Вы находитесь на странице: 1из 186

Colin Crouch

we lose
at chess

Gloucester Publishers pic www.everymanchess.com

First published in 2010 by Gloucester Publi shers pic (formerly Everyman Publi sh ers
pic ) , Northburgh House, 10 Northburgh Street, London EC1V OAT
Copyright 2010 Colin Crouch
The right of Colin Crouch to be identified as the author of thi s work h as been as
serted in accordance with the Copyrights, Designs and Patents Act 1988.
All rights reserved. No part of thi s publication m ay be reproduced, stored in a
retrieval system or transmitted in any form or by any m ean s, electronic,
electrostatic, magnetic tape, photocopying, recording or otherwise, without prior
permission of the publish er.
British Libra ry Cataloguing-in-Publication Data
A catalogue record for thi s book is avail able from the British Library.
I S B N : 978 1 8 5 744 6 3 6 4
Distributed in North America by The Globe Pequot Press, P.O Box 480,
246 Goose Lane, Guilford, CT 0643 7-0480.
All other sales enquiries should be directed to Everym an Chess, Northburgh House,
10 Northburgh Street, London EC1V OAT
tel : 020 7 2 5 3 7887 fax: 020 7490 3 708
email : info@everym anchess.com ; website: www.everym anchess.com
Everyman is the registered trade m ark of Random House Inc. and is used in this
work under licence from Random House Inc.

Everyman Chess Series

Chief advisor: Byron J acobs
Commissioning editor: John Emms
Assi stant editor: Rich ard Palliser
Typeset and edited by First Rank Publishing, Brighton .
Cover design by Horatio Monteverde.
Printed and bound in the US by Versa Press.

Co n t e n t s



Test One


Test Two


Test Three


Test Four


Test Five


Test Six


Test Seven


Test Eight


Test Nine


Test Ten


Test Eleven


Test Twelve


Test Thirteen


Test Fourteen


Test Fifteen


Your Move?



While writing up reports for various m ag azines, I h ave been thinking about the
idea of 'ordin ary chess', of g ames which are not technically perfect, but m ay still
be of interest to players, spectators, and hopefully to readers. For this to work well,
the writer h as to take the annotation s seriously. The idea i s that while the g ame is
interesting anyway, just think of wh at spectacul ar ideas might h ave been thought
of if the pl ayer could h ave found the occasional improvem ent. Often in chess, bril
liancy is just around the corner.
I h ave used a similar perspective in this book, but with a different, almost op
posite perspective. I am writing up 'ordin ary games', my own, with th e thought of
systematically going through them, spotting any mistakes of my own (and there
are several), and finding better moves. I am aiming to find ways of cutting out
mi stakes, thereby improving both my pl ay and th at of the reader.
Many g ames h ave been played in local and national league events, and I dedi
cate this book to those who continue to keep chess clubs going, in wh at is often
quite a difficult time. These days I am cautious about playing in long tourna
ments, and al so quickpl ay tournaments, sometimes travelling from one end of
Britain to the other. It is good to play in my local club, Harrow, where there are
often fifty chess pl ayers in a single evening, sometimes close to sixty if there are
visiting team s. My thanks to colleagues.
Colin Crouch,
Harrow Weald,
April 2010

I n t rod uct i o n

Thi s i s a book of my own g ames. It i s definitely not a compilation of my best wins,

attempting to impress the reader. On the contrary, what I am trying to do i s to
identify all my serious mistakes over a period of several months. I am not quite
sure whether thi s exercise has been tried in public. Attempts h ave certainly been
m ade to an alyse the losses of great World Champion s, such as Capabl anca and
Fisch er, but there seem s to be very little publi shed autobiographical work of a
pl ayer's own losses.
We can feel sure th at strong grandm asters will h ave an alysed their games in
depth, in order to examine any weaknesses in their own play. It is a matter of sur
vival at top level . If you do not find your own weaknesses, you opponent will be
more than h appy to demon strate wh at you h ave done wrong . Of the earlier World
Champion s, one can imagine that Mikh ail Botvinnik would have been extrem ely
methodical in going through his post-mortem an alysi s, uncovering both his own
mistakes and his opponent's mistakes, and learning from all this. It is enough to
rem ember that he was World Ch ampion from 1948 to 1963, a formidable stretch,
and th at while he lost m atches again st Smyslov and Tal, he successfully won the
return m atches. He was al so a great teacher, and a pioneer of ideas in computer
ch ess.
At lower level s, one might argue th at games at purely am ateur stren gth might
be of only minor interest, since there are often many mistakes, but a pl ayer might
not understand why some moves are weak, and other moves are better. Of course,
in saying this, my aim is not to try to condemn am ateur chess. I am hoping to pro
voke interest into trying to encourage ordinary am ateur pl ayers and ambitious

Why we L o s e a t C h e s s
young players t o think about how t o pl ay better.
I am somewhere between the amateur level and the strong grandmaster level .
I am an International Master, with a good understanding of the basic ideas of
chess, but also the ability to m ake serious mistakes over the board, even ag ainst
much weaker opponents. A Dutch 1 M once characterized me as "a chess artist,
rather than a practical chess player", a reasonable comment, but it would be nice
to discover how to be practical . Possibly in writing this book, it would h elp me to
be more practical . But is thi s an excuse, or is it more a case th at while I h ave learnt
how to write a decent game of chess, I h ave not really learnt how to play a good
game of chess?
In term s of life and death in chess, as player and writer, there is something
even more important to me. It is a m atter of chance that I am still alive, in 2010,
rather than dead. It is a matter of chance th at I am merely partially sighted, rather
than blind. It was going to take a lot of h ard work to recover after my stroke in
2004, and I still h ave only partially recovered. I could not read for some time, al
though now I can read slowly, preferably on large print.
Fortun ately, I could see a chess board, just about, and I could therefore play
chess after my stroke. My thought processes were still slow after brain damage,
and at first I felt doubtful as to whether, if I pl ayed, I could play at over F I D E Elo
2000 strength (about 175 ECF rating). My memory was h owever largely intact, al
though it was going to be a lot of time before I could connect different thoughts.
Aphasia i s still a problem. I can understand wh at i s being said by others, but I
could not always string the words together when trying to read.
I needed to keep my mind active, and learn to think con structively again, in
chess or in anything el se. If anything, my thought processes became much more
focussed, as I felt th at I could not waste time. It is difficult enough that I found I
h ad to take naps in the afternoon, and th at my thoughts were no longer able to
fizz. I needed to think carefully about wh at to do next.
Chess was by now much more important in my life, even though I was playing
much less. I could no longer think in term s of playing lots of weekenders all over
the country (Scotl and and Wales, as well as England), and I h ave become increas
ingly reluctant to pl ay two g ames in a day. My g ames mentioned in this book, at
Bury St Edmunds and Kidlington, give good examples as to why. Creativity i s use
ful, but if you want to be successful, you need to focus on good technique, and you
h ave to respect tiredness.
I still wanted to show I could play good chess. Playing a standard nine-round
tourn ament was slightly beyond what I was capable of doing with comfort, but
m aybe in time I could try thi s agai n ? I h ave to admit th at almost five years after

In t ro d u c t i o n
m y stroke, I still h ave not summoned m y courage i n playing a nine-rounder, apart
from pl aying the occasional Braille event, and a small Middlesex versus junior in
tern ation al in London . I got too tired.
I al so wanted to get back into writing, and for a while I did not even go through
my own g ames afterwards. It was so embarrassin g . I had to tell myself th at thi s
was t h e result o f tiredness, because o f concentrating o n my books, rather th an a
sudden deterioration in my brain. At the time, I was working on my book on Tal,
Kasparov and Stein, and I felt confident that while I probably did not analyse com
pletely accurately in any position (who can ?), I was not yet gaga. So I continued
writing, while h oping th at my pl aying strength could improve again.
I h ad a g ap in my calendar in 2006/2007, before attempting serious chess
analysis on top grandmasters, and in this g ap I was writing up my own g ames,
maybe for publication, but primarily for my own interest, and learning again h ow
to play good chess, and how to write. Essentially the result is this book, although I
was able to revise my comments in 2009, not just because my earlier play and un
derstanding might not h ave been as good as I would h ave liked, but al so, m ore
importantly, because thi s was only an initial draft.
Back to Playing Chess

Clearly I wanted to play chess again, but I was not seriously out of touch enough to
think that I would be able to reach my peak in chess. My h ope was th at I could still
play chess, and not decline too fast. I h ad in fact lost almost a hundred Elo points
just before my stroke, and this was at the time a mystery for me. Now, though, it is
all very clear. There was accumulating dam age to my brain before the stroke.
Even so, I wanted to show that with con structive thinking I could recover most
of my peak, despite the slowing down of age, and other problem s. I am not too
surprised th at I h ave not fully achieved thi s yet.
Other readers, looking to improve their chess, will inevitably be thinking in dif
ferent ways. In particular, the young player, h aving reached a degree of experi
ence, will calcul ate quickly, and learn openings speedily, but will not yet h ave the
detailed knowledge of positional accuracy. These days I would not be able to calcu
late lines a dozen moves deep, with sub-variations, but an ambitious teenager
would see thi s as the core to chess improvement. If you can calculate quickly,
when your opponent can calculate less quickly or deeply, you h ave a clear practical
For those over thirty, the player will h ave to modify thought processes. The
general procedure would tend to be that, now you cannot calcul ate everything, it
is best to use your knowledge and experience to cut out extraneous thought proc-

Why we L o s e a t C h e s s
esses. For me, I h ave been forced t o take a slightly more extreme approach, as
brain dam age means th at I am not able to think quickly.
This is recognized in this book. I h ave tended much more th an before to cut
down the number of moves to try to analyse, and concentrate instead on thinking
about just a few moves, systematically. In other words, do not try to thin k of a long
list of possible moves beforehand, trying to assess each position, but start off with
a couple of moves to con sider at first, while keeping a quick note of other lines.
The idea i s to examine two moves first, normally the most plausibl e moves. N atu
rally if there is an immediate tactic which must be calcul ated, then examine it. If it
can be rejected immediately, move on to other lines. Wh at I am thinking about is,
for example, a double piece sacrifice which brings the king out into the open, but
can quickly be seen as un sound. If, h owever, there might still be possibilities, then
look at it again. You never know, you might h ave a brillian cy.
Once we h ave made a quick scan of immediate checks, captures, and other bril
liancies, al so cutting out any immediate big threat by the opponent, we probably
h ave a few moves to be con sidered. These might be attacking moves, or defen sive
moves, or positional moves. You need to keep your eyes open .
For simplicity, in this test book, we give three alternatives; move A, which I am
regarding as the most obvious, then move B, the m ain alternative. If there are two
moves to be con sidered, and other moves are irrelevant, that makes life simple.
One must, of course, keep in mind th at there are al so possible altern atives, start
ing with 'Something Else', move C, and then m aybe D onwards.
Even if you feel you cannot an alyse in great depth over the board, you still h ave
to m ake a deci sion what you are going to play. U sually it is best to concentrate on
the most n atural move, A. If you are confident that it is a good move, and any
other move (C) seems senseless, you should play it without spending too much
thinking time. If, however, you decide that your initial move is not fully satisfac
tory - m aybe because you feel th at it is promising, but th at there should be some
thing better, or maybe because it i s ultim ately bad - then you need to think of al
tern atives.
Remember that for much of the time, the first move you think of i s often the
best. Thi s is because you h ave already been thinking about th at idea when looking
at the previous position. Maybe your opponent h as pl ayed the reply th at you h ave
expected, or m aybe there was a slightly unexpected altern ative, but your possible
reply might well seem good and natural . Or m aybe there might be som ething bet
Think of choice B next anyway, but remember that you must retain your as
sessment of the evaluation of choice A. Is your first move ending up as slightly bet-


In t ro d u c t i o n
ter for you? O r just equal ? Or, perhaps the m ost common assessment, slightly
worse, or at least m aking you feeling uncomfortable? Or is your initial thought
quite simply bad?
Then an alyse B, but with a quick flick though to see whether moves C, D, or
maybe even beyond could be worth trying later. As some guidance, if you are
thinking of analysing a fourth move, or even beyond, you are at ri sk of entering
time trouble, sooner or l ater. If you are juggling six possible moves in a given posi
tion, you will h ave to calcul ate much more th an three times as much than when
there are only two moves to be con sidered. You h ave done your basic calculation s,
and you must decide whether A or B i s better, and then, for example, whether B or
C i s better, then perh aps B or D. In the fin al stage, when you compare B and E,
wh at you would really not h ave wanted in retrospect i s to find that B i s better, and
th at in trying to analyse lines D and E, m aybe also F as well, you h ave wasted time
on the clock.
It is a m atter of judgement to decide whether m oves D and E should be ig
nored. Maybe D looks interesting at first, but a couple of moves l ater, your pawn
structure is sh attered, and you do not think th at it is worth defending the line. On
the other h and, E might be genuinely tactically interesting, and requires more
thought. We must not forget, either, th at an F try might well be worth examining,
even though first tim e round it did not seem so effective.
Quite clearly there are difficult deci sion s to be made. It is difficult to generalize
on how players should find the best thought processes, and find the best move.
The ideal is that a player should be able to calcul ate with complete accuracy, but
of course a hum an player cannot achieve this in over-the-board play. If anything,
it m akes chess far more interesting if the hum an player h as to find good, or in
deed excellent, moves without the h elp of a computer. It is an exercise for the
mind, and top players should quite properly be accorded great respect when find
ing accurate play and in spired brilliancies.
Most of us h ave great imperfections in our chess, but we do not give up the
g ame in respon se to them. We need to develop strategies to find ways of finding
good moves when we cannot calcul ate everything, and when we do not h ave full
understanding of positional chess. I do not pretend that I h ave found the an swer.
All I can do is to try to pinpoint ways in which mistakes, and indeed my own mis
takes, are m ade.
This is a preliminary investigation . I h ave indicated in this book 60 mistakes
over several month s that I h ave m ade. The main point is th at most of these mis
takes are not the result of highly complicated and difficult play. Just because I am
a master, I can still pl ay rubbi sh chess. At least half of these mistakes could easily


Why we L o s e a t C h e s s
h ave been avoided by better thought processes. This is easy enough t o establish, to
the extent that in the test positions, A, B, and 'something else' (C}, I give improve
ments on each of my own games.
List of Exercises

1.1 Crouch-Oryakhal (White's 5th)

1.2 Crouch-Radovanovic (White's 6th)
1.3 Nurmohamed-Crouch (Black's 7th)
1.4 Wall-Crouch (B7)
2.1 Buckley-Crouch (B9)
2.2 Crouch-Rose (Wll)
2.3 Wall-Crouch (Bl0)
2.4 Crouch-Radovanovic (W12)
3.1 Nurmohamed-Crouch (B12)
3.2 Crouch-Peacock (B12)
3.3 Morris-Crouch (B12)
3.4 Crouch-Lewis (W13)
4.1 Lauterbach-Crouch (B13)
4.2 Crouch-Gait (W14)
4.3 Sen-Crouch (B14)
4.4 Sowray-Crouch (B14)
5.1 Crouch-Hutchinson (W15)
5.2 Crouch-Granat (W16)
5.3 Crouch-Peacock (W16)
5.4 Hebden-Crouch (B16)
6.1 Morris-Crouch (B17)
6.2 Crouch-Hutchinson (18W)
6.3 Hebden-Crouch (B18)
6.4 Lauterbach-Crouch (B18)
7.1 Crouch-Peacock (W19)
7.2 Crouch-Lewis (W20)
7.3 Sowray-Crouch (B20)
7.4 Crouch-Peacock (B21)
8.1 Buckley-Crouch (B21)
8.2 Pert-Crouch (B21)

8.3 Wall-Crouch (B22)

8.4 Nurmohamed-Crouch (W22)
9.1 Crouch-Cox (W24)
9.2 Crouch-Gait (W2 5 )
9.3 Crouch-Granat (W25)
9.4 Randall-Crouch (B25)
10.1 Morris-Crouch (B25)
10.2 Nurmohamed-Crouch (B26)
10.3 Lauterbach-Crouch (B26)
10.4 Crouch-Gait (W27)
11.1 Crouch-McKenna (W27)
11.2 Crouch-Okike (W27)
11.3 Buckley-Crouch (B27)
11.4 Pert-Crouch (B27)
12.1 Crouch-Cox (W28)
12.2 Crouch-Roberson (W28)
12.3 Cutmore-Crouch (B30)
12.4 Crouch-McKenna (W3 1)
13.1 Randall-Crouch (B31)
13.2 Gregory-Crouch (B32)
13.3 Crouch-Granat (W34)
13.4 Gregory-Crouch (B34)
14.1 Morris-Crouch (B35)
14.2 Cutmore-Crouch (B35)
14.3 Crouch-Roberson (W37)
14.4 Crouch-McKenna (W38)
15.1 Crouch-Granat (W39)
15.2 Buckley-Crouch (B46)
1 5 . 3 Crouch-McKenna (B48)
15.4 Pert-Crouch (B50)

Thi s table is not merely a brief contents page; it is al so a summ ary of research,
and a starting point for further examination . There are 60 identified mistakes in


In t ro d u c t i o n
thi s book, and no doubt I h ave overlooked some further points. I h ave not bothered
giving every single slip in each game if I played particularly awfully, for example,
in my horrible game again st McKenna (he h ad so m any chances of beatin g me), or
even in the l ater stages again st Pert, when I m ade so m any bad moves around the
time control, but arguably the worst of these m i stakes was missing an unexpected
ch ance of finding a fortuitous draw in the endgame.
Ready to start?

