
13 :ac8 14 i.b3 i.e6
•.. 19 h6
..•
Shirov - Kramnik, Lucerne 1 993 111
26 . . . .:fd8 is not easy to refute. However, Black's only chance was to defend the
my analysis convinced me that after 2 7 position after 38 . . . axb4 39 axb4 l:.c3 40
'iVg3 ! lL!f6 ( 2 7 . . lL!c5?
. 2 8 l:xd8 l:xd8 29 %:tg2 %:tfc4 41 %:te2 %:txc2 42 :Xe6 + �xg5 43
'iVc7 + lL!d7 30 h4 wins for White) 28 .:xd8 .:xg7 + , although White is still clearly on
1:.xd8 29 h4lL!xe4 30 lL!xe4 'i¥xe4 3 1 'iVxg5 top.
'iVd4 (and not 3 l . .J:tg8? 32 l:.fl + �e8 33 39 �xa5 l:txa3 + 40 �b6 %:tac3 4 1 %:tel
1:.dl) 32 'iVg6 + �g8 33 'iVxe6 + �h8 34 'ifel At the time I thought that this was the
Black has no real compensation for the clearest way to win, but in fact there are
pawn. still some snags. Now I would prefer 41
27 'iVg4 �g8 28 h4 %:tf4 l:c7 l:xc7 42 �xc7 %:tc3 + 43 �d7 l:.xc2 44
I must admit to having overlooked this l:Ibl ! and Black's position is hopeless, for
move when playing 24 g5, but White is example 44 . . . �f7 45 g6 + ! �xg6 46 b5 �f7
still better. 47 b6 .:d2 + 48 �c6 %:tc2 + 49 �b5 (the
29 'iVxg5 'iVxg5 30 hxg5 l:.cf8 (D) move that I had missed in my original cal
30 . . lL!xe4
. 3 1lL!xe4 l:.xe4 32 .:d7! is very culations) 49 . . . .:c8 50 b7 followed by 5 1
promising for White, e.g. 32 ... l:e2 33 l:xb7 �c6 and wins.
1:.cxc2 34 g6! and White wins. 4 1 ...�xg5 42 .:xg7+ �f6 43 %:tc7! e5
3 1 b4? (D)
Quite a weak move. After 3 1 b3! lL!xe4 43 . . . %:txc7 44 �xc7 �e7 (or 44 . . . l:c3 + 45
32lL!xe4 .:xe4 33 g6 ! White has excellent �d6) 45 l:e5 l:f5 46 .:e4 is an easy win for
winning chances. White.
31. lL!xe4 32 lL!xe4 l:.xe4 33 a3 �h7?
•• 44 b5 .:Xc7
With this bad move Kramnik offered a I had expected 44 . . . �e6 and was going
draw. Mter 33 . . . g6 ! White would still have to play 45 l:e2 anyway. If then 45 . . . �d5,
retained slight hopes for a win, although 46 l:.d2 + �e6 4 7 l:txc3 %:txc3 48 �a5 is de
a draw would have been the most likely Cisive.
outcome. 45 �xc7 �e6
1 12 Fire on Board
Game 50
Shirov - Ch ernin
Of course 45 . . J1c3 + 46 <itd6 l:txc2 47 Groningen 1993
l:txe5 is lost for Black.
46 l:te2 l:tc3 + 47 �b8 l:.c4 These annotations were made in January
The most logical line ofresistance would 1994 and have been published in various
have been 4 7 . . . Wd5 48 b6 e4 49 b7 e3 50 magazines.
l:th2 (not 50 <ita7? <iifc6! ) 50 ... �c6 5 1 l:th6+ Sacrifices on h5, d5 and f5 in a row and
<itb5 (or 5 1 . . . <itd7 52 Wa7 l:.a3 + 53 <itb6) the rook manoeuvre . . . some people have
52 l:te6 and White wins. expressed the opinion that this is my best
48 b6 e4 49 b7 Wd5 50 l:Ih2 game. I agree it is good but do not con
This is the killer, whereas 50 l':.d2 + ? sider it my finest.
�c6 5 1 l;Ih2 e 3 5 2 l':.h6 + ? Wd5 5 3 �a7? 1 e4 e6 2 d4 d5 3 lbc3 dxe4 4 lbxe4
would have suddenly led to a draw after i.d7
53 . . . e2 54 b8'i¥ l:.a4 + ! 55 �b 7 (or 55 l:.a6 I am always happy to play this system
l:txa6 + 56 �xa6 e1 'if) 55 .. J�b4 + 56 l;Ib6 as White, since Black loses some time
e1'i¥ ! ! (not 56 . . . l:txb6 + 5 7 �xb6 e1'i¥ 58 compared to the Caro-Kann.
'ii'd8 + �c4 59 'ii'd3 + <iifb4 60 'ii'b3 mate) 5 lbf3 i.c6 6 i.d3 lbd7 7 0-0 lbgf6 8
5 7 'ii'd6 + �c4, etc. lbg3
Nowadays 8 lbeg5 is more popular, but
I can see nothing wrong with simple de
velopment.
8 i.e7 9 b3 0-0 10 i.b2 i.xf3?!
•••
If 50 . . . �d4 then 5 1 �a7 l;Ia4 + 52 <itb6 After the game Chernin said he did not
l;Ib4 + 53 <itc6 l:.c4 + 54 �b5 and White like 14 . . . g6 because of 15 d5 ! (note that 15
w1ns. h4 i.g7 16 h5 'ii'a5 is quite okay for Black)
Shirov - Chernin, Groningen 1993 1 13
I had imagined that this was impossi lt:Jg6 31 �b1 'i¥c3 32 l:Id1 (D)
ble, and only now realized that Chernin 32 'i¥f6?!
.•.
would get three pawns for the piece. 32 . . . l:Ig8 would probably have caused
20 d5 cxd5 2 1 �xf6 bxc4? me some problems (I had about eight
Chernin misses his only opportunity to minutes left at this stage) , but during the
equalize. Mter 2 1 . . .lt:Jxffi 22 'i¥xf6 bxc4 23 post-mortem analysis we found that 33
1 14 Fire on Board
<it>h6 38 'iVh4 + , etc.) 36 g4 !? with a clear Curiously enough Judit repeated the
plus. same line against me in a blindfold game
33 l':.d7! in Monaco in 1996. Unable to remember
Now it's all over. our previous encounter properly I contin
33 l::[e 7 34 l1xe7 'iVxe7
••• ued 8 . . . ltJd7 and won again.
Not 34 . . . ltJxe7? 35 'iie4 and White wins 9 h4 e6! 10 h5 d5
immediately.
35 i.. xg6 :gs 36 iYd4 + l:.g7
It is a pity that the game did not finish
with 36 . . . iYg7 37 l::[ h 5. w
37 l::[h 5 e5 1-0
Black resigned, not waiting for the ob
vious 38 :xe5 .
[AS - I especially enjoyed this game be
cause of the way White's rook destroyed
Black's defences.]
Game 51
J. Polga r - S hirov
Linares 1994
1 1 'iVd2
This annotations are based on my notes I was more afraid of 1 1 'iVe2, a direct at
published in lnformator 60. tempt to use White's lead in development.
This was the best tournament in my ca But I think that Black is still solid after
reer (so far) , but I started with lh/3. For the precise 1 l . . . ltJf6 ! ( 1 1 . . . ltJe7?! leaves
tunately it didn't kill my fighting spirit Black's position too passive after 12 h6
and I won the next four games. This one i.f8 13 i..h 3 ltJd7 14 .i.g5 ! ) 12 h6 .i.f8 13
is the last in the series. Despite shaky i.. g5 i..e 7, as 14 i..xf6 i..xf6 1 5 exd5 cxd5
J. Polgar - Shirov, Linares 1994 1 15
(other knight moves might also be inter The bishop must be taken before it can
esting) 1 9 l2Jxe4 dxe4 20 i.h3 'ti'd5 2 1 c4! be sacrificed on g6.
'i'xd2 + 22 �xd2 a5 !? (22 . . . i.xh6 23 �e3 30 l2Jf4 l2Jxb1 31 flxb1 l2Jd3 32 l2Jxd3
f5 24 exf6 �f7 25 i.g4 i.f8 is also okay for White has to exchange another active
Black) 23 bxa5 i.c5 ! 24 flxb 7 0-0 the posi piece as 32 l2Jh3?! fails to 32 . . .'ti'a8! 33 l':.a1
tion should be about equal. lla2 34 flad1 l2Jb2, with a clear advantage
15 i.d3 for Black.
The plan chosen by Judit seems too 32 flxd3
•••
Giving up one, then another pawn, and sequent play can be improved upon.]
not getting anything in return. The right 6 c5 'iVc7
idea was 33 l:thc1 f5 ! 34 exf6 �xf6 35 l:Ia1 ! The other way, 6 ... 'ii'xb3 7 axb3 liJa6 is
with an unclear game. dubious in view of B l:a4!
33 l:Ixc3 34 l:Ia7 l':.c4!
••• 7�f4 'iVc8 8 e3 liJf6 9 'iVa4!
Now Black is winning. White can only Something that I had not analysed, and
hope for time-trouble tricks. actually quite strong.
35 l:Iha1 l:xb4 36 'ii'f4 c5 37 l:Ixe7!? 9 ...liJbd7 10 b4 a6
"VJilxe7 38�f6 'iVc7! A sad necessity. 10 ... ltJh5 1 1 �e5 f6 12
To meet 39. �g5 with 39...f5. �g3 is very unpleasant, while 10...�g4 11
39 :as .:xd4 40 'iVg5 l:Ia4 4 1 l:d6 b5 �xf3 fails to 12 bxc6!
l:aa8 42 l:Ic6!? 1 1 h3
Beautiful but insufficient. Black also [AS - There was no need to spend a
wins after 42 �dB f6 or 42 'i¥h4!? 'iVb7! 43 tempo on this move since after 1 1 .i.e2 .i.e7
�dB f5 44 'iVf6 c4 45 l:Ib6 'iVd7 46 :d6 12 'ii'b3, 12... ltJh5 is not so good due to 13
.l:taxdB 47 l:Ixd7 l:Ixd7. .i.e5, with a slight plus.]
42 'i¥a7!
••• 1 1 ...�e7 12 'i¥b3 0-0 13�e2 i.e4!?
Not 42 ...'i¥xc6?? 43 .i.e7, intending 'iVf6 I spent a long time making up my mind
winning. between this move and 13 ...�dB 14 0-0
43�d8 i.c7 15 i.xc7 'iVxc7 16 a4 e5. In fact this
Offering another poisoned piece. The gives Black a reasonable game, so perhaps
rook sacrifice 43 .l:.a6 would now fail to I should not have rejected it.
43...'i¥xa6 44 .i.e7 f5 45 'i¥f6 l':.xe7 46 'i¥xe7 14 0-0
�a 7 winning. Now Black has no opening problems.
43 f6 44�xf6 c4 45 'iVg4 'iVf7 46 'i¥d4
••• Mter 14 ltJxe4 ltJxe4 15 0-0 �dB he is also
l:Iec8 4 7 'i¥b6 l:Ixc6 48 'i¥xc6 l:Ib8 0- 1 okay.
Game 52
Kra m n ik - Shirov B
After 19 . . . i.a5 (Karpov's suggestion) The last chance for a 'normal' game
Black is quite okay. Now he suddenly faces was 23 . . . h5, although after 24 g3 i.h6 25
big problems. �g2 White is a little better. Anyway, Black
[AS - The mark ' ? ?' is a bit too harsh. would have had good chances of survival
The position after 19. . . e5 is not yet so clear. in this variation, while in the game he im
Besides, 1 9. . . i.a5 does not equalize due to mediately slips into a lost position.
20 ltJe2 with the idea of 20. . . b6 21 ':fell, 24 g3!
when White is on top.] Now, with the queen on c2, this is
20 b6! i.bS strong.
With apologies to the buried rook on 24 iVxh3
•..
Not wanting to die a slow death. Mter 3 1 l::t h3 Black has 3 1 . . . l::t g6, but
[AS - In fact, Black could have tried the text seems decisive. However . . .
21 . . . e4 22 i.e2 ttJh7, with the idea of 3 1. ':e4!!
••
after 32 . . . �e3 + 33 .l:;txe3 .l:;tg6 34 �g3 36 :�xb8 37 fxg4 'VIIi h 2! 38 .l:;tf3 %:txg4
••
as 33 fxg4! (also indicated by Kramnik) The fatal mistake. After 40 b8� �xb8
33. . . :Xg4 34 �g1 :exf4+ 35 �f3 gives 4 1.l:;tf2 White could still have fought for a
White a decisive advantage. All this proves draw.
that I was just lucky in this game. By the 40 .:xg2 4 1 'iVxg2 .l:;txg2 42 %:txg2
•••
time, Kramnik decided that the draw af 4 g3 d5 5 Ji.g2, with unclear prospects,
ter 36 l:.c8 (or 36 .l:;tc7) 36 . . . 'iVg3 ! 37 b8fii would have been more consistent, but
�h2 + 38 'i;g1 .l:;te1+ 39 �xe1 'iVxe1 + 40 White clearly underestimated Black's an
'l;xh2 �h4 + was not enough for him. swer.
A. Sokolou - Shirou, Lyons (French league) 1 994 1 19
4 d5! 5 d3
••• 9 e4!
•..
Taking the pawn is not good, for exam The third pawn sacrifice in a row - and
ple 5 exd5 liJf6 6 i.. c4 (6 dxc6 ltJxc6 7 i.b5 the soundest!
i.d7, intending ...ltJb4) 6...0-0 7 ltJge2 i.g4 10 iff2
8 'iVg3 i.. xe2, intending ... cxd5, with a Again no. This time accepting the pawn
slight advantage. would have given Black too much devel
5 liJf6 6 h3
••• opment and good attacking threats after
I had expected 6 e5 ltJg4 (6 ...liJfd7!? 7 10 dxe4 dxe4 l l liJxe4 ltJxe4 12 'ifxe4 l::te8
h4! h5 ! is also very interesting) 7 h3 liJh6 13 'iVf3 liJd7.
8 g4 and considered 8... f6 to be acceptable 10...exd3 1 1 cxd3 l::t e8
for Black. With the text Andrei tries to Threatening ...d5-d4.
control all the important squares but 12 �d1
somehow he lacks sufficient development. This ugly move is practically the only
As usual I started looking to refute such a stubborn defence. Alternatives would have
provocative set-up, even if this involved been even easier for Black, e.g. 12 d4 c5 !
sacrificing. or 12 i..e3 c5 ! 13 i..xc5 liJc6 14 0-0-0 b6 15
6 e5!
••• i..e3 (or 15 i.. a3 d4) 15 ...l::txe3 16 �xe3 d4,
6...b5!? 7 g4 b4 8 liJd1 i..b 7 was also in with a clear advantage in every case.
teresting, but the other option is simply 12 liJbd7??
•••
is bad for Black because of 18 ... lbc4? 19 Just to make the end more spectacular.
lbef6 + ) 1 7 .i.g5 ! lbce3+ 18 <it>c 1 'ti'a5 19 2 1 . . J�xe4 22 'iVg5 + <it>f8 would do the job
lbxg7 c;t>xg7 20 'li'd2! ? with an approxi as well.
mately equal game. 22 <it>c2?
16 b6! !
.•• White avoids mate after 22 lLlxf2? lbe3 +
Black gets a decisive attack just in 23 c;t>e1 lbc2, but doesn't notice that 22
time. �e1 would have been his last try. How
17 lbh5 .i.a6! 18 'ii' xd4 ever, Black wins easily with 22 . . . lLlxe4 23
The d3 pawn was impossible to defend, i.xa6 lbb4 24 i.e2 lLlc2 + 25 <it>fl lbxa1.
for example 18 .i.g5 f6 or 18 lbxg7 <it>xg7 22 lbb4 + 0- 1
•••
Game 54
S h i rov - Magem
Madrid 1994
These annotations were made in May
1994, but not published until now.
My Catalan compatriot Jordi Magem
has a narrow but well worked-out open
ing repertoire and he is a tough defender
with Black. Fortunately in this game he
stuck to the official theory, not knowing
the latest developments, which I was told
by Bologan in time for this game.
1 e4 c6 2 d4 d5 3 e5 i.f5 4 i.e2 e6 5
lLlf3 c5 6 0-0 lbc6 7 c3 cxd4 8 cxd4
lbge7 9 a3 i.e4 10 lbbd2 lLlf5 1 1 b4
1 1. a5
.•
is 12 . . J�d8 !?, as suggested by Magem af I was pleased when my opponent played
ter the game. this move, as I was more worried about
12 g4! 19 . . . 0-0. Mter 20 'ii'd3 i.b6 ! 2 1 �d2 'ife7
This move, first played by Bologan White does not achieve anything, so he
against Khalifman (Bundesliga 1994), should play 20 'W/d2, with a slight plus.
practically refutes 1l. . . a5. 20 fxe6 fxe6?
12 i.xf3
••• Practically the decisive mistake. Mter
Khalifman tried 1 2 . . . lDfxd4, but after 20 . . . 'W/xe6 2 1 g5 lDb6 22 i.g4 'iVe7 23 'iff3
13 lDxd4 lDxd4 14 lt:Jxe4 dxe4 15 i.e3 0-0 24 h4 White seems better to me, but
lbxe2 + 16 'ifxe2 'ifd3 1 7 'ifb2, White's in the position is still full of fight. I guess
itiative became extremely dangerous. that Magem must have overlooked my
13 lt:Jxf3 lDh4 14 lt:Jxh4 next move.
The best. Now Black's queen on h4 pre 2 1 b6!
vents him from developing normally. I played this without hesitation. Magem
14 'ifxh4 15 b5
••• had just 15 minutes to reach move 40 and
must protect his king against all White's
pieces. I don't think such a defence is pos
sible.
15 ...lt:Jb8
I had planned to meet 15 . . lt:Je7
. with 16
i.f4, when White is clearly on top.
16 f4 i.e7 2 l . l!f8
••
This takes space away from the queen, This loses by force, but 2 l . .. i.xb6 22
but I can't suggest anything better, for ex i.b5 is very unpleasant for Black since
ample 16 . . . lDd7 1 7 b6! or 16 . . .'1!Vd8 1 7 f5 22 . . . 0-0-0 fails to 23 '1!Vc1 + �b8 24 i.g5.
i.e7 18 fxe6 fxe6 19 i.e3lt:Jd7 20 b6! and 22 'ifc2 l:.xf1 23 l!xf1 i.xb6 24 i.b5
White has a strong attack. g6
1 7 i.e3? What else? 24 . . . '1!Vxa3 25 i.g5 wins.
A silly move, after which the position 25 '1!Vc1 !
becomes much less clear. After the simple The decisive penetration.
17 �g2 lDd7 (or 1 7 . . . i.d8 18 a4!) 18 l':.f3 25 �h4
•••
i.d8 19 a4! Black would not have been left 25 . . . i.d8 loses to 26 i.h6 'W/h4 27 l!f8 +
with anything to hope for. �e7 2 8 i.g5 + .
17 lLJd7 18 �g2
••• 26 h3 i.d8 2 7 a4?!
An alternative idea was 18 g5, but I I didn't see that 27 i.xd7+ �xd7 28
don't see how I could have proved an ad l!f7 + �e8 29 l':.xb 7 wins immediately.
vantage after 18 . . . h6 19 g6 f5. Anyhow, there is no hurry because Black
18 i.d8 19 f5 'W/e7
••• is in zugzwang.
122 Fire on Board
27 .i.e7
.•. 16 l:Ie1 'Wid6 17 'iVxd6 i..xd6 18 i..e3 0-0
After 2 7 . . . l:tb8 White has several ways 19 l:Iad1 i.. e 7 20 i.. xc5 i.. xc5 2 1 ttJxe6
to win, for example 28 .i.h6 i..e 7 29 'Wic7 l:Ifc8 22 h3 :ab8
l::. d8 30 i.. d2! .i.b4 3 1 'iVb6 ! 'Wie7 32 i.. g5, This, Lautier's move, had already been
but the text loses even more quickly. tried several times (for example Nikolic
28 VJI/c7 l::t d8 29 'ii' b 6 1-0 Bareev, Munich 1994) and Black had never
Black resigned as there is no defence to experienced any problems. But Nikolic
30 'iVxe6. had something new in mind.
23 g4 i.. f3 24 l:td2 b4 25 ltJa4 il.a7 26
Game 55 g5 ttJd5
Nikolic - S h irov
Horgen 1994 w
These annotations were made during the
preparation of this book, based on my
analysis in lnformator 61 .
1 d4 d5 2 c4 c6 3 ttJc3 ttJf6 4 e3 e6
With this move order White can answer
4 . . . a6 with 5 'iYc2 !?
5 ltJf3 ltJbd7 6 il.d3 dxc4 7 i.. xc4 b5 8
il.d3 il.b7
I spent around 15 minutes on this move.
After 8 . . . a6, 9 e4 c5 10 e5 has caused Black
some trouble recently.
9 0-0 27 g6!
9 a3 is a more popular continuation This natural move is a novelty. Nikolic
nowadays. Bareev saw 27 ltJd4.
9 ... a6 10 e4 c5 1 1 d5 27 h6
..•
In fact I later found the game Mednis This is much stronger than 22 ..txe5
Jansa, Budapest 1978, on my database, in dxe5 23 'iVxe5 f6 24 'ii'e 6 + �g7 with some
which Black continued 1 7 . . . bxa3 18 bxa3 compensation for the pawn minus.
�h5 and lost. So Benjamin's reply is the 22 f6?!
•••
.I B B l::txffi is threatened.
25 �g7
w Bii'B
•••
l':.b6 l::txd4
If 29 . . . ..txe4 then 30 l':.e1 ..txg2 + 3 1
�g1 .:de8 32 l':.xe7 + l':.xe7 3 3 l:.xffi wins.
22 �g4! 30 cxd4 ..txe4 31 l::t e1
Shirov - Yudasin, Moscow 1 994 125
3 1...f5
I would have preferred the game to end
with another queen sacrifice : 3 1 . . . i.xg2 +
32 �g1 i.d8 33 .l:b8 'ii'f3 (after 33 ... 'iic 6
34 .l:xd8 .l:xd8 35 .l:e 7 + <it>f8 36 l:.f7 + <it>e8
37 'ii'xh6, mate is unavoidable) 34 .l:xd8
'ii'xf4 (34 . . . .l:xd8 35 'i/c7 + ) 35 .l:e7 + .
32 'iie5 + i.f6 33 .l:[xf6 1-0
Black resigned in view of 33 ....1:xf6 34 12 i.d3 i.d7!?
Wie7 + . But here is the difference. Four and a
half years ago Leonid had played 12 ... b6.
Game 57 13 0-0 lbc6!?
Even more surprising. This rare move
S h i rov - Yudasin was introduced into practice by Karpov in
Mosco w 1994 his match against Korchnoi in Baguio
1978.
The game was annotated in December 14 i.c2 lbe7 15 l:.ae1
1994 and published in New In Chess and Probably a novelty. 15 .l:fe1 was tried in
other magazines. Rechlis-Portisch, Manila Interzonal 1990,
The playing conditions and especially but it is not easy to evaluate whether the
the meals at the Moscow Olympiad were rook stands better on a1 than fl . Also in
so bad that in the opening rounds I was teresting is 15 �d 1 !? which, according to
completely slow-witted. Fortunately a loss Yudasin, leads to an unclear position after
in round five shook me out of my torpor 15 . . . 'i/a5 16 'i/g4 g6 17 �g5 �d8 ! 18 'iih 6
and I managed to win four games with ltJg8!
five draws in the latter rounds, although 15 b6
•••
this was insufficient to avoid losing 12 Another possibility was 15 ... 0-0, when I
FIDE rating points. intended 16 �d1 'iia5 17 l:.e3!?
My best achievements in the Olympiad 16 iYd1 bxc5 17 dxc5 iia5 18 'ii'd4
were my games against Lalic (a draw af Nothing good comes from 18 tLixd5?
ter both players missed opportunities to exd5 19 'ii'xd5 i.e6 (Rechlis-Portisch, but
win - see the Foreword by Jon Speelman) with the white rooks on a1 and e1) 20 'iig5
and the present game, which I hope read tiJg6 with the idea of 21 f4 0-0.
ers will find amusing. 18 0-0! 19 .l:e3
•••
33 !tf8
••• I didn't waste a second on this move. In
33 ... 'ii'd 6 34 cxd8'iV+ 'ii'xd8 35 ltJc8 serious chess I would also probably have
i.xc8 36 b8� is hopeless. considered 12 . . . ltJxe4!? 13 :xe4 d5 . Mter
34 bS'iV f4 35 �el 'ii'd6 36 c8'ii' 1-0 14 l:.e2 i.. xb1 15 ltJxb1 ltJxb4 Black might
Completing the picture. Black resigned even be slightly better, so in the post-mor
in view of 36...'ii'xb8 37 'ii'xe6 +. tem Ivanchuk suggested 14 ltJe5 ! ltJxe5 15
:xe5 i.. xe5 16 dxe5 i..xb1 1 7 ltJxb1 e6 18
Game 58 ltJd2. At the time we decided that White
has an advantage, but now I think that af
lva nc h uk - S h i rov ter 18 . . . c5 ! 19 bxc5 'iVa5 Black gets excel
Monaco (rapidplay) 1995 lent counterplay.
13 h3 i..xf3 14 i.. xf3 e5
This annotations were made in May 1995, 14 . . . ltJd7 15 e5 ! ltJa7 16 e6 ! fxe6 17 i..g4
and published in the tournament book. is in White's favour.
Usually rapid chess games are not worth 15 d5 ltJd4 1 6 i.e3 ltJd7 (D)
publishing. However, sometimes they can 16 . . . ltJxf3 + 1 7 'ii'xf3 ltJd7 18 'ii'e 2 f5 19
be very exciting with many interesting ttJxb5 f4 20 i.. d2 looks insufficient for
ideas . The quality and precision may go Black.
down, but the entertainment value often 1 7 i.. g2
increases. Hopefully, this is just such a During the game I was quite afraid of
game. 1 7 i..e 2 f5 (I see nothing better, because
1 tiJf3 tiJf6 2 c4 g6 3 g3 i.. g7 4 i.. g2 1 7 . . . :es 18 i..fl f5 19 exf5 ttJxf5 20 ttJxb5
0-0 5 0-0 ltJc6 6 ltJc3 d6 7 d4 a6 yields White a pawn for nothing) 18 exf5
This position had already occurred in a ltJxf5 19 .i.xb5 e4 20 i.. x d7 or 18 . . . gxf5 19
game between the same opponents only i.. xd4 exd4 20 ltJxb5, when in both cases
one month previously (Linares 1995). On White is clearly better. However, after
128 Fire on Board
and I should have thought for longer than Also losing is 31. . . 'iVxe4 32 l:t.xe4 ttJd3 33
I did. :c2 c5 34 bxc5 b4 (or 34 . . . ttJxc5 35 l:t.b4)
23 dxc3?