We h ave 60 exercises for the reader to con sider. It i s important to rem em ber that
thi s is not a quiz. We are not asking you to try to dig out your memory.
More to the point is asking readers to think about new and original positions,
to try to find the best move, and to avoid identifiable mistakes. I am setting out
my own errors, at m aster level, and you are invited to m ake improvements. Also,
at another level, you can think about how you can take full advantage of mistakes
by your opponent. Not all mistakes get punish ed. Quite often there could easily be
half a dozen slips in normal pl ay. If one of the pl ayers win s quickly, that often
mean s th at the opponent's mistake is so serious th at a reason able pl ayer should
be able to move quickly.
When playing through the g ames, take them seriously, but not excessively so.
Trying to analyse in great depth, with the h elp of your own brain and computer
suggestion s, can be extremely absorbin g . Som e of the position s in this book a re
analysed in great depth, sometimes spilling over extra days. The writer likes to aim
for perfection, but of course this does not always happen. The reader might by
daunted by the thought that it takes a coupl e of pages of inten se analysis to show
th at one interesting move eventually turn s out to be better than another. The
point is, though, that over the board one can only see a fraction of what could
have h appened, and quite often it is possible to say that a player is "lucky" if the
critical move turn s out to be good, and "unlucky" if an obscure move turn s out to
be an unexpected refutation . A player who loses m ay feel he is unlucky, but it is
still a loss and it is important to cut down your unluckiness by working out how to
avoid mistakes.
For the reader, think about seeing a new position in a g ame at the board. It i s
just a n ordin ary position, so w e are not asking you t o find a brilliant sacrifice. W e
are asking you t o find a good, ordinary move, avoiding any pitfalls.
At an initial glan ce, you will quickly decide whether you think your position is
better, or equal, or worse. Or m aybe you just won't know what is h appening and
will need to look further before you can m ake any sort of judgement. Once you
have sorted out your background information, you can try to decide wh at your


Why we L o s e a t Ch e s s
next move is going t o b e . I f there is genuinely only o n e sen sible move t o m ake, you
can play it immediately. Most of the time, you h ave to think, which is h ard work.
You are being asked to find a good move, and for the first part of the exercise,
to make a decision quickly. What would you play if you h ad to do something in the
next fifteen seconds? In this exercise, you h ave three option s, although the third
option is given as "som ething el se", which might be a choice of h alf a dozen rea
sonable alternatives, particularly in "quiet" position s. Or, on the other h and, there
are no reason able altern atives, and it is really only a choice between A and B. As
far as possible, I h ave tried not to give a big clue about the best move. Of the three
options given, A, B and C, one is the move I actually pl ayed, but was a mistake. The
other suggested option i s a natural move, which m ay h ave been right or wrong .
You need t o bear in m i n d th at there i s a third possibility, which might o r might
not be good.
For the initial assessment, decide which move you would play, and write it
down . We are not yet asking you which move you would decide to play in a tour
nament. You will h ave m ore tim e to think l ater. Wh at you are being asked for now
are your first impression s.
The next stage is to decide which move you would pl ay in a tournament g ame.
If you genuinely think that you h ave decided on your choice after a couple of min
utes, write down your move. In run-of-th e-mill positions, you would h ave on aver
age three minutes, or less, to decide anyway. In critical and difficult positions, you
will want to think for much longer. It often h appen s that a player m akes moves
quickly early on in the g ame, and then suddenly slows down . In a critical position,
a player will be aware th at the best move m ay keep an advantage, while a second
best move leads to only equality, and a worse move, attractive but leading to a
tactical oversight, may end up as being bad. Take your time, but do not waste
time. The clock ticks, and it does not h elp if you are spending 40 minutes on a
move if you then have to m ake your last ten moves before the time control in only
one minute. In a tourn ament, at some stage you will h ave to m ake a deci sion, and
it i s often best, if not necessarily desirable, to save some time for thinking l ater.
Use wh atever time you want in this exerci se, then m ake a deci sion and write
your move down . Do not be worried if you h ave ch anged your mind since your
initial assessment. It m ay well be th at you h ave given yourself extra time to allow
extra clear thinking. Maybe you h ave decided that the initial assessment was
wrong, inaccurate, or un subtle, and you h ave corrected your thought processes. A
more worrying aspect might well be th at you h ave chosen th e correct move i m
m ediately, but after some more thought you h ave introduced a n extra error
through over-sophi stication. Indeed, on the day after I finished the initial draft, I


In t ro d u c t i o n
m ade precisely thi s mistake.
Another possibility is th at you chose an inaccuracy, and pl ayed it too quickly.
This can be very common in the opening if a player saves time by relying on "natu
ral moves", and the problems come later. I h ave h ad to remind myself that it is no
bad thing to spend ten minutes in the opening, rather th an pl ay the first ten
moves in a couple of minutes.
After going through the exercises, the next stage i s to read closely through my
annotation s, in which I h ave m ade use of computer engines and of course a con
siderable amount of hindsight. I think I h ave learnt a lot from pl aying through my
own g am es, and I hope th at you will learn a lot in playing through them too, poor
though they m ay be when judged at the top level . After that, it is up to you to start
thinking about your own games, and to work out how to improve your play in
later encounters.
Good luck!


Te st On e
1.1 White to play

1.3 Black to play

A) 5 cxb S
8) 5 Ji.e3
C) Something el se?

A) 7 . Ji.cS
8) 7 ... Ji.e7
C) Something el se?

1.2 White to play

1.4 Black to play

A) 6 e3
8) 6 ttJes
C) Something el se?

A) 7 . .f6
8) 7 .. a6
C) Something el se?



Why we L o s e a t C h e s s

My opponent, and for a time a club

colleague, i s what could be described
as a 'n atural ' player, or an 'experimen
tal ' player. He has shown little interest
in the opening, or in studying the end
g ame, but can at times be dangerous
in tactical middlegames, and m any
club members h ave suffered quick
losses to him.
It was disconcerting that I was al
ready worse as White, within h alf a
dozen moves.
1 d4 e6 2 c4 c6!?

leads to a French Advance where Black

h as exch anged prematurely, compared
with 1 e4 e6 2 d4 dS 3 eS cS 4 c 3 . After
4 ... exds ! , h owever, even the World
Ch ampion of the time could not find a
convincing edge in M.Tal -A. Bisguier,
Bled 1961. Following S eS tDa6 6 tDC3
tDo 7 tDge2 tDe7 8 tDf4 tDfs Black
seem s to h ave comfortably equalized,
but Tal l ater ground his opponent
down in the endg ame. Thi s was
reached via a Caro-Kann move order (1
c4 c6 2 e4 e6 3 d4 dS, etc).
So 3 ... dS i s playable, but Qryakh al
found something more original . He
played 3 a6!? in Basm an styl e, but
perh aps even more effective th an the 1
e4 a6 2 d4 b S St G eorge's Variation .

Thi s is by no mean s a bad line for

Black, and if I develop with 3 tDc3 or 3
tDf3, Bl ack can quickly move into a
Semi-Sl av with 3 ... dS. I do not know
whether h e would h ave tried some
thing el se, although I suspect he would
h ave done.
My instinctive respon se was 3 e4, so
th at if 3 ... dS 4 cxdS cxdS, then S e S

It's not so easy for White to set up

any clear advantage, and indeed the
computer gives only a plus of '0. 19'. All
I can sen sibly do i s develop my pieces,
and try to ensure th at I will be able to
keep a plus. No heroics.
4 tDf3
This cannot be bad, developing the
knight to a solid square. However, 4

Test 1. 1

H i l l i ngdon League




Te s t On e
a4?, to prevent Black from playing ... bS,
would be an over-reaction . White
would h ave weakened his b4-square.
Black naturally plays 4 ... bS, leading to
our first test position .

I pl ayed 5 e3? (B) rather carelessly,

and after S bxc4 6 xc4 (m aybe 6 e S
d S 7 exd6 xd6, but Bl ack is still equal)
6 ... d S ! 7 exd s cxd S 8 b3 I started to
appreciate that while my pieces were
better developed, Black h as the better
pawn structure, and should be at least
equal . Un accountably Black then tried
8 h6?, with a clear loss of tempo, and
then after a few later inaccuracies
White was able to play for an edge, and
later won .
White should instead h ave m ade a
clear decision over his c-pawn.
I did not enjoy the thought of giving
away an extra tempo with the pawn
before exchanging, namely S cS d6 6
cxd6 xd6. After 7 C2 b7, however,
Black's ... c6 move h as got in the way of
his pieces. So s cSI (C) is an improve
ment, possibly with a slight edge.
There's al so S cxbS axb S (A), which I

slightly distrusted at the time, with

Black keeping the extra pawn in the
centre (the c-pawn, as opposed to
White's a-pawn). However, Black h as
various slight pawn weaknesses on the
queen side, m aking the assessment
finely balanced. Bl ack m ay well want to
pl ay ... dS, but it needs to be well tim ed.
Flipping my assessment during the
g ame, I would now prefer S c S (C) to S
cxbS (A), but either move would be
preferable to S e3 ? ! (B).
The rest of the game is of less inter
est, from my point of view, as my op
ponent did not play accurately, and I
quickly went from a slight disadvan
tage to equality, then an edge and a
win, with 9 0-0 iLl f6 10 iLl es d6 11 f4
0-0 12 iLl d2.

Test 1.2

London Open



One of those slight slips in the

opening which i s easily missed - here
by both sides. The trouble is that both
players will want to play quickly in the
opening, to avoid the danger of time
pressure later on, but if the players do
not slow down as soon as something
unusual h appen s, one player will un
expectedly be worse, even after only a
few minutes.
Play started with 1 d4 d s 2 c4 c6 3
iLl f3 iLl f6 4 iLl c3 dXc4 5 a4, and then
S ... g4, instead of the usual S ... fS .


Why we L o s e a t C h e s s
7 dxe s 'ilVxd1+ 8 'Dxd1 'Dfxd7 9 .1i.xC4
.1i.xf3 10 gxf3 'Dxes 11 .1i.e2, since Black
is slightly ahead in development. If my
opponent h ad played the best move, I
might in the end h ave tried this line,
but it leads to no advantage.

Test 1.3

Thames Valley League 2006

Since m y stroke, a coupl e of years
earlier, I h ad not re-examined thi s posi
tion, and was taken slightly by surprise.
I played the natural 6 e3? (A), but 6
'Des! (B) is more effective, avoiding the
pin . Of course, I was well aware of the
knight thrust, but was concerned that I
would be caught in a m ain line which I
h ad not revised.
My opponent played 6 ... e6?, which
suggests th at he too was not fully
aware of the theory. We will see how
play continued in Test 2.4.
In stead 6 ... e s ! is fully equal . I am
sure that had I quickly skimmed
through some likely lines before my
move, I would h ave decided that since
we are still in standard opening play, 7
.1i.xc4 .1i.xf3 8 gxf3 (8 'ilVxf3 ? ! exd4 loses
a pawn, without sufficient compen sa
tion) 8 ... exd4 9 exd4 should be at least
playable for White, and would not h ave
an alysed much further. After all , saving
five or ten minutes in the opening can
often be useful later on.
I doubt whether I would h ave
wanted to think, in advance, of playing


Thi s was the start of a nightmare

g ame in which my mind was not in
focus, I became dizzy, and m ade several
mistakes. Fortun ately I became more
alert during the end of the session,
woken up by the thought that I could
easily lose, and was able to negotiate
some defensive tactics when he h ad
opened up lines again st my king . There
was no possible doubt, h owever, th at
during the later stages, h e h ad a clear
chance of a win .
The opening moves started 1 e 4 c5 2
'Df3 e6 3 d4 cxd4 4 'Dxd4 'Df6 5 'Dd2
(too quiet) 5 ... 'Dc6 6 'D4f3 'ilVC7 7 c3.

Te s t On e
White has lost a lot of tim e with his
knights. Should Black already be play
ing for an edge? Thi s seem s a little too
early. White has played ineffectively,
but he starts off with the advantage of
the first move, and it often takes more
than a couple of quiet moves by White
to allow Black to be already better.
Usually the m ost practical approach as
Black in such a scen ario i s to pl ay quiet
but forceful moves early on, and to al
low the opponent to try to prove that
he is equal.
In stead, I quickly played the naive
attack 7 ... i.. C 5?! (A). It's a developing
move, but it can h ardly be expected
that my opponent will be un able to
find a good defence to cover the f2square. As we will see in Test 3.1, a few
moves l ater, I soon found out, in the
main line and in a few variations, th at
the bishop is too exposed, and th at
White can create counterpl ay with one
of b4, ttJb3, or (after ... dS) a pawn ex
change on dS, and if Black recaptures
with the knight, then ttJe4. Black would
not necessarily be worse, but h e has
lost his initiative.
A reality check h elped by using the
computer i s th at the position after 7 c3
is close to equal . The hum an pl ayer
might well want to argue that Black
still keeps the initiative, though, by
playing a Scheveningen set-up, with
7... i..e7 (B), and ... d6, when Black i s
solid in his centre a n d his pieces can
develop quietly and securely. White's
pieces cannot be more effectively

placed here than in standard openings.

Possibly White can still equalize, or
possibly Black with sen sible play can
demon strate an edge. Who knows?
What i s more certain th at Black has
played better than White in the open
So 7 ... iLe7 i s the suggested m ain
line, but 7...d6 or 7... dS are fully ac
ceptable too, under 'something else'
We sh all return to this g ame l ater.

Test 1.4

B ritish League (4NCL)


After 1 e4 c5 2 ttJf3 e6 3 c3 ttJf6 4 e5

ttJd 5 Wall played the slightly unusual 5
ttJa 3 .

Going through with the computer a

few days later, I saw that he had tried
this before. Maybe there i s a lesson
h ere? Or maybe not. If I had taken the
opportunity of quickly using the oracle
an h our before the game, I could have


Why we L o s e a t Ch e s s
tried t o work out a possible improve
ment. U sually I prefer not to bother.
Chess is a tiring game, and I prefer not
to turn a four-hour session into a five
hour session, through pre-game analy
sis. In swiss tournaments or team
events, it is not possible to examine in
leisure the openings a day before. An
extra hour resting in bed, or a gentle
stroll is often better th an pre-match
Play continued S ... d6 6 tiJC4 dxes 7
tiJcxes, and here I h ad to decide over
the board what to play next.

Here the chance of preparation in

advance would possibly h ave h elped
me to the extent th at I would not h ave
been too anxious of a check on the a4e8 diagonal . Indeed, the computer in
dicates that thi s check i s not too dan
gerous, and that Black could quite sim
ply h ave developed.
I pl ayed 7 ... a6?! (B), aiming to con
solidate in the centre, but also giving
away a tempo with a pawn move. Black
is close to equality, nevertheless, as we
will see in Test 2 . 3 , but it would h ave


been useful to end up with complete

7 .. i.. e 7 (C) is the most natural de
veloping move, after which I probably
would h ave been slightly worried by 8
i.. b S+ tiJd7 9 d4 0-0 10 0-0. White keeps
slightly more space in the centre, and
probably a small but irritating edge
after, for example, 10 ... cxd4 1 1 i.. x d7
tiJxd7 12 'iYxd4. In a 4NCL encounter
several years earlier, Wall played more
directly in the centre, with 8 d4 0-0 9
i.. C 4 cxd4 10 'ixd4 f6 1 1 tiJd3 tiJc6 12
iVe4, and his pieces in the centre were
irritatingly effective in G .Wall-N. Davies,
Briti sh League 2000.
7... tiJd7 (C) is also playable and
n atural, but after 8 tiJxd7 i.xd7 9 tiJes
i.. d 6 10 tiJxd7 'ixd7 1 1 d4 White i s bet
ter placed in the centre, even if only
Sometimes it m ay be acceptable for
Black, even very early on, to take the
initiative and play for an edge, even
when there is no obvious serious mis
take by White. Thi s might well h appen,
for example, if the player with White
tries slow manoeuvring in the opening,
and Black sets up some well-timed
counterpl ay.
The computer suggestion is 7 .../61
(A), and if 8 i.. b S+ then, as a second
choice, 8 ... e7 ! . I am sure that h ad I
been a teenager, I would h ave seen thi s
a s the obvious a n d natural line, but as
one gets older, the inclination is often
to try to cut out the more bizarre ideas,
since quite often they turn out to be

Te s t On e
wrong . Strange moves still need to be
con sidered, though, as well as n atural
moves, and it is best not to eliminate
too quickly some prom i sing ideas.
If then 9 ct:Jg4 e s 10 ct:Je3 e4 11
ct:Jxds+ 'iVxds 1 2 'iVe2, and Bl ack is
happy with his position . Once his king
has taken the time to escape to f7 after
12 . . . e6 (though not immediately
12 . . . 'it>f7 ? ? 13 C4), his other pieces and
pawn s will be more active.
White therefore can think about
aiming for a sacrificial attack, with 9
d4! ? Then perh aps 9 . . :as (there are
other tries) 10 d3 ! ? fxes 11 g s+ 'it>e8
12 ct:Jxes g6 1 3 'ilVf3 'ilVC7 14 0-0-0 (or
maybe 14 bs+ ct:Jc6 15 0-0-0 first),
and we reach a familiar sacrifice of
knight versus pawn, where Black's king
is in the open and his pieces are not yet
fully developed.

I leave thi s as an exercise in annota

tion for the reader, whether with the
traditional aid of board and set, or with
the h elp of the computer.
A possible line is 14 . . . cxd4 ! ? l s l::th e l
ttJxc3 16 bxc3 'ilVxC3+ 17 'it>bl 'i!Vb4+

with a perpetual, and thi s i s a natural

way of finishing. H owever, h ere
16 . . . a3+? 17 'it>d2 'ilVxC3+ 18 'it>e2
would h ave been unwise, as White's
king i s much safer than Black's, with
rooks and minor pieces now being
Bl ack could try in stead 14 . . . i.d6 1 5
..ibs+ ct:Jc6, and the position would
seem to be m ore or less incalcul able
over the board after 16 ..id8 'ilVg 7 17
ttJxc6 ..id7. Try it! H owever, pl ayers are
allowed to accept sacrifices, and the
simple 16 . . . 'it'xd8 1 7 ttJf7+ 'it'e8 18
ttJxh 8 C4+ 19 'it>bl d7 i s good for
There are m any other lines to be
con sidered from move 9. Earlier, White
al so h as 8 'ilVa4+.

This time 8 . . . 'it>e7 ? ! takes things too

far, since White has 9 4!, and if
9 . . . g s 10 ttJg 6+!' So instead 8 . . . ttJd7 9
ttJxd7 iLxd7, and in my previous notes,
I suggested that there is "comfortable
equality for Black". This is a shorthand
way of saying that Black m ay already
h ave gone beyond equality, and be


Why we L o s e a t C h e s s
slightly better, but there i s no real
point in analysing every line in depth .
H ere, for example, 10 i.bs i.xbs 1 1
'iVxbS+ 'iVd7 12 'iVxd7+ 'it'xd7, and 10
iVe4 i.d6 11 d4 fS 12 'iVe2 cxd4 13
tbxd4 '*'f6 both seem promising for
In summ ary, 7 .. .f6 ! seem s the best,
or at the very least the most promising,


although Black must be aware that

extreme tactical complication s m ay
arise. With best play, Black should end
up with an edge, but we need to re
m ember that few players can play per
fect chess in a long and inten se battle,
and the likelihood is that the better
player will win .

T e s t Two
2.1 Black to play

2.3 Black to play

A) 9 . . cxd4
B) 9 ... ..te7
C) Something else?

A) 10 . 'iVC7
B) 10 ... SLe7
C) Something el se?

2.2 White to play

2.4 White to play

A) 11 'iVe2
B) 11 a3
C) Something el se?

A) 12 1tJxg 6
8) 12 0-0
C) Something else?



Why we L o s e a t C h e s s
Here I played 7 c5, and I com
mented th at "7 ... iLe7 appears to be
more common. Then after 8 iLd3 iLxd3
9 'ii'x d3 0-0 White can play 10 e4 c 5
with possibl e inter-tran sposition with
my m ain line, although White has an
extra recapture on d5 after 11 d5 exd5
12 exd5, with a slight edge for White.
My thought with an early ... c5 is th at if
White plays e4, Black keeps the option
with ... cxd4. It might not be needed,
but it is there ! "
The computer suggests that there
mig ht h ave been several equalizing
choices at the diagram position . I chose
wh at seem s to me the most natural .
Pl ay continued 8 iLd 3 iLxd 3 9 'ii'x d 3.

Test 2.1

D. Buckley-C.Crouch
British League (4NCL) 2006
Play started with 1 d4 ttJf6 2 c4 e6 3
ttJf3 b6 4 a3 .ib7 5 ttJC3 . Fairly standard
so far. The next few comments were
written at the time.
5 ttJe4
"Here I realized that I h ad not played
this variation since my stroke, and I was
hazy with my opening. Never mind, I
remembered the knight exchange was
the move I had previously played, and I
recalled that there was some slight
problem (in retrospect, it was that
White had too easy a time to keep the
draw). I just played a move quickly.
5 ... iLe7 is standard. A quick flick through
the next couple of games suggests that
6 'ii'c 2 c5 7 e4 is reasonable, and has
been played several times. I looked at
this, and felt that maybe White has a
slight edge, nothing spectacular, but I
was h appy to simplify."
6 ttJxe4 iLxe4 7 e3

We h ave now reached the position

in the test. It is so important as Bl ack to
find a clear equalizing move in the
opening, if given a chance. There i s of
ten a great temptation for the higher
rated pl ayer as Bl ack to try to keep the
tension alive, but so often the pl ayer
with White has the chance to keep a
modest edge. After th at, it would take
m uch m ore work to try to win a game


Te s t Two
th an if the position i s equal . Wh atever
happen s, the only way of playing for a
win is to rely on some sort of mistake
from the opponent, landing him or her
with a slight inferiority. If the opponent
is slightly better, it would take a larger
technical error to end up worse, than if
the opponent is only equal, and even a
microscopic slip becomes more signifi
cant. In a quiet and accurate position,
the lesson would seem to be th at it is
generally best to play quiet and accu
rate moves.
9 cxd4! (A) i s simplest and best, but
probably only to a slight extent. Then
10 tDxd4 tDc6 should be equal, and
could easily transpose later to the stem
game - see Test 8.1.
9 ill. e 7?! (B), as I pl ayed, is more
provocative. My notes at the time sug
gested th at 10 dS exdS 11 cxdS d6, fol
lowed by ... ill.f6, might give White a
slight edge, but not of any great sig
nificance. Looking at the position later,
10 dxcS ! is uncomfortable for Bl ack. If
10 ... bxcS, Black's pawn s are difficult to
mobilize, while 10 ... ill. x cs involves a
loss of tempo with the bishop. After 11
0-0 Black's bishop will probably h ave to
return to e7 at some stage. White is
slightly better.
In th e opening Bl ack is usually
slightly worse, with White h avin g
more options thanks to th e advantag e
of th e move. It i s often best for Bl ack
to try to simplify th e central pawn
structure, to try to cut down th e op
pon ent's option s.

Test 2.2

Kid l i ngton


We start with standard King's In

dian play with 1 d4 tDf6 2 c4 g6 3 tDc3
ill. g 7 4 e4 d6 5 tDf3 0-0, then 6 e3.



Thi s i s not quite th e main line, but

could easily tran spose into 6 ill. e 2 eS
and then 7 ill. e 3 ! ?, which I h ad tried a
few times, just before my stroke. This
line is still fashionable, and partly as a
result of this, I wanted to switch my
pl ay slightly, not engage in opening
theory. Apart from anything else, I
could barely read in the previous cou
ple of years.
He played 6 e5, and I could then
transpose, if I wanted, into main stream
theory with 7 ..te2 ! ? Then if 7 ... tDg4 8
ill. g s f6, all of 9 ill.h 4, 9 ill. d 2 and even 9
ill. c l h ave been tried. Ag ain, I could not
remember why, three years earlier, I
h ad decided which of these was best.
Another line, experimented by Bent
Larsen many years ago, is 7 dxe s dxe s 8


Why we L o s e a t Ch e s s
'iVxd8 'iVxd8 9 CLlds, but the slightly
unlikely 9 ...l:1d7 ! equalizes easily.
So I tried 7 dS, and for what it is
worth, the computer gives this as a
reasonable edge for White. Al so, we are
starting to go beyond mainstream
opening theory.
After 7 ... CLla6, it is possible th at my
reply, 8 .i.d3, was inaccurate. There are,
however, alternatives l ater on which
might still give White chances of a
slight edge, so perh aps it is too early to
give question marks. The bishops on e3
and d3 look good and imposing in the
centre, except of course that Black's
knights will hit back with ... CLlb4 and
... CLlg4. Black n aturally starts with
8 ... CLlg4, then 9 .i.gs f6.

CLlxe4 ..ixb2 16l!bl fxe4 17 ..ixe4 'iVg 7

18 hxg4 .i.xg4 19 iVC2 .i.f6 20 .i.xf6
l:1xf6 2 1 CLlh 2 .i.h s 2 2l:1b 3 ! ?, keeping an
edge. In other words, my opening
might h ave been unusual, but it was
certainly not daft. There is an impor
tant distinction between playing a
completely bad opening, and making a
slip slightly l ater.
I played instead the inferior 10 ..id2,
which i s pl ausible but no more than
equal . It allowed Black's pawn push,
10 ...fS, to come too quickly.

Then I completely mi sjudged the

tempo of pl ay, with 11 a3?? (B). A few
natural developing moves by White
and after 11 CLlcS 12 .i.C2? (White
needed to give away the bishop-pair),
my position collapsed. He exch anged
with 12 ... fxe4, then came 13 CLlxe4?!.
Battl e-weary, I now saw the sacrifice on
f2, but I could not see any way to avoid
it. 1 3 .i.xe4 provides more resistance,
but even h ere 1 3 ... CLlxf2 ! 14 'it>xf2
CLlxe4+ 15 CLlxe4 'iVh4+ 16 CLlg 3 (16 e3
leads into the game) 16 ... e4 will not
allow White to l ast for long.

H ere I probably should h ave played

10 ..ih4! ?, slowing down the .. .fs push .
This move i s well known in analogous
positions, with the light-squared
bishop on e2 rather than on d3. It
would take things too far to analyse
thi s in detail, but a possible computer
generated line run s 10 ... 'iVe8 11 0-0 fs
12 exfs gxfs 13 l:1el e4 14 h 3 'iVg 6 1 5


Te s t Two
Bl ack
exch anged
13 ... ctJxe4 14 ..ixe4 ...



... and now 14...ctJxf2 !.

Perhaps the gentlemanly finish
would have been to allow the queen sac
rifice with 15 ..txf2 'iVh4+ 16 'it>e3 ..ih6+
17 'it>d3 'iVxe4+ 18 'it>xe4 (18 'it>C3 lasts
only a little longer after 18 ... l:1.xf3+ 19
gxf3 'iVd4+ 20 ..tc2 ..if5+ 21 ..tCl 'ivxc4+,
and mate next move) 18 ... ..if5 mate.

In stead, I wanted to keep some

hopes alive. Not for long, though. After
the moves 15 ..ig5 ctJxdl 16 ..ixd8
"Lixb2 17 ..ixe7 l:1.f4 18 ctJd2 ctJxe4 19
..te2 ctJxd2 20 ..txd2 l:1.f2+ it was time to
resign (0-1).

So what should I h ave done at move

1 1 ? It was a matter of pride to try to
prove that, a couple of rounds l ater,
ag ainst Andrew Lewis, my opening was
okay, and that it was only a silly miscal
culation o n my part that l e d m e t o a
loss. Indeed, I could h ave improved a
move earlier, with 10 ..ih4! (in stead of
10 ..id2). Or, of course, I could h ave
tried one of the several 11th-move al
ternatives. So again st Lewis, I played 1 1
'iie 2 (A), but I l ater lost the thread of
this game too, and then lost.

Test 2.3

B ritish League (4NCL)


Not, this time, a question of sh arp

tactics, but rather one of finding the
best move order. We resume the Wall
(rouch g ame (1.4), with 8 d4 ctJd7 9
At the time I felt White's move was
probably too direct, and I was more
worried about his g aining the bishop-


Why we L o s e a t C h e s s
pair with 9 tDxd7 i.xd7 1 0 tDes cxd4
(10 ... i.e7 11 s 0-0 12 dxc S i.xcs 1 3
.i.d3 is promising for White) 11 tDxd7
't!Vxd7. Black has a few local difficulties
on the king side, so maybe he can
switch to the queen side after 12 'iVxd4
a-a-a ! ? : for example, 13 i.e2 "iJlC7 14
0-0 i.d6 15 h3 i.h 2+ 16 h l .i.es 17
a7 llVb8 18 't!Vxb8+ i.xb8. The bishop
pair is to be respected, but Black's
knight, bolstered by his pawn on a cen
tral square, may h old the bal ance.
in stead
9 ... tDxes 10 dxes, and now I h ad a
choice to make.

only then go ... i.cS : for example, 1 3 ... b S

1 4 .i. e 2 .i. b 7 i s comfortably equal, and
indeed White could h ave to face a few
problems on the 'Meran diagon als',
with open bishops on b7 and cS.
In the g ame, play instead continued
with 10 ... i.e7 11 es g6 12 g4 'ii C 7
13 f4 c4 14 a4!? (preventing ... bS)
14....id7, and it i s clear th at Bl ack i s
developing much slower than in the
alternative line. Black is still not far
away from equality, nevertheless, as
we will see in Test 8 .3 .
Thi s all might seem very technical to
the reader, but the accumul ation of
minor technical points may quickly
accumul ate to points on the board.