••• 35 c6 b3 36 l:t.xc3 i.xc3 3 7 c7. Since in
Natural and bad. Correct was 23 . . . c6 ! Monaco they use the Fischer clock (which
24 i.xc6 (24 ltJe4 'iVf5 seems insufficient obviously improves the quality of rapid
Sian Castro - Shirou, Leon 1995 129
games) , the rest of the game was not too This is the critical position. Some time
difficult for Vasily. ago there was the game A. Sokolov-Salov,
32 <itg2 �c4 33 'ii'e6 + 'ti'xe6 34 l:xe6 Nikolaev 1983, which White won beauti
�f7 35 l:c6 i.. e5 36 �f3 �d2 + fully after 15 . . . exf5 16 g6 hxg6 1 7 l:xg6
36 . . . i.. d 6 would have offered more re �e5 18 l:xg7 + <it'xg7 19 l:g1 + �g6 20
sistance without changing the final re exf5 l:h8 21 i.. d4 + i.. f6 22 fxg6 fxg6 23
sult. 'ti'g4 :h6 24 i..xf6 + <it'h7 25 l:e1, etc. Even
37 <ite3 �c4 + 38 <ite4 i.. xg3 39 l:xc3 more convincing was the game Sion Cas
i.. d6 40 �d5 i.. xb 4 41 .:f3 + �e8 42 lhg6 tro-Rivera, Cordoba 1991, which contin
\td7 43 l:xh6 .i.e7 44 h5 �d6 45 l:txd6 + ued 15 . . . b4 16 'ti'h5 �e5 1 7 l:g3 exf5 18
i.. xd6 46 l:f7+ �e8 47 �e6 b4 48 h6 exf5 bxc3 19 g6 cxb2+ 20 �b1 hxg6 2 1
.i.e5 49 h7 .i.h8 50 l':.f3 1-0 fxg6 �xg6 2 2 'ti'xg6 i.. f6 2 3 'ii'xf6 �e6 24
i.. xe6 fxe6 25 'ti'xe6 + 'ti'f7 26 l:xg7+ 1-0.
Game 59 When I saw this game on my computer I
felt sure that Sion Castro would repeat
Sian Cast ro - Shirov this line in our encounter, although I
Leon 1995 didn't have much time to learn all its se
crets. I recalled that in the mid- 1980s
The game was annotated in May 1995 Latvian IM (now GM) Zigurds Lanka had
and published in New in Chess. found the move 15 . . . l:fc8 !?, which he suc
Before this game I had expected a cessfully introduced in a little-known
Velimirovic Attack because it suits Sion rapidplay game against Andrei Sokolov,
Castro 's sharp style and he had enjoyed but when I checked my computer I discov
some success with it. Even so, I still didn't ered that this move had also been played
manage to prepare properly. in Brunner-Wittmann, Graz 199 1, which
1 e4 c5 2 �f3 d6 3 d4 �f6 4 �c3 had continued 15 . . . l:Ifc8 16 l:g3 �e5 1 7
cxd4 5 �xd4 �c6 6 i.. c 4 e6 7 i.. e3 i.. e 7 l:h3 �g6 18 'ii'h 5 �f8 19 i.. d4 i..xd5 20
8 'ii'e 2 a6 9 i..b 3 'ti'c7 1 0 0-0-0 0-0 1 1 'i!Vh6 i..xg5 + 21 'ii'xg5 f6 22 �h6 + �h8 23
l: hg1 �d7 1 2 g4 �c5 1 3 �f5 b5 1 4 i.. d5 iVxf6 e5 24 'ti'f3 i..e 6 25 i.. xc5 'ii'xc5 26
For those readers who are unfamiliar l:g3 b4 2 7 �f7 + i..xf7 28 'ti'xf7 l:c7 29
with this variation, I should point out ifd5 and it seems that here Black could
that these sacrifices are well-known the have obtained a decisive advantage with
ory and that it has been proved that nei 29 . . . :ac8. Having looked at this I decided
ther piece can be taken. that my 'preparation' was perfect and set
14 .i.b7 15 g5
••• off for the game.
15 ... l:fc8 16 l:g3 �e5
Here I realized that after 17 f4! I would
probably have been in trouble. The fol
lowing analysis doesn't offer Black very
much: 1 7 . . . exf5 (or 1 7 . . . b4 18 i..xc5 dxc5
19 �xe7 + 'V/I/xe7 20 i.. xb7 'V/I/xb7 2 1 fxe5
bxc3 22 l:xc3) 18 i.. xc5 lbg6 (18 . . . dxc5 19
fxe5) 19 i..e 3 b4 20 i..xb7 'V/I/xb7 21 �d5
fxe4 22 f5, with a clear advantage for
White in each case. Perhaps Black can try
17 . . . �g6, but this looks somewhat risky
after 18 �xg7 <it'xg7 19 f5 with a strong
attack. Anyway, I had no time to calculate
all this as Sion Castro quickly played
1 7 i.. xb7
130 Fire on Board
afraid of 18 liJxe7 + ( 1 8 f4 exf5 19 fxe5 This turns the game in Black's favour.
4Jxe4 is fine for Black) 18 ... 'ifxe7 19 f4 liJc4 22 !ld2 looked safer.
20 .i.d4 and White has some dangerous 22 bxc3 23 4Jb6 4Jb5! 24�f6
•••
threats, for example 20 . . . 4Jd7 2 1 :h3! b4 I was very afraid of 24 4Jxc8 !lxc8 25
(or 21 . . .4Jf8 22 f5! exf5 23 4Jd5 with an at 'ifc2 (25 a4 c2 ! offers Black excellent at
tack) 22 'iVh5 4Jf8 23 �xg7! and now: tacking chances), but in the post-mortem
a) 23 . . . �xg7 24 f5 ! f6 (or 24 . . . exf5 25 it became clear that after 25 . . . 4Jc5 ! (not
exf5) 25 'ifh6 + �h8 26 gxf6 'iff7 27 4Je2 25 . . . 4Ja3? 26 !lxc3 ! or 25 . . . d5? 26 a4 ! and
with a strong attack; White is on top in both cases) 26 .i.xc5
b) 23 . . . bxc3 24 �xc3 and again White dxc5 Black probably has the advantage.
has powerful attacking chances. Still, this was better than the text.
However, instead of 19 . . . liJc4 Black can 24...'�c7 25 4Jxc8 !lxc8 26 a4
play 19 . . . 4Jed7! and White's attack is not
strong enough, e.g. 20 f5 b4 2 1 f6 'iVe8! 22
fxg7 bxc3 23 Vi'h5 cxb2 + 24 �b1 liJxe4 25
!lh3 liJc3 + 26 �xb2 4Jxd 1 + 27 �c 1 �xg7 B
26 e5!?
•••
Game 60 1 1 0-0
It is not necessary to defend the e4
S h i rov - Tim man pawn because 1 1 . . . iixe4? fails to 12 l:Iel.
Bie/ 1995 [AS - A couple of weeks after the game
the Dutch GM Roberto Cifuentes demon
The game was annotated in August 1996 strated to me that 12 1:le1 can simply be
and published in New In Chess. answered by 12 .. /�{8 1 3 .i.b5 'VJ/ig6, when
During my preparation for this game I White has no real compensation for the
had anticipated that Jan would answer 1 pawn. Therefore his only chance would
e4 with 1 . . . e5, but since I had recently had have been 12 ltJd4! ?, but this is still not
a bad experience with it as White I was convincing. I have to admit that 1 1 0-0 is
unsure which variation to choose. On the a complete bluff and the right continu
way to the game I passed a blitz tourna ation would have been 1 1 'V/Iif4 with an un
ment that was taking place on large gar clear game.]
den boards (and in which some strong 1 1... .i.d7 12 liJd4!?
players were competing) . The sight of 12 liJbd2 i.xc3 13 l:tb1 ltJffi didn't ap
these players moving big pieces and then peal to me.
running to press the clock made quite a 12 0-0-0 13 ttJd2 ltJc6
•••
from that point on I applied maximum 25 . . . 'ifxc3 26 ltJxd8 .l:txd8 27 i.b2) 26 i.f4
concentration. 'ii'xc3 (26 . . .'�ie6 fails to 2 7 'ii'b 5) 27 i.xd5
16 b6
••• and wins; or
I had expected 16 . . . l:.dg8, after which b) 23 l:.a1 ltJa5 24 l:xa5 ! bxa5 25 ltJc6
1 7 l:xb7 ltJxd4 18 'it'a6 ltJxe2 + 19 �h 1 i.xc6 26 i.e3! �xe3 (forced) 27 �xc6 +
fails to 19 . . . 'iVe6, when the black king can �b8 28 fxe3 ltd6 29 l:b1 + �c8 30 'ii'b 7+
escape to e7. The correct response is 1 7 �d7 3 1 i.xd5 with a clear edge.
ltJ2f3 with a slight advantage. 23 �xa5! 'ii'xd4!
17 a4! You don't see this kind of mutual queen
This is no time to look back! This pawn sacrifice very often, do you? Of course
advance is objectively the best move. 23 . . . bxa5? loses immediately to 24 ltJxe5
1 7 �b8?!
••• i.xe5 25 ltJc6.
Playing with fire. After 17 . . . l:.dg8 18 a5 24 �a2! �xc3
ltJxa5 1 9 'iVa6 + �d8 20 �xa7 �e8! Black Both 24 . . . 'ii'd3 25 i.e3 and 24 . . . 'iVh4 25
can transfer his king to safety, for exam g3 also lose quickly.
ple 21 ltJb5 i.xb5 22 'i¥b8+ 'ifd8 23 i.xb5 + 25 i.e3!
cJi?e7 24 'it'xd8 + l:xd8 25 i.b2, and White It was not too late to fall into a clever
has only a small advantage. trap: 25 i.b2 'ifxf3 ! ! 26 gxf3 l:hg8 + 2 7
18 a5! ltJxa5 19 'ii'a6 �aS? �h 1 dxc4 2 8 i.xf6? i.b7 and Black wins!
This seems to have been the decisive Although White has now sacrificed three
mistake. It was obligatory to play 19... i.xd4 pawns, he has tremendous threats. Black's
20 cxd4 �a8, although after 21 i.b2!, with monarch is worth more than his queen.
the idea of 2 1 . . . i.c8 22 �b5 !, White keeps 25 i.b7
•••
w
Game 61
Sa lov - S h i rov
A msterdam 1995
The game was annotated in September
1995 and published in New In Chess .
Before this game I was not in a very
ambitious frame of mind: I was Black and
all my five previous Black encounters
with Valery had ended peacefully (despite
various degrees of struggle) .
After 22 . . . ltJxc4 White would have had 1 c4 e5 2 ltJc3 i.b4 3 ltJd5 i.e7
a choice between: When I played this move for the first
a) 23 �xc4! i.c8 (23 . . . i.g4 24 l:.a1 i.xf3 time, against Ehlvest at the 1990 Manila
25 ltJc6 wins) 24 �a4 ! a5 (24 . . . i.b7 25 Interzonal, I lost in twenty-odd moves
ltJb5 a6 26 i.f4 also wins) 25 ltJc6 i.d7 (or and people looked at me as if I needed a
Salov - Shirov, Amsterdam 1995 133
doctor, but nowadays it is perfectly re I saw this move right after playing
spectable . Recently it was christened the 12 . . J::tfd8 and at first I was cursing myself,
'Shirov variation' by New in Chess in one though I found a good reply after some
of their yearbooks, but this is mistaken thought. Instead of 13 i.b5, it was dan
and I would like to put the record straight. gerous for White to play 13 l::tc 1 l::t ac8 14
The move 3 . . . i.e7 was discovered by the 'i¥xc5 'ii'h 4!, while 13 i.d3 h6 14 i.e4 l::tac8
Moldavian trainer Viacheslav Chebanen would have led to an unclear position.
ko; the earliest game in which it appears 13 ttJa5!
..•
i.b5 e6, etc., but with colours reversed. The only move. 14 i.xd7 loses immedi
The game Psakhis-Mek, Hertzliya 1993, ately to 14 . . . 'ii'xe5.
went 6 . . . 0-0 7 l2Jc3 c6 8 d4 d6, but Black 15 i.xb5 16 i.xa5 l2Jf4!?
••.
was unable to equalize completely. I couldn't see anything clear after 16 ... b6
7 cxd5 ltJxd5 8 a3 0-0 9 ltJg3! 17 i.c3 :
In the aforementioned Sicilian line a) 1 7 . . . ltJf4 18 'ti'f5 ! ltJxg2 + (18 . . . ltJd3 +
Black's knight usually finds itself on c6 19 Wfl) 19 Wd1 is also very complicated;
after an exchange of the bishop on b5 for b) 1 7 . . . ltJxc3 18 dxc3 i.d3 19 'ti'b2.
Black's other knight. Here it looks better Since I was afraid of losing my advan
for White to place his knight on g3 rather tage in that line, I decided to sacrifice the
than c3 . exchange. Perhaps I was overestimating
9 c5 10 b3 ltJc6 1 1 i.b2 i.e6 12 'ifc2
••• my chances, but there was already little
l:Ud8!? time to think. When I played 16 . . . ltJf4 I
This move may not be so bad but it does had about twenty minutes left for 24
look a little risky. 'Normal' was 12 . . J::tac8 moves.
to meet 13 i.b5 with 13 . . . f6. 17 i.xd8 l::txd8 18 'ti'c3!
13 i.b5 A surprise. I had been counting on 18
0-0-0? ltJd3 + 19 'itb1 ltJxf2 when Black is
on top. Now 18 . . . l2Jd3 + is silly due to 19
�fl.
18 .. .'ti'e6! (D)
The critical position. Here Salov also
started running short of time. Obviously
in mutual time-pressure it is much more
pleasant to attack than defend and the ex
tra material doesn't count for so much at
the moment. After the game Valery ad
mitted that he had been feeling quite pes
simistic at this stage, and this may help to
explain his next error.
134 Fire on Board
24 -tes 25 tbe4
•..
19 f3?
Now Black is clearly on top, although
he still has to be very precise. Of course
my next few moves were not based on
much calculation and I am very glad that
they were okay. Instead of 19 f3?, White
had to play 19 <it>d1 ! �d3 20 .:n �e5 ! 2 1
:g1 (2 1 f4 i.xfl 2 2 fxe5 i.xg2 is clearly
better for Black. ) 2 l . . . :d3 ! 22 'iVxc5 , al
though after 22 . . . :xb3 ! , with the idea of 28 'iVg3?
23 'iVxa7 h6 ! , Black's attack looks very The last mistake, with his flag hanging.
dangerous. Still, it is very difficult for me I was less sure about the consequences of
to give a deep analysis or clear assess the line 2S 'tWxa7 tbxc3 29 dxc3 'iVxb3 30
ment of this position. l:Ib2 (30 l:.c2? .ta4 3 1 l!b2 'iVxa3 wins)
19 l:.d3 20 'iVxc5 �xg2 + !
••• 30 . . . 'iVxc3 + 3 1 <it>b1, but the post-mortem
This is stronger than 20 . . . :xe3 + 2 1 and my home analysis showed that al
�d 1 ! (not 2 1 'iVxe3? tbxg2 + 22 �f2 t2Jxe3 though after
23 dxe3 'ii'xb3 with a clear edge for Black) a) 3 l . . .�d3 + ?! 32 <it>a1 (but not 32 �a2?
and now: �d5 + 33 �b1 i.d7 and wins) 32 . . . 'iVxf3 33
a) 2 l . . . .:xb3? 22 'iVd4! i.e2 + 23 tbxe2 l:Ie1 f6 (33 . . . :as 34 l:Ie3! 'iVd1 + 35 .:b1
'iVxe2 + 24 �c2 ! tbe6 25 �e4! ifxe4 + �d4 + 36 l!b2 leads to an unclear posi
(25 . . . 'ii'b5 26 .:hb1 :xb1 27 :xb1 'ti'c5 + 2S tion) 34 �e7! l!aS 35 l!e3 'iVh1 + 36 l!b1
<it>b2 is also better for White) 26 fxe4 t2Jd4+ ifxh2 37 :d3 ! White has some counter
2 7 �d1 :b2 2S :c1 h6 29 :n ! and White play, Black can play
has good winning chances; b) 3 l . . . i.c6 ! 32 'iVe7 (the only move)
b) 2 l . . .'YWxb3 + 22 'iVc2 �xg2 23 'iVxb3 32 . . . 'iVd3 + 33 <it>a1 :as 34 'iVb4 (again
l!xb3 24 a4 i.c6 25 l!g1 with an unclear there is nothing else) 34 . . . %!xa3 + 35 :a2
endgame. l!xa2 + 36 <it>xa2 i.d5 + 3 7 <it>b2 'iVxf3 3S
21 <it>d1 .:e1 'iVf2 + 39 �c1 h6 and Black should
Of course not 2 1 <it>f2? l!xd2 + 22 �g1 win sooner or later.
tbh4 and White is mated. 28 t2Jxc3 29 dxc3 'iVxb3
•••
Game 62 15 ltJxe5
I didn't like the position after 15 dxe5?!
S h irov - Piket 'tic5, with the idea of 16 . . . a5 and 17 . . . a4.
A msterdam 1995 Black can annoy White a lot.
15 ltJxe5 16 dxe5
•••
The game was annotated in September It is odd that this exact same position
1995 and published in New In Chess . could have arisen in Bogolyubov-Alek
Sometimes analysing the game gives hine, World Championship match 1929,
even more pleasure than the game itself. after the moves 1 1 . .. e5 12 a3 i.. xc3 13
This is what I felt while uncovering the bxc3 'iVxa3 14 ltJxe5 tbxe5 15 dxe5 l:te8,
mysteries of the variation 23 .. .'ifc5 24 had Bogolyubov played 16 0-0 instead of
i.. g5 ! . I should mention that no playing 16 i.. g3, after which he lost hopelessly.
program will ever be able to suggest 24 The plan I found over the board is more
i.. g5 as the strongest move, although I logical because the bishop is very strong
was still using Fritz in my analysis. on the h4-d8 diagonal.
1 d4 d5 2 c4 c6 3 lLic3 lLif6 4 ltJf3 e6 5 16 'i/Vc5
•••
20 . . . �g4 2 1 'ii'xg4 I:txd3 then 22 e6! is very a) 24 . . .lLic4 is met by 25 'ii'd 7! 'ii'f2 26
good for White. .tlgl 'iiigS 27 �xh6 gxh6 2S 'ii'g4 + <it>fS 29
19 f4�c4 20 !U3 I:txh6 with a decisive attack;
A move with strong psychological over b) 24 . . . 'ii'fS and now the correct way is
tones: no-one likes to see the enemy pieces not the direct
encroaching on their king. b1) 25 f5? lLic4 26 'ii'f4 ltJxe5 27 f6 l:.e6
20...�xd3 2 1 !:txd3 (D) 2S lU1 (2S fxg7 'ii'xg7 29 �xh6 I:txh6 30
I:txh6 + 'ii'xh6 3 1 'ii'xe5 figS! is unclear)
2S . . . I:taeS and Black holds on, for example
29 fxg7 jVxg7 30 �f6 'ti'g6 3 1 !Ig3 !:txf6 32
B 'ii'xf6 'iWxf6 33 :xf6 a5 with excellent com
pensation for the exchange; but
b2) 25 l:.fl ! Just bringing one more
piece into the action. Black is now help
less, for instance 25 . . . �gS (25 . . . l:.e6 26 f5
.tlxe5 27 f6 wins) 26 'ii'e 2! l:.e6 (or 26 . . . hxg5
2 7 'ii' h5 f6 2S fxg5 l:.xe5 29 'iVh 7 + �f7 30
!:txf6 + winning) 27 f5 l:.xe5 2S �f4 !:ta5 29
f6 g5 30 �d6 'ii'xd6 3 1 !Ixh6 and mate is
unavoidable;
c) 24 . . . �gS Clearly the toughest de
fence. Now White should go for a long and
2 1.. .'iVc4? sometimes not even forced line starting
A serious mistake. After 21. .. <it>h7 I don't with the piece sacrifice, 25 �xh6 ! , fol
see anything concrete for White, although lowed by 25 . . . gxh6 26 f5 ! 'iWfS 27 !Ixh6
I still prefer his position. l:Ixe5 2S l':.h5 ! (but not 2S 'ii'g5 + ?! jVg7 29
22 tiel �h7 23 I:th3 I:te6 'ifh4 ltJd5 ! with an unclear position)
This move allows White to win almost 2S . . . f6 29 'ii'a2 + ltJd5 (not 29 . . . �g7? 30
by force. I:te3 and wins or 29 . . . 'iff7? 30 l::t hS + �g7
Correct was 23 . . . 'ii'c5 as the queen is a 31 l:.h7 + and wins) 30 I:te3! �f7 31 :h7 +
good defender on fS. White then contin <it>eS 3 2 l:.eh3 !Ixe4 (or 32 . . . ltJf4 33 l%.3h4
ues 24 �g5 ! !IdS 34 h3 l:.xe4 35 :hs and wins) 33 'iVb1 !
Black has now several possibilities but Initially I finished my analysis at this
it seems to me that he cannot defend him position but later I realized that Black
self, e.g. still has a resource, i.e. 33 . . . lLixc3 34 !Ixc3
Stefansson - Shirov, Clichy (European Club Cup) 1 995 137
'Yi'b4! (not 34 . . Jie7 35 l:Ich3 :xh7 36 :xh7 It would have been more precise to play
and White wins outright) Now White has 33 'Yi'f6 + �c5 34 'Yi'e7 + , winning a piece.
to find the very precise 35 'Yi'dl ! (the rook 33 lt'Jb6 34 h4 �c5 35 h5 a5 36 h6
•••
This loses immediately, but 27 . . . l:If8 I didn't feel like taking the 'poisoned'
wouldn't have changed the result because pawn with 7 . . . i..xd4 because after S 0-0-0
of 2S .:.f5 ! �g7 (2S . . . lt'Jc4 29 'Yi'f2 �g7 30 i..c5 9 'Yi'g3 White has good compensation.
l:.g3 + �h7 31 l:.f6 also wins) 29 l:.xe5 lt'Jc4 8 a4 b4 9 lt'Jce2 d5
30 l:tg3 + �h7 3 1 :xe6 tbxd2 32 l:te7 a5 33 After 9 . . . lt'Jf6?!, 10 h4! (and if 10 . . . h5 1 1
.:.d3 tbb3 34 l:If3 lt'Jc5 35 l:tf5 b6 36 l:Ifxf7 + i..g5) gives White good attacking chances
l:.xf7 37 .:.xf7 + �gS 3S :c7 tbxe4 39 l:.xc6 10 h4!?
b5 40 c4 b4 41 .:.a6. A new move. 10 e5 lt'Jd7 would have
28 l':.xh6 + 'Yi'xh6 29 l':.xf7+ �g6 30 been quite unclear while 10 0-0-0 has
.:.f6 + �xf6 3 1 'Yi'xh6 + �e7 32 'Yi'g7 + been played before and is similar to the
�d6 33 'Yi'xb7 game.
138 Fire on Board
16 i..xe6!
I think that this is stronger than 16 d5
after which, however, I would have had a
tough (taking into account less than half
an hour left) choice between:
a) 16 . . . i..xe2 1 7 dxe6 0-0 ( 1 7 . . . i.. x dl??
18 exf7 + wins for White) 18 :del ( 18
exd 7? c4 is better for Black) 18 . . . i.. xf3 19
e7 'ii'c8 20 exf8i¥+ �xf8 ! (not 20 . . . i.. xf8?!
2 1 'ii'xf3 c4 22 i.. a2 b3 23 i.bl c3 [or
23 . . . �c5 24 c3] 24 cxb3 i..b 4 25 i.. d3 with
a plus for White) 2 1 'ii'xf3 c4 22 i.. a2 b3 23
i..bl �e6 24 c3 �xf4 25 'ii'xf4 �g4 26 'ii'g3
i.. h 6 + 2 7 f4 'ii'c 7! 28 l;Ihfl (28 'ii'xg4
1 0 dxe4 1 1 'ii' xe4 �f6 12 'iVf3 i.. b 7
••. i.. xf4 + 29 �dl 'ii'd6 + 30 �e2 h5 3 1 'ii'f3
13 0-0-0 �bd7 :e8 + 32 i..e4 f5 33 �fl fxe4 34 lixe4 l:.f8
Black is intending 14 . . . �d5 and 15 . . . c5. 35 :d4 'ii'x d4! 36 cxd4 i.. c l ! 37 �e2 l;Ixf3
14 'it'h3 c5 15 �f3 38 l;Ixcl l;Ig3 39 �f2 :d3 is equal) 28 . . . f5
Interesting was 1 5 dxc5, after which I with good compensation for the exchange;
mistakenly intended to play 15 . . . �d5?! , or
which brings White a huge advantage af b) 16 . . . �d5! 17 i..xd5 and now not
ter 16 i.. d6 'ii'f6 17 �d4 l;Ic8 18 �gf3 l;Ixc5 bl) 17 . . . 'ii'f6? (which I had been calcu-
( 18 . . . tbxc5? 19 i.e5) 19 lihel ! Stronger is lating during the game) 18 i.. xa8 i.xe2
Stefansson's post-mortem suggestion of ( 1 8 . . . 'ii'xb2 + 19 �d2 i.. c3 + 20 �e3 'ii'xc2
15 . . . �e4 ! and it is not easy to see any 2 1 �c l ! wins) 19 l:.xd7 'ii'xb2 + 20 �d2
thing better for White than 16 i.. xe6 fxe6 �xd7 21 i..b 7 'ii'c3 + (2 1. . .b3 22 �xe2 bxc2
17 c6 'ii'f6 ! 18 cxd7 + �f7 19 'ii'b 3 and 23 l:.cl 'ii'xb7 24 �d2 is also winning for
now: White) 22 �xe2 'it'c4 + 23 �dl 'ii'xf4 24
a) 19 . . . �xf2 is not good because of 20 'ii'h2 and White is better; but
�h3 ! tbxd l 2 1 �g5 + ! (2 1 lixd l? h6 is b2) 17 . . . exd5 !, when neither . . .
good for Black) 2 1 . . . �e7 22 l:.xdl i.. d5 23 b 2 1 ) 18 l:.hel ?! 0-0 19 l;Ixd5 'ii'f6 20
l;Ixd5 'ii'xb2 + (23 . . . exd5 24 'ii'xd5 'ii'xb2 + l;Ixd7 (20 'ii'xd7 l;Ifd8 21 'ii'xd8 + l;Ixd8 22
25 �d2 wins for White) 24 'ii'xb2 i..xb2 + i..e 5 i.. h 6 + 23 �bl 'ii'b 6) 20 . . . 'ii'xb2 + 2 1
2 5 �xb2 exd5 2 6 i.. e 5 and White has a �d2 (2 1 �d l b3 + ) 2 1 . . . i.. c8 ! winning for
very pleasant endgame. Much stronger is Black; nor . . .
b) 19 . . . �c5 ! 20 'ii'c4 (20 'ii'a2? i..xg2 2 1 b22) 1 8 l;Ixd5 i.. c4 (intending 19 . . . i.. e 6
l:.h2 b3 2 2 cxb3 i.. d5 and Black is on top) with a slight plus)
20 . . . 'ii'xb2 + 2 1 �d2 i..c3 + 22 tbxc3 'ii'xc3 + . . . is what White was hoping for.
23 Wixc3 bxc3 + 24 �xc3 i..xg2 25 l':.h2 i.. d5 16 0-0
•••
and Black doesn 't seem worse to me in a) 25 �xc2 l2Jf4 26 'iVf3 (26 '*fe3 l2Jxd5
spite of the fact that he is a pawn down. 2 7 i.xd5 l:Ixc4 + 28 i.xc4 '*fxb2 + 29 <it>d3
2 1. l2Jh5
•• l:Id8+ ; 26 'iVb3 lDe2; and 26 i.xf4 'iVxf4
I spent a lot of important time calculat are all good for Black) 26 . . . l2Jxd5 ! 27 'iVxf6
ing variations such as 2 l . . . l2Je4 22 'iVe3 (27 i.xd5? l:Ixc4+ ) 2 7 . . . l2Jxf6 28 l2Jxa5
l2Jd6 23 g3 and 2 l . . . c3 22 b3 l2Je4 23 'iVe3 l2Jxg4;
l2Jd2 24 g3, only to realize that they would b) 25 l:Id2 tbf4 26 'iVe3 (26 'iVf3 l2Jxd5 is
not have left me much hope of anything also good for Black) 26 . . . l2Jxd5 2 7 i.xd5
good. It was practically at the last mo l:Ixc4 28 i.e5 'iVxe5 29 'iVxe5 i.xe5 30
ment that the right idea occurred to me. i.xc4 i.f4 3 1 �xc2 i.xd2 32 �xd2 l:Id8 +
22 i.h2 b3! ! (intending 33 <it>c3 l:Id1).
The strongest move i n the game. When I might also add that 24 'iVe3?! 'iVxh4 25
I made it I had just five minutes left for l:Ih1 i.xe5 26 'iVxe5 'iVxg4 leaves White
the remaining eighteen moves, whereas with no real chance of survival, while 24
my opponent still had more than fifteen. gxh5? 'iVxf2 loses immediately.
Of course, I needed a high level of concen 24 'iVxh4 25 .:hl i.xe5!
•..
tration to play well under these circum Accuracy is still necessary. 25 . . . l2Jf6??
stances. I should also mention that 22 . .'iVf6
. was impossible because of 26 i.f4 and
23 l2Jxc4 'iVxf2 24 tbxa5 would have been 25 . . .'ti'g5 + 26 'iVe3 is quite unclear.
slightly better for White. 26 i.xe5 'iVg5 + 27 'iVe3 'iVxg4
23 g4? Black is winning because of his mate
This turns things in Black's favour. rial advantage and threats against the
White should have continued 23 l2Jxc4 (af more vulnerable white king. The pair of
ter 23 c3? 'iVf6 24 l2Jf3 l2Jf4 Black is win- bishops and the activity of White's major
ning) 23 . . . i.h6 + 24 l2Je3 (24 �b1 ?? l:Ixc6 pieces are no longer of any genuine sig
wins) 24 . . .bxc2 25 �xc2 'iVb6 and the posi- nificance.
tion is unclear. 28 i.c3
23 'iVf6!
••• Other tries would also have been hope
less, for example 28 f3 'iVg2 29 l:Ih2 'iVfl +
30 �d2 l:Ifd8 3 1 i.d4 c3 + ! !
24 'iVf3?!
Making things still worse . However, af
ter 24 l2Jxc4 bxc2 ! (but not 24 . . . l2Jf4? 25 32 bxc3 (or 32 'iVxc3 l:Ixd5) 32 . . .b2 33
'iVxb3 l2Je2 + 26 <it>b1 l2Jd4 27 'iVc3 l2Jxc6 28 'iVe5 'iVc1 + 34 <it>e2 'iVxc2+ ; 28 i.d6 'iVf5 29
'iVxf6 i.xf6 29 dxc6 l:txc6 30 l2Jxa5 l:Ia6 3 1 'iVd2 l:Ifd8 30 i.e7 l:Ixc6 31 i.xd8 l:Id6; or
b4, when White has a slight plus) White is finally 28 cxb3 cxb3 29 'iVxb3 l:Ifd8 30 l:Id1
clearly in an inferior position: l2Jf4.
Shirov - Leko, Belgrade 1995 141
2S 'iVf5 29 �b1
•.•
Forced.
29 bxc2 + 30 �a2 ltJf4! 3 1 l:th4
•••
b) 25 �d2! (which in fact I was intend Black is a piece and a pawn up, but his
ing during the game) is much better, army is so badly co-ordinated that he can
when Black has a choice between: not stop White's attack, for instance:
b1) 25 . . . �b3, which loses to 26 l:xb3 b32 1) 29 . . . liJe5 30 l:xe5 'ifd7 3 1 'ifc4 +
axb3 27 �xb4 cxb4 (or 2 7 . . .hxg5 28 tDxc5 ; �h8 32 l':.e6 ! �g7 (or 32 . . . �xd6 33 l:xg6;
27 . . . tiJb6 28 �c3) 28 ltJxd6; 32 . . . �b5 33 'ife4) 33 l:xg6 l:f8 34 �f5; or
b2) 25 . . . hxg5 26 �xb4 cxb4 27 tiJf6 + b322) 29 ... 'ifg7 30 'ifc4+ �h7 (30 ... �h8
'iixf6 28 'iVxd5 + �h8 29 'tWxa8 with the 31 tiJf7+ �h7 32 l:.g3! wins) 3 1 'tWxc6 l':.d8
same outcome; and the clearly stronger 32 'iVc7 'iVxb2 33 �xg6 + ! �xg6 34 .l:g3 +
b3) 25 . . . �c6 . But then Sulipa's idea of �h5 (or 34 . . . �h7 35 'ifxd8) 35 tiJb7 ! (35
26 tiJh7! ! really does work! To prove it I 'ifxd8? 'ifc1 + 36 �h2 .i.xd6 37 'tWe8 + �h4
had to study this position thoroughly 38 'ifxd7 'iff4! is unclear) 35 . . . ltJe5 36
(with the help of Fritz4, to be honest, al liJxd8 'ifb1 + 3 7 �h2 'iff5 38 ltJc6 tDxc6 39
though I had to find the main ideas by 'tWxc6 'iff4 40 �g1 and White soon gives
myself! ) and here is the analysis: mate soon;
b31) 26 . . .'tWd5 27 tiJhf6 + ! ! (the decisive b33) 26 . . . 'ti'f7. Probably best. Now af
piece sacrifice ; 2 7 'ifg4 is much less clear ter 27 l:.g3 Black should play not 27 . . . �h8
in view of 2 7 . . . �h8 28 �c3 ltJe5 29 'ifg3 28 tiJxf8 �xe4 (28 . . .l:xf8 29 �c3 wins for
�e7 30 f4 'ifd 1 + 3 1 l:e1 'tWh5 with the White) 29 tiJxd7 �xb1 30 �c3 'ifxd7 3 1
idea of 32 fxe5 �h4 33 'iVxg7 + �xg7 34 'ifxb1 :e8 3 2 :xg7 'ifxg7 3 3 �xg7 + �xg7
e6 + 'iie 5) and now Black has two ways to 34 'iVd 1 where he is losing; but 27 . . . d5 28
take the knight but they lead to the same liJxf8 J:lxf8 29 �xh6 dxe4 30 .l:xg7 + 'iVxg7
thing: 3 1 �xg7 �xg7 32 'ifg4 + , and although
b3 1 1) 2 7 . . . liJxf6 28 liJxf6 + gxf6 29 White holds better chances it is not clear
�g4 + 'iVg5 (or 29 . . . �h8 30 l:g3 'iff7 3 1 whether he should win. Still, Leko's move
'iih4) 3 0 'ife6 + �h8 3 1 l':.g3 .i.d5 (3 1 . . . 1':.e8 (24 . . . hxg5) is better.
32 'iff7) 32 �xg5 �xe6 33 .i.xf6 + and 25 ttJxg5
White wins ; or
b312) 2 7 . . . gxf6 28 'ifg4 + �h8 29 :g3
'iff7 30 'ifh4 h5 3 1 �c3 �g7 (or 3 1 . . . tiJd5
32 ltJg5 'ife8 33 'ife4) 32 liJxd6 'ifd5 33 tiJf5 B
25 'iff6??
•••
work because of 29 . . . 'ti'd5 30 :e6 'ti'd1 + 3 1 A logical move when you don't know
�h2 .itxe6 3 2 fxe6 'ti'h5, when Black wins) much theory. Now the knight on a5 and
27 . . . .itb3 28 'iff3 repeating the position. the bishop on b 7 are not the best pieces on
26 %Ie6! .itxe6 27 fxe6 g6 the board.
2 7 . . . lbe5 loses by force to 28 'ii'h 5 �h6 14 l:tac8 15 .itb1 lbh5 16 lbf1 lbf4 1 7
•.•
�f7 would have prolonged it. Now both Consistent but allowing Black a nice
the g6 pawn and the black king are terri opportunity. 19 %Ib1 should have been
bly weak. preferred, with a slight plus, and if now
30 d5 3 1 lbf3 �g7 32 .itg5 'iVxb2
••• 19 . . .lbc4 then the simple 20 ..txc4.
Or 32 . . . 'ifd6 33 .itf4 and wins. 19 .'ti'c5! 20 %Iac1 'ii' b 4
••
This is a mistake. After the game Tim White's plan is very simple - to grab
man pointed out that he had a very inter the pawn on b4. He only needs to be care
esting double pawns sacrifice, i.e. 23 ... tiJb7 ! ful about . . . f7-f5 .
24 'ti'c6 l;Id8 25 'ti'xa6 ltJc5 2 6 'ti'xb5 'ti'c3, 30 .ic3 31 tiJb1 i.a1 32 liJe1 f5
•••
when White has no advantage, for exam At home I discovered that Black could
ple: have put up tougher resistance by sacri
a) 27 .ie2 ltJxe4 28 .ic4 .id7 29 'ti'b6 ficing a piece, i.e. 32 . . . i.d4 33 liJc2 i.c5 34
tiJg5 30 tiJ1d2 liJxf3 + 3 1 tiJxf3 g5 32 .ie2 tiJd2 f5 35 a3 fxe4 36 .ixe4 bxa3 3 7 b4 a2
.if5 with compensation; 38 bxa5 (38 bxc5 dxc5 is less clear because
b) 27 .ic4 .ia6 (27 . . . .ixh3? 28 e5 .id7 the black king might become very active)
29 'ti'b6 wins for White) 28 'ti'c6 (28 'ti'b6 38 . . . .id4 39 tiJb3 .ib2, but it seems to me
.ixc4 29 e5 .ixd5 30 exf6 'ti'xf6 is equal) that White is still winning after 40 �d3
28 . . . .ixc4 29 bxc4 'ti'xc4 and again Black i.c8 41 �c4 .id7 42 �b4.
has sufficient compensation; 33 liJ c2 fxe4 34 .ixe4 .if6 35 tiJxb4
c) 27 .ib1 .ixh3 28 b4 (28 tiJ1h2 .ic8 is h6 36 .id3 .id4 37 tiJc2
also fine for Black) and now not Going for the piece, e.g. 37 . . . .ie5 38 b4.
c1) 28 . . . .id7? 29 'ti'b6 ltJa4 (29 . . . .ig4 30 37 .ia7 38 tiJc3 1-0
•••
bxc5 .ixf3 31 e5 ! ! dxe5 32 gxf3 'ti'xe1 33 But now that is not even necessary, as
c6 wins for White) 30 'ifxd6 .ig4 31 'ifxf4 the text is a lot easier. It is not every day
.ixf3 32 lle3 ! 'ilfxb4 (or 32 . . . 'ir'c1 33 l:Ixf3 that I win a positional game like this.
'ilfxb1 34 e5) 33 l:Ib3 'ife1 34 'iVxf3 liJc3 35
i.d3 ltJxa2 36 l:.b7 with a clear plus for Game 66
White; but
c2) 28 . . . tiJd7 29 tiJ1h2 ltJe5 30 gxh3 Va n der Sterren - S h i rov
liJxf3 + 3 1 tiJxf3 'ifxf3 32 'ti'd3 'iVh5 with Bun desliga 1995/96
excellent compensation.
24 %:tel 'iVxc2 25 l:.xc2 The game was annotated in December
Now White is clearly better. 1996 and published in various magazines.
25 b4 26 l:Ic7
.•. Although nowadays I am less success
Black's problem is still the same as in ful in the Bundesliga than in my first sea
the opening - the knight and the light son, every year I manage to score 2 out of
squared bishop. 2 in at least one weekend. Thanks to this
26 .ib7 27 tiJ1d2 �f8 28 �f1 :e7
••• game (and incredible luck in the Saturday
28 . . . h6 29 l:Id7 would have led to a simi one!) I kept up the tradition.
lar position. 1 d4 d5 2 c4 c6 3 tiJc3 tiJf6 4 tiJf3 e6 5
29 l:.xe7 rtJxe7 30 �e2 e3
Normally Paul doesn 't mind playing 5
.ig5, but perhaps he didn't wish to enter
the Botvinnik variation at 9 a.m.
B
5 tiJbd7 6 .id3 dxc4 7 .ixc4 b5 8
•••
i.d3 a6
Switching from my usual 8 . . . .ib7.
9 e4 c5 10 e5 cxd4 1 1 liJxb5 axb5 12
exf6 gxf6 13 0-0 'ifb6 i 4 .ie4
Nowadays 14 'ilfe2 is considered more
critical.
14 .ib7 15 i.xb7 'iVxb7 16 tiJxd4
•••
l:.g8 1 7 f3 .ic5!?
In one sense this move is a novelty, al
though it ultimately leads to a known
Van der Sterren - Shirov, Bundesliga 1995/96 145
28 l:txc2 0-1
••• Here I saw a fascinating sacrifice, and
White resigned as he loses the house. after about five minutes' thought I de
cided that it was worth a try. The funny
Game 67 thing (but not for me) was that a few
months later I had an opportunity to play
S h irov - lva nchuk exactly the same idea in a different posi
Tallinn (rapidplay) 1996 tion against Kramnik (Monaco rapidplay
1996) , but scoring just half a point from
These annotations were made during the two such promising games is very annoy
preparation of this book. mg.
It's curious that I have had several in 15 e5 f5 (D)
teresting rapidplay and even blindfold 16 lbd5!!
games against Vasily recently. The time 16 i.. f3 .i.b7 17 i..xb7 'i¥xb7 18 :d2 or
control of this one was 25 minutes each. 18 lbe2 would have led to an unclear game.
1 e4 c5 2 lbf3 e6 3 d4 cxd4 4 lbxd4 Not very exciting.
a6 5 lbc3 'ifc7 6 i.e2 16 exd5 17 e6 0-0!
•••
Shirov - Adams, Wijk aan Zee 1996 147
correct one, but the king goes to the wrong 23 . . . h6 would have yielded White a
square. 19 <ith l ! was to be preferred, as slight edge in the endgame arising after
then any attempt by Black to free his po 24 fic7 i.. g5 25 'iixd8 i.. x d8. The pawn
sition would rebound, for example: sacrifice is better.
a) 19 . . . fxe5 20 fxe5 l:txfl + 2 1 l:txfl l:tf8 24 i..xh7 + <itb8 25 l:tel i.xe3
22 l:tcl intending 22 . . . 'iid8 23 g5 ! i.xg5? 25 . . . i.. f4? 26 i.. f2 fig5 2 7 h3 is better
24 'ii'h 5 h6 (or 24 . . . i.. h 6 25 l:tgl <ith8 26 for White.
i..xh6 gxh6 27 i..xh7) 25 'iVg6 �6 (25 ...�f7 26 l:txe3
26 'iVh7 + Wf8 27 .l:c8 'iVxc8 28 'ti'h8 + We7
29 i.. xg5 + hxg5 30 'iVxc8 and 25 . . . i.xe3 26
'iVh7 + Wf7 2 7 i.. g6 + both win for White)
26 i.xg5 hxg5 2 7 'iVh 7 + Wf7 28 'iVh5 +
We7 29 'ti'xg5 + and White is winning; or
b) 19 . . . .:tac8?! 20 f5 ! fxe5 2 1 fxe6 'ti'xe6
22 i.f5 and White has a clear plus. In
stead, Black should also bide his time by
c) 19 . . . l:tf7! , after which White might
continue 20 l:.gl !?, intending 2 1 llg3 with
an attack.
19 fxe5 20 fxe5 l:txfl 2 1 l:.xfl l:tf8
•••
22 .:tel 'ifd8
Here I probed deeply into the position
and discovered that my advantage had
Shirov - Gelfand, Wijk aan Zee 1996 149
26 :f4?
.••
Game 69
Tempting but not best. Instead 26 ... 'iig5
27 %:tg3 'tif4 28 i.d3 �g8 (28 . . . 'tixd4? 29 S h i rov - Gelfa nd
:h3 + <&t>g8 30 i.h7 + <it>f7 3 1 'iVg6 + �e7 Wijk aan Zee 1996
32 'iVg5 + wins for White) 29 i.h7+ (29
'ii'c 3? ! l2Jb8! is slightly better for Black) The game was annotated in February and
29 . . . � h8 would have forced me to repeat published in various magazines.
moves. Before this game I had just 3 out of 8;
27 %:tg3 l:.xd4? and besides I had developed a severe cold,
This loses. Black had some other possi which is clearly not the best companion if
bilities but it seems to me that White you're playing in Wijk aan Zee, with its
should always stand better. Here are the strong winds and poor heating in the
variations: playing area where doors sometimes open
a) 27 . . . 'i¥h4 28 i.d3 <&t>g8 (or 28 . . . l2Jf8 by themselves. I was just dreaming of sur
29 h3) 29 'i¥c7 l2Jf8 30 h3 with a clear plus vival and getting back to Spain, but some
(30 . . Jlxd4? is met by 3 1 %:tf3) ; how in this condition I managed to score a
b) 27 . . . g5 28 'ii'g6 %:txd4 with three vari hat-trick.
ations: 1 e4 c5 2 l2Jf3 d6 3 d4 cxd4 4 l2Jxd4
b1) 29 'iVh5 <&t>g7! is unclear; lbf6 5 l2Jc3 a6 6 i.e3 e6 7 g4
b2) 29 i.g8 %:td2 + 30 <it>h1 (30 <it>fl :xh2 This move has become rather popular
is also unclear) 30 . . . l:.d1 + 31 <it>g2 %:td2 + recently, especially since a nice victory by
with equal chances; Topalov against Kasparov at the Moscow
b3) 29 'tixe6! %:td2 + 30 <it>fl ! (not 30 Olympiad in 1994.
<&t>h3 �xh 7 31 %:tf3 'ii" h 8 ! ! 32 'tif7 + 'iVg7 7 e5
•••
33 e6 lbe5 34 'ii"h 5 + �g8 with a draw) This leads by force to very sharp and
30 . . . :Xh2 3 1 i.f5 and White is on top; concrete play. The fact that Kasparov
c) 2 7 . . . l2Jf8 28 'tid2 g5 29 i.b1 'i¥b6 30 played 7 . . . h6 in the aforementioned game
k:tf3 ! llxg4 + 31 �h3 .:h4 + 32 <&t>g3 with a may indicate that the variation with 7 . . . e5
clear advantage for White. is quite risky for Black.
28 'i¥f2 ! 8 l2Jf5 g6 9 g5 gxf5 10 exf5 d5 1 1 gxf6
After the game Michael admitted that 1 1 'i¥f3 used to be more popular but I
he had overlooked this. His position is think it's weaker.
now lost. 1 1 d4 12 i.c4! 'i¥xf6
•••
28 l:.b4?!
••• Another critical continuation is 12 ...'i¥c7,
28 . . . �xh 7 29 'i¥xd4 would have offered after which White normally continues 13
slightly more resistance, although White 'ii'd3 with good attacking chances.
is still winning. 13 l2Jd5 'ii"c6 14 i.xd4 i.b4 +
29 i.g6 'i¥h4 This check is well known to be the only
Or 29 . . . 'i¥e7 30 llh3 + <&t>g8 3 1 i.f7 + move. 14 . . . 'tixc4? 15 l2Jf6 + �e7 16 i.xe5
'tixf7 32 %:th8 + �xh8 33 'i¥xf7 llxb2 + (or and 14 . . . exd4? 15 'i¥xd4 are definitely in
33 . . . llxg4 + 34 <&t>h3) 34 <&t>g3 and White advisable for Black.
wins. 15 c3 'i¥xc4 16 i.e3 i.a5 (D)
30 'i¥f7 l:.xb2 + 31 <it>f1 l2Jf6 32 'i¥f8 + ?! A novelty which sent me into long
It was a lot simpler to play 32 exf6 thought. 16 . . . i.e7 1 7 l2Jb6 'i¥c6 18 %:tg1 was
'iVx:ffi + 33 llf3, but in time-pressure I found played in the game Th. Thorhallsson-J.L.
the other winning idea first. Arnason, Reykjavik 1992, which appears
32 l2Jg8 33 l:c3
••• to favour White.
Now 34 'iVxg8 + <&t>xg8 35 %:tc8 + is threat 1 7 l2Jf6 + <&t>e7 18 l2Jd5 +
ened and Black has no useful checks! I already felt that it wouldn't be easy
33 %:tb1 + 34 i.xb1 'i¥xg4 35 'i¥f3 1 -0
••• to calculate everything with the modern
150 Fire on Board
also have made a long trip, but only to its c) 30 i..xe5 ttJxe5 3 1 tbxe5 l:.g8. The
grave, i.e. 24 ..th6 + �e7 25 'i¥g5 ! ttJd7 best White can get in any of these vari
(25 . . . ..tb7 26 l:d1 ! ) 26 ltJxd7 + <it'xd7 27 ations is a slight advantage.
0-0-0 + <it'c6 (27 . . . ..td6 28 'i¥f6) 28 'iVg2 + 30 ttJh6!
e4 (or 28 . . . �b6 29 'ti'xa8 'ti'xa2 30 i..e3 + I am very proud of this move, refusing a
<it'a5 3 1 'iVe4) 29 .l:d4 'ti'xa2 30 'ti'xe4 + �b6 pawn and still playing for the attack, de
3 1 i.. e 3lta5 (3 l . . . lt'a7 32 .l:d8 + wins) 32 spite time-pressure. 30 lLlxe5 and 30 i.xe5
l:ta4 + !! 'ti'xa4 (32 . . . bxa4 33 'ti'b4 mate) 33 simply transpose to 29 . . . ltJd7.
b4 + 'ti'xb4 34 'i¥xb4 mate. 30 i.. g6
.•.
29 i.. xc7
3l. ttJb4!!
••
to 30 .l:d8 + <it'g7 3 1 i.xe5 + f6 32 i.. xf6 + After making this move Gelfand had al-
<it'xf6 33 l:xh8, and if 33 . . . i.f3 34 l:.f8 + , most no time left.
but 2 9 . . . ttJd7!, seemingly the same as the 34 h5!
text, would in fact have been much better. 34 l:.c6 f6! equalizes.
Black loses a pawn, but retains consider 34 :Xc7 + 35 <it'd2 i.e4 36 .l:gl + <it'h8
•••
able drawing chances, for example: The king goes into the corner, but after
a) 30 tbxe5 ttJxe5 3 1 ..txe5 (31 i.. d6 + 36 . . . <it'f8 3 7 .l:g8+ <it'e7 38 .l:xa8 <it'xd6 (or
�g7 32 i.. xe5 + f6) 3 l . . Jig8; 38 . . . ..txa8?! 39 ttJf5 + ) 39 l:xa6 + White's
b) 30 i.. d6 + �g7 3 1 i..xe5 + ltJxe5 32 task would have been even more straight
ttJxe5 l:.he8; forward.
152 Fire on Board
38 �e8?!
.••
Down to his very last seconds, Gelfand Since Kasparov revived this old system,
doesn't find 39 . . Jlc4 ! which in fact would it has become worth analysing. It's so
have been a good practical try. Now the concrete that in some lines one needs to
correct variation is 40 f3 l:tf4 4 1 l':.d8 dig deep into the endgame to draw the
l:txf3 + 42 �e2 l:txf5 43 l:txe8+ (43 h6 i..c 6! right conclusions.
44 l':.xe8 + i.. xe8 45 l':.d8 l:tg5 46 l:.xe8 + 11. .'ili'xg5
•
pieces against Smyslov in 1979 and had allows Black an easy draw after 27 . . . �b2 !
very good winning chances there, but (but not of course 2 7 . . . �d2?? 28 �fl �d3
Smyslov continued 19 ... �e7 which is prob 29 lDc6 or 27 . . . �d3?? 28 lDc6 l:Ide8 29
ably weaker. lDb4 + �d2 30 l:Ied 1 + �e2 3 1 l:Id3 �h6 32
l:Ic2 + �e1 33 l:Ib3) 28 l:.b1 + and now
either
a) 28 . . . �a2 29 l:Ia1 + (29 4Jc6?! d3 ! 30
w 4Jxd8 Itxd8 31 l:.xb5 d2 32 l:Id1 �c3! gives
Black plenty of compensation) 29 . . . �b2;
or
b) 28 . . . �c3 29 l:Ibc1 + �b2 forces White
to give perpetual check.
20 �b4 �g7 2 1 a4
2 1 l:Iad 1 would have been very risky in
view of 2 l . . . d3 22 lDc1 d2 23 lDe2 l:Id3 24
tLlf4 l:Ihd8 25 4Jxd3 l:Ixd3 26 c6 g5 ! , when
although Black is a rook down his chances
are not at all worse.
2 1. �d7 22 axb5 axb5
•.
Here 23 l:ta6 would simply be met by 29 . . . i.e5? would have lost by force to 30
23 . . . l:Ia8. l:Ixb5 + �e6 3 1 l:Ixe5 + �xe5 32 i.xc7 +
23 ...�e6 24 l:.fe1 + �d5 33 �xd8 .r!xd8 34 c7 l:Ic8 35 �fl �d6
I was trying to figure out what would 36 �e2.
happen after 24 l:Id3 and I concluded that 30 l:Ixb5 l:.b8! 31 �b4+
in the line 24 . . . �d5 25 .i.xc3 (25 4Ja5? l:Ia8 31 l:.b7?! is met by 31 . . . l:Ixb7 32 cxb7 c5.