Test 2.4

London Open

I played 10 ... .ie7?! (B), which i s of

course a natural developing move, but
there are a coupl e of small problems.
First, if Black l ater plays ... 'iVC7, he can
not cover f7, and second, in delaying
the bishop move, h e can try in stead to
play ... c4 and then ... .ics, without loss
of tempo.
lO..:f1c7! (A) is more accurate, and if
11 f4 c4 12 'iVh s g6 1 3 'iVg4 I h ave not
committed myself to ... .i.e7. Indeed, I
can defend on the queenside first, and



We h ave discussed thi s position be

fore (1.2). Normally Black will h ave
played ... i.fs rather than ... .ig4.
The g ame continued 7 i.xC4 .i.b4 8

Te s t Two
b3 a s 9 t2:ies h S 10 t2:id 3, with White
having good chances of keeping a
slight edge, with attacking flexibility in
the centre. Still, 10 ... 'iVb6 keeps play in
reasonable bal ance. Black tried instead
10 ... t2:ia6 and I replied 11 t2:if4.
11 xa6 was tempting, but would
not h ave led to much after 11 ... bxa6 12
t2:ixb4 6 1 3 t2:ica2 t2:ids 14 d2 .l:[b8.
Black played 11 ... g6, reaching the
quiz diagram, and the first serious mis
take, though there h ave been a few
minor slips in the earlier pl ay.

12 t2:ixg6? (A) is a remarkably lazy

move, saying in effect th at " I h ave done
my job, I 've gained the bishop-pair,
now I can rest on my laurels." As play
went on though, it became obvious
th at White h ad great difficulties on the
kin g side, with most of his main pieces
being temporarily stuck on the queen
side, and Bl ack quickly being able to
pressurize White's king side.
12 0-0 (B) is the obvious altern ative,
not allowing Bl ack the open line for his
TOok on the h -file. White safely castles,
although Bl ack can set up some slight

pressure with 12 ...C7 .

12 i..d2!? (C), followed quickly per
h aps by l:tCl, seems the most accurate.
There will be some ten se manoeuvring
to follow by both sides. White should
be able to keep an edge, but certainly
no more than on move 1.
In stead, after 12 t2:ixg6 hxg6 13
O-O?! gS! White faces unnecessary
problems on the h-file.

The immediate problem i s the g8h 2 diagonal . I played 14 t2:ie2 (14 f4

immediately is
probably better)
14 ... d6 15 f4 ( 1 5 t2:ig 3 at the time
seemed to passive to me) lS ... t2:ib4 16
d2 C7 17 e1,


Why we L o s e a t C h e s s
starting t o set up the barriers o n the
king side. The reader will no doubt h ave
the impression that I am not h appy
with most of my moves in thi s game,
but somehow my pieces work together.
Radovanovic continued his devel
opment by bringing his rooks into play
immediately with 17 ... 0-0-0?, no doubt
thinking about doubling on the h -file,
but he could have improved.
Indeed, White is under pressure af
ter 17 .. .'!i:JbdS ! , and if 18 .Jtg 3 tt:Je4.
White is so busy, understandably so, in
cementing his pawn s on the stonewall
dark squares that he h as great diffi
culty covering the light squares. It's not
I was fortun ate, after he castled, to


find a way of aiming for a quick per

petual by way of 18 .Jtg3 tt:JhS?!
{18 ... gxf4! still leaves Black better} 19
fxgS ! tt:Jxg3 20 hxg3 .Jtxg3 21 l:lxf7
.Jth2+ 22 h1 {23 fl? ? xf7+}
22 ... .Jtg3+ 23 g1 i.h2+ 24 h1

and a draw with 24 ...i.g3+ {Vz-Yz}.

Te st T h r e e
3.1 Black to play

3.3 Black to play

A) 12 ... bxC3
B) 12 ... i.b7
C) Something el se?

A) 12 ... d4
B) 12 ... aS
C) Something else?

3.2 Wh ite to play

3.4 White to play

A) 1 3 'itf2
B) 13 i.d2
C) Something else?

A) 13 exfs
B) 13 0-0-0
C) Something el se?


Why we L o s e a t C h e s s

Test 3.1

Tha mes Va l l ey League


Readers will h ave recognized the

opening moves of the game (from Test
1.3). We start from move 7.

Black has just played the in accurate

7 ... iLc s. Then White pl ays the natural 8
iLd 3. I vaguely con sidered 8 ... liJg4 9 0-0
liJce s 10 liJxe s liJxe s, but this seems
anti-positional . After 1 1 iLC2 Bl ack has
no attack to carry on with, and White
will develop quickly.
In stead 8 ... dS, or 8 ... 0-0 9 0-0 first,
and then 9 ... dS, is playable, with good
equalizing chances, but un exciting af
ter White's previous slow pl ay.
I wanted more, and overpressed
with 8 ... bS?!, a second doubtful move.
White quietly played 9 0-0. I h ad
hoped to provoke him with 9 iLxb S ! ?
iLxf2+ 1 0 'It>xf2 'iVb6+ 1 1 liJd4 liJxd4 12
cxd4 'iVxb s 1 3 e S liJds 14 liJe4 0-0 with
un clear, and m aybe equal, pl ay, but he
was evidently not interested.


After he quietly castled, I continued

my bad-brain day with 9 ... b4?! . There is
no point in opening up lines on the
queenside, especially as the queen on
C7 i s on an awkward square if White's
c-pawn is exposed. 9 ... a6 is more the
matic, waiting to consolidate the
pawn s on the queen side, and develop
ing the pieces.
Nurmoh amed played natural devel
oping chess, with gain of tempo, with
10 liJb3 iLb6 11 ..id2, and it is now my
pawn that is under attack on b4, rather
th an mine attacking his. I tried to keep
my pawn s together, with 11 ... a S . After
11 ... bxC3 12 ..ixC3 ..ib7 Bl ack's position
is playable, but I h ave lost tim e with
the b-pawn . White h as a slight edge
Then 12 l:tC1, adding pressure to the
b-pawn .

12 ... bxc3? (A) was another miser

able move, allowing my opponent to
open up lines. A pawn exchange a
move earlier might just about h ave
been acceptabl e, but now Bl ack has lost
a tempo by inserting ... as.

Te s t Th r e e
Yes, I can g et tired in the evenings,
but I should all the sam e pl ay logically
and imaginatively. Th at said, 12 ... .L.b7?
(B) 13 cxb4 axb4 14 xb4 .l:!xa2 1 5 1i.a3
leaves the rook stranded, and is not to
be recomm ended either.
In stead 12 .. :Wid6 or 12 .. :Wia7 (C)
would h ave stopped the position mov
ing out of control, although undoubt
edly White keeps a slight edge.
After the exch ange on c3, and 13
1i.xc3 a7 14 ctJe5 ctJxe5 15 1i.xe5
White clearly h ad a substanti al edge.
For some reason, I h ad not expected
that after 15 ... b7, he h ad 16 f3 ! . I
suppose that I h ad vaguely expected
th at White would h ave wanted to play
eS, and then of course the queen on f3
would h ave self-pinned the pawn, but
of course the queen h as itself become
highly m obile with, for example, at
tacks with g 3, or, m ore immediately,
working with 1i.xf6 to win a pawn. I felt
I h ad to cover the threat on f6, with
16 ... a4 17 ctJd2 1i.d4.
I h ad expected, or perh aps hoped,
that White would allow the exch ang e
of bishops with 18 g 3 1i.xe s 19 xe s
0-0 and perh aps a minimal edge for
White, but White sacrifices a pawn in
stead, with excellent compen sation
after is 1i.d6! . It is difficult to say why I
did not con sider this sacrifice, which
keeps the bl ack king stuck in the cen
tre. The pawn dropping on b2 was only
of secondary importance. More likely
perh aps was th at I was so intent on
trying to take control of the long di-

agonal that I mi ssed the significance of

another dark-squared diagonal .

Even h ere, I was not yet worried,

and in any case all I could do was grab
the pawn , lS ... 1i.xb2, and see what was
going to h appen . Taking the pawn,
with g ain of tempo, is useful, but at
least as important from my point of
view is th at, with the help of a later
... a3, I can keep the bishop on the long
I doubt wheth er I genuinely be
lieved at the time that Black was fully
equal . It was more the case th at I knew
I h ad done something seriously wrong,
but th at it was no tim e to resign, and I
was pl aying on as effectively as possi
ble, and waiting for any slight slip by
my opponent.
White played vigorously with 19
.l:!C7 .l:!cs 20 .l:!xcS+ 1i.xcs 21 ctJC4 a 3 .
Even h ere I felt I had chances o f hold
ing, and playing for more, and it so
happened that I was later able to win
after scrambling through some tactics.
For the next test (8.4), I will ask how
later White should h ave won .


Why we L o s e a t Ch e s s

Test 3.2

Kid l i ngton


Pl ay started with a sh arp theoretical

opening, 1 d4 4:Jf6 2 c4 c5 3 d 5 e6 4 4:JC3
exd 5 5 cxd 5 d6 6 e4 g6 7 f4 Jig7 8
..ltb5+ 4:Jfd7 9 a4 a6 10 ..ltd3 h4+.

(Briti sh Ch ampion ship, Hove 1997), a

draw which could h ave gone either
way, but th at was about it. Then sud
denly, when I started to resume play,
the Modern Benoni seemed fashion
able ag ain, and I started to recognize
that Bl ack's play was again of interest.
Certainly my bash-bang attack again st
Peacock was less than productive.
I played 11 <j,(f1, something of a
novelty according to the database, but
in fact I h ad tried thi s many times in
weekend congresses during the 1980s.
The usual move i s 1 1 g3, but I wanted
to avoid the weakening of the king side
pawn structure. I tried this move three
times in a few m onth s, each tim e with
11 0-0 12 4:Jf3, a sure sign th at the
Modern Benoni i s again fashionable.
Arguably the most sophisticated of
these openings was the 12 ... d8 ! ? of
C.Crouch - N . Povah, British League 2006.
This at first looks like a complete loss of
tempo, but White n eeds to think about
where he is going to place his king ; the
f1-square should only be used tempo

White intends to ram through th e

centre, with a quick e 5 . This line is ex
tremely dangerous, and h as often led
to a Bl ack collapse. Indeed, for m any
years the Modern Benoni with 2 ... c 5
a n d 3 ... e6 h ad become deeply unfash
ionable. Black, if wanting this system,
h as generally played 2 ... e6, and if 3
4:Jf3, only then 3 ... c5, avoiding any
quick f4 system . If 3 4:Jc3, Black would
be un able to justify the Benoni or the
Queen 's Indian. Thi s leaves only the
Queen 's Gambit, or the Nimzo-Indian,
3 ...Jib4. H alf of Black's repertoire h as
been cut out.
For a decade before my illness, I do
not think I h ad encountered the critical
line. Pl askett tried it again st m e once


Te s t Th r e e
The natural an swer i s 1 3 '.tf2, but
Povah demon strated why he returned
with the queen to the starting square.
Not 1 3 ... 6?, as White gains space
quickly with 14 a s ! , but rather 13 .. .'i4Vc 7 !
1 4 !:t e l c 4 1 5 C 2 , with a n extrem ely
unclear position. He tried arguably the
safest option, first l S ... 6+, an d then
immediately offered a draw.

I thought about thi s for a long time,

looked at all the other games, but really
I was only going through the motions.
So a draw.
Very few players, at below the 2 700
level, are capable of playing a fully ac
curate game against relentlessly good
play by the opponent, and m ature
players are fully aware of it. A norm al
1M might be devastating again st a
2000 opponent, but would be likely to
lose against a 2 700 opponent. The 2000
player will tend to make far more mis
takes th an an 1 M, while a 2 700+ GM
would wipe out most of the errors th at
an 1 M will produce. If a pl ayer offers a
draw in a wild position, th at mean s he
is fully aware that h e i s capable of m ak-

i n g a mistake, beyond which h e could

be capable of understanding over the
board. The opponent (myself) would be
wondering whether he can set up a
good position, or whether his opponent
can pl ay with complete accuracy. Al so,
h e is un sure whether he i s capable of
pl aying with complete accuracy him
self. Sometimes it seem s wisest to take
a draw, and a safe h alf point.
If the impression might be given
that it is only the 'ordinary' I Ms, and
pl ayers at lower level s, who quickly ac
quiesce to a draw, then take a note of
two exceptional World Champion s of
attack, Tal and Kasparov, who played
against each other only rarely (the
chronology was wrong), but h ad some
wild g ames, and then suddenly agreed
draws, just when play was about to
become even sharper.
Maybe it is not even a question of
fear, either. A full-blown struggle i s
often physically exhausting, and ulti
mately th e players decide to con serve
their energies.
I looked at the fin al position in some
detail, both during the game and at
home, and decided first th at perhaps I
should have reached a slight edge, af
ter 16 '.ttl ! (16 '.tg 3 4Jf6 17 h 3 4Jh S+ 18
'.th 2 was con sidered by both sides, and
seem s wildly unclear, with no convinc
ing attack for Black, but no obvious
way for White to find full security ei
ther) 16 ... l:Ie8 17 as, and Bl ack's queen
is slightly out of position .
But maybe Bl ack could have tried


Why we L o s e a t C h e s s
something better o n move i s ? Or
maybe not? I leave this as open-ended,
and to the theorists.
Back to the Crouch -Peacock game,
and 12 "iVe7.

t o examine a dozen moves. 1 4 'iVh 3 ! ?

seem s sen sible, putting pressure on the
b6-pawn, and h ence m aking it difficult
for Bl ack to bring his knight on d7, and
therefore develop the other queenside
pieces. A slight edge to White would be
the conclusion.

Test 3.3

B ritis h League (4NCL)


After 1 iLlf3 iLlf6 2 c4 b6 3 iLlC3 ii.b7

4 d3 e6 5 e4,
Here I mistimed my pl ay, aiming to
consolidate on the king sie with 13
f2?! (A), but allowing Bl ack to con
solidate with 13 ... itd4+! 14 iLlxd4 cxd4
15 iLle2. We will con sider this position
l ater - see Test 7.1.
It is only because of the f2 move
that Black can play with effect, and a
check, after ... itd4. Therefore White
should h ave continued his develop
m ent on the queen side, with 13 itd2!
(B) (13 'iVc2 ! ? is to be con sidered too,
but is less promising), and White keeps
an edge.
The aggressive 13 .. .fs 14 exfs gxfs
15 iLlg s damages Black more than it
dam ages White. The quiet 13 ... b6
leaves Bl ack more flexible, and if 14
<Jif2 ? ! Black can again equalize with
14 ... itd4+. There are several possibili
ties for White, according to the com
puter, and it h ardly seem s worthwhile


I pl ayed S .. dS, trying to put White

under pressure in the centre before he
can complete his development. These
days I would tend to prefer s ... d6, fol
lowed by ... cs, setting up a H edgehog
structure, but with White being at least
a tempo behind if h e aim s to play d4.
I suspect th at the choice of s ... ds,
rather th an s ... d6, is an indirect result
of my poor h ealth . In those days, and
indeed still now, I would not h ave been
enthusiastic about pl aying long, ten se

Tes t Th r e e
gam es, and therefore preferred to keep
the position fairly open . In H edgehog
set-ups, there is often plenty of close
ran g e m anoeuvring, with both pl ayers
havin g to m ake detailed calculation of
whether the position should suddenly
be opened up, or whether the ten sion
should be kept. Notoriously, the break
through in the centre sometimes only
takes place well after the first tim e
control . Not really what I would h ave
After some thought, my opponent
in the g am e tried 6 "iVa4+!?, but I was
not worried about playing ... c6, espe
cially if White spends a tempo provok
ing it.
Maybe 6 cxdS exds 7 eS might be
met by 7 ...ttJfd7 8 d4 iLe7 9 iLd3 ttJf8 ! ?,
followed by ... ttJe6 or ... ttJe6. I am, how
ever, not wildly enthusiasti c about thi s
l i n e a s Black.
After White's check, both pl ayers
developed with 6 ... c6 7 iLg5 iLe7 8 iLe2
0-0 9 0-0 ttJbd7 10 :t:fe1 h6 11 iLh4 ttJC5
12 "iVC2.

No detailed comments need to be

m ade, except t o say that each player

would h ave thought about whether
eith er player could gain something by
altering the central pawn structure,
either with an exchange, or a pawn
push with eS for White or ... d4 for
Bl ack. So far, n either pl ayer would h ave
seen any significant improvement.
Unfortun ately just h ere I became
lazy. I played 12 ... a s ? ! (B), aiming to
con soli date the knight on cS (White's
pawn can no longer go to b4), but there
were other ways of pushing away the
knights, with 13 e S ttJfd7 14 i.xe7
"iVxe7 15 d4 ttJa6, and then a pawn ex
chan g e with 16 cxdS, keeping a clear
edge. Indeed, Black's queen side pawn s
later became weakened, as a result of
my ... as pawn push, as we will see in
Test 6.1.
Careless. In stead, Black should play
th e superior 12 d4! (A).

Now 1 3 ttJdl as is basically equal,

but with the likelihood of a ten se posi
tion al struggle, rather than a quick
Of the 'something else' moves,


Why we L o s e a t C h e s s
1 2 ... dxc4?! 13 dXC4, o r the correspond
ing exchange on e4, gives White the
better placed bishops and pawn s in a
symmetrical pawn structure. This is no
encouragement for Bl ack.
The computer suggests 12951 (C),
which I h ave to admit I h ad not really
con sidered.

Test 3.4

Kid l i ngton


We reach, by tran sposition, the

opening of the Crouch- Rose game (2.2)
two rounds earlier, 1 d4 t2Jf6 2 c4 g6 3
t2JC3 g7 4 e4 d6 S t2Jf3 0-0 6 e3 eS 7
d S t2Jg4 8 gs f6 9 d2 fS 10 d3 t2Ja6.
I now tried 11 'iVe2!?, avoiding the dis
astrous 1 1 a3 ? ? t2JC5 1 2 i.. c 2 fxe4, with
a quick win for Bl ack.
After 11 t2JcS 12 c2 as, I h ad the
choice between sharp or steady chess.

It seem s at first to weaken Black's

pawn structure on the king side. Then
13 g3 d4 14 t2Jdl would, of course,
tran spose into the possible line 12 ... d4
13 t2Jdl g5 14 g 3 . Bl ack's idea then is
14 ... t2Jxe4! , and if 15 dxe4 d3, which is
m aybe equal, or m aybe Bl ack could
even try to claim an edge with the two
bishops. In stead, 15 e 5 leads to pos
sible sh arp play after, for example,
15 ... t2Jf6 16 t2Jxd4 t2Jfd7 1 7 g 3 f5 . Th e
computer suggests it is equal, but my
instincts favour Black slightly. I leave it
to the reader to decide.
Certainly this line i s entertaining,
and it offers several chances for the
opponent to play inaccurately. Unfor
tunately it was I who m ade the first
serious inaccuracy.


Probably the simplest way of h an

dling the position is 13 exf5 (A) 13 ... ..ltxf5
(Black does not want to keep these bish
ops on the board) 14 xf5 gxf5
(14 ... .l:rxf5 15 0-0 seems to give White a
slight edge, but not 1 5 h 3 ? t2Jxf2 ! 16
'it>xf2 e4 17 t2Jxe4 4+ 20 t2Jg 3 .ixb2,
and Black is better; a familiar, if more
complicated line than in the Crouch
Rose game) 1 5 h 3 t2Jf6, with unclear play
to follow. The likelihood is that Black is
at least equal, with some of White's

Te s t Th r e e
light squares needing attention.
13 0-0 (C) 1 3 .. .fxe4 14 tDxe4 tDXe4
15 Xe4 tDf6 is safe enough for White,
but is only about level . Quite probably
there are other ways of keeping equal
ity. The computer gives about a dozen
alternatives, but it is unlikely that
m any of these lines would be con sid
ered as good by critical hum an an aly
sis. We could, for example, disregard
without an alysis such moves as 1 3
bl, 1 3 C l o r 1 3 dl, all highly un
con structive waiting moves.
13 O-O-O?! (B) is entertaining and
con structive, but it al so leaves weak
nesses on the queenside, in front of the
king . It breaks open the positional bal
ance, a n d while it i s not clear th at
White i s necessarily worse, White also
needs to play very accurately - which I
failed to m anage.
In the g ame Lewis played 13 .. a4,
then I tried 14 l:tdg1, again ambitious,
opening a square on dl, and thinking

about advances with the g- and h

pawns. Then some clarification in the
centre with 14 ... a 3 15 b3 fxe4 16 tDxe4
tDxe4 17 xe4, and I felt th at I was
probably not better, but at least I was
not doing worse, provided I can cement
the e4-square. H owever, I now h ad a
m assive shock.

Indeed, after 17 ...l:tf4! I did not pl ay

well, as we will see in Test 7.2. Some
h ow such an exch ange sacrifice is
much easier to see for the sacrificer
than it is for the defender.


Te st F o u r
4.1 Black to play

4.3 Black to play

A) 13 . . .'i'd7
B) 1 3 .. .lkB
C} Something el se?

A) 14 .. l2JcS
B) 14 .. .'iVcB
C} Something el se?

4.2 White to play

4.4 Black to play

A) 14 i.d3
B) 14 iLg s
C} Something el se?

A) 14 ... b S
B) 14 ... eS
C} Something el se?


Tes t Fo u r

Test 4.1

B ritis h League (4NCL)


Physically, I was in considerable

pain, as that morning I misjudged the
stairs at the hotel, and tumbled down a
dozen steps. This was one of my more
unwelcome problems resultin g from
the loss of sight in one eye. The result
ing damage to my legs, after this and
dozens of other falls and slips, plus fur
ther fear of falling again, has reduced
my mobility quite significantly in the
last few years.
Thi s gam e h as been the occasion
th at I was in my m ost severe physical
pain I faced in those years, but the real
agony was after the g ame. Yet while I
was actually playing, I was able to con
centrate. It was only afterwards th at I
could not function, and I was fortunate
that my match captain was able to give
me a lift home, as oth erwise it would
have been an extremely uncomfortable
evenin g .
I m ade m y mi stakes during the
game, but I did not m ake the sort of
string of almost incomprehensible
moves that I h ave m ade in occasion,
notably ag ain st McKenna (11.1) in thi s
book, a n d o n m any other occasion s
before and after. The really bad games
were through drowsiness rather than
through pain. If your body wants to
sleep, your brain needs to regenerate,
and there is no pain, just a snooze.

My first discernible slip, i n term s of

chess, was minor; no big blunder, but
inaccuraci es can soon add up to prob
The opening started with 1 e4 c5 2
ct'lf3 e6 3 d 3 ct'lc6 4 g3 d 5 5 ct'lbd2 ..td6 6
..tg2 ct'lge7 7 0-0 0-0 8 c3 a s 9 a4 b6 10
.l:!.e1 ..ta6 11 'iVc2 ct'lg6.

I felt comfortable with my position

here, avoiding any possibility of the
thematic advance by White with e S .
Possibly Black could h ave con sidered
11 ... 'iVd7 ! ? immediately, White being
un able to create pressure after 12 eS
.JLC7. Either way, Bl ack should be equal .
White exch anged pawn s with 12
exd 5 exd 5, then 13 ct'lb3. Bl ack now has
to decide wh at to do with the queen
and rooks. Black has started off with
the standard opening pl an of develop
ing with bishops and knights, plus of
course some appropriate pawn pushes,
and then castling into safety. This
mean s th at quite often the last pieces
to move are the queen and the queen 's
rook. If anything, the classic pl an is to
move the queen first, and then to de-


Why we L o s e a t C h e s s
cide later which square the rook should
move to.

On this occasion, the queen move,

13 ..:fid7! (A), is indeed best. The rook
would then probably move to e8, ch al
lenging the open file. O r maybe Black
m akes a second queen move before
h and, as White has to con sider the pos
sibility of ... 'fifS. 14 d4 is therefore a
possibility, but after 14 ... c4 1 5 tDbd2
l:tae8 Bl ack is ahead in development,
and indeed is a tempo ahead when
compared (see below) with the g ame.
13...'fif6 (C}, ag ain with the idea of
... f5, is not as effective, as 14 d4
leaves Black's queen unexpectedly on
the edge, with even a big threat of
.Jtg 5 .
1 3 . tDces (C} 14 tDxe 5 tDxe5 i s en
tertaining, with thoughts of an ex
change sacrifice on the long diagonal
with 1 5 .Jtxd5 tDxd3 16 l:tdl c4 17 .Jtxa8
'fixa8 18 tDd4. White could avoid thi s
with 1 5 c 4 dxc4 16 dxc4, with a slight
After not looking through all the al
ternatives, Black played lazily with


13 . l:tc8?! (B). I was presumably vaguely

thinking about weaknesses on the long
diagonal, and I wanted to hit White's
pawns with ...d4 or ... c4, but as the game
showed, this never really happened, and
I lost time with my rook. She played 14
.Jtg5 d7 15 d4!, taking the initiative. I
was surprised at first, but soon I appre
ciated that my central pawns had been
weakened. White has successfully cov
ered the strategically important d4square, with the help of two sturdy
knights. Sometimes a player can forget
the basics, counting up the number of
pieces attacking, and the number of
pieces defending.

I pl ayed 15 ...c4, with some reserva

tion as I did not like blocking the
queen side pawn s and leaving a back
ward d-pawn . I did not like altern ative
pawn structures either, the usual iso
lated pawn or h anging pawn s being
potentially weak.
15 ... h 6 ! ? 16 .Jte3 tDb4 is a more
imaginative way of h andling the posi
tion, and after 17 d2 tDd3 18 l:tedl c4
19 tDcl tDxcl 20 l:tdxcl the pawn struc-

Tes t Fo u r
ture is the same as in the m ain g am e,
but a knight on either side h as been
exchanged. When analysing thi s varia
tion just after the g am e, I assumed th at
this idea for Bl ack was not very effec
tive, as I h ad exch anged my active
knight for White's passive knight. Now
I am of the opposite point of view.
Black has a backward pawn on d5,
which requires protection, and gener
ally it i s easier to defend with one
knight against one knight, rather th an
two knights again st two. Bl ack should
be close to level .
Maybe thi s is one to be added to the
list of test position s? The trouble is that
if I were to try this con scientiously, I
would h ave to question thi s g am e
every third m ove o r so, and thi s i s po
tentially misleading. The game is not as
bad as it looks, and my opponent has
deserved the praise of outpl aying me in
the early middlegame, rather than the
suggestion th at I h ave played a whole
string of bad moves.
Pl ay continued with 16 ttJC1, then in
reply 16 ... l:i.ce8.

If I h ad thought about it at the time,

thi s would be an admission th at Black
h ad lost a tempo with his earlier ... l:i.a8c8. Over the board, however, you must
take things move by move.
We go under the microscope later
(Test 6.4). It is worth remembering that
so far, pl ay h as been close to equal
throughout, although neith er player
has approach ed fully accurate chess.
My opponent, for example, has allowed
me to equalize in the early opening
with little difficulty.

Test 4.2

H i l l i ngdon League


We opened with a Semi-Slav, with 1 d4

ds 2 c4 c6 3 ttJf3 ttJf6 4 ttJC3 e6 S e3. I
briefly thought about the wild stuff
with 5 Jtg 5, and then 5 ... dxC4 (or, quite
often these days, 5 ... h6) 6 e4 b5 7 e5 h 6
8 .t h 4 g 5, but then I realized th at I had
not looked at thi s properly for nearly
twenty years. It seem s unlikely that he
h ad looked at this in detail either - one
only looks at such lines so eagerly in
their teen s or twenties - and this
llli ould probably h ave been quite safe
on my part. 5 .tg 5 is something of a
g amble.
With age and declining speed of
thought, I eventually decided th at I
would not want to go for the sh arpest
lines. So it's 'only' to be a Meran .
After S ... ttJbd7, I accordingly contin-


Why we L o s e a t C h e s s
ued with the quiet line, 6 "iC 2, avoiding
the tempo loss with 6 d3 dXc4 7 xC4
bS, which is still highly theoretical, and

Then 6 ... d6. Now White's usual

choices would be the ultra-quiet 7 b3
and b2, or the g ambit line with 7 g4.
But after my illness, I h ave not h ad a
chance of updating much of my open
ing repertoire, and so h ave relied on
lines I h ad tried many years earlier.
Thus I went 7 i.d2, followed by queen
side castling. As I wrote at the time, " I
can 't pretend th at I was thinking for
too long to m ake my decision . I just
played natural moves."
After the opposite-castling se
quence 7 ...0-0 8 0-0-0 Black h as tried
8 ... cS 9 e4 cxd4 10 ttJxd4 dxc4 11 xc4,
and even the gambit 8 ... b S . Either
seem s possible. My opponent played
solidly instead, keeping his pawn s in
tact with 8 ..."ie7. Then 9 b1, taking
the king away from the centre. Neither
pl ayer quite seem s ready to open up
th e centre just yet. Play continued with
9 .l::t e 8 10 h3 a6.


In my notes, I queried the next

move, 11 e4, probably on the some
wh at superficial basis that the com
puter suggested that 1 1 g4 was slightly
better, while my own m ove was only
equal . Th e computer only gives guid
ance, rather than any certainty, and
indeed if one were to wait a little
longer, it would give 1 1 e4, as played in
the g ame, as better, whereas 1 1 g4 is
only equal . For example, 11 ... dxC4 12 g s
ttJdS 1 3 xC4 b S 1 4 ttJxdS cxdS 1 5 d3
A.Kakageldyev-5.Moh amm ad,
Asi an Team Champion ship 2000, when
Bl ack m ay even h ave been slightly bet
ter, before losing a minor piece end
50 my own move was acceptable.
Bl ack exchanged in the centre with
11 ... dxe4 12 ttJxe4 ttJxe4 13 "ixe4 eS,
and now it i s time for th e question .
I pl ayed 14 iLd 3? (A), with little
thought, partly because of an uncom
fortably quick time limit, and al so
partly because I was too busy with
other things, like rescuing a club which
was about to collapse (and I am not

Te s t Fo u r
talking about my current local club,
H arrow).

Of course, ultim ately my mistake is

mine. The problem is that after
14 ... 4Jf6 15 h4 ( 1 5 i.g 5 4Jxe4 16
fi.xe7 4Jxf2 17 i.xd6 e4! win s m aterial
for Black) 15 ... e4 16 hel Black h as
16 ... d8 ! , which I missed when trying
to calculate earlier, and which my op
ponent missed altogether.

Now th at the e-pawn is no longer

pinned, White's bishop and knight are
both forked. Another pin would follow
after 17 i.xe4 xe4 18 xe4 i.f5 ! . In
stead, 1 7 i.c2 exf3 18 i.g 5 .l:[xel 19
llxel is an attempt to gain some coun -

terpl ay again st the king, but the sim

plest way to add more geom etry i s
19 ... i.f5 ! 20 i.xf5 'iVa5, keeping the
extra piece in comfort.
White could attempt an improve
ment with 17 c5, with the idea th at if
17 ... iJ.. C 7, Black will no longer h ave the
option of playing ... i.f5 and ... a5 . This
may well still be playable, but a simpler
option would be to help defend the
king side with 17 ... iJ..f8. Then if 18 i.c2
exf3 19 xe8 xe8 20 iJ.. g 5, and Bl ack
again uses the ... i.f5 resource, with
20 ... 'iVe2 21 iJ.. xf6 i.f5 ! . When just after
my stroke, a coupl e of years before this
game, I was trying to rethink all my
chess thoughts, I soon recognized
ag ain, but with better emph asis per
h aps, that chess i s not just the art of
calcul ation, it i s al so the art of geome
Finally, 17 i.g 5 exd3 18 i.xf6 gxf6
19 xe8+ 'iVxe8 20 'iVxf6 'iVe7 soon win s
with comfort for Bl ack. White h as no
geometry with the bishops.
Instead of entering all these tactics,
in favour of my opponent, I should
h ave considered 14 i.g5 (B) 14 ... f8 1 5
dxe5 4Jxe 5 16 4Jxe5 xe 5 17 4 i.c5+
18 i.d3, with equality. Or perh aps 14
dxe5 (C) 14 ... 4Jxe 5 1 5 i.C3 4Jxf3 16
'iVxf3 i.b4, again about equal .
14 e1!? (C) is probably the best of
all, even though it would appear at first
to involve too m any successive moves
with king and rook. White is, however,
adding pressure to the e-file, and this is
difficult for Bl ack if he cannot continue


Why we L o s e a t C h e s s
h i s development quickly.