26 l:.fd1 l:Ixa5 2 7 �xa5 �c4 wins for 31...�e6 32 l:Ie1+ �f6! (D)
Black) 25 . . . �c4 26 l:Ifd1 dxc3 27 lDa5 + 32 . . . �f7? was wrong, not because of
�xc5 28 lDb 7 + �c4 29 4Ja5 + (29 4Jxd8?? a) 33 l:.e7+ �f6 34 l:Ibe5 �f8! (note
c2) 29 . . . �b4 30 lDc6 + �c4, a draw is the that 34 . . . l:Ixb4?? allows a simple mate af
most that White can get. ter 35 g4 �h6 36 h4) 35 l:I5e6 + �f5 36
24 �d5 25 �xc3
••• l:Ie5 + �ffi and the game is drawn; but
I rejected 25 4Ja5 in view of 25 . . . c2 26 b) 33 l:Ib7! l:Ixb7 34 cxb7 winning.
l:td2 c1'1W 27 l:.xc1 �h6, which is unclear. 33 �e7+
25 ...�c4 26 .i.a5! My original idea was to win the ex
The only practical try to win the game. change with 33 l:Ib7 l:Ixb7 34 cxb 7 l:Ib8 35
26 lDa5 + �xc3 2 7 l:Ic1 + is tempting but l:tb1 l:txb7 36 �e7 + �xe7 37 l:.xb 7, but
154 Fire on Board
then I realized that after 37 . . . Wd6 Black's .l:Ie6 46 1:c3!, after which just good tech
compensation is at least sufficient. nique is required to win the game.]
33 �f7 34 1ld5 1Ihc8?!
.•. 45 1!e6
.•.
Black was only a small step from the 45 . . .'�'f7 46 �f3 is also hopeless for
draw that would have been reached after Black.
34 . . . 1Ihe8! 35 1Id7 Wg8 36 Wfl d3 37 1Ixd3 46 l:d8 + �f7 47 l:d7+ 1Ie7 48 1Ixe7+
1Ib6. �xe7 49 g5! 1-0
35 1Id7 Wg8 36 g3! Black resigned in view of the obvious
Looking for more practical chances. 49 . . . Wd6 50 h4 Wxc6 5 1 f5 �d6 (5 l . . . gxf5
36 1Ib6
••• 52 h5 �d6 53 g6 hxg6 54 h6 and wins) 52
36 . . . 1Ie8 was still good enough. f6 and White wins easily. It's strange that
37 1Ic1 l:tb3 38 i.. c5 d3? I had already won exactly the same pawn
38 . . . 1':. c3?? 39 1Ixg7 + needs no comment ending against Akopian (Oakham 1992)
but 38 . . . i.. h6 would probably have saved with the only difference that pawns were
the game. After the text move White is on the queenside in that game.
clearly on top. [AS - The real difference from that
39 l:td1 l:cb8 40 <it'g2! game is in fact that here Black has a pawn
Black was hoping for 40 1!1xd3? .l:txd3 on c7. Thus when White 's king is on c6
41 l':.xd3 1Ib1 + 42 Wg2 1Ic1 43 1ld5 Wf7 ! Black puts his to dB and there is no zug
4 4 1Id7 + <it'g8 with a draw, but once White zwang since White 's pawn is already on
avoids this the game is over. h4 and there is no way to give Black the
40... ..tf8 41 ..txf8 1Ixf8 42 .:1xd3 turn. Therefore Timman simply resigned
J:txd3 43 l:txd3 l':.f7 44 f4 1Ie7 (D) in a drawn position. This draw was dis
Setting the last but clever trap. covered by an amateur from Switzerland.]
45 g4!
After the natural 45 Wf3? I would have Game 71
had to demonstrate my queen endgame
abilities in the line 45 . . . J:te6 46 1Id8 + �f7 N u n n - S h i rov
47 l':.d7 + l:e7 48 l':.xe7+ Wxe7 49 Wg4 Bundesliga 1996
<it'd6 50 <it'g5 <it'xc6 5 1 <it'h6 Wb7 52 Wxh7
c5 53 �xg6 c4 54 f5 c3 55 f6 c2 56 f7 c1 � The game was annotated in February
5 7 f8'i¥ 'iic2 + , which is not the best idea 1996 and published in various magazines.
for a sudden death time-control. This game was difficult for me. I was
[AS - In fact the pawn ending is drawn not in a perfect state of health and shortly
while the queen 's one should be winning before the game I suffered from a nose
after all. But I think that the best is 45 �f3 bleed. I was quite afraid it would recur
Nunn - Shirov, Bundesliga 1996 155
during the game but as it went on, I As usual I was being quite ambitious. I
started feeling better and when I sacri saw that the simple 15 . . . i.. xb5 16 axb5
ficed the rook c:xc4) it was already clear 4Jb6 would be very good but I wasn't sure
to me that the blood would be just in the that I would be able to establish a clear
game. advantage after 1 7 'ii'd 3 !, with the idea of
1 e4 g6 2 d4 i.. g7 3 4Jc3 c6 4 ..tc4 d6 answering 1 7 . . . a4 with 18 i.. c4. The text
5 �f3 e6 6 4Jge2 b5 7 i..b3 a5 8 a3 i..a6 might objectively be even better but it's
Despite my terrible loss against Judit not as strong as it looks. Now White finds
Polgar at the Donner Memorial (Amster some very good moves (most of which of
dam 1995) I still chose to employ this set course I hadn't seen) and the game be
up. It's hard to believe that with such a comes rather sharp and complicated.
bishop on b3 White should get the advan 16 'iVd3! c4! 17 i.. xc4 4Jxa4 18 :a1!
tage. I had overlooked this. 18 :b1 4Jc5 19
9 0-0 i..xf6 i..xf6 20 'tWf3 0-0 21 �a1 �fc8! 22 c3
Judit played 9 d5, and after 9 . . . cxd5 10 'V/iid 8! with a clear advantage would have
exd5 e5 1 1 4Je4 'V/iic 7 12 c4 bxc4 13 i.. a4 + been more according to my plan.
4Jd7 14 4J2c3 �e7? 15 4Jxd6! I was crushed
with incredible speed. Of course, this time
I hoped to improve somewhere but Nunn
decided not to test my preparation. B
and now I had to take such risks! After was definitely my day since I quickly
several minutes of uncertainty I decided found a move that justified all my pre
to take the plunge. vious play.
23 l:txc4!!
••• 30 'ifd3!!
•••
Variations such as 23 ... i..xb5 24 cxb5 .l:a8 But not 30 . . . a4?, which fails to 31 ttJc3
25 .l:fa1 would just demonstrate White's 'ii'xc6 32 dxc6 with a clear plus for White.
positional power. But now Black's pawns Now White 's pieces lack co-ordination,
are going to be really strong. his back rank is vulnerable and 3 l . . . .l:c8 is
24 i..xf6 threatened.
24 'ii'xc4 'ii b6 would ultimately lead to 3l l:tcbl?
the same thing. Short of time, Nunn loses immediately.
24 i..xf6 25 'iYxc4 'iYb6 26 fi/c7 fi/xb5
.•. He had to try 3 1 l:txb3!, when I'm not sure
27 'iixd6 �g7 that during the game I would have re
The less concrete 27 . . . i.. d8!? would sponded correctly. Home analysis con
probably also have kept Black's advan vinced me that 3 1 . .. fiixe2 would have given
tage. The text leads to positions in which White good drawing chances, whereas
such a small thing as the placing of White's 3 l . . . 'ifxb3! should maintain Black's edge
pawn on h2 and not on h3 will sometimes intact. Here are the variations:
become a decisive factor. a) 3 l . . . 'ifxe2 32 l:tf3 i..e 7 (32 . . . i.. g5? 33
28 .:tel! "V/I/c7 l:.f8 34 "V/I/xe5 + f6 35 l:tc7 + �h8 36 h4
The unprotected first rank leads to 'iYd l + 3 7 �h2 'iixf3 38 l:txh7 + �xh7 39
White's defeat in the line 28 l:tfa1?! b3! 29 'ifc7 + �g8 40 gxf3 is better for White) 33
.l:xa5? b2 30 'ifxa6 b1'if + . 'ifc7 'ifb2 and now:
2 8 l:td8! 2 9 'ifc6
••• a1) 34 h4 and:
29 'ifc5 b3 30 l:.b2 a4 looks clearly in al l) 34 . .. l:td7? 35 'ifxd7 'ifxc1 + 36 �h2
Black's favour since the queen's ex 'ifc5 3 7 'ife8 (37 h5 !?) gives White com
change, 3 1 'ifxb5 i.. xb5 32 .l:a1 .l:c8, is pensation;
catastrophic for White. a12) 34 . . . l:.f8 35 .l:cc3 (or 35 l:tc6 i.. d8
29 b3!
••• 36 fi/d6 i..b 7 37 .l:c4 i.. xh4 38 .l:c7 i.. c8)
Not 29 . . . a4? since in the endgame aris 35 ... i..xh4 36 fiixe5 + �g8 wins for Black;
ing after 30 l:txa4 'ifxc6 31 dxc6 i.. xe2 32 a2) 34 h3 .l:d7 35 'ii'xd7 'ifxcl + 36 �h2
c7 .l:c8 33 .l:xb4, only White has winning 'ifc5 37 'iYe8 i.. f6 and wins;
chances. a3) 34 .l:dl? .l:e8;
30 .l:b2 (D) a4) 34 .:tel ! l:te8 35 'iYxa5 i.. b 5 with a
Here I had a feeling of horror, as if I had slight plus for Black in view of 36 �al
completely misplayed everything. But it 'ii'd4! ;
Korchnoi - Shirov, Madrid 1996 157
b) 3 l . .. 'iWxb3 ! 32 'ifxa6 a4 (32 ... i.g5 33 last pawn and, with a certain degree of
'iVc4! is only slightly better for Black) 33 luck, this worked.
liJg3 ! and now: 1 c4 e5 2 g3 f5 3 i.g2 liJf6 4 d3 i.b4+
b1) 33 . . . a3 34 l':.c7 (34 liJfl? i.e7 35 l:!c7 5 liJc3 i.xc3 + 6 bxc3 d6 7 liJf3 c5
i.d6 36 l:lb 7 'ifc2 37 'iWa7 IUS wins for I was not very familiar with the open
Black) 34 . . . a2 35 l:la7 'tib1 + 36 liJfl 'ii'xe4 ing and had already begun to improvise.
37 'tixa2 'iWxd5 and Black only has a slight 8 0-0 liJc6 9 tbe1 0-0 10 tbc2 i.e6 11
plus; liJe3 'iWd7 12 liJd5 liJe7!?
b2) 33 . . . i.e7! 34 l':.c7 (or 34 'ifa7 i.d6 The knight was annoying me.
35 'iWa5 l:lb8, intending 36 'iWa6 I:[b6) 13 'iWb3!? tbexd5 14 cxd5 i.f7 15 c4
34 . . . i.d6 35 I:[b7 and: i.h5
b2 1) 35 . . . 'iWc2 36 'ifb6 ! ! i.c5 3 7 l:lxf7+! Although White has a nice set-up in the
�xf7 38 'iWe6 + <it>f8 (38 . . . <it>g7 39 'iWxe5 + is centre and good prospects on the queen
equal) 39 'ii'f6 + <it>e8 40 'tie6 + i.e7 4 1 side, I was still quite optimistic about my
'ii'g8 + <it>d7 (4 l . . . i.f8 42 'iWe6 + is again attacking chances on the kingside. Easier
equal) 42 'iWe6 + <it>c7 43 'ifxe7 + with a said than done!
draw;
b22) 35 . . .'ti'd1 + ! 36 liJfl and now:
b22 1 ) 36 . . . a3 3 7 'iWb6 ! I:[b8 (37 . . . a2? 38
'iWa7 ! is better for White) 38 I:[xb8 i.xb8
39 'ifxb8 a2 40 'ifxe5 + �h6 41 'iVf4+ with
a draw;
b222) 36 . . . 'tid4 ! , with a clear plus. In
this position Black should definitely win
due to his a-pawn but good technique is
still required.
31. l:.c8
••
The game was annotated in May 1996 This came as a surprise . In fact, White
and published in various magazines. could have played 19 i.b2 with the same
The day before this game I had drawn basic idea but avoiding the complications.
against Illescas from a much superior po Black's answer would then be something
sition and, with a fifty per cent score, I had like 19 . . . 'iWc7 20 d4 liJd7 2 1 i.h3 with un
now practically lost all chance of a good clear play. Of course, my calculations
place. Still, I felt like fighting until the were concentrated on the line 19 i.xf4
158 Fire on Board
the threat of 28 . . . 'ti'g3 makes Black's ad counterplay after 1 1 i.f3 ltJbd7 12 a3 i.b7
vantage clear. 13 f5 :ac8 14 g4 d5 !? 15 exd5 e4!
26 :xg2 27 <it xg2 :xe4
•••
0-1
Game 73
S h i rov - Ge lfa n d
Dos Hermanas 1996
The game was annotated in June and
published in various magazines.
My only victory in Dos Hermanas was a 11. dxe5 12 i.g5!? ltJbd7 13 .i.d3
••
truly interesting game but one with many i.b7 14 'ti'f3 h6 15 i.d2!
mistakes. Still, it gave me some sense of After 15 i.h4 ltJb6 (intending 16 . . . ltJh7)
creativity which makes me less pessimis it would have been difficult to generate an
tic for future. attack.
1 e4 c5 2 ltJf3 d6 3 d4 cxd4 4 ltJxd4 15 ttJb6! 16 'ti'g3
•••
6... e5 7 ltJb3 i.e7 8 0-0 0-0 9 <ith1 Boris is being ambitious. After the game
"ii'c7! he admitted that he rejected 16 . . . ltJh5 in
This very natural plan has occurred view of 1 7 'ifg4 ( 1 7 'iVh3?! was in fact my
surprisingly seldom in tournament prac intention, but this just yields Black a
tice. Black avoids the line 9 . . . b5 10 a4 but slightly better position after 17 . . . ltJf4 18
nevertheless wants to place his bishop on .i.xf4 exf4) 17 . . . ttJf6 18 'ti'g3 (18 l:xf6?
b7. i.xf6 19 i.xh6 is completely unsound ow
10 f4 ing to 19 . . . i.c8 20 'ti'g3 i.e6, when Black is
The flexibility of Black's 9th move on top) 18 . . . ltJh5 and White has nothing
would have been seen after 10 a4 i.e6, better than to repeat the position.
when White has already weakened his 17 l:f5!
queenside. Going for broke!
10 b5
••• 17...i. c8
This is the point. This allows White a very dangerous
11 fxe5 attack by force. I had actually expected
Finding myself in a new situation I 1 7 . . . b4 18 l:.afl ! ( 18 ltJd5 ltJbxd5 19 exd5
went in for an unusual plan. Despite i.xd5 is no better for White) 18 . . . bxc3 19
Black's excellent development and con i.xh6 g6 (otherwise Black gets mated, for
trol over the centre, White still hopes to example 19 . . . gxh6? 20 :xf6 i.xf6 2 1 :xf6
attack the enemy king one day. The pawn �h7 22 'ir'h4; or 19 . . . ltJe8? 20 i.xg7+
assault doesn't seem very promising. For ltJxg7 21 :xf7 :xf7 22 :xf7) and I think
example, in the game Zarnicki-Sadler, that the best way now is 20 l:xf6 ! (20 bxc3
Buenos Aires 1995, Black had excellent ttJbd 7 is probably slightly better for Black)
160 Fire on Board
20 . . . i.xffi 2 1 :xf6 cxb2 22 :n with a very 27 �g6 + <it>f8 28 'ii'h6 + <it>g8 29 :fl!
unclear game. Also worth considering is :e8 30 'iVg6+ �f8 31 "ii'h 6+ <&t>g8
1 7 . . . tbc4, but I believe that by playing 18
:afl ! tbxd2 (18 ... ttJxb2? 19 i.xh6 gxh6 20
l:.xffi i.xffi 21 l:.xffi is disastrous) 19 ltJxd2
White gets better prospects.
18 :xe5!
Of course!
18 i.d6 19 i.f4 g5
•••
32 i.g6?!
I saw the variation 32 i.h7+ l:.xh7 33
�g6+ <it>f8 34 'ii'xh7 i.xg2 + 35 <it>xg2 'ifc6 +
36 :f3 'ii'xffi, but forgot that then I would
have had 37 ltJe4. After 3 7 . . . 'ii'g 7+ 38
'ii'xg7 + <it>xg7 39 l:.xf4 (to meet 39 . . . ltJc4
with 40 ltJc5 !) Black's chances of survival
Taking the exchange would allow White are minimal. After the text White is still
total domination. better but the position gets much more
20 l:.c5! complicated, especially in time-trouble.
It's funny when the only move in the 32 'ii'c4 33 i.xf7+ 'ii'xf7 34 'ti'g5 +
•••
position is also the most spectacular. 34 "ii'xf4 would have guaranteed me the
20 gxf4 21 'ii'h4 i.xc5 22 'ifxf6 +
.•• advantage, but I was still looking for a
<&t>g8 forcing way.
So far everything had been forced, but 34 <it>h8 35 'iVh4+ 'iVh7 36 f7 :r8 37
•••
Now Black gets into serious trouble. Afterwards Gelfand was unhappy about
24 . . . tbd7 ! had to be played and after 25 this move and instead proposed 38 . . . ltJc4.
'iff5 not White's best answer then is 39 ltJe2 !, with
a) 25 . . J�d8 26 'ifh7 + <it>f8 27 'ifxh6 + a clear plus.
<it>e8 (27 . . . <it>e7 28 'iVh4 + <it>e8 29 ltJe4 39 h4?
'ii'xe5 30 :e1 also wins for White) 28 ltJe4! 39 ltJe2 ! would have been especially ef
and White's attack is decisive; but fective now.
b) 25 . . . :e8! 26 'ifh7 + <it>f8 27 'ii'xh6 + 39 <&t>h7 40 <it>g1?! 'ii'h6?
•••
<it>e7 28 'ii'g5 + (28 i.e4 ltJb6 ! is far from As often happens, the final move of the
clear) 28 . . . <it>f8, when I see nothing better time control is the worst of the game. Af
than 29 'ii'h6 + with a draw. ter 40 . .. tbd6 White would already have to
25 'ti'xh6 f5 26 exf6 hlf7 look for a draw. Still, it seems that he can
The only defence. 26 . . . 'tWc7 27 'tWg5 + ! achieve it by continuing 4 1 'ii'e 5 l:txf7 42
<&t>h8 28 l:.e1 was hopeless. 'ii'xd6 'ti'xd6 43 l:txf7+ <it>g6 44 :xb7.
Shirov - J. Polgar, Dortmund 1996 161
by the forced 46 . . . 'i¥h6, leading to the 'ifdl + 58 �fl 'iVd4+ 59 �hl �g6 60
same thing. 'ii'e6+ 1-0
46...�h8 47 'i¥f5 �h6 48 �e5+
48 liJd5 ! was the right move. Game 74
48 'i¥g7 49 'i¥c7?!
S h i rov - J. Po l g a r
•••
50 'i¥xb7 bxc3 5 1 bxc3 liJd6 allows Black The annotations on this game were made
some counterplay. in July 1996 and have not been published
50 i.d5 51 h6
••• before.
It is not clear whether White has any My encounters with Judit have a
winning chances in the endgame arising af strange history. Finally I stopped my se
ter 5 1 liJf4 i.xf7 52 liJe6 i.xe6 53 �xg7 + ries of losses against her (it was six in a
�xg7 54 h6 + <it>g8 55 h7+ �xh7 56 �xf8 row!) maybe because this year I played ex
<it>g7, but perhaps I should have chosen 5 1 clusively with the white pieces against
'i¥c5 !? i.xf7 52 'i¥xb4, retaining an edge. her. All the five last games (Dos Her
51. 'i¥g5 52 'i¥c5 (D)
.• manas, Dortmund, Vienna, Yerevan and
5 2 �f4? ! �xf4 53 liJxf4 ltJe7 54 liJxd5 Tilburg) were Sicilians (pity that Lev
liJxd5 would just have been unclear. Polugaevsky has passed away, he would
52 ltJe7 53 liJg3?
•.• certainly have enjoyed it) and I always
162 Fire on Board
be completely lost since there is no direct Perhaps the only defence. 19 . . . l:Id8 20
mate and the material deficit is already fii h 5 ! is extremely unpleasant for Black.
quite large. However, it became clear to 20 fiixf2
me that White' s attack is strong despite
being rather vague.
20 ttJd7?!
•..
since now 25 'ifd6 + 'ife7 is not very prom possibility was 25 l:.d8 + �e8 26 lbc5 !?,
ising. but after 26 . . . �f7!? (26 . . . ..txf5 2 7 %Ixa8
21 �e4!? ..th6 28 %Ib8 is unclear) 2 7 'tib3 + �f8 28
This move took me a long time to find, %Ixc8 %Ixc8 29 �e6 + �f7 30 �xg7 + �f8
and after making it I was already in dan 3 1 �e6 + �f7, he has to go for perpetual
ger of serious time-trouble. The problem check.
was that I couldn't find a win after 21 25...�e7 26 l:txg7 + �8 27 l:tc7! �e8!
'ifh4 ..tg7 22 �d5 'ifd6 23 :hg1 �xf7, but Of course, not 2 7 . . . ..txf5? 28 �xf6 and
in fact it is reached in one move, i.e. 24 White wins easily.
�c7 ! ! , when the line 24 . . . 'ifxc7 (or 24 ... 'ifh6 28 lidS %Ig8!
25 'ifc4 + ) 25 l:txg7 + �xg7 26 'ife7 + �h6 After the game Judit stated that she
2 7 :gl speaks for itself. However, instead had missed the line 28 . . . ..txf5 29 :xa8
of 23 . . . �xf7 Black has the much stronger ..txe4 30 :xe8 + Wxe8 3 1 l:.c8 + , when she
23 . . . h6! 24 'tig4 %Ih7, and I don't see how was considering something her like 22nd
White can get more than a draw after 25 move . . . However with the text she threat
�f4 (25 'iYg6 'ifxg6 26 fxg6 %lh8 2 7 �c7 ens both 29 . . Jlg1 + and 29 . . . ..txf5 .
%Ib8 28 �e6 + �e7 29 �xg7 �f6 ! should 29 �g3
be okay for Black) 25 . . . 'ifffi ! 26 �e6 + The only way to parry these threats.
�xf7 2 7 'ifc4 �b6 28 'ifc7 + (28 �g5 + 29 %Ig4
•••
Game 76
S h i rov - Ada m s
Ti/burg 1996
18 'iff6!
••• 1 e4 e5 2 �f3 �c6 3 �b5 a6 4 �xc6
A very precise move. Now White is Not exactly a new opening in my reper
forced to exchange pawns before complet toire, as I used to play it from time to time
ing his development. more than ten years ago. But now, of
19 dxe5 course, I had to study it all over again.
166 Fire on Board
on in the present game. I didn't notice The other way would be 20 . . . a5 2 1 :c3
this remark in time, maybe because after b4 22 :e3 �b5 + and Black has enough
reading Short's notes to his game against for the pawn.
Timman (New In Chess No.8 of 1995) I 21 flc3 tld6!
got a feeling that New In Chess was some An innovation. Previous experience has
how more for entertainment than for been 2 1 . . . a5 22 a4 b4 23 :Ic5 �xb3 24
chess annotations, even though Nigel did flb5 + �a6 25 cxb3 with a clear advantage
enrich my vocabulary. for White, Kelleher-Adams, New York 1996
14 b5 15 ttJd5
••• and 2 l . . . b4 22 l:te3 l:tc8 23 ltJd4 �c4 + 24
�el :hd8 25 ltJe2 with an edge for White,
as in J.Benjamin-A.Ivanov, USA ch 1993.
21.. Jlc8 22 :xc8 l:txc8 23 ttJd4 also fails to
equalize.
15 ltJe7!
•••
i:e3 i.xb3!
Forcing a drawn rook ending. My hope
was to meet 26 . . . i.b5 with 27 l2Ja1 ! and by
attacking his pawn get some winning
chances, for example 2 7 . . . i:d2 + 28 �g3
i:c4 29 i:d3 ! i:cxc2 30 l2Jxc2 i.xd3 3 1
i:xb4+ �c5 3 2 i:d4 with a promising posi
tion. 33 gxh4?!
•••
30 l:Ib 7 + �d8 3 1 h4 i:d7 32 i:b8 + �e7, Black must now avoid exchanging one
with a draw. pair of rooks, for example 35 . . . h3 36 i:b1 +
30 l:Ib7 i:a6 31 l':.lb3 g5?! �a3 3 7 i:a1 + l:Ia2 38 l:Ixa2 + �xa2 wins
Here I started thinking that I might get for White after 39 l:Ic5 ! ! .
some winning chances. Of course, 3 1 . . . �d6
would still be an easy draw.
32 h4 h6? (D)
32 . . . g4! was correct, when to avoid a
forced draw White must continue 33 �f4!
(33 i:xh7 l:Ig3 34 �f4 l':.xf3 + 35 l:txf3 gxf3
36 �xf3 �d6 3 7 �f4 �e6 is drawn)
33 . . . gxf3 34 l:Ixf3 although it seems that
Black's counterplay after 34 . . . h5 ! is suffi
cient.
33 i:d3!
The most difficult move in the game,
intending to cut Black's king off from his
pawns. 33 hxg5 hxg5 34 l:Id3 ! would also
have yielded White some practical win
ning chances, so it was not easy to choose. 36 l':.h l f5 3 7 exf5
168 Fire on Board
7 .'iih5
••
Of course, White could have chosen Now 15 ... 'ti'b6 16 g5 would yield White
simple development by 8 .i.e3 or 8 i.f4, some counterplay after 16 ... ltJh5 1 7 ltJd2 !
with an equal game. or 16 ... ltJd7 17 'ti'd2 .i.d6 18 ltJe4.
8 ttJxd4 9 ltJxd4 'ti'xg5 10 h4!
••• 16 'ti'xa6 bxa6 17 .:td4 ltJd7!
To be honest, I had overlooked this, Not an easy move to find. Black is
counting only on 10 ltJcb5 'ti'c5 1 1 0-0-0 a6 ready to give up his a-pawns since White's
which is very sad for White. pawns on the kingside are also vulner
10 ifh5!
•.• able.
Gaining an important tempo. 18 l:Ia4 ltJe5 19 ttJd2
11 f3 19 l:.xa6 Wb7 20 .1:a4 ltJxf3 2 1 ltJa5 +
1 1 'ti'xh5 ltJxh5 12 ttJcb5 :c8! is winning WeB offers White no chances.
for Black as 13 tbxa7?! loses to 1 3 . . . :d8 ! 19 Wb7 20 <it>d1 i.e7 21 h5?!
•••
• •
B
•
started to play ridiculous moves. I rejected An endgame piece sacrifice is quite un
24 . . J�h4 in view of 25 f5 (25 g5 i.. d6 wins usual. Its main motivation was White's
for Black) , but the endgame resulting af clock, although during the game I thought
ter 25 . . Jih1 + 26 l:Ie1 l:Ixe1 + 27 �xe1 gxf5 that Black is already almost winning.