14 ... exd4 15 'iVc2 'iVd8 16 .l:i.xe8+

'iVxe8 17 d3 leaves White better, his
development being more advanced; he
will soon recover the pawn . Or 14 .. .lbf6
15 'iVc2 e4 16 g 5 'iVd8 17 tiJd2 f5 18
tiJb3, with a ten se middlegame and an
edge for White.
Sadly my opponent was close to his
last year in active chess, his physical
mobility deteriorating rapidly. This
would h ave been a memorable, final
good win against 1 M opposition h ad h e
seen the queen retreat in time. He still
later won, but after further errors, no
doubt not h elped by the quick time
limit. We con sider the next stage in
Test 9.2.

e7 7 d4 cxd4 8 'iVxd4 tiJc6. Possibly

8 ... d6 could be con sidered as slightly
more accurate, with a H edgehog set
up, with ... tiJbd7.
White's queen retreated with 9
'iVd 3. Maybe 9 'iVf4 is more threatening,
as attempts to try to equalize quickly
with a queen exch ange, with 9 ... iVb8,
are not fully convincing after 10 l:td1.
My g am e continued with 9 ... 0-0 10
l:td1 tiJb4!. Thi s should be equal . After
11 'iVd2, Black equalizes with 11 ... 'iVc8
12 b3 d5 13 a3 tiJa6 14 cxd5 tiJxd5 1 5
b2 tiJxC3 16 i.. x C3 l:td8, and a dull
draw after, for example, 17 d4 C5
18 'iVg 5 i.. x d4 19 tiJxd4 i.. x g 2 20 'it>xg 2.
H e pl ayed in stead 11 'iVb1?! and I
began to sense that I could pl ay for
rather more th an a quick h alf point. His
c-pawn was weak, and so I played
11 ... 'iVc8 12 a3 tiJa6.

Test 4.3

London Open


The opening is reasonably well

known, an English with 1 tiJf3 tiJf6 2 c4
b6 3 g3 b7 4 g2 c5 5 0-0 e6 6 tiJc3


H ere my opponent thought for a

long time, and I started to appreci ate
that h e could already be in difficulties.
H e g ave up a pawn with 13 i..f4?! to
activate his pieces.
While h e was thinking, I h ad ex-

Tes t Fo u r
pected 1 3 'ilVa2 ! , keeping the pawn on
c4 safe, and al so the queen . Thi s should
end up in symm etrical equality ag ain,
after 1 3 ... dS 14 cxdS tiJxdS 1 5 j.,d2
tiJxc3 16 j.,xC3 j.,dS 17 1 tiJcS . There
are other possible quiet alternatives,
such as 1 3 j.,g s h6 14 j.,xf6 xf6 1 5
'ilVc2 lId8, m aybe slightly favouring
Black (one usually wants to keep the
bishop-pair in an open position), or 1 3
'ilVd3 tiJ c S 1 4 'ilVc2 tiJce4 1 5 b3 dS 1 6
tiJxe4 tiJxe4 17 j.,b2 dXc4 18 bXc4 tiJcs,
but again White has to work h ard to
gain full equality.

White pl ayed 14 :C1, leading to the

quiz position . The question i s one of
careful and accurate defence, and al as
my calculation skills soon folded badly.

Black's problem is that his queen i s

under pressure, with no quick escape
from the c-file. He needs to sort this
one out as soon as possible. 14...'ilVb3??
(C), the only immediate escape from
this file, is worse than useless, as 1 5
tiJd2 leaves the queen trapped.
14 :WicB?! (B) is just about playabl e,
but after 15 tiJdS tiJxdS (lS ... 'ilVd8 16
tiJxf6+ j.,xf6 17 tiJg 5 j.,xg 5 18 j.,xb7
leaves White m aterial up) 16 lIxc8
J:!.xc8 17 j.,d2 White's queen slightly
outweighs Black's rook, knight and
pawn . Thi s is, of course, the first line
for Black to be con sidered while analys
ing, but it should only be played in
emergency, and only if any other move
is even worse.
14 ... tiJb8!? (C) i s almost the last non
blundering move that a player might
expect, and this retreat may be possi
ble, but is unlikely to be the best. If 15
b4, Bl ack must avoid l S ... dS?? 16 tiJd2

Sen played imaginatively, and as

the continuation shows, he an alysed
better th an me, but if I h ad analysed
well myself, I would h ave kept on top.
13 ...'ilVxc4! is correct! I remembered
my recent games ag ain st H ebden and
David Buckley (Tests S A and 8.1), and
reminded myself that pawn snatches
are often vital in chess, and th at failing
to taking the pawn is often a m ajor
concession . This move i s good. It was
on ly the follow-up that was poor.
13 ... lId8 is equal, but nothing more.


Why w e L o s e a t C h e s s
a6 17 b S as 1 8 tLib3, and the queen
is trapped, in favour of lS ... aS 16 bxas
l:txas 17 'iNxb6, which is level . I doubt if
many players would want to examine
this line over the board, but these days
if the computer suggests something, it
is worth con sidering.
14 :iYC5! (C) i s better and the cor
rect choice, pointing the queen in an
other direction, with the escape line
being to either hS or fS .

tIe more with l S .. J:Uc8, and if 16 b4

'iNfs 17 'iNxfs exfs 18 tLid4 xg 2 19
xg 2 g 6, when White's queen side h as
been weakened a little further th an in
the previous line.
I played in stead 14 tLi C 5? (A), but
thi s is poorly thought through.


Then 1 5 tLidS ? ? 'iNxds 16 tLih4 e4 i s

just a n illusion .
The somewh at superior 1 5 h 3 ! cuts
down Black's queenside option s. Black
can then simply pl ay lS .. :fS 16 'iNxfs
exfs 17 tLid4 xg 2 18 xg 2 (18 tLixfs
xh 3 favours Black) 18 ... g6, and this i s
more than enough t o demon strate that
14 .. :cS ! i s better th an any alternative.
Black is not worse, but could well be
better. His extra pawn i s not looking
threatening, but looking closer, one
sees th at White might h ave a few
weaknesses on the c-file and the light
squares on the queenside.
Possibly Bl ack could even press a lit-


Th e computer at first suggested

th at thi s was good, so it is not a totally
ridiculous idea, but on further analysi s,
it then shows the refutation, admit
tedly far from in stantly, that I mi ssed.
My opponent alas saw it, pl aying 15
If now l S ... tLib3 ? 16 tLies, and White
win s m aterial .
If l S ... xf3 16 xf3 tLib3 1 7 tLidS !
exdS 18 l:txC4 dxc4 19 xa8 l:txa8 20
l:ta2, and White i s ahead on m aterial,
with queen versus two knights and two
pawn s. The m aterial di sadvantage is
still rel atively slight, and perh aps thi s is
the best chance for Black to set up a
rearguard action . H e h as an advanced
c-pawn which can be protected by an
other pawn with 20 ... dS, so White h as
to work h ard.

Te s t Fo u r
In stead, I pl ayed the move I previ
ously intended, 15 . ..luce4?!. He pl ayed
16 ctJe5, and tactically Black's position
is no longer under control .

If, for example, 16 ... ctJxc3 17 ctJxC4

ctJxb1 18 .1i.xb7 l:tab8 19 .1i.g2 dS 20
ctJb2, and Black's knight is hopelessly
trapped, much as in the g ame.
In stead, 16 .. :iVxC3 17 l:txC3 ctJxC3 18
"YiVd3 .1i.xg 2 19 'it>xg 2 ctJcdS 20 .1i.d2 is
the best of a bad batch, with rook,
knight and pawn for the queen, theo
retically level m aterial. Black's minor
pieces are, however, poorly coordi
nated, blocking each other, and poten
tially under attack from the e4 push .
Moreover, Black's extra pawn is ineffec
tive in attacking term s. I should h ave
pl ayed thi s, but my position is not
16 ... "YiVd4?! h astened the collapse.
After 17 e3 ctJxc3 18 exd4 ctJxb1 19
xb7 l:tab8 20 ..ig2 my knight i.s again
Afterwards I felt slightly guilty
about not resigning quickly ag ain st my
young opponent, but ag ain st this,

there was a quick time limit, and h e

was already slightly (but not seriously)
short of time.

I carried on with 20 ... ctJxa 3 21 l:txa 3

a s 22 ctJd 3 l:tbc8 2 3 l:tac3 l:txc3 24 l:txc3
axb4 25 l:tb3 l:tc8 26 ctJxb4 l:tC4 27 ctJd 3
l:txd4 28 .1i.e3 l:tC4 29 l:txb6, but it was
essentially a m atter of time. White's
extra bishop, when given the opportu
nity, would chew up some of Black's
rem aining pawns, and I had to give up
on move 43 (1-0).

Test 4.4
P Sowray C Crouch

London League


give no analysis of the opening

moves, just to say instead th at I felt
reason ably h appy in th e opening, with
1 e4 c5 2 ctJf3 e6 3 d4 cxd4 4 ctJxd4 ctJf6
5 ctJC3 d6 6 g4 h6 7 h3 a6 8 ..ig2 "YiVC7 9
.1i.e3 ctJbd 7 10 f4 ctJb6 11 "YiVe2 ctJc4 12
0-0-0 .1i.d 7 13 .1i.f2 l:tc8 14 l:td 3 .
This h a s been played before, but not
very often. Among the 'something else'


Why we L o s e a t C h e s s
options i n this quiz, one of the possi
bilities here is 'draw agreed, M. Katona
E. Kovacs, Hung arian Team Champion
ship 1995'. Thi s does not help us very
much, neither in theoreti cal term s nor,
since Sowray did not offer a draw, in
practical term s. I was on my own, try
ing to find a decent move.

does not do all that much, and wastes

Bl ack h as been careful not to h ave
played ... Jie7 early on, so that if the
centre open s up quickly, jumping with
the bishop over the currently blocked
d6-pawn, Black has ch ances of g aining
a tempo with the bishop to b4 or cS, or
even sometimes as far as a3. Indeed,
14 . . eS! (B) is clearly the best, and then
15 fxe s (if 15 lLlfS, then simply
lS ... exf4) ls ... dxe s 16 lLlb3 (16 lLlfS 'iVas
with advantage to Black) 16 ... Jib4 17
Jiel (17 lLldS lLlxdS 18 exds Jib s fa
vours Bl ack) 17 ... Jixc3 18 Jixc3 Jibs
leaves Bl ack's pieces far better coordi
nated than White's.
In the g ame, Sowray lined up the
thematic knight sacrifice on the e-file,
with 15 :el!?, as we will see in Test 7 . 3 .
1 5 b3 was al so good. White h as gained,
either way, some extra time for his ini
tiative. However, if I h ad played my
best move m ore quickly, I would h ave
g ained the initiative myself.

I pl ayed 14 ... b5?! (A), a natural Sicil

ian move, but h ere it i s not preci se. In
many lines it is an immediate respon se
for Bl ack to push th e knight on c3 to a
worse square, and in so doing, to sof
ten the defence of the e4-pawn. Here it


T est F ive
5.1 White to play

53 White to play

A) 1 5 .l:tadl
B) 1 5 l:tfel
C) Something el se?

A) 16 lbxd4
B) 16 .l:tel
C) Something el se?

5.2 White to play

5.4 Black to play

A) 16 'it'b3
B) 16 l:td4
C} Something el se?

A) 16 ... it'xb2
B) 16 ... l:td8
C) Something else?


Why we L o s e a t C h e s s

Test 5.1

B u ry St. Ed m u nds 2006

1 d4 lLlf6 2 c4 e6 3 lLlC3 b4 4 e3 0-0 S
d3 cS 6 lLlge2
At this stage, I was not really up to
playing standard theoretical chess, and
I was often worried that I h ad quite
simply forgotten much of my opening
knowledge. Here I h ad played this move
a few times, a few years ago, but I was
aware th at I h ad decided that I was
planning to abandon this line, with no
real chance of playing for an edge. 5
lLlf3 would h ave given more option s.
Play continued, with approximate
equality, after 6 ... cxd4 7 exd4 dS 8 0-0
dxc4 9 xc4 a6 10 a3 ..td6 11 h 3 . White
avoided 11 g 5 ? xh 2+, but it is clear
th at White will sooner or later want to
play g 5 .
Bl ack now used the queen side long
diagonal with 11 ... bS 12 a2 b7 13
..tgs, and h e continued his develop
ment with 13 ... lLlbd7 14 'iWd 3 ::I.c8.


Now I h ad to decide wh at to do with

my rooks. Assuming th at White does
not want to block the other rook, the
basic choice here i s (from left to right)
15 l1acl, 15 ::I.adl, 15 ::I.fdl or 15 ::I.fel.
Quite possibly the other rook will de
velop in the centre soon after, or quite
possibly White h as other plans.
The most natural plan i s to develop
with ::I.acl and l1fel, but which comes
first? In my earlier notes, I suggested 15
'g,fe1 (B), but there is a slight degree of
inflexibility in th at if White were
quickly to play bl (for in stance, in
one line the computer suggests
15 ... lLlb6 16 bl ! ?), then the rook on al
h as been blocked out of play. Probably
thi s will not be a significant defect, and
White still keeps a slight edge, but even
so, if given the option, it is best to keep
Th erefore 15 ::I.ac1! (C) is better, and
Bl ack still h as to demon strate complete
Moving one of the rooks to dl is
positionally unin spiring if the pawn
centre is closed. Thus I was not im
pressed with 15 ::I.fd1 . It i s now gener
ally accepted th at the defence of the
isol ated d-pawn should be covered
lightly, so th at the other pieces remain
I decided to play the other rook to
dl, is ::I.adl?! (A), with the intention in
the m ain line of tactical play with the
pawn to d5. It i s imaginative, but as so
often with imaginative tactical ideas, a
hidden resource m ay suddenly emerge

Te s t F i v e
for the defender. In Test 6.2 we sh all
see wh at h appen s l ater. In the m ean
time, it might be of interest to the
reader to con sider wh at the pl ans and
counter-plans would be after lS .. :C7
16 f4.

Test 5.2

Britis h League (4NCL)


Black pl ayed a slightly unusual, but

playable, opening with 1 d4 d s 2 c4
.Us 3 cxd s iLxbl.

I immediately pl ayed 4 :xbl?!, and

then almost immediately regretted it. 4
"iVa4+! is more accurate, and only then
4 ... c6 5 :xbl. 4 ... b S ? 5 "iVxb S+ c6 6 dxc6,
of course, gives nothing.
I h ave decided not to give the dia
gram position as one of the tests, be
cause if the reader i s already alerted
that something is amiss, he or she will
look more carefully and find the more
accurate move. The problem for the
player over the board is to spot

whether t h e obvious move, here a re

capture, is automatically the best.
After 4 ..."iVxd s, I cautiously pl ayed S
a 3 . There are a few g ames, I l ater found
out, in which White collapsed after 5
"iVa4+ lLlc6, and if 6 lLlf3 eS 7 dxe s
iLb4+ 8 iL d 2 "iVe4! (gaining a tempo
with the attack on the rook) 9 iYdl
0-0-0. It was not difficult for me to
work thi s line out in advance, but after
my quieter move, White's advantage of
the first move h as gone.
Black was not interested now in hit
ting the bl-rook, with ...iYa2 or ... iYe4,
and in stead developed quickly with
s ... lLlc6 6 e3 eS. I was relieved th at I
was able to exch ange the queen s with
7 dxes iYxdl+ 8 \t>xdl 0-0-0+.