(27 ... lL\e5 ! ? is also possible and should be 35 l:Ixc6
enough to win) 28 gxf5 exf5 should be Black wins after 35 lDxf4 lDe5 .
hopeless fo; White.
·
35 f3 36 l:If6?
•••
c 3 ) 3 7 ltJd2 (this i s probably the right It took me some time to realise that
way for White to proceed) 3 7 . . . �b7 (the 54 . . . �d5? 55 ltJxd6 �xd6 56 c4 ! might be
line 3 7 . . J:th1 + 38 �c2 f2 39 ltJe3 l:te1 40 only a draw because Black's king or rook
l:txd6 cxd6 4 1 ltJxf5 �d7 42 ltJg3 is also a will be stuck defending the c5-pawn and
draw) 38 l:txd6 cxd6 39 liJxf31:t xb2 and al so it will be very difficult to make pro
though Black can keep on fighting, it's go gress. Fortunately at this point I suddenly
ing to be a draw. saw the correct set-up.
36 1:t hl + 37 �d2 f2 38 l:tf8 + �d7
••• 55 liJf2 l:th4 56 ltJd3 �g5!
39 l:txf5 fl if 40 l:txfl hlxfl 41 ltJb4 i:. d6! The point. White has to abandon the
A very important move after which fortress he's built.
White has to turn to passive defence. 57 �c3
42 ltJd3 57 ltJxc5 + �d5 58 ltJxa6 l:txc4 59 ltJb4 +
42 ltJxa6?! �c6 loses because the white �c5 60 ltJa6 + �b5 61 a4 + :xa4 62 ltJc7 +
knight will not return, for example 43 �a5 is winning for Black.
ltJb8 + �d5 . 57 �d5 58 b3 l:th2
•••
However, after the text Black's task is A move without a clear idea that wins
not at all easy. The ending is unusual and the game immediately!
at the board I couldn't decide whether it 59 b4?
was good or bad for me to exchange the Of course, White should play some
bishop for one of his knights. It is also not thing like 59 ltJel . The best winning plan
clear how Black should arrange his pawns. seems to be 59 . . .1:t h3 + 60 ltJd3 �f6 + 6 1
I am not sure that my subsequent play �d2 hlh2 + 6 2 �d1 �g5, intending to
was entirely correct; during the game I penetrate with the king to c3.
spent some time moving back and forth 59 1':. h4 0-1
•••
searching for the right plan until sud White's defence collapses and he there
denly the correct idea became clear. fore resigned.
42 1:t f5 43 �c3 1:t h5 44 ltJd2 �c6 45
•••
ltJa4 �d5 51 ltJc3 + �e6 52 ltJdl c5! 53 This game was annotated in November
ltJe3 l:te4 54 ltJc4 1996 and published in various magazines.
It's curious that four of five games I
won in Tilburg finished in the endgame.
Three of them I consider interesting and
B offer for the reader's consideration. The
technique I demonstrated in these games
is very far from perfect (although against
Adams I played rather well) but I think
that after analysing them I penetrated a
little bit further into the secrets of these
endgames.
1 e4 e5 2 lbf3 lbc6 3 �b5 a6 4 i.a4
ltJf6 5 0-0 �c5!?
When playing the Ruy Lopez with
Black, I have been the constant follower
of Vladimir Malaniuk. I have followed the
54 �e7!
••• Ukrainian grandmaster in using both the
172 Fire on Board
' normal Arkhangelsk' and then the 'Ark Logical but slow. 16 g4! was better.
hangelsk with 6 . . . .i.c5 ' . Now, having seen 16 l:.a8! 17 :b1
•••
that both Malaniuk and Onischuk had Both 17 :xa8 iVxa8 18 g4?! bxc3 19 bxc3
played 5 . . . �c5 at the Olympiad, I decided 'iVa1 + ! 20 <it'g2 ltJd8 and 1 7 :an bxc3 18
to give it a try. bxc3 l:.a3 yield Black enough counterplay.
6 c3 b5 7 �b3 . 17 ltJa5! 18 �a4
..•
The game has transposed to 6. . . .i.c5, Not 18 cxb4?! ltJxb3 19 ttJxb3 'ifb8 with
but this move-order avoids the line 7 a4. an edge for Black.
7 d6 8 a4 �g4
..• 18....:b8!
I was surprised to learn afterwards After 18 . .. c5?! 19 c4! the knight on a5 is
that this logical move is new. 8 . . . �b7 has trapped offside.
been played before, but I didn't like it be 19 l:.a1
cause after 9 d4 �b6 White can try to find After 19 c4 b3! Black has counterplay.
something more useful than 10 .:el. 19 bxc3 20 bxc3 l:.b2!
••.
32 �g7?
•••
33 �d3
Now White is again just a little worse.
33 ... �e6 34 'iVb2 'iVc7 35 �c2 h6! 36 B
<it e2?
Another error. The king has nothing to
do in the centre and now Black's queen
has a chance to penetrate. 36 'iVb4 would
be correct, meeting 36 . . . �g5 by 37 �d2.
36 �g5! 37 �xg5
•••
This should have been the decisive ma Once again Black is completely win
noeuvre. ning and this time there are no mistakes.
174 Fire on Board
73 c5 g4 74 c6 g3 75 c7 g2 76 e4 'iVf1!
88 'ti'd5 'ti'f3 + 89 <it'e1 'iie3 + 0-1 Van Wely continued 10 . . . b5 and Polgar
10 . . . ttJc6. The text (with a different move
Game 79 order) occurred in Shirov-Anand, Buenos
Aires 1994. Then I continued 1 1 g5 ttJfd7
S h i rov - J. Po l g a r 12 i.h5 !? g6 13 i.g4 and won a compli
Tilburg 1996 cated battle. Nowadays 12 i.d3 seems
more critical (instead of 12 i.h5) and this
The game was annotated in November was the way Sutovsky played his games.
1996 and published in New In Chess. Why didn't I follow his example? Because
Before the present game I had lost two I remembered that during my game
games in a row and had dropped from the against Anand I was seriously considering
lead almost to the middle of the field. A the alternative . . .
free day helped me to relax, but who 11 f5!?
knows what the story would have been if . . .but then I decided not to play it. This
Judit hadn't fallen into some of my old time I saw less objection to the idea.
analysis. Anyway, I am satisfied that after 11. i.f8?
••
so many setbacks (after this game I also This is what Anand suggested in our
lost to Piket) I still managed to finish the post-mortem but later the same evening I
tournament in third place by winning in found a crushing piece sacrifice. Of course,
the last round. I kept my discovery more or less secret,
1 e4 c5 2 ttJf3 d6 3 d4 cxd4 4 ttJxd4 but when I learned that it had occurred in
ttJf6 5 ttJc3 a6 6 i.e2 the rapid game Wolff-Rychagov (Hastings
Mter some bad games with 6 i.e3 I de PCA qualifier, 1995), I no longer expected
cided to switch to a 'less aggressive' set-up. to be able to play it myself. Fortunately
6 e6 7 0-0 i.e7 8 f4 0-0 9 i.e3 'ti'c7
••• the Wolff game didn't get into the main
(D) databases.
A provocative line. As Vishy Anand once We shall see that the text move just
stated there is nothing to be expected loses. As I remember, the main reason why
Shirov - J. Polgar, Tilburg 1 996 175
( 15 . . . lbe5 16 �f7 + �h8 1 7 �xe8 iVxe8 18 A long time ago Vladimir Bagirov taught
lbd5 wins) 16 �f7 + <ith8 17 �xe8 'iVxe8 me (I believe that he himself was taught it
1 8 �xc5 dxc5 19 tbd5 J:.a7 20 'iVh5 g6 2 1 by Vladimir Makogonov) about brilliancy
iVf3 �g7 2 2 e7 �g8 2 3 'ti'b3 �h8 24 lbf6, in chess - 'A brilliancy is when you attack
and Black's fate is rather similar after a piece and there is nothing to defend it'.
14 . . . :Xe7 15 fxe6 tbb6 16 iVf3 'ti'd8 17 �xb6 This is exactly the case with the rook on h 7;
'iVxb6 + 18 :If2. once again the co-ordination of Black's
15 fxg6 l:Ie7 16 lbd5! pieces speaks for itself.
I don 't know whether White has other However White had a second winning
ways to win the game but this one is sim line which would have been even better:
ple and effective. 19 g6! :Ig7 ( 19 . . J�e7 20 �g4 wins) 20 �g4
16 exd5 17 'iVxd5 + �h8 (D)
••• :Ixg6 2 1 �d4 + ! (I considered the move
18 gxh7! 19 g6 during the game but I probably
After long thought, I didn't go in for overlooked this very important check)
another sacrifice. In fact 18 l:Ixf8 + ?! lbxf8 2l. . . lbf6 (2 l . . . lbe5 22 1hf8 + <ith7 23 �xe5
19 l:Ifl would probably also have won but l:Ixg4 + 24 <ith l dxe5 25 :If7 + l:Ig7 26 :Ixc7
I preferred to keep it simple. The following J:.xc7 2 7 'iVxe5 wins) 22 :Ixf6 :Ixf6 23
176 Fire on Board
23 g4 l!e7!
With this precise move Black controls
the d7-point and should hold the game.
24 d5
24 g5 tbh7 25 h4 l:tg8 leads nowhere.
24 l!c7 25 c3 tbh7 26 �h2 c5 27 h4
•••
15 gxf6 16 liJd1!
•••
30 g5 hxg5 31 hxg5 l:.d8 32 .i.f5 lLlf8 al) 44 . . . :b7 45 'ii' h 5 c4 46 'ii'e2 .U.ec7
33 'ii'h6 + �g8 34 'ii'f6 l:Idd7 ( 46 . . . l:Ibc7 4 7 bxc4 l:te8 48 c5 I:.xc5 49 g7
A sad necessity. wins) 47 'ii'g4 l:tb8 48 'ii'e6 + �h8 49 'ii'xd6
35 �g3 lbh7 36 .i.xh7 + �xh7 37 �g4 J:lbc8 50 bxc4 with an easy win.
b5 38 'ii'h6 + �g8 39 �f5 f6 a2) 44 . . . c4 45 bxc4 b3 (45 . . . l:tb7 46 c5
This desperate move is the only chance. dxc5 4 7 'ii' h3 l:te8 48 d6) 46 c5 b2 4 7 'ii'h l
l:tb7 48 c6 bl 'iW 49 'ii'xbl l:txbl 50 rJdxe7
winning the rook for the c-pawn.
b) 4 1 . . . c4!? (intending . . . a5-a4) 42 b3!
b4 (42 . . . a5 43 bxc4 bxc4 44 'ii'h 5 a4 45
'iid l l:.a7 46 'ii'e 2 l:Iac7 47 'iib2 l:Ib7 48 'iia3
and 42 . . . cxb3 43 axb3 a5 44 c4 bxc4 45
bxc4 a4 46 c5 both win for White) 43 bxc4
bxc3 44 'ii'h2 I:.e8 45 �g5 and again White
w1ns.
42 'iih5
Having thought for a long time, I found
the winning plan if Black plays passively.
Howeyer, I was still afraid there might be
some tricks.
42 l:Ig7 43 'ii'f5 l:tge7 44 �g5 I:.g7 45
•••
considerable practical problems had he Here, too, the pawn moves deserved at
played 4 1 . . . a5 or 4 l . . . c4 trying to block tention but probably would not have
the queenside. In both cases White would saved the game:
only be able to win with very accurate a) 47 . . . a5 48 a4 bxa4 49 'ii'e6 + ! (49
play - here is a brief analysis of those 'ii'fl ?? l:txg6 + 50 �xg6 :g7 + is a draw)
tries: 49 . . .�f8 50 'iWg4 and wins.
a) 41. .. a5!? 42 a4! (not 42 'ti'h5? a4 and b) 4 7 . . . c4 and now:
the blockade is set up) 42 . . . b4 (42 . . . bxa4 bl) 48 b3? b4 49 'ii'fl ( 49 'ii'e6 + �f8 50
43 iih5 a3 44 bxa3 a4 45 �dl l:ta7 46 'ii'e2 'ii'f5 + �e8 ! is unclear) 49 . . Jlxg6 + ! 50
wins) 43 cxb4 axb4 (or 43 . . . cxb4 44 'ti'h5 �xg6 :g7 + 5 1 �h5 I:.h7 + 52 �g4 l:.g7 +
followed by 'ii'e2, penetrating on the queen 53 �f3 l:tf7 + 54 �e2 l:Ixfl 55 �xfl bxc3
side) 44 b3! and now: 56 �e2 cxb3 5 7 axb3 �f7 58 �d3 �e7 59
Shirov - Kasparov, Erevan Olympiad 1996 1 79
�xc3 �d7 60 Wb4 �c7 6 1 �a5 �b7 62 b4 Black resigned in view of the obvious
�a7 63 b5 axb5 64 �xb5 �b7 with a draw. 60 . . J:Id8 6 1 'il'e6 + �f8 62 'iWf6 + 'ite8 63
b2) 48 'iWe6 + ! �f8 49 'iWf5 + Wg8 (alter g7.
natively, 49 . . . We8 50 'iWf6 wins) 50 b3 b4
5 1 bxc4 bxc3 52 c5 c2 53 c6 ! c1 'iW 54 cxd7 Game 81
iYxe3 + 5 5 Wh5 'iWe2 + 56 �g5 and White
wins. S h i rov - Ka s p a rov
48 c4! Erevan Olympiad 1996
Now I was again confident that White
has a winning position. The threat is 49 The game was annotated in October 1996
b4! . and published in various magazines.
48 b4
••. Mter the game Kasparov stated that I
48 . . . bxc4 would still call for precise saved this game by a long series of only
play, e.g. : 49 'iWfl :h7 + 50 Wg5 ! (not 50 moves. I thought exactly the same about
gxh7+?? lhh7+ and Black draws) 50 ... l:Ih2 him. Who is right? Probably both.
5 1 'iWf5 ! :e7 52 Wf6 l:Ie8 53 'ii'd 7 l':.f8 + 54 1 e4 c5 2 l2Jf3 d6 3 i..b5 +
�e6 l':.h6 55 �xd6 :xg6 + 56 Wxe5 and This made Kasparov think for a while.
White wins since he has too many passed I remember that some time ago, having
pawns. lost to Ivanchuk, he stated that White
49 'il'f3 :b7 only plays 3 i..b 5 + when he wants a draw.
Or 49 . . . :h7 + 50 Wg5 and wins. I wouldn't be so sure . . .
50 'ii'd l :bs 51 'ii'a4 :bs 52 b3! 3 i..d7
•••
after the logical continuatioh 12 . . . ltJh6 1 3 19 . . . 'iVxc4 20 ltJxf6 seems very attrac
ltJb3 ltJxb3 14 axb3 ltJf7 15 f4 0-0 i s quite tive for White.
complicated and appears about equal to 20 l:te1 'ti'g6
me.
12 JWa4! 13 f4 ltJh6 14 e5 0-0-0
.•
21 fxg5
Having just twenty minutes left to
reach the time-control at move 40, I re
jected the tempting 2 1 f5 because the po
sition after 2 l . . .'ifg7 (2 1 . . .'iff7 22 ltJxg5
fxg5 23 i.xg5 l::tdg8 is also unclear) 22
ltJg3 ltJxc4 23 ltJh5 'ti'f7 24 l:te6 ltJe5 didn't
15 ltJf5?!
••• appeal to me at all, for example 25 h3
After the game Kasparov was very criti l::t h e8! or 25 ltJxf6 h5! 26 .:txe5 (26 gxh5?
cal of this move and suggested 15 . . . .l:thf8 ! l::t h 6 wins for Black) 26 . . .'ti'xf6 27 l::te6
instead. H e was completely correct in that 'ti'xc3 and Black is much better. However,
after 16 e6 f5 !, intending to bring the White has a stronger line, namely 25 'ti'e2!
knight to e4 via f6, Black has the advan l:Ihe8 26 'ti'e4! and his compensation for
tage. the pawn seems quite good. Thus 2 1 f5
16 g4 would have been a real, though risky, win
Risky but necessary. 16 e6 h5 is very ning try while the text leads to a forced
annoying for White as he can not do any draw.
thing against the plan . . . r:Jilc7, . . . .:tb8, . . . a6, 21. l;Ihe8!
••
22 ltJxd6 + ! .:txd6 2 3 .:txe8 + 'ti'xe8 24 Black, but after the text it's White who
�f4! has to be careful.
24 iYxh4 'ti'e4 25 .:tal ltJxc4 can only fa
vour Black. Having made the intermedi
ary bishop move I thought I was on top
but Kasparov had a surprise ready.
24 ltJxc4!
•••
Game 82
Ye J i a n g c h u a n - S h i rov
Erevan Olympiad 1996
The game was annotated in October 1996
and published in various magazines.
It's not a bad feeling to beat the first
Here I was completely sure that my op board of the team representing a nation
ponent had miscalculated something and of 1.2 billion, is it?
that the game was over. But . . . 1 e4 c5 2 ltJf3 d6 3 d4 ltJf6 4 lL\c3
26 'ti'e4!!
••• cxd4 5 ltJxd4 a6
26 . . . ltJhf3 + 2 7 �hl 'ti'e4 28 'ti'g2! or Having played the white side of the
26 . . . ltJdf3 + 2 7 �fl would simply lose for Najdorf several times against Gelfand, I
182 Fire on Board
Here I realised that I had done some The most solid and logical continu
thing wrong. White's threats seemed very ation. White's reply is forced.
dangerous and after due consideration I 15 .1:d3
decided to take radical measures. 15 �f3 .1:xc3 16 bxc3 �xe4 looks very
11. h6?!
•• dangerous.
The theoretical move is 1 1 . . . d5 , al 15 . 0-0 16 �f3 tiJf7 17 t£Jh4?!
. .
only move) 27 liJf8 + �h8 28 liJg6 + (not The only move, but sufficient to parry
28 lbxd7? liJxe1) 28 . . . �h7 29 4Jf8 + with a White's threats.
draw. It was possible to try 25 . . . i.f6 26 0-1
4 Th e Botvi n n ik variat i o n
Although I have done a lot of deep open time limit in serious competitions) that I
ing analyses in my life, I still cannot con have played in the Botvinnik variation,
sider myself a strong theoretician. This arranged in chronological order, plus a
doesn't, however, mean that I don't like few others. I will also explain something
working on theory. And if you were to ask about the history of the line, but the reader
me which opening I like to investigate should forgive me if I sometimes reduce
more than any other, I wouldn't hesitate my commentaries to brief descriptions -
for even a second - the Botvinnik vari the last word on the Botvinnik variation
ation ! To find the truth in this opening hasn't been said yet and I might still play
one needs to analyse certain lines very it with either colour, who knows . . .
deeply and always make very cool assess My first experience with this opening
ments because many of the positions go came in January 1987. I was playing in
against standard chess principles. the Soviet Junior Championship and at
The complexity of the Botvinnik vari that time I had just switched from 1 e4 to 1
ation has always greatly attracted me, d4. When I was to due to play Gata Kam
and it was always traditional to analyse it sky with White, I found out that he some
in Latvia since Tal often played it. Of times goes for the Botvinnik. Of course, I
course Bagirov was also an expert on it - didn't really know any theory then, but
after all it was he who had to oppose Polu after looking at a few Informators I de
gayevsky's famous idea in lengthy analy cided to join the fight.
ses (Bagirov used to be Polugayevsky' s
main trainer) before it was played against
Torre in Moscow 198 1.
S h i rov - Ka m s ky
I should also mention that Alexander USSR Junior Championship,
Shabalov (now living in the USA) played Kapsukas 1987
many exciting games in the Botvinnik
variation and several other Latvian play These annotations were made during the
ers participated in theoretical analyses. preparation of this book.
In fact the trio Tal-Bagirov-Shabalov 1 d4 d5 2 c4 c6 3 ttJc3 ttJf6 4 ttJf3 e6 5
were making valuable discoveries as early i.g5 dxc4 6 e4 b5 7 e5 h6 8 i.h4 g5 9
as the beginning of the 1980s; the game ttJxg5 hxg5 10 i.xg5 ttJbd7
Kasparov-Tal (USSR 1983) was one of the
fruits of their joint work. Their other im
portant discovery, in 1984(!), was . . . l:.d4! ! ,
which came into practice only in the game w
ber 1989 and first appeared in Shakh Now I don't see any compensation for
maty Riga. the two pawns.
1 c4 c6 2 d4 d5 3 lLlc3 ltJf6 4 lLlf3 e6 5 14 �e2
�g5 dxc4 6 e4 b5 7 a4 �b4 8 e5 h6 9 [AS - 14 dxc5 also seem to be losing for
�h4 White in view of 1 4. . . i.xc3 1 5 bxc3 lLld7 1 6
9 exf6 hxg5 10 fxg7 .1:g8 1 1 g3 is more 'iV{3 'iVxf3 1 7 lLlxf3 lLlxc5 and Black is win
promising. ning.]
[AS - Recent tournament practice has 14 ltJd7 15 �f3
•••
demonstrated that 11 h4 (instead of1 1 g3) 15 'iVf3 can be met by 15 . . . 'iVxf3 16 �xf3
1 1 . . . g4 12 lLle5 �g7 1 3 h5 is probably lLlxe5 17 �xa8 (or 17 dxe5 fibS) 17... lLld3 + ,
White 's best try. In fact this position looks with advantage to Black.
too risky for Black, but ofcourse 7. . . �b4 is 15 cxd4! 16 lLlxd7 �xd7 17 �xa8
..•
not obligatory.]
9 g5 10 exf6?!
•••
resistance. But now on 18 'ii'e2 Black has playing the King's Indian as well.) I had a
the unpleasant 18 . . . d3. lot of exciting games in the Meran system
18 bxc3 d3 19 l:Ixd3 (in which White plays 5 e3 instead of 5
Mter using up nearly all of his remain .ig5), but I wanted another kind of blood!
ing time, White was unable to find any Shabalov and I did not ignore the fashion
thing better. 19 <it'b2 is met by the decisive for the Meran system and in 199 1 we
19 . . . b4 20 l:Ic1 'ii'xf2 + 2 1 <itb1 b3, and oth started working on a crazy line (of course,
erwise there is no defence against the I mean the variation 5 e3 lbbd7 6 '+i'c2
threat 19 . . . 'ii'xc3 + , 20 . . . 'ii'b3 + and 2 1 . . . c3. .id6 7 g4!?) which later became known in
19 cxd3 20 'ifc5 'ii'f4+ 21 <it'b2 'ii'xa4
••• Russia as the 'Shabalov-Shirov Gambit'.
22 .if3 However, at the beginning of our investi
White cannot save the game by 22 .ic6 gations I felt quite sceptical about it, so it
'ii'c 2 + 23 <ita1 .ixc6 24 'ii'xc6 + <itf8 25 was no wonder that when I had the choice
'ii'c 5 + �g8 and Black has an overwhelm in August 1991 I still went for the Botvin
ing advantage. nik variation with White.
22 'ii'c2 + 23 <ita1 d2
•••
S h i rov - Stis i s
London 199 1
These annotations were made during the
preparation of this book.
1 d4 d5 2 c4 e6 3 lbc3 c6 4 lbf3 lbf6 5
.ig5 dxc4 6 e4 b5 7 e5 h6 8 .ih4 g5 9
lbxg5 hxg5 10 .ixg5 lbbd7 11 g3
When this game was played I consid
ered this move to be more precise than 1 1
exf6, which allows the lines 1 1 . . . .ib7 12
g3 c5 13 d5 .ih6, which we have already
seen and will see again later in this book,
and 13 . . . lbxf6.
24 l:Ib1?!
A slightly tougher defence was offered
by 24 l:Id1 , after which I was intending
24 . . . l:Ig8 ! 25 'ii'd4 'ifc1 + 26 <ita2 e5 ! 2 7
'ii'xd2 .ie6 + 2 8 .id5 'ii'xd2 + 29 l'hd2 1hg2,
when the ending is hopeless for White.
24 l:Ig8 25 'ii'd4 l:Ig5 26 l:Id1
•••
not popular and nobody really wanted to Not a pleasant surprise. I knew that
go in for it. (Of course, the Semi-Slav was this move was not so bad as considered by
not my only opening against 1 d4 - I was the theory of those times, but I had never
188 Fire on Board
studied it seriously before this game. A afraid of losing the c6 pawn - it is more
few months later I suggested to Kramnik important to exchange the light-squared
that the line was worth a try and he bishops.
started working on it as well. In Linares 20 i.xc6?!
1993 he played 1 1.. Jig8 against Beliavsky Natural but not the best. 20 f4, first
and won. played by Bareyev against San Segundo
12 h4 (Madrid 1995), is much more ambitious.
The most ambitious and critical con (Another important game with this move
tinuation. The line 12 i..xf6 lbxf6 13 exf6 is Mecking-San Segundo, Linares (open)
'ii'x f6 14 i.. g2 i..b 7 is equal according to 1995.)
theory. 20...i..b7 21 :h7+
12 �xg5 13 hxg5 lbd5 14 g6 fxg6 15
••• I still didn't smell any danger as I
'tWg4 'tWe7 thought that a rook and two pawns would
So far everything seems forced. Moves in the end be stronger than two active
like 15 . . . 'ii'a5 have been severely punished bishops. In any case I didn 't like 2 1 i.. xb7
in the past. .:xb7 at all, since as soon as Black gets his
16 'ii'xg6 + knight to d5 his position will be extremely
At that time I didn't consider any alter pleasant.
natives to capturing the pawn. Both 16 21. �g6 22 :xd7 i..xc6 23 �a7 i..f3!
••
i.. g2 and 16 :h8 later came into vogue This came as a shock. I was mainly
and the reader also will see them in this counting on 23 . . .b4 24 �d2 bxc3 + 25 �xc3
book. i.. d 5 26 :ra4, with a slight pull, whereas
16 .'iff7 17 'ii'xf7 + �xf7 18 i.. g2
•• after 23 . . . i..f3 it becomes clear that White
Here my knowledge ended. I wasn't is in danger. He is practically forced to
sure whether White was really better but move his king away since Black threatens
still felt optimistic. 24 . . . i..h 6, 25 . . . i.g5 and 26 . . . .l::t h 8.
18...lbxc3! 24 �d2?!
A novelty at that time. Other moves are However, 24 :a6! �f5 25 �d2 was more
clearly worse. precise, when Black is only slightly better.
19 bxc3 Now Black seizes a clear advantage.
19 i..xc6 �b8 20 bxc3 would ultimately 24 ...i..h6 + 25 �c2 :fs! 26 �b2 :f5?!
lead to the same thing. This appears to be a bit too slow. Cor
19 .:b8!
••. rect would have been 26 . . . i.. d2 ! 27 �c2
This is the idea. Black wants to play (27 a3 i.. d 5) 2 7 . . . i.. g5 with the idea of
. . . c6-c5 as soon as possible and he is not 28 ... .l::t h8.