I very vaguely remembered that I

had reached this position before, again
with the slight error on move 4. On
checking up later, I found the game
C.Crouch-I.Sakovich, Decin 1996, with
probable equality after 9 i.. d 2 lLlxes 10
lLlf3 iLd6 11 lLlxes iLxes 12 iLe2 lLlf6 13
iLf3 l:the8 14 \t>e2. I could try to pretend
that my bishop-pair leaves me slightly


Why we L o s e a t C h e s s
better, but i n fact i t i s not s o significant.
I later won, but this was far from forced.
Several years later, I played 9 'it>c2,
then came 9 . .4:Jxes, with possible
threats with ... 4:Jg4. This was difficult to
assess. My general preference in such
position s would be to prefer the bishop,
but in compensation Black is slightly
ahead in development. Probably the
position remains equal with best play
on both sides, but any slight slip on ei
ther side would start to change the bal
ance. It is not an 'easy' equality.
I played 10 e2. There h ave been
other games in which White h as tried
10 f3 or 10 h 3, but this seem s slow. I
considered 10 4:Jf3, but was a bit wor
ried about 10 ... 4:Jg4, continuing with 11
.tC4 4:Jxf2 12 lIft 4:Je4 13 4:Je5 4:Jh 6 14
4:Jxf7 4:Jxf7 1 5 .l::t xf7. Equal, probably.
Black's knight is an irritation, but
White has the bishop-pair.
Over the next few moves, both play
ers develop and consolidate with
10 ...4:Jf6 11 4:Jh3 h6 12 b4 .td6 13 .tb2.
The computer suggests th at early on
Black is slightly better, but th at later
th e position is equal . I would suspect
th at the l ast few moves were equal
throughout. There followed some more
quiet m anoeuvring with 13 ... lIhe8 14
lIbdl 4:Jfd7 15 4:Jf4 g6, and we h ave
now reached the quiz position.
Black's last couple of m oves h ave
seemed almost random, at least from
White's point of view, and I started to
feel more confident. It is often around
thi s stage th at a pl ayer might relax his

vigilance. So it proved.


My assessm ent was th at it was a

good time to move one of my pieces,
rather than trying to push a king side
pawn . 16 'it>b3? (A) looks natural, but I
overlooked or underestim ated 16 ... a s! .
attacking White's queenside pawn s. I
did not like the idea of allowing an i so
lated pawn on b4, and so played 17
bxa s 4:Jc6 18 4:Jd S . Now Black gains a
tempo with check, with 18 ... 4:Jxa s+, as
a result of White's unfortunate king
m ove. So a retreat, 19 'it>a2 .tes, and
Black h as set up a useful passed pawn
on the c-file, while also neutralizing
White's bishop-pair.

Tes t Five
While thi s is far from deci sive, I was
starting to feel under pressure, and
later m ade mistakes. We shall return to
this exercise at a l ater stage (Test 9.3),
with an accumulation of minor errors,
plus a time trouble blunder, leading to
a clear loss.
So what possibl e improvements
could there be for White ? Thi s all seem s
open -ended, with several possible ideas
ending up reaching a position of likely
One n atural possibility is 16 cj;b1!?
(C) 16 ... lt:ib6 17 It:ids It:ibC4 18 i.xe s
It:ixe s, equal . At the time, I wanted
16 It:idS (C) puts some slight pres
sure on the opponent, but with quiet
play, for example, 16 ... c6, it is unclear
that White can achieve anything con
Maybe the best move is 16 'f1.d4! (B),
preparing to play the other rook, either
to dl, doubling, or to set up pressure
on the c-pawn with :tCl.

There are various quiet moves for

Black, m aybe with equality, or maybe

with a fractional edge for White. 16 ... a s

would still h ave been possible, as in the
game, but after 1 7 bxas It:ic6 18 .l:!.d2
It:ixas 19 :thdl White is ahead on
tempi, and possibly with a slight edge.
Thus m any of the 'something else'
moves are playable, and quite likely to
be better than the move I actually
played, but my preference is for 16

Test 5.3

Kid l i ngton


Returning to a g ame we con sidered

back in Test 3.2, here Black played
lS ... lt:icS.
I pl ayed the natural 16 It:ixd4? (A),
but this soon allows Black to equalize
with 16 ... lt:ixe4+ 17 itxe4 'iWxe4 18 lt:if3
itg4. We continue this exercise at a
later stage - see Test 7.1.
16 l1e1! (B) is more accurate. If
16 ... lt:ixe4+ 17 itxe4 'ilVxe4, I h ad only
con sidered 18 'ilVxd4?! 'ilVxd4+ 19 It:ixd4


Why we L o s e a t C h e s s
tLJd7, with equality, since both d-pawn s
drop, with a symmetrical opposite
coloured bishop position . However, 18
tLJc3 ! substantially improves. After
18 ...f5 19 xd4 White's comfortably
centralized queen causes problems for
I was al so concerned with 16 ... i.. g 4,
but 17 1:[a3 tLJxe4+ 18 i.. x e4 xe4 19
'Yi'xd4 xd4 20 tLJxd4 leaves White with
the more active pieces in a queenless
The best line for Black i s 16 ... 4+!
17 'iit> g l tLJxd3 18 g3 tLJf2 19 gxh4 tLJxdl
20 1:[xdl i.. g 4 21 Wf2 tLJd7 22 ktxd4 f5 .

White is a pawn up, but thi s is only

a doubled i solated h -pawn, which h as
littl e impact. Black still h as to take
som e care in holding the g ame, but
after 23 tLJC3 tLJa5 24 1:[a3 fxe4 25 tLJxe4
tLJxe4 26 l:txe4 1:[f5 27 l:tg 3 i.. d l 2 8 1:[g 5
i.. d l 29 i.. d 2 ktC5 h olds the bal ance for
Bl ack.
At least this way it is Black who h as
to keep equality. With my own move
for White, I h ad to work h ard myself to
achieve the same goal .


Test 5.4

Metropol ita n Open


The final round of a weekend tour

nament, with ch ances of equal first,
and some prize money, though much
less, in real term s, than in the seventies
or eighties.
I did not relish the thought of play
ing a second g ame in a day, given the
likelihood of dizziness, and tiredness
during the end of the session . A pro
longed g ame ending up in a blitz finish
would h ave been a nightmare for me,
so I wanted to pl ay quickly, trying to
eliminate any elaborate play. My op
ponent has always believed in playing
quickly him self, finding decent moves,
if not necessarily the best. He m akes his
mistakes, like all of us, but it is often
very difficult for his opponent to find
any refutation when th e time ticks on .
My psychological mistake in this
game was to pl ay too quickly against
quick but provocative play.
1 d4 tLJf6 2 tLJf3 b6 3 i.. g 5 i.. b 7 4
tLJbd2 c5
Some thirty month s earlier, playing
again st the same opponent el sewhere
(M.Hebden -C.Crouch, Coventry 2004). I
avoided thi s pawn exch ange, playing
instead 4 ... e6 5 e4 h 6 6 iLxf6 xf6 7
iLd3 d6 8 e2 e7 9 0-0 g 6 10 e5 i.. g 7
1 1 1:[fel 0-0 12 c3 tLJd7 1 3 iLa6 i.. x a6 14
xa6 dxe 5 15 tLJxe5 1i.xe 5 16 dxe 5,
with perh aps a minimal advantage for

Te s t F i v e
White, ending up as a draw.
Thinking back to the earlier tour
nament, it seem s such an incredibly
long g ap between 'before' and 'after',
with a stroke intervening. It seem s so
strange th at after pl aying in Coventry,
first equal, then a hundred-mile walk
ing h oliday, then pl aying another tour
nament in Oxford, my health was in
such a sudden collapse. And indeed the
physician s were puzzled too. There
were some extrem ely elaborate diag
noses, quite often contradictory. I do
not want to discuss in great depth
wh at was h appening at the time, as
this is something of a digression of
analysing two g ames of chess. I tried
hard to understand what was going on
in the thought processes and degree of
understanding of the medical doctors,
and I was looking for contradictions
and possible mistakes. A difficult proc
ess, but necessary. After all, medics, like
chess pl ayers, can often m ake mistakes,
and I wanted to assess early on
whether a mi stake was being m ade.
We can add th at economists and
politician s too are highly capable of
making serious mistakes, something
fairly clear these days for the gen eral
public. Does anyone believe th at we are
in a stable economy now?
Anyway, pl ay carried on with 5 c3 e6
6 e4 h6 7 xf6 'ilVxf6 8 d3 cxd4 9 cxd4
CDc6 10 e5 'ilVd8 11 a3 d6 12 'ilVe2 e7
13 0-0, and if anything, I might well
have played better in 2006 th an in

Thi s h as, o f course, been played be

fore. In an earlier game, I . Rausis
J . Plaskett, Port Erin 1998, Plaskett
played 13 ... 0-0, not the sort of move
that I would generally want to con sider
myself, not really liking castling wh en
moving the king towards the risk of
immediate attack. Black was better,
and later won, after 14 'ilVe4 g6 15 'ilVe3
g 7 16 CDe4 dxe s 17 dxe s 'ilVds 18
liJf6 ? ! 'i!Vb3 19 l1acl l1ad8, but 18 liJO is
equal .
I chose 13 ... dxe5, and if Plaskett's
immediate castling looks to me slightly
strange, then my own plan would to
m any oth ers seem even more unusual,
allowing the possibility of my king h av
ing to stay in the centre. Quite often it
is not a question of whether a player
thinks th at one type of move is 'better'
or 'worse' than the other, but rather a
case of h alf-remembered experiences
of encounters from many years ago. I
tend not to like going into early mid
dlegames where my king gets stuck in
the corner when I do not h ave extra
pieces defending. Moreover, I tend to


Why we L o s e a t C h e s s
feel uncomfortable about m y king get
ting forced away from the back ranks. I
have h ad a few bad experiences in the
My plan was 14 dxe5 ttJd4 15 ttJxd4
"iVxd4, so far, equal throughout.

I had expected 16 b5+ 8 17 b4

reckoning this as a draw, and maybe
this is objectively the best. Instead Heb
den gave away a pawn, with 16 l:i.acl?!.
Over the years, indeed over the decades,
Hebden has developed the tournament
strategy of playing reasonable moves
quickly, hoping his moves are okay, and
above all, making sure that he does not
allow himself to be in time trouble. This
can be an effective way of playing, espe
cially in weekend games, when his op
ponents get tired and short of time, and
he himself remains fresh for the second
game of the day.
Hebden 's last move was played
quickly, and I h ad simply not expected
it. Worse, I h ad decided beforeh and
that if my opponent was going to play
quickly, I would do the same. I did not
want to get too tired, especially when


getting short of time. The nightm are

would h ave been that, as has occasion
ally h appened, I was to reach a position
where he h ad an hour on the clock, and
I h ad five minutes for the rest of the
game. In better years I could h ave h an
dled it, but right now, I couldn 't. It did
not help that on the Saturday before
h and, I h ad received a phone message
'urgent' concerning a club event. This
was spectacul arly non-urgent, and I
should have ignored it. Any m atch cap
tain will know about this type of stress
before a game.
So the general picture was th at I
was out of sorts, and th at I would h ave
been more than h appy, especially as
Bl ack, to find a comfortable way of
finding quick and steady equality. And
maybe thi s was at the root of the psy
chological error I m ade, quite a com
mon one.
I played 16 l:i.d8? (B), aiming for
equality, but ending up worse.

White played 17 b5+ 'iM8 18 l:i.C7,

and I was starting to struggle, as we will
see further in Test 6.3. U sually I have a

Te s t F i v e
counterattacking style of play, not aim
ing for quick solidity as Black, but in
stead keeping a sharp eye (these days,
alas only my left eye) for any possibility
of taking over the initiative after any
thing that seems slightly inaccurate.
In more normal circumstances, I
would h ave gobbled the pawn,
16 :xb2! (A), without much hesitation.

Bl ack i s better. T o continue the at

tack White h as to play 17 .l:!.C7 i.ds 18
b S+ <t>f8, but his attack seem s specu
l ative. If, for example, 19 a4 ..td8 20
.l:!.d7 g s 21 f4 iVd4+ 22 h l ..txf4, and
White cannot go any further.
In stead, 19 tf-fcl adds another at
tacker into play. The computer sug
gests 19 .. .f6 ! ? and 19 ... ..tcS ! ? as possi
bilities for an edge, and both moves
look reason able enough. Moreover, if I
wanted a quiet and safe equality,
19 .. :ha3 would h ave done. Then 20
.l:!.7c3 (20 .l:!.c8+? ..td8 21 tf.xa8? ? iVXC1+)
20 ... iVb2 (20 ... iVb4?? 21 .l:!.d8+ ..td8 2 2
.l:!.xa8 win s for White, n o w th at Bl ack
does not cover the cl-square) 21 .l:!.lc2
iVal+ 2 2 tf.cl iVb2 repeats.


Test S i x
6.1 Black to play

63 Black to play

A) 1 7 .. :e6
B) 17 .. JUe8
C) Something el se?

A) 18 .. :Vxd2
B) 18 .. :xb2
C) Something el se?

6.2 White to play

6.4 Black to play

A) 18 4'lxe4
B) 18 dS
C) Something el se?

A) 18 ... h 6
B) 18 .. .f6
C) Something el se?


Te s t S ix

Test 6.1

British League (4N C L)

before the time control, before ulti

m ately losing. The remaining moves
will be con sidered later in Test 10.1.


Black should have played with much

more grit, with 1 7 JJj.e8! (B), with the
defensive idea of ... ti.Jf8 and ... .1i.. c 8. The
other knight, when given the opportu
nity, can join in with ... ti.JO. White's
edge is negligible. I would like to think
that had I been in better health, I would
have easily seen this idea.
The queen exchange was a di saster
for Black. Exchanging queens can be
useful in defending the king when un
der attack, but is often less than useless
when defending weak pawns. I needed
to defend on the kingside with my
pieces, rather than exchange my queen.

As we saw in Test 3 . 3 , White h as a

better pawn structure, and started to
set up pressure with his pieces, with 17
Here I pl ayed 17 'iVe6? (A), which I
described at the tim e as "a weak and
lazy move." Certainly I was feeling very
tired that day, and could not even con
template, on Remembrance Day, walk
ing almost next door to the old Coven
try Cathedral, heavily bombed during
the War.
In th e position itself, I was too wor
ried about pressure on my kingside,
and underestim ated my possible
queenside pawn weaknesses. After 18
.id3 'iVxfS 19 i.xfs .l:!.ad8 20 .l:!.acl ti.JC7
21 iLxd 7 ! .l:!.xd7 2 2 ti.Ja4 ti.Ja8 my posi
tion was crumbling.
There were twi sts and turn s later
on , and I was even briefly able to get
back into the game, if temporarily, just

Test 6.2

B u ry St Ed m u nds


Continuing from Test 5.1, my plan is

highly ambitious, with thoughts of a
pawn breakthrough with either d5 or f5.


Why we L o s e a t C h e s s
Either i t works or i t doesn't. To the best
of my calculating capability, I felt I was
doing well, but I missed a resource sev
eral moves later in a critical position. It
turns out that there were good alterna
tives for Black anyway, so my earlier
play was too ambitious.

Black played 16 .. :c6. It would be

prem ature to break the diagonal by
force with 17 d s ?, as after 17 ... exds 18
ttJxds i.. c s+ 19 h l ttJxds 20 i.. x ds
'iVxds 21 'iVxds i.. x ds 22 1:!.xds ttJb6 the
pawn structure is roughly symmetrical,
but Black's pieces are far better placed,
and so he h as an advantage.
So 17 1:!.f2.