Shirov - Oll, Tilburg (rapidplay) 1 992 189
S h i rov - 0 1 1
Tilburg (rapidplay) 1992
This was the second rapid play-off game
in a knock-out match. An exciting tourna
ment situation - but the game itself is
practically not worth annotating. Lembit
Oll, who is generally a great expert in
opening theory, couldn't stand the ten
28 i.d5?!
••• sion and played carelessly quickly at the
After the game my opponent admitted most critical moment.
that at this point he was still playing for a 1 d4 ltJf6 2 c4 e6 3 ltJf3 d5 4 ltJc3 c6 5
win. In my opinion, by now White is not i.g5 dxc4 6 e4 b5 7 e5 h6 8 i.h4 g5 9
worse and Black should have preferred ltJxg5 hxg5 10 i.xg5 ltJbd7 11 exf6
28 . . . i.c6! 29 f4 (or 29 :Ixc4 :Ixf2 + 30 <ita1 i.b7 12 g3 c5 13 d5 'ti'b6
l':.fl + 3 1 <itb2 i.c1 + 32 �b3 i.d5 33 :Id7)
29 . . . i.xa4 30 :Ixa4 :Ih5 31 :xc4 :Ih2 + 32
<itb3 lig2 and the game is drawn.
29 f4 :Ih5 30 :14a6! :Ih2 +
Black couldn't prevent the white king
from escaping, since 30 . . . i.f8? would have
failed to 3 1 g4 :Ih2 + 32 <itcl .
3 1 <ita3 :Ic2?
Probably the decisive mistake. Black
could still have reached a draw by playing
3 1 . . . i.f8 + 32 :Id6 (32 <ita4 :Ia2 + 33 <itb5
:Ib2 + 34 <ita5 :Ib3 35 f5 + �xf5 36 :If7 +
�e4 3 7 :Ixf8 :Ixc3 38 :If4 + <itd3 can only
be dangerous for White) 32 . . .:Ih8! 33 <itb4
<itf5.
32 �b4 i.f8 + 33 :Id6! Undoubtedly the most complex branch
Simple and powerful. If Black takes the of the Botvinnik variation, and the one
rook he won't be able to stop the pawn, that most players have focused their at
but otherwise he will soon lose his bishop tention on.
since White threatens 34 f5 + . The rest is 14 i.g2 0-0-0 15 0-0 b4 16 ltJa4 'ti'b5
agony. In 1992 the move 16 . . . 'iVa6 hadn't yet
33 11b2 + 34 <itc5 l1b8 35 g4! lieS+ 36
.•• been seen.
�b5 i.xd6 37 exd6 i.f3 38 g5 i.d5 39 17 a3 ltJbS 18 axb4 cxb4 19 'iVd4!?
d7 l1d8 40 �c5 i.e4 41 �d6 <itf7 42 l1c7 I had analysed 19 'i¥g4 a lot over the
i.f5 43 g6 + 1-0 years, but when this game was played I
Although I won this game I wasn't very wasn't convinced that it led to an advan
happy about the opening, because my tage for White (today theory claims that
much lower-rated opponent has equalized White is better in that line). On the other
190 Fire on Board
19 J�xd5
••
was different. Perhaps this helped me to which continued 23 . . .'iVd5 24 'iVe2 lbxf3 !
win, who knows . . .] 25 .:e8 + (or 25 �xf3 .:e8) 25 . . . l:txe8 26
17...i.xg2 18 �xg2 'ti'c6 + 'iVxe8 + �c7 and Black soon won. In my
Black now picks up the pawn on e6. In opinion the best continuation is
the game Aseyev-Bagirov, Helsinki Open c) 23 :e2, with the possible sequel
1992, White short-circuited himself here 23 . . . 'ti'd5 24 'iVf5 + �c7, and now both 25
with 19 'ti'f3?? (intending 19 . . . ifxe6 20 i.f4 :es and 25 .:fe1!? tbd3 26 Ile7 +
'ti'a8 + ) 19 . . . .:h2 + , winning the queen. i.xe7 2 7 .:xe7+ �c6 28 'ti'xd5 + :xd5 29
19 f3 'iVxe6 20 'iVc2 tbe5 21 :ae1 llxf7 lead to unclear play. If Black doesn't
want to take any risks he can force per
petual check in the last variation with
29 . . .lle8 30 l:.xa7 (there is nothing better)
B 30 . . J�e2 + 3 1 �h3 tbf2 + , etc.
23 ...'iVd5
21..J�d4!?
Tal and Shabalov had discovered this
move as long ago as 1983, but it wasn't
tried out in practice until nine years later,
in the game Bareyev-Oll, Moscow Speed
Tournament 1992 . [AS - Here I should 24 'iff5 + �c7 25 Ile2
mention that as the representative of Esto Now 25 i.e3? is bad due to 25 . . . tbd3,
nia, which is a neighbour to Latvia, Lem and 25 i.f4 is met by 25 . . . :xf4! 26 'iVxf4
bit Oll has always had close chess contact tbd3 27 l:.e7 + (27 'ti'e4 lbxe1 + 28 .:xe1
w ith Shabalov and I.] Belyavsky-Piket, �c6! is good for Black) 27 . . . �c6 28 'iVe4
Amsterdam 1989, went 2l. .. 'iVh3 + 22 �g1 :g8 ! ! 29 g4 i.xe7 30 fxe7 (not 30 'ti'xe7?
tbd3 23 lle2 i.d6 24 'iVxc4 i.xg3 . White .:xg4 + 3 1 �h3 .:g8, threatening . . . tbf4 + )
then had the choice between perpetual 3 0 . . . 'ti'xe4 3 1 fxe4 :xg4 + 3 2 �f3 (or 32
check with 25 'iVa6 + (and 'iVb5 + , 'iVc6 + , �h3 l:.xe4) 32 . . . l:.g8, when in view of the
etc.) and the game continuation 25 'iVg4 + badly placed knight at a4 Black has the
'iVxg4 26 fxg4, with complicated play. advantage.
22 h4 i.d6 23 a3 25 ...�c6!! (D)
This is the first new move of the game, Steinitz understood the value of the
but it doesn't even guarantee White equal king as a strong piece, and in the present
play. Other tries are: game Black's monarch will perform ster
a) 23 i.e3? occurred in Rublevsky ling work. Nonetheless when I made this
Savchenko, Helsinki Open 1992, but it move I couldn't help remembering the
backfired after 23 . . . l:.dxh4! 24 gxh4 :xh4, game Shirov-Georgiev, Biel 1992, in which
when Black had a strong attack; my king didn't feel at all comfortable on
b) 23 :e4 was effectively punished in h5, and indeed eventually came to grief
the already mentioned game Bareyev-011 there. On the present occasion, however,
192 Fire on Board
;;.; ;;
i.f4) 32 . . . 'iVxf3 + 33 iVxf3 I;Ixf3 34 l:.e7, /!" " %
"/.? /' / / / ? /
With his last move before the time con tbxg5 hxg5 10 i.xg5 tbbd7 11 exf6 i.b7
trol, Nikolic throws away the draw he has 12 g3 c5 13 d5 i.h6
almost achieved and which can be reached, This move could hardly have surprised
as Ivan Sokolov pointed out in the post Yusupov, since he had played it himself
mortem, by 40 <ifi>f3 ! :Ih3 4 1 <it>g4 .l:txg3 + against Bagirov as long ago as in 198 1 !
42 �h5 .:h3 + (42 . . . �f5 43 :Ia5 + i.e5 44 14 i.xh6 :xh6 15 �d2
<ifi>h6 ! is also a draw) 43 <it>g4 :Ixb3 44 .:a5 ! Bagirov continued 15 i.g2, the main
Black cannot improve his position. White line at that time, but it doesn't yield White
answers . . . i.e5 with lateral rook checks, any advantage.
and if the rooks are exchanged, for exam 15...'�xf6 16 0-0-0
ple after . . . :Ie3-e5, White draws with <it>h5 As the reader may recall, I continued
and g5-g6, eliminating the last black 16 tLie4 against Kamsky, but Yusupov fol
pawn. lows a theoretical recommendation.
4o .:gs 41 .:g4
••• 16 i.xd5
•••
4 1 l1a5 doesn't save White because his Following our old Latvian analysis. Af
king is too far back and he is unable to ter the game Yusupov pointed out the
play g5-g6 to get rid of the last black move 16 . . . �f8!? and the new era of the
pawn: 4 l . . . i.e5 42 l:a6 + <ifi>f5 43 :a7 <it>g6! 13 . . . i.h6 line had begun.
followed by . . . .l:tb8. 17 tbxd5 exd5
4l. i.e5 0-1
•• Here Yusupov sank into deep thought
Since . . . �f5 will be decisive. and I realized that Black's position was
In 1993 I again had a 'break' from the not at all as bright as it had seemed in
Botvinnik system with White, because I 1987.
had found some new ideas in Meran with
7 g4 (which is actually no less exciting!)
and later I won two games with it. How
ever, the Botvinnik still remained one of
my main weapons with Black. The next
game was played two months later in the
last round at Linares 1993, and it was
very important from a sporting perspec
tive - if I could win I would catch Karpov
and Anand and lie behind only Kasparov:
Before the Linares tournament I had
found out that the course of the Wijk aan
Zee game between I. Sokolov and 011 was
probably losing for Black (since confirmed
by tournament practice), so I decided to
switch to an already almost forgotten 18 .i.g2
line, which as the reader will recall, I had I was quite afraid of 18 �xd5 l:.d8
analysed with Tal and Bagirov in 1987. ( 18 . . . tLib6 19 l1e1 + �f8 20 �xc5 + �g8 21
'iVxb5 ! , intending 'ife5, which is slightly
better for White) 19 f4, but the game Gof
Yu s u pov - S h i rov stein-Kacheishvili, played one year later
Linares 1993 in Groningen 1993, showed that this is
probably not all that dangerous for Black.
These annotations were made during the That game went 19 . . . tbb6 20 llel + <it>f8 2 1
preparation of this book. �xc5 + 'iVd6 2 2 'iVxd6 + l:Ihxd6 2 3 i.e2 c3 !
1 d4 d5 2 c4 c6 3 tbf3 tbf6 4 tbc3 e6 5 24 l:.d l (24 bxc3? tba4 25 l:.d1 tbxc3 26
i.g5 dxc4 6 e4 b5 7 e5 h6 8 i.h4 g5 9 llxd6 llxd6 gives Black a slight pull)
194 Fire on Board
Now 20 �xd5? would have been wrong 29 . . . l:le5 !? 30 �b2 b4 would have been
in view of 20 . . . l2Ja4! with an attack. another winning try, but I wanted to
20 �g8?!
••• avoid any possible risk.
This appears to be a serious mistake. 30 �d3 l:l2e5 31 �b2 b4 32 l:ta1 a5
There is just one difference between this 33 �c2!
move and 20 . . . �g7, but it seems to be a Despite his time-trouble Yusupov de
very important one. The reader will un fends very well.
derstand what I mean eighteen(!) moves 33 l:ta6 34 f4 l:le3 35 l:ld4!
.•.
42 lie4 %-%
Here we agreed a draw, which meant
that I finished fourth since Karpov and w
Anand had also drawn. Nevertheless, this
was a great success for me.
I again had to wait a long time for my
next opportunity to play the Botvinnik,
because I had another pet line with Black
(the Slav with 4 . . . a6) and I was still em
ploying the Meran with White. In October
1993, at the World Team Championship
in Lucerne, I saw Kamsky defeat Kram
nik in the variation 5 i.. g5 h6 (instead of
5 . . . dxc4). A few days later I had Black
against him and I suspected that he would 17 f4
feel confident in the Botvinnik with White. 1 7 g4 was played in the game Piket
I didn 't prepare much for the game, but I Kaidanov, Tilburg 1993, and Black was
remembered Yusupov's suggestion in Lin OK. Today the most critical continuation
ares and thought it would be great to try is 1 7 f3 . One can find some games with
it. this on a database.
The game appeared so impressive to 17...lL!b6! 18 i.. g2
other players that at Tilburg 1993, which 18 lbe4 'ii'g7 worked out well for Black
started only two weeks after Lucerne had in the game E . Vladimirov-Bareev, Til
finished, the move 16 . . . <it>f8 occurred no burg 1993 . Perhaps 18 g4 was still worth
less than three times! trying.
18 exd5 19 'ii'f2
•••
play it against the same opponent after Accuracy is the most important thing
six and a half years but with the colours now. Variations such as 19 . . . d4?! 20 i..xb7
reversed. Like Yusupov, Kamsky didn't l:lb8 2 1 lL!e4 'ii'e 7 22 f5 ! would have yielded
wish to try 16 lbe4. White unnecessary counterplay.
16 0-0-0 �f8! 20 ttJxb5?!
A novelty which later on was recog After this I consider White's position to
nized the best in lnformator57, though be lost. After long thought Kamsky re
of course it belongs to Artur Yusupov. jected both ways of capturing the d-pawn,
196 Fire on Board
and it's true that neither 20 .i.xd5 lbxd5 25 l:.f3 d4 and 24 �d2 lbb6 25 lbb1 lbc4 +
2 1 lbxd5 �a6, threatening 22 . . . �xa2 and both win for Black) 24 . . J:b8! (with the
22 . . . l:Id6 ; nor 20 lbxd5 .i.xd5 2 1 i..xd5 c3 ! idea of . . . i.. a8, . . . l:Ihb6) and it seems to me
with the idea of 22 b3 c4! 23 bxc4? c2 ! 24 that there is no defence, for instance 25
iixc2 lbxc4 offers White much hope; but i.. xd5 (if White continues 25 l:Ihe1 i.. a8!
probably his best chance was still 20 g4. 26 l:Ie3, intending to meet 26 . . . l:thb6 with
20 lb a4!
••• 27 lbd2 and 'ii' h 7, then the simplest is
Threatening 2 1 . . . c3. 26 . . . d4! 2 7 cxd4 i.. xg2 28 �xg2 cxd4)
2 1 iVc2 25 . . . i..xd5 26 l:Ixd5 l:Ixb1 + 27 �xb1 �xc3 +
Both 2 1 lba3 and 2 1 lbxa7 would have 28 iVc2 (28 �d1 �f3 + ) 28 . . . 'ii'a1 + 29 iVb1
been answered by 2 1 . . J:tb8 ! , and the line 'iVxb1 + 30 �xb1 lbc3 + and Black wins.
2 1 i.. xd5 i..xd5 22 l:Ixd5 iVc6 23 lbxa7 23 lbxb2 24 �f5
•••
�xd5 24 l:.d1 �b 7 25 lbxc8 �xeS also ap I had to see precisely the line 24 i.. xb7
pears hopeless for White. �xa3! 25 �b3 (25 i..xc8 lbd3 + [25 . . . lbc4+
26 �b1 l:.b6 + 27 �a1 'ti'b2 + 28 'ti'xb2
cxb2 + 29 �b1 lba3 mate, as given by the
Swiss IM Beat Ziiger is even more accu
rate] 26 �b1 l:.b6 + 2 7 'it>a1 :b2) 25 . . . �xb3
26 axb3 l:c7 27 l:d8 + �e7 28 l:tb8 l:Ib6 29
�c2 c4! and Black wins.
24 l:If6 25 'ti'b7
•••
variations like 28 i.c6 + cst>xc6 29 �xf6 + [AS - The reader has already seen 1 8
�b5 and 28 i.e6+ cst>c6 29 i.d7+ (29 i.d5 + i.g2 in the game Shirov-Stisis . The next
Wb5) 29 . . . �c7 speak for themselves. game will feature another critical posi
28 cst>d6! 29 i.xb7 + l2Jxd1 + 30 cst>xd1
••• tion.]
�xa2 3 1 'ii'g2 �b1 + 0-1 18 l2Je4!? i.b4 + 19 cst>e2
Here Kamsky's flag fell, but of course Another surprise. If White doesn't want
he would have had to resign anyway be to exchange knights, he can play 19 cst>d 1
cause of 3 1 . . . �b1 + 32 cst>e2 lieS + . This in order to meet 19 . . . c3 with 20 b3.
game was awarded a special prize for the [AS - Probably the best answer to 1 9
best game of the tournament and was ob �d1 is 19 . . . c5.]
viously the golden moment for me in the 19 ••• c3! 20 bxc3
Botvinnik variation. However, for a while Now 20 b3? fails to 20 . . . i.a3 !
I continued to employ it with success. The 20 ...l2Jxc3 + 2 1 l2Jxc3 i.xc3 22 l:.d1
next game was played one month later, in
the German Bundesliga.
Lo b ro n - S h i rov
Bundesliga 1993/94
These annotations were made in Decem
ber 1993 and first appeared in the Ger
man magazine Schack 64.
1 d4 d5 2 c4 c6 3 lt:Jc3 lt:Jf6 4 l2Jf3 e6 5
i.g5 dxc4 6 e4 b5 7 e5 h6 8 i.h4 g5 9
l2Jxg5 hxg5 10 i.xg5 ltJbd7 1 1 g3
Less fashionable than 1 1 exf6 nowadays.
[AS - Nowadays 1 1 exf6 is still played
more often.] Here I saw what White had in mind.
1 1 . .:tg8! 12 h4 .:txg5 13 hxg5 l2Jd5
•• The attack with connected rooks is poten
1 4 g6 fxg6 15 'ii'g4 �e7 1 6 �xg6 + �f7 tially very dangerous for Black, especially
1 7 �xf7 + cst>xf7 (D) if White manages to realize the plan i.g2-
The critical position of the line with e4, f2-f4, g3-g4-g5 , etc. So I was obliged
1 1 . . . .:tg8 . Now White usually chooses be to react forcefully before my opponent's
tween 18 i.g2 and 18 0-0-0. Lobron's move pieces could become too active.
came as a surpnse. 22 ••• b4! 23 i.g2 i.a6 + 24 �e3 l:g8!
198 Fire on Board
Black must parry the threat of25 l::t h 7, i.. xf6 34 J:.xa7 i..b 2 ! , intending 25 . . . i.. a3,
so he doesn't mind exchanging a pair of Black is also winning, but 29 l::t h 7 + ! J:.g7
rooks. 30 l::t 1h6 (the same idea as in the variation
25 i.. xc6 t2Jb6 26 l::t h 7!) 30 . . . i..b2 3 1 �d3 would still have
yielded White some drawing chances.
28 i.. xa2 29 i.. d7 tbc7 30 l::tf6 + �e7
•••
3 1 i.. c6 b3 32 d5
The last try. 32 l::t h 7 + �d8 33 l::t d7 +
�c8 34 l::tff7 i.. a5 is curtains.
32 ••• exd5 +
Here 32 . . .b2 33 d6+ �d8 34 l::tf7 t2Jd5 35
.:.d 7 + �c8 36 .:.Xa7 b1 'iV + 3 7 .:txb1 i..xb1 +
38 �f3 l::tf8 + 39 �g4 (39 �g2 l::t xf2 + 40
�g1 l::ta2) 39 . . . i..f5 + 40 �g5 J:.g8 + 4 1 �h6
tbb6 with the idea 42 l::tc 7 + �d8 43 l::tb 7
i..xe5 44 .:.xb6 i..xd6 would also have been
winning, but the text doesn't spoil any
thing.
33 i..xd5 t2Jxd5 34 l::t a6!
The second critical position. White's At least this prevents the immediate
pawns are weak, so the two extra pawns . . .b3-b2.
don't fully compensate for the advantage 34 ... t2Jc7 35 l::txa7 �d8 36 f4 :Xg3??
of two pieces against a rook. Terrible. In time-trouble I forgot that
26 l::t h6? after 36 ...�c8 37 J:.h7 Black has 37 ... i..bl + ,
During the game I had the feeling that when the game i s completely over.
26 l::t h 7 + J:.g7 2 7 l::t h 8 might have been a 37 l:th8+ t2Je8 38 l::ta 8+ �d7 39 l::thxe8
more accurate way to play. Mter some b2
home analysis I would say that in the line
with 2 7 . . . i.. c4 (I see nothing better) 28
l::tdh1 (a more precise move order for White
is 27 l::tdh 1 i.. c4 28 l::t h8) 28 . . . i..xa2 29 l::t 1 h6
t2Jd5 + 30 i.. xd5 i..xd5 3 1 l::t c8 ! l::tg4 32 f3!
l::t xg3 33 l::tf6 + �e7 (or 33 . . . �g7 34 l::tc 7+
�h8 35 l::t f8 + l::tg8 36 l::tf6 l::tg3) 34 l::tc 7 +
�e8 35 l::tff 7! White is not in any danger of
losing, although he has nothing better
than a draw after 35 . . . a5 (36 .:th7 �d8 37
l::t hd7 + , etc.) .
2 6 i.c4 2 7 J:.dhl
•••
S h i rov - M o rovic
Fe rna n d ez
Las Palmas 1994
23 4Jb4
•••
My idea was to answer 24 ...i.xe4 with 3 7...4Jb6 38 l:.cc7 i.h6 39 l:.a6 is the
25 axb4! i.f5 (forced) 26 'i¥g5 4Jc8 (26...'ii'g7 same.
2 7 l:.ah 1 ; 26 ... 4Jc6 2 7 iff6! 'iVxf6 28 exf6, 38 .:c8 i.h6 39 .:g8+ �h5 40 .:h7 1-o
intending l:.a6) 2 7 ifh6! (threatening 28 This game was a kind of 'Pyrrhic vic
l:ta6 or 28l:ta5) 27 ...�e7 28 'ii'h4+ <itd7 29 tory' for me since after this my results in
'ii'f6! �e8 30 l:.a6! ifxf6 3 1 exf6 <iiff7 32 the Botvinnik system worsened. But the
.:c6, winning. reader will see that the opening was not
25 i.xd3 cxd3 26 .:d1 always the reason.
Simple chess. White wins a pawn, after In July 1994 I played a category 15 tour
which the two rooks and two pawns will nament in Pardubice (Czech Republic),
be clearly stronger than the three minor and my last round opponent was Alexan
pieces. der Khalifman, who was having a disas
26 iff5
••• trous tournament and had no special
The exchange of queens only favours ambitions in the last round. A draw in
White, in any case it wasn't easy to avoid some quiet opening would have sufficed
it, for instance 26 ...4Jd5 27 l':.xd3 �d7 28 to tie for first place, but for some reason I
'ii'h 3! i.e7 29 'iVh7 and 29 ...'ii'f5? is impos again chose the Botvinnik variation. The
sible because of 30 'i¥g8. opening worked out well but then ...
27 'i¥xf5 gxf5 28 l':.xd3 i.d5 29 l:.c3
tbc6 30 <ite3 �f7 31 l:.h7+! i.g7 32 l:.h2
�g6?! Khalifman - Shirov
Morovic doesn't put up much resis Pardubice 1994
tance in the final stages. He had to play
32 .. . �e8 with the idea of 33 .:Ihc2 <itd7, These annotations were made during the
but I think that after 33 g4! White should preparation of this book, based on my
win quite quickly. notes in lnformator 61.
33 l:.hc2 4Ja5 34 l':.c7 (D) 1 d4 d5 2 c4 c6 34Jf34Jf6 44Jc3 e6 5
Now it's curtains for Black. i.g5 dxc4 6 e4 b5 7 e5 h6 8 i.h4 g5 9
34 a6
•.• 4Jxg5 hxg5 10 i.xg5 4Jbd7 11 g3 .:.g8
34 ...i.h6 + 35 f4 tbc4 + 36 <ite2 i.e4 3 7 12 h4 l':.xg5 13 hxg54Jd5 14 g6 fxg6 15
l:tc3 ( 3 7 l':.2xc4 bxc4 3 8 l':.xc4) 3 7...a5 3 8 'i¥g4 'i¥e7 16 l':.h8?!
lvanchuk- Shirov, Novgorod 1994 201
1 7 dxe6 'ti'xe6 is also critical - see the 30 ... lbf4 + 31 �xf4 'ti'xd1 32 'ti'xd1 l:.xd1
next game. 33 h6 a5 34 g4 would have led to exactly
17...�xd5! the same thing.
This is the move that Shabalov told me 31 g4
about. 1 7 . . . b3, which had been played be 3 1 h6 lbf4 + 32 �xf4 'ti'xd1 33 'iVxd1
fore, isn't a serious idea. ltxd1 34 g4 is an alternative move order.
18 �xd5lbe5 19 axb4 3l. lbf4+ 32 �xf4 'ti'xd1 33 'ti'xd1
••
have been the last chance. This move came as quite a shock to me.
38 .i.e3 During the game I thought it must be a
White is winning now and the rest novelty.
needs no commentary. 17 dxe6
38 /it'b5 39 h7 .:h8 40 i.xd4 exd4 41
•• 1 7 i.f4 e5 18 i.g5 i.h6 didn't seem at
g6 fxg6 42 f7 �c6 43 �d3 �d7 44 l:.e2 all clear to me.
a4 45 l:.e8 axb3 46 :xh8 b2 47 l:ld8+ 1-0 17 'ifxe6 18 l:le1lbe5!
•••
After two such bitter losses it would In fact only this is the new move, but it
have been easy to completely lose confi is much better and more natural than
dence in the Botvinnik variation. How 18 ...'iff5 , as played in Agzamov-Gen. Ti
ever, six months later I was due to play an moschenko, USSR 1982.
eight-game match against Jeroen Piket, 19 'ifxd8+
who also employs the Botvinnik with both I rejected 19 i.xb7 + �xb7 20 'ifxd8 be
colours, and I needed to adjust my prepa cause of 20 . . . lbf3 + 2 1 �g2 'i¥c6 ! 22 l:le7 +
ration. However, the opening occurred i.xe7 2 3 'ifxh8 (23 'V}JJxe7 + �aS wins for
just once in the match. Black) 23 . . . lbxg5 + , followed by 24 . . .i.xf6
with a very strong attack for Black. In
stead of 20 'ifxd8, 20 'ife2 is correct, for
Shirov - Piket example 20 ...'V}JJh3 2 1 'tixe5 'tixh2 + 22 �fl
Aruba 1995 'ifh3 + 23 �e2 'ifg4 + 24 �fl with a draw.
19 ...�xd8 20 i.xb7
These annotations were made during the
preparation of this book, based on my
notes in lnformator 63.
Like yours truly, Jeroen has a great
deal of experience in the Botvinnik vari
ation with both colours. However, in the
third game of our match he had avoided 5
i.g5 as White, preferring 5 e3. Here was
another chance ...
1 d4 d5 2 c4 c6 3lbc3Lbf6 4lbf3 e6 5
i.g5 dxc4 6 e4 b5 7 e5 h6 8 i.h4 g5 9
lbxg5 hxg5 10 i.xg5 lbbd7 11 exf6
i.b7 12 g3 c5 13 d5 'ifb6 14 i.g2 0-0-0
15 0-0 b4 16lba4
204 Fire on Board
20 cJ;c7??
•••
22 'iVxa4 23 .i.xe5+
••• 44 'ii'f4+ �b6 45 .i.f3 1-0
Probably even better was 23 a3 ! .i.d6 If 45 ...'iVb1 then 46 .i.e4.
(23...'iVd7 24 axb4 wins for White) 24 axb4 For a long time I had expected Alexan
'iVxb4 25 .i.xe5 and White's attack should der Belyavsky to enter the Botvinnik vari
prove decisive. ation with White against me, since he is a
23 cJ;b6 24 :ad1
••• well-known expert on it. However four (!)
Now 24 a3 'iVd7 is less clear. times he chose other openings until he fi
24 l:Ih5 25 l:Id8 l:.xe5
••• nally went for a real discussion in Ljub
Black's last chance. ljana in December 1995.
26 l:Ixe5 c3 27 bxc3 bxc3 28 l:Ib8+?!
28 .i.e4 was a lot easier.
28 Wa5 29 l:xf8 'iVd1 + 30 .i.f1 'iVd6
•••
Belyavsky- Shirov
31 l:tfe8 Ljubljana (European
Technically more precise was 3 1 l:e7 c2 Club Cup) 1995
32l:txa7+ �b4 33 l:.fa8 c1'iV 34 l:Ia4+ Wc3
35 .:c4 + �d2 36 .:xc 1 cJ;xc1 37 :a4 �b2 The game was annotated in December
38 l:.f4, winning. 1995 but the present version appeared
31. c2 32 l:Ie1 'iVd1 33 .:8e7! cJ;b6 34
•• only in ChessBase Magazine 51.
h4 cHi 1 d4 d5 2 c4 c6 3 l2Jc3 l2Jf6 4 t2Jf3 e6 5
Desperation, but Black cannot do any .i.g5 dxc4 6 e4 b5 7 e5 h6 8 .i.h4 g5 9
thing. tt:Jxg5 hxg5 10 .i.xg5 ttJbd 7 11 exf6
35 l:txd1 'iVxd1 36 l:txf7 c4 37 l:te7 c3 .i.b7 12 g3 c5 13 d5 'iVb6 14 Ag2 b4 15
38 f7 'iVf3 0-0 0-0-0 16 .:b1
Belyavsky- Shirov, Ljubljana (European Club Cup) 1995 205
Nowadays this move is considered What is this? A rook down and then the
more critical than 16 ttJa4. bishop out of play? Unfortunately the e7
16 .'iYa6 17 dxe6 i.xg2 1 8 e7 i.xf1
•• pawn compensates for everything.
19 'i¥d5 i.h6 20 i.xh6 i.d3 23 Jlh5
••
Nothing new so far. But now Belyavsky Another interesting try was 23 ...ttJxffi!?
plays an interesting novelty which might 24 �d1 'ii b5!
be a fruit of his homework. Probably the only move to prevent 25
'iia8+ ttJb8 26 �d8 +, since 24...'i¥a4? would
have lost to 25 b3 cxb3 26 1Va8 + ttJb8 27
l:.d8 + l:txd8 28 exd8'iV + �xd8 29 'i¥xb8 +
�d7 30 'i¥b7 + �e6 3 1 'i¥e7 + �f5 32 'iixf7.
25 f4?
I believe that most players would have
gone for a move repetition after 25 'i¥a8 +
ttJb8 (25 ...'i¥b8? 26 'iVc6 + 'iVc7 2 7 'iVa6 +
�b8 28 'iixc4 and White wins) 26 'i¥e4 (26
'ii f3 �e5 27 h4) 26...ttJd7, but it is well
known that Beliavsky almost never goes
in for them. In this case it seems that he
was wrong. It is true that now the rook on
h5 has no squares and is indefensible, but
there is a trick that should turn things
2 1 ltJe4!? into Black's favour.
The known course is 2 1 'iia8 + ttJb8 22 25 c3! 26 bxc3
•••
RitRZR �-
B w�
�
;-'"'"�
���
w�.0
•� �
w r� a B
:Y/0}%:
�-� � /.un � '" , ,y
26 c4??
•••
iVU- •� n
�w� B v/%
Unbelievable. I saw that after 26...bxc3
27 �b1 I would have 27...c2! !, but for some
• • • • mysterious reason I rejected it. White
would have been in a hopeless situation,
.,.iV. • for example 28 �xb5 (after 28 'i¥xc2 'iic6
• • • 0 Black defends everything while an extra
A
0�����- ��-
/.'
-�
"� ���
�,� �
� ,�
�
0 � " " � /, ' '�
rook plus the position of the enemy bishop
yields him an easy victory) 28 ...c 11V + 29
-�·
�A � •
B �� �g2 'i¥d2 + 30 �f3 'iid 1 + 31 �g2 'iid5 and
the game is over.
206 Fire on Board
27 cxb4
Now White is not worse anymore. How
ever, during the game I still thought dif B
ferently, counting on my queen and rook
which are now well connected. Due to this
over-optimism I was unable to put up much
resistance to the strong white pawns.
27 'i;c7
.•.
Formally a new move but the idea of sac The choice wasn't a pleasant one. Vari
rificing the exchange was already known ations like 27 ...'tWxd l+ 2S l:Ixd l �xb7 29
as 20 ... �bS is bad in view of 2 1 'ti'd4! (Ag l:Id4 l:Ixg7 30 l:Ixc4 a5 31 l:If4! �b6 32 h4
zamov-Chandler, Belgrade 19S2). I doubt �b5 33 b3; and; 27...'iWxg7 2S l:Ixa7 �bS
that anyone would think twice about 29 ltdal lteS (29 ... 'ti'd4 30 tba5!! 'ti'c5 3 1
playing it. l:Ib7 + �cS 3 2 ltxf7) 3 0 tbd6 l:Iel+ 3 1 l:Ixel
21'iWg7!! �xa7 32 tbxc4 speak for themselves.
A positional queen sacrifice, after which 28 t2Jd6+ cJi;bS 29 ltdb1'VJIJxg7?
the attack of the white rooks and minor This loses immediately. Tougher re
pieces becomes extremely strong. When sistance would have been promised by
the move appears on the board it looks so 29...'iVd2 30 tbxc4 'tWc3, but my home analy
simple ... Vasily told me after the game he sis indicates that White gets a decisive ad
discovered 2 1 'ti'g7 over the board. Im vantage by playing 3 1 :a4! (3 1 t2Je3 'tWc5
pressive. 32 :a4 a5 33 :hal b3 is unclear) 3 l. .. b3
2 1 tbxc5 �xc5 22 �xc5 'tWxc5 23 �h3 32 l2Ja5! (32 tbe3 �aS 33 lta3 'tWxg7 34
l:IhdS 24 'tWg7 cJi;c7 25 �xd7 ltxd7, with :axb3 is only slightly better for White)
compensation, was what I had wanted. 32 ...�aS (32 . . . b2 33 ltb4 + �aS 34 lt4xb2)
21 �xg7 22 fxg7 l:Ig8 23 t2Jxc5 d4
••• 33 :a3! (33 :xb3 'tWel + 34 �g2 'tWdl! al
I was trying to overcome the difficulties lows counterplay) 33 .. . 'ti'xg7 34 tbc6 �b7
with concrete play, but clearly underesti 35 t2Jxa7 and it's all over. I should also
mated White's 26th move. Other options mention that I saw the line 29 ...'iYc3 30
seemed to me like a kind of 'slow death', l:Ixa7! cJi;xa7 (30 . . . l:Ixg7 3 1 :b7 + cJi;aS 32
for example: 23 .. J�c7 24 lbxb7 l:.xb7 25 l:I7xb4) 31 t2Jb5 + �a6 32 t2Jxc3 bxc3 33
:fdl :d7 26 �d4 (intending h4-h5-h6 ; :b4! too late.
23 ...f5 24 l:.xa7! (24 t2Jxd7?! 'iVxd7 25 i.d4 30 l:.xb4+ cJi;c7 31:a6!
:xg7 is only a little better for White) Now Black must give up his rook to
24 ... l:Igxg7 (24 ...l:.dxg7 25 ltfal) 25 l:.fal; avoid being mated. Of course, further re
23 . . . l:.xg7 24 �h3! �c6 25 :xa7; and sistance is impossible.
23 . . .i.c6 24 l:.a6, intending 25 %Ual. 3t...:bs 32 l:.xa 7+ �xd6 33 l:Ixb8
24 �xb7+ ltxb7 25 t2Jxb7 'tWb6 (D) 'ti'g4 34 ltd8+ �c6 35 l:Ia11-0
The only move. 25 . . .�xb7 26 �xd4 a5 And since the passed pawn is going to
27 ltfel loses quickly. fall ... Black resigned.
26 �xd4!! P. S. Although 'the Botvinnik chapter'
I was mainly counting on 26 �f4 �xb7 should have been completed with this
27 :fel a5!, with counterplay. game, new problems have since arisen. At
26 'tWxd4 27 l:.fd1'tWxb2
••• Monaco 1996 I just couldn't resist trying
208 Fire on Board
Kramnik- Shirov
Monaco (blindfold) 1996
These annotations were made during the
preparation of this book.
1 c4 c6 2 d4 d5 3 �c3 �f6 4 �f3 e6 5
i..g5 dxc4 6 e4 b5 7 e5 h6 8 i..h4 g5 9
�xg5 hxg5 10 i..xg5 �bd7 11 g3 i..b7
12 i..g2 iV b6 13 exf6 0-0-0 14 0-0 c5 15
d5 b4 16 l:Ib1 29 fxg3 l:.xe4 30 �d6 .:e6 3 1 �xf7 + �e8
Somehow this move doesn't seem very 32 �g5 l:Ixf6 33 �e4 l:tg6 are not espe
logical to me - but it's terribly dangerous. cially dangerous for Black. Now he also
has to look for an endgame, but a much
less favourable one this time.
28 �c7 29 �xc4 'ii'd4 30 b3 'ti'e4
.•.
19 �xf1
This move hasn't been seen in tourna
ment praxis for a long time, but of course
nobody could miss it in home analysis.
19 iVc6
•••
Very strong! Variations such as 28 'ilfxc4 After this error Black can only hope
l:txe3 29 fxe3 iVxg3 + ; or 28 �xc4 'iVxg3 + for a miracle. Both 34 . . . l:Id4 and 34 ... c4!?
Azmaiparashvili - Shirov, Madrid 1996 209
would have promised him some drawing 52 lha2 53 �6 l:.b2 54 .:h6+ �xg3
•••
Azmaiparashvili - Shirov
Madrid 1996
The game was annotated in May 1996,
and published in various magazines.
1 d4 d5 2 c4 c6 3liJf3liJf6 4 l2Jc3 e6 5
39 f4! i..g5 dxc4 6 e4 b5 7 e5 h6 8 i..h4 g5 9
The rest of the game doesn't require l2Jxg5 hxg5 10 i..xg5 liJbd 7 11 exf6
much commentary. White is winning, but i..b7 12 g3 c5 13 d5 'i¥b6 14 i..g2 0-0-0
in blindfold (as well as rapid!) chess, there 15 0-0 b4 16 �bl 'ii'a6 17 dxe6 i..xg2 18
are always some practical chances- luck e7 i..xfl 19 'i¥d5
was with me on this occasion. This move, which was once considered
39 ltJcl 40 ltJc4 f6 41 �a6! l:.e4 42
••. practically the refutation of the 13 . ..'i¥b6
�f3 l:Ie2 43 ltJe5+ �g7 44 �a7+ �h6 line, in fact seems to lead to a draw ac
45 liJd3! �c2 46liJxcl �xcl 47 �e4 �g6 cording to the present game. Of course,
48 f5+ �h5 49 l:Ih7+ �g4 50 �g7+ 19 �xfl causes Black more suffering.
�h3 51 �d5 l:.c2 Funnily enough Kramnik went in for this
position with Black(!) against Kamsky at
Dos Hermanas 1996, but he also came
very close to losing.
52 l:Ig6??
A terrible mistake. The simplest way to
win would have been 52 �c7 l:Ixa2 53
l:Ixc5 �xg3 54 �c4. 19 .•. i..h6
210 Fire on Board
19 ... ..txe7 20 fxe7 .tldg8, played by Kram 27 'ti'xb7+ <it>xb7 28 l:.e7+ .l::td7 29
nik against Kasparov, doesn't equalize. ..txb8 'it>xb8 30 l:.xd7 .i.xd7 31 Wg2 <it>c7
20 .i.xh6 32 �f3 <it>d6 33 <it>f4 .i.c6!!
The latest fashion. An old alternative is You have to find such moves at home if
20 exd8� + , played by Yermolinsky against you play the Botvinnik variation! If White
Ivanchuk in 1986, but that game didn't can get his king to g5 and pawn to f4 then
achieve any advantage for White. Black can only resign. The bishop ma
20 ..td3 21 'i¥a8+
••. noeuvre prevents this set-up.
As the reader will remember, 2 1 l2Je4, 34 �g5 i.f3 35 �f5
as played in Beliavsky-Shirov, Ljubljana White cannot force Black into zug
1995, is not especially dangerous for Black. zwang as the latter has space for his king.
2 1 exd8'iV+ is bad according to Yermolin 35 �d5 36 g4 �d6 37 h5 'it>d5 38
.•.
pionship, where the Yerevan IM, after When I adjourned the game, I was still
adjourning 12(!) games, one after another, hoping to reach a draw, but my analysis
2 12 Fire on Board
definitively established that there was no position just before the adjournment ses
way of saving the game. I should mention sion. But in the very next round my col
that the simplest way for White to win lection of adjournments was increased. I
was by 93 �f4 tLig6 + 94 �e4 �g5 95 .l:g7 was now also faced by a difficult defence
�h6 96 :a7 �g5 97 :a3 tLie5 (if 97 . . . �h4 in my game with Psakhis, the analysis of
98 �f5, while after 97 . . . f5 + 98 �f3 the which took up all my strength and time.
win is technically straightforward) 98 When the Psakhis game was resumed I
:g3 + �h4 99 �f4. But it so happened came close to a draw, but several blunders
that my opponent did not find this con reduced all my efforts to nought . . . So my
tinuation, and he assumed that the most mood before the next adjourned game was
probable outcome was a draw. not a vecy happy one.
There followed 48 �gl
93 �e4? �h4 94 l':.h8+ �g5 95 �d5? The sealed move.
A fundamentally incorrect plan. By 48 �g4 49 'iWdl+
•••
playing his king to g3, White would still A surprise. I had thought that after
have retained winning chances. 49 'iWf3
•••
pectations, so I didn't bother to go into Not 5 7 'iVxh5 'iVd4 + , when the pawns
details, since I was hoping to return to the are lost with check.
Balashov- Shirov, USSR Championship Semi-Final, Klaipeda 1988 213
57 h4 58 b4
••• �f6 + �c4 50 l:Ih2 .i.d1 5 1 �e5 �b3 52
In the variation 58 a4 �g6 + 59 Wfl �d5 b4 53 c4 bxa3 54 bxa3 �xa3 55 �xc5
'11:Vxe4 60 �xd6 + �e3 Black has a clear win. .i.b3 56 l:Ih3! �b2 5 7 �b4, etc.
There followed 47 �d5 48 .:as?
..•
h4 54 f6 + <&t>xf6 55 <it>d6 h3 56 e7 h2 5 7
e8'i*' h1'iV with a draw.
[AS - Instead of 49. . . ltJc5, stronger is
49. . . a5 50 {6 ltJe5! 51 �xe5 <&t>xd7, but after
52 <it>d5 b4 53 axb4 axb4 54 e5 b3 55 e6 +
�eB 56 �d6 b2 57 f7 + <it>{B 58 e7 + �xf7
59 <&t>d7 bl 'ii' 60 eB�+ <&t>f6 61 'iVe6+ �g5
62 'ii'e3 + White still draws, because Black
can't avoid perpetual check.]
44 b3!?
A very interesting position arises after
44 e5 ltJxb2 45 d5 ltJd3 + 46 <&t>e4 ltJc5 + 4 7
�d4 (47 �f4? a5!) 47 . . . ltJb7 48 d6.
·�
54 �e7 b1 'iV 55 f6 + �g6 56 f7. Black is a
queen up, but it is not difficult to see that
.l·�·�· one of the white pawns will promote, and
Black has at best a draw, e.g. 56 ... 'ii'b 7 +
• ·�· 57 d 7 'ii'b4 + 58 �e8 'ii'b8 + 59 d8'ii' 'ii'xe5 +
• • 60 �e7 'ii'b 8 +) 52 �e 7 b2 53 f6 + �g6 54
f7 b1 'iV 55 f8'ii', and White has everything
• • •
in order, e.g. 55 .. .'�f5 56 'ii'g8 + <&t>h6 57
54 e7 l'£Jf7 55 d7! (after 55 e8'iV a1'iV Black White makes a perpetual check - this is
should win) 55 ... a1'iV 56 d8'iV 'iVc3+ 57 <it>d7 what I missed five years ago) 57 <it>b6
iid4 + 58 <it>c8 with a draw. �d4+ 58 iic5! (but not 58 <it>xa6?! 11Jxd6
b) 48 . . . h5! 49 Wd5 h4 50 <it>e6l'£Jd8 + 5 1 59 'iVd7+ <it>f6 60 'iVe6+ cst>g5 61 'iVg6+ cst>f4
<it>d7 h3 5 2 cst>xd8 (the alternative is 5 2 e6 62 f6 �d5! 63 'ifih6+ <it>g4 64 'ii'g6+ [or 64
h2 53 e7 l'£Jf7 54 e8� h1 'iV 55 �c8!, and �g7+ <it>f5!] <it>f3 65 <it>b6 l'£Jc4+ 66 �c7
here Black wins by 5 5 ...'i¥d5! 56 'i¥xa6 'iYd6+ 67 <it>c8 'iYc6+ , followed by 68... l'£Jd6!
'iic4! 5 7 cst>e7 [the only move] 5 7 . . l'£Je5!
. 58 or 58 Wc7?! �a7 + 59 cst>c8 a5! 60 d7 �d4!
d7 'iic5 + 59 cst>e6 [or 59 'iid6l'£Jc6 + 60 cst>e6 with excellent winning chances for Black
l'£Jd4 + ] 59 ...l'£Jf7! 60 d8'iVl'£Jxd8 + 6 1 <it>d7 in both cases) 'iVd3 61 <it>c7 Wf6 62 d7 and
l'£Jf7) 52 . . . h2 53 <it>e7 h 1 � 54 f6 + cst>g6 5 5 the maximum Black can get is queen and
f7 knight against queen.]
44 l'£Jd2
•••
45 4Jxb3 46 d6
••• On 52 We5 there follows 52 ...4Jxd7 + 53
exd7 b4, and wins.
52 4Jd5 0-1
•••
Earlier I had intended 4 1 �e8 + �f7 42 i.. a2 + �c3 56 f6 �c2 5 7 f7 b1'ii 58 i.. xb1 +
d6, but in this calm situation I now saw �xb1 59 f8'ii' a2. But then in analysis the
that after 42 ...b3 43 �d3 b2 + 44 �c2 Leningrad player Alexei Yuneyev discov
l:tc3 + I would lose. I could, of course, have ered the simple, but by no means obvious
played 42 �d6 + with a probable draw, but 49 i.. f5 + ! �d6 50 �c2, when questions, as
I wanted to win! Therefore I played they say, are superfluous.
4l �d4 45 d7 �e7 46 f5 l:xd7
Mter some thought Black replied The only move.
41. l:al
•• 47 i..xd7 �xd7 48 �f4!
Also possible was 4 l. ..b3 42 �c3 (42 It's all over. The white knight holds the
�c4 l:a1 and 42 �e4 + �f5 43 d6 b2 44 queenside pawns, while the king picks up
�d2 l':.g3 45 d7 l:g8 also draw) 42 ...h3 43 the h4 pawn and is in time to defend the
�e4 + �f7! (43 ...�f5? 44 d6 h2 45 �c5 f5 pawn. The game concluded:
wins for White) 44 d6 h2 45 �g5 + �f6 46 48...�e7 49 �g4 a5 50 �xh4 a4 51
d7 �e7 47 �e6 b 2 + 48 �xb2 l:Id3 49 �g5 a3 52 �d2 �d6 53 f6 a2 54 �b3
d8'ii' + l:Ixd8 50 �xd8 �xd8 5 1 �c3 a5, and �d5 55 f7 �c4 56 �al 1-0
it is not difficult to see that the position is
drawn. Practical Problems of
42 �e4+ �f5?
This move loses. Correct was 42 ...�f7 43 Opposite-Coloured Bishops
d6 l:Id1 + ! 44 �e5 b3 45 �g5 + (45 i.. d5 + ?
l:Ixd5 + ! 4 6 �xd5 b 2 47 �d2 h 3 4 8 d7 This article was completed in the Autumn
�e7 49 �c6 �d8 wins for Black) 45 ... �f8 of 1990 and published in Shakhmaty
46 �e6 + �f7 47 �g5 + (47 d7? b2 48 d8'if Riga.
l:Ixd8 49 �xd8 + �e7 50 i..e4 h3 ! 5 1 �c6 + Endings with opposite-colour bishops
�f8 also wins for Black) 47 ... �f8, with a can, in my opinion, be regarded as a little
draw. explored field of chess theory. Amateurs
43 d6 l:dl+ 44 �e3 (D) are of the opinion that these endings 'al
44...�e6 ways' end in a draw, and that wins can be
During the game I could not see a win regarded as exceptions. It will be under
after 44 ... b3 45 d7 b2 46 �d2! l:xd2 4 7 stood that I have in mind endings where
i.e4 + �e6 48 �xd2 �xd7. Indeed, noth one side has a material advantage of not
ing is promised either by 49 �c2 �e6 50 more than one pawn. But in practice, one
�xb2 h3 5 1 �c3 h2 52 �d4 a5 , or 49 �e3 side wins so often that the question sug
a5 50 �f3 a4 5 1 �g4 �e6 52 i.b1 �d5 53 gests itself: which is the exception, and
f5 a3 54 �xh4 (54 f6 �e6) 54...�c4 55 which the rule?
Shirov - Vyzhmanavin, Lvov Zonal 1990 2 19
However, I am not intending to conduct The draw in this case is obvious, since
a statistical study here. I should merely White cannot even advance his e-pawn.
like to mention that such endings have a Now let's again place a white rook at f6,
great similarity with the middle game, be and a black one at a8.
cause the stronger side, if he doesn't have
a forced win, is obliged to engage in ma
noeuvring play.
I will attempt to describe one such end
ing. Or more precisely, one of my most im
portant games, played in the sixth round
of the Lvov Zonal (February 1990).
Shirov - Vyzhmanavin
Lvov Zona/1990
co-ordination of his forces, and will be And the plan was a simple one- to play 42
able to dictate matters. Black should have f5 and exchange the white f-pawn for the
played 39 ...Ae7! 40 l:tc6 :da2!, with the black g-pawn. After this Black is left with
idea on 4 1 :b 7 of playing 41... :2a 7 42 three weak pawns - d4, f7 and h6, and
:cc7 l:xb7 43 :xb7 �f8, while if 41 :b5 White can mount a combined attack, in
�g7. His defences would then have been which both rooks and his bishop partici
very difficult to breach. pate, and, in the distant future, perhaps
40 l:c6 Aa3 also his king. To this it should be added
After this, my opponent's last move be that the black pieces cannot immediately
fore the time control, I no longer had any be included in the defence. I had no doubt
doubts about winning. In analysis (after that my plan should lead to a win, and
the game) I also discovered a win against therefore I did not seek any alternatives.
40...Ae7, which I should like to demon After winning the game, I was still sure
strate: 4 1 :b8 + Cft;g7 42 :cc8! (threaten about the correctness of 4 1 l':.b5. But now
ing 43 l:tg8 + �f6 44 :hs Cft;g7 45 :bg8 + I see that it would have been stronger and
and 46 :xh6, so Black's reply is forced) simpler to play 41 :b8 + ! �g7 42 :cc8! (as
42...h5 43 l':.g8 + Cft;h7 44 l':.h8 + �g7 45 Af3! in the 40...Ae7 variation).
(of course not 45 Axh5? gxh5 46 :bg8 +
�f6 47 l:h6 + Cft;f5 48 �f3 f6!, and Black
stands no worse)
B
f6 (49 .. Jla5? 50 l:te8 + ) 50 l:txh5, when the After thinking for some ten minutes, I
rest is, so to speak, the 'gathering in of nevertheless failed to find an immediate
the harvest'. win. In order to keep a sufficient reserve
4 1 llb8 + ! was undoubtedly much bet of time, I decided to make a move that
ter than 4 1 l:tb5!? However, from a practi was useful in all respects. The white king
cal perspective I don't consider my choice will be well placed at h3.
to be a blunder. After evaluating the pos 50 h4! <it>g7?
sible consequences of 41 l:.b5, I was, as I The decisive mistake in a poor position.
have already mentioned, firmly convinced Understandably, Vyzhmanavin did not
that I would gradually win. And at the like the idea of 5 1 llg8 ! , which I was in
same time the move 4 1 llb8 + demanded tending to play next move. Two other con
the calculation of concrete variations, and tinuations came into consideration, even
in attempting to work out everything I though they too would not have saved the
could have ended up in time-trouble. If game:
during a game I see one way to win, I pre a) 50 . . . i.c3 5 1 l:tg8! (but not 5 1 <it>h3
fer not to look for another. �g7! 52 :ee8 i.d2 or 5 1 l:td8 l:te6 ! 52 l:tf4+
4l. �g7 42 l:td5!
•• �e7, and Black can still resist) 5 1 . . . l:.a5
An important nuance. The bishop is (the threat was 52 llf4 + <it>e7 53 i.h5 lU6
driven to b2, from where it cannot control 54 lle4 + l:.e6 55 l:te8 + ! <it>xe8 56 llxe6 +
f6 , and where it will block the path of the and 57 l:.xh6, with a two-pawn advantage)
rook at d2, which comes into play too late. 52 lleg4 ! ! (this move creates the almost ir
42 i.b2 43 f5 gxf5 44 l:txf5
••• resistible threat of 53 l:.h8) 52 . . . l:t7a6 (evi
Threatening both 45 i.h5 followed by dently best) 53 l:th8 �e7 54 l:te4+ �d7 55
46 .:df6, as well as 45 llff6. Therefore llf8! llf6 (55 . . . f6 56 %leeS) 56 i.g4 + , and
Black's reply is forced. it's not difficult to see that, after driving
44 lldl 45 i.f3!
••• the black king to the queenside, White
Now 45 i.h5 l:tc1 ! 46 %lcf6 .l:tcc7 or 45 picks up the kingside pawns in the end.
llff6 i.c1 is no longer effective. White in b) The toughest defence, in my opin
tends first to drive the black king away ion, was 50 . . . lle6! 5 1 l:tf4+ ! <it>g7 (51. . .<it>e7?
from g7, and then to concentrate on the 52 i.h5!). After this I was intending 52
main weakness - the h6 pawn. i.e4! , and didn't calculate any further.
45 lldal
••• Now I can say that on the possible move
45 . . . llc1 fails to 46 l:tb6 i.a3 (or 46 ... l:.c2 52 . . . i.c3 (with the idea of 53 . . . i.d2) White
47 l:.f4) 47 l:.f4. replies 53 l:td8 ! , and Black is still ex
46 l:tf4 l:.la6 47 llg4+ �f8 48 l:tc8 + tremely restricted. The ' active' 53 . . . l:ta2
�e7 49 l:te4 + <it>f6 doesn't work on account of 54 l:td7!
222 Fire on Board
Wisely considering the h5 pawn to be The winning idea would have been to
his main enemy, Prie rejects the natural go immediately for forced variations, e.g.
4. . . 'iVf1 5 �e6! 'iVc4 + (5... 'iVh3 + 6 �xd6 is 1 .t b6! �g6 (the only move, as 1 ... lt:Jd 7?
lost; while 5...'ifxf3 6 'iWxf3 + tbxf3 fails to just loses another pawn after 2 .i.c7; while
7 a4!!, and the pawns on b6 and a5 will l . .. �e6 2 .i.a7! �e5 [forced] 3 .i.b8 + �f5
fall) 6 'iVxc4 �xc4, because 7 i.d4! seems 4 .i.d6 leads to zugzwang) 2 �e2! (2 .i.xc5?
to be winning in the line 7 ... �xa3 8 f4 lt:Jxh4 + 3 �e2 lt:Jxg2 4 .i.xd4 h4! is an easy
�c2 9 .i.b2! a4 10 f5 a3 1 1 .i.c3 b5 12 h6. draw for Black) 2... �xh4 3 g3! leaving
5 'iVaS+! Black with unpleasant choice between:
It is better to avoid 5 �e6 'iVxh5. a) 3 ... lt:Jg6 4 i.xc5 �e5 5 �f3! fxg3 6
5 ... �e7 6 'iVb7+ �d7 7 �g6! fxg3 h4 7 �g4! hxg3 8 �xg3 �h8 (or
Now it's clear that the pawn will pro 8 ... �f4 9 .i.d6 + ) 9 �g4! (not 9 .i.f8 �g6
mote one day. Black has no chance of a 10 .i.g7 + ? �f5 1 1 .i.xd4 �f4, with a draw)
perpetual, so the game is over. 9 ... �g6 (9 . . �f7 10 i.f8 �f6 1 1 �f4) 10
.
7 'ti'h3 8 .i.d4! d5
••• �g5 �h8 11 .i.xd4 + ! �xd4 12 �f6 win
Or 8. . .'iVe6 + 9 �xg7 'iWf7 + 10 <it>h6. mng; or
9 �xg7 'iWxf3 b) 3 ... f3 + 4 �d2 �g6 5 .i.xc5 �e5 6
9.. .'iVxh5 10 'iVxd5. <it>c2! and again Black has several possi
10 .i.f6+ �d6 11 h6 'ife3 12 h7 �xf6 bilities, but none of them work:
13 gxf6! 'iVg5+ 14 �f7 'iVh5+ 15 �g8 1-0 b1) 6 ... h4 7 gxh4 �xh4 8 .i.f8! �f6 9
Black resigned in view of 15 .. .'iVg6 + 16 �b3 �f5 (9 ... �g2 10 c5; 9. �g6 10 .i.c5
. .
'iVg7 'ii'e8 + 1 7 'iVf8 + and wins. �e5 1 1 .i.a7!) 10 �a4 tbe3 1 1 c5! lt:Jd 1 12
c6 �e6 13 �b5! lt:Jxf2 14 �a6 �xd3 15 c7
and wins; or
Torre - Shirov b2) 6 . . . lt:Jh8 7 i.f8! �f6 [forced] 8 �b3
Manila /nterzona/1990 �f7 9 �a4! lt:Je5 (9... �g5 10 c5! �e6 1 1 c6
�h3 12 �b5!) 10 .i.c5! �xd3 1 1 .i.xd4 +
These annotations are based on my notes �e6 12 �b5 and wins.
in lnformator 49. l. lt:Jg6
..
for 8 .i.b6! obtaining the already known For these annotations I should like give
position. special thanks to German IM Karsten
8 .i.g5 �e6 9 �f3 �f5?! Miiller, who was my second in Munich
9 . . . �e5! would have been more precise, and not only analysed this position with
but it seems that the text doesn't lose me when the game was adjourned, but
either. also made a very deep investigation of its
10 .i.d8 �e5! 11 .i.c7+ �f5 12 �e2 mysteries later on. His work was checked
ltJxh4! by German chess columnist and trainer
Of course, not 12 . . . �e6?! 13 �fl!, in Claus-Dieter Meyer, who then published
tending to get his king to h3, when White an article on it in the German magazine
gets all his winning chances again. Schack 64 (12/1994), and in these annota
13 g3 f3+! 14 �d2 ltJg6 15 .i.d6 ltJe5 tions I will often refer to that article.
16 .i.xc5 lbc6 I was utterly lost a large part of this
The knight both defends d4 and stops game, but it was finally adjourned in this
White's passed pawn. White's only win curious position, where Black had to seal
ning chance would be to go to b5 with his his move. He is still a lot of material
king, but he cannot do so because of the (three pawns) up, but in fact the ending is
f2 pawn. already dangerous for him since White's
17 �e1 attack is now very strong indeed! Never
1 7 �c2? �g4 18 �b3 �h3 19 �a4 �g2 theless I wasn't that optimistic about
20 �b5 �xf2 2 1 �xc6 �xg3 22 .i.xd4 h4 my winning chances, thinking that a draw
just loses by one tempo. would be the most likely result. It was
17...�g4 18 �f1 h4 19 gxh4 �xh4 20 more Karsten than me who was finding
.i.d6 �h3 21 �g1 ltJa5 22 .i.b4 ltJc6 23 magnificent tries in the various vari
.i.d6 ltJa5 24 �f1 lbc6 25 .i.c5 �h2 26 ations.
i.d6+ �h3 lf2-V2 l...�g8?
Here my opponent in a rather odd way When the game was resumed I thought
said 'Yes', offered his hand and stopped that this move was still sufficient for a
the clock. All this might have been inter draw but now it's clear to me that Black
preted as resignation but, of course I gets into serious trouble with it. The main
didn't apply to the arbiter for a win. alternative would have been 1 . . Jid3, try
ing to get rid of White's bishop as soon as
possible. White must then play 2 !:txg7 + ,
Shirov - Lautier and after 2 . . . �h8 he has three possibili
Munich 1993 ties, but none of them with a real hope of
success:
a) 3 .i.f6 lld6 4 llg6 + (4 l:.d7 + ?! !:txffi 5
�xf6 g3 can only be dangerous for White
B as the black pawns are already too close)
4 . . . � h 7 5 llg7 + with a draw by repetition;
b) 3 �g6 f4 + 4 �h6 !:txd4 (4 . . . llh3 + is
OK as well- see variation 'c') 5 cxd4 g3 6
l:tf7 �g8 7 l:.xf4 g2 8 llg4 + �f7 9 !:tf4 + (9
�g5?! .i.d5! 10 �f4 �f6 is not advisable)
9 . . . �g8 10 !:tg4 + with the same outcome;
c) 3 �h6 (the most ambitious try)
3 . . . %:.h3 + ! (this time giving up the ex
change might cause Black problems, for
example 3 . . . a5 4 l:.g5 + l:txd4 5 cxd4 b4 6
d5 i.xd5 7 !:txf5 i.g8 8 !:th5 i.f7 9l:txa5 c3
Shirov - Lautier, Munich 1993 225
10 bxc3 b3 1 1 l:Ia8+ i.. g8 12 l:Id8 g3 1 3 escape from the mating net so easily! A
l':.d2 i.. c 4 14 �g5 i.fl 1 5 �g6 i.. c 4 16 a4 sample line, indicated by C-D. Meyer, is
and White wins according to C-D. Meyer) 5 . . . i.. e 6 6 'itf6 i.. d 7 7 l:Ic7 �e8 8 l:Ia7! 'itd8
4 �g6 f4 + 5 �g5 i.. f3 (5 . . . .l:d3? was beau (8 . . . i.. c6 9 l:Ie7 + 'itd8 10 i.. b 6 + �c8 1 1
tifully refuted by Karsten, who shortly l:Ic7 + 'itb8 1 2 :Xc6 'itb7 1 3 .l:d6 should be
before resumption found 6 l:Ie7 + l:Ixd4 7 winning for White in the long run)) 9
cxd4 g3 8 .l:xe4! g2 9 l:Ie1! f3 10 �g6 f2 1 1 i.b6 + 'itc8 10 l:Ic7+ 'itd8 1 1 l:Ic5+ �e8 12
l:Ie8 mate . Need I say that it was this line l:Ie5 + �f8 13 i.. c5 + 'itg8 14 .l:e 7! l:Id3 15
that put me in a good mood before going .l:g7 + �h8 16 i.. d4 i.. c6 17 l:Ic7 .l;txd4 18
to play?) 6 'itxf4 i.. d 1 7 l:txg4 + 'ith7 8 cxd4 i.. e4 (18 . . . i.. d5 19 .l:c5 i.. e4 20 d5) 19
1;tg7 + 'ith6 9 .l;ta7 1;th2 10 :xa6 + �h7, and �g6 f4 + 20 �h6 and we get the same as
Black should hold his own (C-D. Meyer). occurred in the game.
2 �g6 i.. c6 3 l:Id3?
.•.
13 �g5 a5 14 �f4 b4 15 axb4 axb4 16 l:.g3 Black has full compensation for the ex
�h7 1 7 'it>e3 �h6 18 �f2 c3 19 bxc3 b3 20 change. In order to drive the black queen
c4 i.xc4 21 l:.xg2 b2 22 l:tg1 i.. a2. But the from her strong position I now offered a
text is a killer. repetition of moves.
9 i..g8 10 d5! f3
••• 39 'ifc3 'ifg3 40 'ife1 'i¥g4?
10 ...g3 is the same in view of 1 1 d6 g2 Kamsky incorrectly avoids the draw.
( 1 1....i.e6 12 l:te5) 12 l:Ig5 f3 13 d7. 41 %:tc5!
11 d6 .i.e6 Now that the black queen stands pas
Black also gets mated if he promotes sively; White plays to win.
his pawns, e.g. 1 1...f2 12 l:lf5 g3 13 d7 g2 41. �h7
.•
14 d8'it' g1'ii' 15 'iVf6 + . 4l. ..f6 42 �h2 e5 was better. Here Kam
1 2 l:le5! i.d7 1 3 l:le7 �g8 1 4 l:Ixd7 f2 sky offered a draw, one move too late.
15 l:.g7+ 1-0 42 �h2 g5 43 hxg5 hxg5 44 l:.c8!
Just in time! Mter 15 .l:tg7 + �f8, 16 d7 44 'ii'g3 'i¥xg3 + 45 �xg3 �g6 only leads
wins, so he resigned. to a draw.
44 'it>g6?
•••
year that I had beaten both of them in the The ending after 46...'iff5 4 7 'ifc7 'i¥e5 +
same event (albeit because they misplayed 48 'i¥xe5 fxe5 49 g4 ! would have been
good positions!). On the other hand, when hopeless for Black - White places his rook
I only had to face one of them in a tourna on b6 and brings his king to the queen
ment, my winning chances were close to side.
zero! 47 'it'c7 �f5 48 'i¥h7+ 'it>e5 49 'iVa7!
[AS - Later on I managed to beat Kam �f5 50 l:.h6! 'i¥f2
sky in a tournament where Kramnik 50 ...'i¥b2 5 1 'ifh7 + �e5 52 'it'g7 'i¥f2
didn't play (Buenos Aires 1994). With the (52 ...�d6 53 'iVxf6 wins for White) 53 'i¥a7
latter things are more difficult, if one ex �f5 transposes to the game.
cludes rapid chess.]
Mitkov - Shirov, Cannes (French League) 1994 227
Now it looks as though Black can hold by Kramnik a long time ago. Objectively
on, for example 5 1 'iVd4 �g4!, but after 13 ...:a4 appears better but one should
some thought I found ... know a lot about it.
51 <it'h3!! 14 lbxa7 4Jxa7 15 c3 4Jb5 16 exf5
This beautiful move threatens both 52 4Jc7 174Je3!
g4 + �e5 53 'ii'c5 mate and 52 �c5 + e5 53 A strong novelty ( 1 7 0-0 would be the
'iVc8 mate. old path). White's idea is to play 18 'iVf3
51. g4+ 52 <it'h2
.• with advantage.
Now the threat is 53 �h5 + �g6 54 17 �b7! 18 �a4+?!
•••
26 . . . 'ii'e4 27 'iWxe4 lbxe4 was better, in Passive defence is completely wrong.
tending . . . d5, . . . �e6 and . . . i.. c5. By playing 37 b4! l1xa3 3S l1b1 White
27 I:.ad1 �f6 28 I:.fe1 l:.g8 could still have reached an easy draw.
2S . . . lbf4 would probably just transpose 37 �xf4
•••
since after 29 'ii'e 4, 29 . . . 'ii'xe4?! doesn't Now the black king is so close that
seem very promising due to 30 l:.xe4 .:.gS things are no longer easy for White.
(30 . . . d5 3 1 .:.x£4 + ! exf4 32 .:.xd5 gives 38 �e2 f5 39 h3?
White an edge) 3 1 g3 d5 32 l:.a4! d4 (not Only weakening his position. The last
32 . . . e4? 33 f3! favouring White) 33 �fl chance lay in playing actively, e.g. : 39 .:.d2
with a slight plus for White. h5 40 b4! (40 .:.d1 h4 4 1 l:.d2 �g4 is
29 'ii'e4! ltJf4! slightly better for Black) 40 . . .l:.xa3 41 l:.b2
When playing this, it was important to �g4! (not 4 1 . . . l1h3? 42 b5 l1xh2? 43 b6 e3
see that I would still be okay despite be 44 :b1!! exf2 45l:tb3! and wins) 42 b5 l:.a7
ing two pawns down in a rook ending. 43 b6 l:.b7 and I don't believe that Black's
29 . . . 'ii' h5?! 30l:td3! intending 30 . . . lbg5 3 1 winning chances are real.
'ii'c6! looks better for White. 39 h6
•••
30 'ii' xf5+ �xf5 31 g3 e4 32 .:.d4 39 . . . h5 would have been even better for
Otherwise Black is certainly not any Black.
worse. 40 l1d2 h5
32 d5 33 �f1 i..c5! 34 gxf4 .i.xd4 35
••• When the first time control finished I
cxd4 .:as realized that my winning chances were
very good but there was a practical prob
lem. I had to catch a plane from Nice in
just a few hours and it was not clear how
long the game would last as the time limit
was 20 moves in one hour plus one hour
for the remainder of the game - this
meant that eight hours could be played!
Should I mention that my plane was to
Buenos Aires where I was going to have
my wedding? The best I could do in this
situation was to try to use my opponent's
time and move quickly. Later on it spoiled
the quality of my play and didn't give me
a chance to make this endgame a real
wedding present. Still, I should not com
36 a3 plain as I did finally win and some time
Both players were in some time-pres later it was really fascinating to analyse
sure and White didn't realize that he it.
should just go for an immediate draw 41 l1d1
with a continuation such as 36 l1b1 .:.xa2 Still playing passively. However, after
3 7 b4 �xf4 3S b5 e3 39 b6l:taS (forced) 40 41 b4 l:.xa3 42 I:.b2 :xh3 (this is why 39
b7 l:.bS 4 1 l:.b5! exf2 42 :xd5 l:.xb7 43 h3 was a mistake!) 43 b5 l1a3 44 b6 l:taS
�xf2. A more complicated line which even Black should also win.
forces Black to play accurately is 36 l1a1! 41...h4! 42 I:.d2 l:.c4 43 l:.d1 ?! l:tc2+
l:.a4 3 7 b3 I:.xd4 3S a4 l:.d3 39l:ta3 l:.d1 + 44 l:.d2 l1xd2+!
40 �e2 l1b1 4 1 a5 d4 42 a6 d3 + 43 �d2 That's the point! The rook endgame
l:.b2 + 44 �e1 l:.b1 + leading to a repeti turns into a pawn and then a queen end
tion of moves. ing which is winning for Black.
36 l:.a4 37 l1d1?!
.•. 45 �xd2 �f3 46 �e1
Mitkov- Shirov, Cannes (French League) 1994 229
66 �d2??
Now White's king gets sidelined and
he finally loses the game. Correct would
be 66 �f3 �h3 + 6 7 'iite 2 'tWh5 + 68 'iitd2
'tWg5 + 69 �e2 'Wie7 + 70 �d3! (70 �f3??
'tWf6 + wins) 70 ... 'tWa3 + 7 1 �e2 'tWa2 + 72
�f3 'tWb3 + 73 �g4! and Black cannot
strengthen his position.
66 'tWel+ 67 �c2 'tWe2 + 68 �cl
•••
had thought more about the game than Very exact. Now everything is in time-
about catching my plane, I would prob both promoting the pawn and arriving at
ably have realized the difference. Now the the airport!
game is drawn but fortunately during the 69 d6
game neither player appreciated this. 69 �d2 �g2 70 �g7 + �g3 wins.
58 a4 'tWg3?! 59 'iVe4!! 69 �g2 70 'iVd2
•••
It's odd, but I didn't realize during the Or 70 'tWg7 + 'tWg3 7 1 'tWb 7 + �h2 72
game how strong this move is; indeed, I 'tWh7+ �g1 73 'tWa7 'tWxd6, winning.
230 Fire on Board
Leko- Shirov
Dortmund 1996
These annotations were made in July 1996
and have not been published before.
Is it always easy to win with two rooks
against a rook and three pawns? Mter
this game I doubt it.
1 e4 g6 2 d4 i..g7
Getting into the middlegame has re
cently become a difficult task for me and
somehow these two solid moves give me a
The time scramble had just finished certain confidence that the game will still
and I had sadly come to the realization be going after the opening.
that my position was very difficult. But 3 ttJc3 c6 4 lbf3 d6 5 a4 tiJf6 6 .i.e2
then I found a nice combination. 0-0 7 0-0 ttJbd7
41 l:e2 <itg6 I went for 7 ... ttJa6 two rounds earlier
Playing quickly as he didn't see my in but then had to suffer to make a draw.
tentions, but I doubt that Black's winning 8 .i.f4 l:e8!?
chances would be very good any way. Surprisingly, this is a new move.
42 l:f1 llc7 43 e6! Black's idea is to continue his develop
White sacrifices first one, then another ment by 9...'iVc7. The immediate 8 .. .'i¥c7 I
passed pawn, but makes a funny repeti didn't like because of 9 e5! lbh5 10 i.. g5
tion of moves which Black cannot avoid. dxe5 1 1 .i.xe7 l:e8 12 d5! as I was not sure
whether my compensation for the ex
change after 12 .. Jixe7 13 d6 'iVd8 14 dxe7
'iVxe7 would be sufficient. Later on I
learned that all this has already occurred
in the game Levitina-Ioseliani, Beijing
1992, which Black won. Still I am not con
vinced that Black is equal in this line.
9 l:e1
Now Black can play his idea. White
could try to cross Black's plan with 9
ttJd2!?.
9 'iVc7 10 e5 lbb5! 11 exd6 exd6 12
•••
i..e3 d5
Black has equalized.
Leko - Shirov, Dortmund 1996 23 1
13 'i¥d2 liJf8!
Tempting White's knight to e5 where it
will not do a lot. I considered 13 ...lt:Jhf6 14
i.. f4 to be slightly inferior for me.
14 liJe5 liJf6 15 i.f4 liJe6!?
As often happens, I decided to go for
complications at all costs. Had I seen every
thing, I would probably have preferred
15 ...'ifd8 with equality, or 15 ...'ifa5 which
Leko was intending to answer with the
unclear 16 lt:Jd3!?.
16 liJxg6 'i¥b6!
Not 16 ...lt:Jxf4? 1 7 lt:Jxf4 i.. h 6 18 liJfxd5
and White wins.
17 liJh4! should then continue 23 ...f5 24 g4 'iVf6!
17 liJe5?! 'i¥xd4 is better for Black. 25 lt:Jxf5 (25 i.c4 + �h8 26 g5 'iVf8 is
17 lt:Jxf4!?
•.• amazingly unclear since White can't mate
Maybe not really bad, but certainly an Black; the same is true after 25 g5 'iff7)
extremely risky decision. The 'normal' 25...i.. xf5 26 'ifxf5 'ifh4! 27 i.. c4 + �h8 28
line would be 1 7...'i¥xd4 18 'i¥xd4 liJxd4 19 'ifh5 'ifxh5 29 gxh5 i.. h6 and although
i.. d 3 i.. e 6 20 i.. e5 liJd7 with approximate White is a pawn up the game should be
equality. drawn.
18 'ifxf4 'iVxb2 19 'i¥d2! 23 i.d7 24 h3?
•••
Of course, I saw this move but underes Now it's simply too late. It was obliga
timated it. As often happens, panic set in .. . tory to play 24 a5! with an unclear game.
19 liJe4?!
••• 24 'i¥f6 25 'i¥g3 c5!
•••
There is nothing wrong with 19 ...'iVb6, Now Black is not only better but win
since after 20 i.d3 i.. d 7 21 i.. f5 l:.xe1 + 22 ning! The a-pawn will be worth a piece.
l:txe1 l:te8 White has no advantage be 26 dxc5 i..xa4 27 %ld6 'i¥c3! 28 :f1!
cause of his stupid knight on c3. The line Leko, however, puts of a stiff resis
23 i.xd 7 :xe1 + 24 �xe1 liJxd 7 25 liJf5 tance.
.i.xd4! 26 'iVe8 + liJf8 only supports this 28 'ii' xg3 29 fxg3 i..e5 30 liJf5!
•••
original intention 2 l...i.e6? would fail to 32 liJe7 + �f8 33 lt:Jd5 l:.e5 is good
22 i.h5!. enough for Black.
22 'i¥f4 'i¥e6 (D) 32 �g7 33 :xf7+ �xh6 34 .l:xd7
•••
Black's position looks pretty dubious l':.ed8! 35 .l:xb7 %Ixd6 36 i..g4 l':.f6?
but in fact things are not so clear because What can be more natural then parry
the bishop pair and the position of the ing White's only threat and moving the
white knight yield Black certain counter rook away from White's passed pawn?
chances. However, now things become much more
23 c4 tricky, while Leko's suggestion 36 ...�g6!
If White wants to fight for an advan (centralizing the king is one of the pri
tage then sharper play is required. I think mary endgame rules!) would have won
23 h3 would be White's best chance. Black fairly easily, for example 37 c5 (37 i.. d 7 e3
232 Fire on Board
chances. iB %-
%!" " :%
:?.' :: -�·
// /' /' / /' / /
40 l:Ixh7+ %:?.'�;;; • • •
Now 40 h4? loses to 40 ...l:.8a7. //'// 1' / / /
All forcing lines were lost for White, for pawns. However, Leko didn't notice my
instance 42 i..b 7 a4 43 h4+ �g6 44 i.. e4+ simple plan, stood passively by and threw
�f6 or 42 h4+ �g4! 43 i.. b 7 �xg3! 44 the game away immediately.
l:Ig7 + �xh4 45 g3 + �h5. With the text, 55 l:.d6 56 l:.f2+?!
••.
Leko tries to put up a more stubborn de White's best chance would be 56 g4!
fence and makes it tricky until the very and although it seems to me that Black is
end. winning after 56....l:d3 + 57 g3 .l:bb3 58
42 a4 43 l:Ic8 a3 44 l:Icl a2 45 l:.al
••• g5+ (58 l:.a6 + �e7 59 l:.a7 + �d6 60
l':.d6 46 �h2 l:.d2 47 h4+ �g4 48 i..f3+ l:.a6 + �c7 61 h5 l:.xg3 + 62 �h4 l:.h3 + 63
�f5 49 �h3 l:Ia5 50 i..g4+ �e5 51 i..h 5 �g5 l:Ib4!) 58 . ..�e6 59 l:Ia6 + �d7 60
�f6 52 i..f3 l:.a4! (D) l:.a7 + �c6 61 �g4 l:.xg3 + 62 �h5 (62 �f5
Black has taken control of all the l:Ib4) 62 ... l:Ib4! , I am not one hundred per
squares from which the white bishop can cent sure about it.
attack the pawn, so now White has to give 56 �g6 57 l:.a2 l:.b3 58 l:Ia8?
.•.
it up. But, astonishingly, it's not yet the White could still reach a similar posi
end of the story! tion by playing 58 �h2 l:Ib4 59 �h3 �h6
53 i..c6 l:Ib4 54 i..d5 60 g4 .l:d3 + 61 g3 .l:bb3 62 l:.a6 + (62 g5 +
Forced. �h5 63 l:Ig2 l:Ib4 wins) 62 ...�g7 63 l:Ia7 +
54 l:Ixd5 55 .l:xa2
••• �f8 64 .l:a8 + �e7 but, as I already indi
It's not so easy to win the game because cated, I think that Black should still win.
when Black activates his rooks, White 58 �h7
•••
gets a lot of checks which drive the black Now it's all over.
king far away; then he can push his 59 l:.a7+ �h6 60 .l:f7 l:.dd3 0-1
Index of Opponents
Page numbers in bold indicate the games where Shirov was White.
Benko 50
Caro-Kann 120, 147
Dutch 14, 33, 48
English 8, 107, 132, 157
Evans Gambit 131
French 12, 75, 85, 112, 176
Griinfeld 26, 65, 66, 81, 94
King's Indian 19, 3 1, 43, 52, 56, 60, 74, 78, 91, 96, 106, 127
Modern I Pirc 68, 1 14, 1 18, 137, 154, 230
Nimzo-Indian 21, 35, 42, 44, 47, 54, 125
Queen's Gambit Accepted 70, 72
Queen's Gambit Declined 38, 40, 58, 135
Queen's Pawn Game 28, 1 16
Reti 92
Scandinavian 168
Semi-Slav 80, 83, 122, 144
Sicilian 30, 62, 87, 89, 101, 103, 109, 123, 129, 146, 149, 159, 161, 174, 179, 181, 227
Slav 99, 164
Spanish 16, 141, 143, 152, 165, 171
Index of Variations (Botvinnik System )
Page numbers are given in italics.
16 ••• 'ii' b 5
16 . . .'ii'a6 201
16 .. .'�'d6 203
17 a3
1 7 dxe6 1 90
17 . . . lL:lb8 1 89
Index of Endgames (Chapter 5)
Advantage of the exchange 2 1 1, 213, 226, 230
Bishop vs. knight 223
Piece vs. pawns 2 14, 230
Queen and bishop vs. queen and knight 222
Queen ending 2 12, 22 7
Rooks and opposite-coloured bishops 2 18, 224
Rook ending 22 7
Two minor pieces vs. rook 2 1 7
FIRE ON BOARD
SH IROV'S BEST GIMES
• 11 ISBN 1-85744-131-1
\
9
'