Bl ack could now h ave considered

the ultra-solid 17 ... h 6 ! ? 18 i..h 4 1:!.fe8 19
fS eS 20 dxe s i.. x es 21 ttJd4, followed
by finding one of the good queen re
plies. I h ave to admit that White's king
side looks far too loose, and th at I
would h ave needed to work h ard to try
to hold the position . Black's kingside is
extremely safe, despite White's at
tempt to start an attack on th at side.
There are, as we shall see, two good
moves for Bl ack, and therefore I cannot
justifiably claim that I am 'unlucky' in
h aving chosen the line starting with 15
1:!.adl, and with the unexpected coun
terattack l ater on . In stead, I misjudged
the position, m aking both positional
and tactical errors.
Black in stead pl ayed the more direct
and obvious 17 ... ttJe4. Now I should
h ave pl ayed the drawish 18 ttJxe4 (A)
18 ... 'iVxe4 19 'iVxe4 i.. x e4 20 dS i.. x ds
21 i.. x ds exds 22 1:!.xdS 1:!.c6, and quite
probably Black would h ave been h appy
with a h alf-point.
I had prepared, imaginatively but
incorrectly, is d 5 ? (B), and play seemed
smooth enough for m e after lS... ttJxf2?
19 dxc6 ttJxd 3 20 cxd7 i.. C 5+ 21 h2
ttJxb2 22 dxcS'iV i.. xcs 2 3 1:!.d2 ttJC4 24
i.. x C4 bXc4 2 5 a4, and I eventually won
with my extra knight versus two
While h e was thinking about his
taking the knight on f2, I wondered
what would h appen after lS ... exd 5,
which at first seem s like a blunder.
I h ad thought th at everything was

Tes t S ix
covered, and I h ad not really con sidered
that there could be a danger for m e,
but then I saw a possibl e problem just
before he m ade a move, and back at
home I saw that it would h ave been a
major problem, and th at I could h ave
lost two games (the other again st
Gregory - Test 1 3 . 2 ) on the same birth

I could h ave h ad a slalom run with

19 tiJxe4 dxe4 20 "tIYxd6, when I felt I
was safe a piece up, but Black has le
thal counterpl ay with 20 ... e3 ! ! .

"tIYxc6 exf2+ 2 2 xf2 ..ixc6.

If White wants to move the rook in
stead, the only try i s 21 Itf3 xd6 22
.l::!. x d6 i.xf3 . Black should win, although
it m ay take time, after either 23 gxf3
.l::!. c 2 24 .l::!. x d7 Itxe2 2 5 ..ie7 Itxb2 2 6
..ixf8 xf8 2 7 l:!.xf7+ e 8 28 .l::!.f5 .l::!. x a2
29 .l::!. e 5 + f7 30 .l::!. x e3 f6 (level m ate
rial, but White's king and pawn s are
dreadful), or 23 tiJd4 ItC1+ 24 h 2 ..ih 5
2 5 .l::!. x d7 e2 2 6 tiJxe2 ..ixe2 (exch ange
up, and Black can squeeze the bishop
with ... ..ic4, after 2 7 ..ie7 .l::!. e 8).
But I won the g ame. Was I 'lucky', in
the sense th at he missed his chance of
winning, or 'unlucky' in th at while I
calculated an interesting line in ad
vance, there was an unclear tactic
many moves on ? It depends on the
strength s of the two players. At very
top grandm aster level, all thi s would
h ave been a string of blunders, and of
course I am aware of th at. For the time
being, I h ad the excuse of illness, and
few things can be worse th an brain
dam age for a chess pl ayer. But I am
starting to TUn out of excuses now ...

Test 6.3

Metropol ita n Open

I h ad missed th at ! And of course h e

missed it too. White is a clear exch ange
down, without compen sation, after 2 1


Despite earlier events (see Test 5 .4),

my position should not collapse. Unfor
tunately i t did. I would h ave suspected
that I h ad done something wrong, but I
still h ave play, h aving a useful bishop-


Why we L o s e a t C h e s s
pair, a good open file, and an active
queen . That said, I al so h ave obvious
disadvantages, with my king side pieces
being clogged up, and White h aving a
rook on the seventh .

Again, I played too quickly, and

m ade the obvious capture, 18 ... "ii'x d2?
(A), assuming that I was at least con
formably equal . I missed something a
few moves along the line. White won
after 19 l:[xb7 "ii'x e2 20 xe2 gs (better
is 20 ... g6, but White is still on top after
2 1 b4) 21 l:[c1 l:[d2 22 f1.

Now I missed a critical pin after

22 ... l:[xb2?? 23 ..ihS!, and I immediately
resigned (1-0), in view of 23 ... C5 24


.uxc 5 ! bXc5 25 l:[xb2 . 23 ... ..ixa3 24 ItcS+

g7 2 5 l:[xf7 m ate i s even quicker.
Black has to try to improve. My im
mediate reaction afterwards was to try
to find an improvement with 18 . :Wixb2
(B) 19 Itxb7 'iix d2, swiping the b-pawn
(not though 19 ... Itxd2? 20 ItbS+ .lidS
21 ItxdS+, with a cross-pin). White is
still much better, h owever, after 20
"ii'x d2 Itxd2 2 1 l:[c1 g5 22 Ikc7 l:[d1+ 2 3
..itl .lixa3 24 .uxa7 ..i C 5 2 5 .uxf7+ eS
26 Itfb7. Black i s a tempo down after
taking the b-pawn, and White ag ain
keeps a strong attack with two rooks
and bishop.
Here 21 ... ..idS provides more resis
tance. Then 2 2 .uxa7 g 6 23 a4.

Can Black hold this? Offh and, I do

not know. In practical term s, White
would be more th an h appy to carry on
trying to chip away for several hours,
or, in a quickplay finish (as h ere), wait
ing for Bl ack's positi on to collapse. Even
if Black manages to fin ally completed
his development, with ... g 7 and a
rook move, he still has problems with
his pawn s on f7 and g6 (also e6, if

Tes t S ix
Black's king i s on g 7), and his bishop i s
not mobile.
This, however, i s a first impression .
If Black were to try 2 3 .. J:td5 ! 24 .l::t d 7 (24
f4? ! g5 gives counterplay) 24 .. J:txd7 25
iLxd7, with the first pair of rooks gone,
Black increases his possibilities of hold
ing. He is of course not equal yet.
Bl ack can al so try to set up a differ
ent pawn structure in the endgame,
with lB iLxg2!? (C) 19 'it'xg 2 "iVxd2 20
'ixd2 l:txd2 2 1 l:!.xa7.

( 2 5 .l::t c 8 'it'g6 i s comfortable for Black)

2 5 ... iLh4 26 l:!.cxb6 l:!.xf2+ 27 'it'g l l:!.a2
28 l:!.xe6 l:!.a2+, with a draw, following a
reason able degree of accuracy by Bl ack.
Thi s would seem to be the most ac
curate line for Black, boldly aiming for
equality with counterplay, rather than
h oping th at the opponent cannot find
a way of keeping a slight edge.
There i s another way for Black,
keeping m aterial on the board with
lB i..oB (C) 19 4:Jb1 ! .

Clearly Black will not be able to take

the pawn (21 .. J:txb2 ? ? 2 2 l:!.a8+), and so
we now h ave a more dyn amic pawn
structure, with strength s and weak
nesses on both sides. White will h ave
excellent chances of creating a danger
ous passed pawn on the queenside,
while Bl ack will need to set up coun
terpl ay on the kingside. It is important
to recognize for Black that he must not
just sit on the extra pawn on the king
side, but that he needs to play actively.
Therefore 21 ... g 5 22 b4 (saving the
pawn, and al so preventing ... iLc5)
22 ... g4! 2 3 .l::tb 7 iLd8 24 .l::t c 1 'it'g 7 2 5 l:!.c6

The knight retreat is unexpected,

and indeed it was pointed out to me by
computer. The knight soon bounces
back though, and after 4:JC3 l ater, Bl ack
h as no control of the d5- and e6squares. White i s better, Black still h av
ing problems with his development.
There are several tries for Black here,
but none seem s to equalize. For exam
ple, Bl ack can start with 19 ... "iVd5
(19 ... iLc5 20 4:JC3 squashes Black's
counterpl ay) 20 f3 iLC5+ 2 1 h 1.
Then 2 1 ... iLe3 2 2 "iVxe3 "iVxb5 is an
attempt to break the natural course of
play, and if the natural 23 l:!.fc1 "iVxb2



Why we L o s e a t C h e s s
2 4 f4 g 5 2 5 4, Black holds the position with 25 .. J:tC7 ! . Once thi s has been
appreciated, 23 lIg 1 ! is quickly seen as
a good move, and if 23 ... xb 2 ? (but
other moves are not very good) 24 f4
f5 2 5 5 lIe7 26 lId8+, and m ate next
Black can try instead 21 ... a5 22 4Jc3
d4 23 lId1 4 24 lIxd8+ xd8 2 5
lId7, and White keeps a steady edge.
The test position i s more compli
cated th an it looks, and time pressure
was beginning to be a problem. I did
not h ave time to analyse in depth, and
found one of the worst moves.
18 ... xg 2 ! i s the best, although thi s
would take good nerves. 18 ... xb2 i s
pl ayable, but not 18 ... xd2?, when I did
not see a tactic later on .

slight pressure for the last few moves,

but my next move,

18 ... h6?! (A), is highly compromis

ing: a pawn weakness. Before too long,
Bl ack decides h e has to pl ay .. .f6 as well
as ... h6, and then he has light-squared
weaknesses in front of the king .
18 ../61 ( B ) i s much better, and is
about equal .

Test 6.4

Britis h League (4NC L)


We resume pl ay from 4.1 with 17

4Je2 4Jd8 18 h4. I h ave been under


Black will be able to keep the pawn

on h 7 . Any h 5-h 6 push by White can
usually be countered quite easily. No
detailed analysi s here. Just play
through the game, and imagine wh at
would h ave h appened if Black h ad de
layed ... h 6 .

Tes t S ix
Another possibility i s l.B ..tDe6?! (C),
with a trap. If 19 iifS .tc8 ! ? 20 iixdS ?
CiJe], and the queen i s unexpectedly
about to be trapped. 21 'ilYxC4 is the
only move, but 2 1 ... .ta6 skewers the
knight. An attractive variation, but the
simple 20 hS keeps an edge for White.
In the g ame, White quietly re
treated with 19 .td2.

Black does not h ave any imm ediate

problems just yet, and the computer
gives it as equal , but there are will be
difficulties ten or twenty moves along
the line. In other words, this i s a posi
tion al battle, with advantage to White,
rather than a tactical struggle.
There are two basic problems with
Black's pawn structure. First, Black i s
suffering from covering what i s in ef
fect an isol ated pawn in the centre. His
(-pawn has moved too far, and cannot
defend the ds-pawn, nor even do any
thing to attack White's pieces and
pawns. Black is forced to work out h ow
to defend the ds-pawn . Second, White
has g ained space on the king side, with
pawn s as well as pieces. This suggests

that Black i s forced to defend his king

side structure as well as the central
pawn . White can think of a possible
attack again st the king .
I pl ayed 1 9 .tb7, in part t o cover
the ds-pawn, al so perh aps to try ... .tc6
later, to cut down any m anoeuvring by
White's queen and rook. It seem s a
slightly lazy move. I was in pain suffer
ing after a nasty fall, as well as longer
term illnesses, and I was not playing
energetically. Of course the only way of
losing a game of chess is to m ake bad
moves, and so somehow the chess
pl ayer, when under pressure, still h as
to work h ard. Lauterbach played better
th an I did in the middlegame.
19 ... CiJe6 is more relevant, and if 20
hS CiJe7 21 CiJes ? ! ..txes 2 2 dxe s CiJcs,
and Bl ack soon m aintains the balance
with his pieces. 21 CiJh4! ? keeps a slight
edge for White, though.
Lauterbach played 20 h S .
. . .

I played 2 0 CiJh8?!, which may

seem strange and unnatural, placing
the knight into a passive corner. It is
not as bad as it looks, and I did not like
. . .


Why we L o s e a t C h e s s
the alternative. Black's mi stake came
later. I felt the knight was, in several
respects, worse with 20 . ..tlJe7 than on
h8, blocking the e-file.

There i s al so the questi on of m a

noeuvrability. The knight on e7 can go
to c8, but then wh at next? At least on
h 8, Black can find a better square with
.. .f6 an d ...elJf7. But, remembering ear
lier comments, Black would h ave done
much better with 18 .. .f6 ! (kicking the
bishop out), rather th an 18 ... h 6 ? ! .
O n specifics, after 20 . ..tlJe7, i f White
plays 21 ..tf4 ttJe6 22 ..txd6 iNxd6 2 3
ttJe s ? ! , following the idea of the m ain
line, then Bl ack equalizes with 23 ... ttJc6,
a good argument for the ... ttJe7 ap
proach . 23 l:!ab 1 ! keeps White an edge
though, forcing Black to work out h ow
he h as to face b3 (or b4), or ttJes, or,
after 23 ... ttJc6, then 24 ttJh4.
I suspect th at I might in stead h ave
been worried about 21 ttJes ..txe s 2 2
dxe s, opening up a square for the
knight on d4. Thi s does not seem all
that effective after 2 2 ... ttJfS ! . If 2 3 ..th 3
ttJd4 24 ttJxd4 iNxh 3 (threatening the


h s -pawn) 2 S iNfs ..tc8 26 iNxh 3 ..txh 3

2 7 f3 ..td7, and Black should equalize,
with the idea of .. .f6.
So the obvious 20 ... ttJe7 is better
than 20 ...ttJh 8, but only because of a
self-pin with ... ttJfS, not so obvious. I
could easily h ave added this as a test
position, but there would be too many
interruptions if there are questions to
be asked on each successive move, for
each minor slip. We need to keep the
flow going. Without the ... ttJfS self-pin,
Black's knights would h ave been clumsy,
the knights on d8 and e7 not working
well together, and not helping the other
pieces. This was what I remember being
worried about at the time.
Back to the m ain line. Lauterbach
played 21 ..tf4, exchanging off Black's
better bishop. Then 21 ttJe6 22 ..txd6
iNxd6 leaves White with a slight, but
annoying, edge.

She played 23 ttJes, which instinc

tively surprised me. As Nimzowitsch
used to say, "the threat is stronger than
the execution". The knight is not attack
ing anything, apart from the easily cov-

Te s t Six
ered lz'lxc4, and Black can hit back, later
on, with .. .f6 followed by ... lZ'lf7. Probably
White should leave the knight at home,
allowing Black to decide whether to try
.. .f6, without any tempo gain, and have
to decide whether it is playable or bad.
23 l:tab1! would be a way to test her op
ponent. Looking at this now, 20 ... lZ'lh 8
was over-elaborate, and does not do the
job, so deserves its ' ? ! '.
After her knight advance, I played

23. . .l:te7, then came 24 'ii'd 2?!. Maybe 2 4

lZ'lg4! l:tfe8 2 5 lZ'l e 3 would have more
effectively justified White's lZ'le 5, keep
ing an edge. I was not sure what White
was doing with her queen move. Obvi
ously White is not worse after the text,
but she could have achieved even more.
Play continued with 24 ...l:tfe8 2 5
Wf1 f6 26 lZ'lg4, and at last I felt I was
fully equal . We sh all resume the game
later in Test 10. 3 .


Test S eve n
7.1 White to play

73 Black to play

A) 19 J::t a 3
8) 19 tLig s
C) Something el se?

A) 20... ii.a3+
B) 2o . f6
C) Something else?

7.2 White to play

7.4 White to play

A) 20 h 3
8 ) 20 .l:!.d1
C) Something else?

A) 2 1 h 3
8 ) 2 1 'iVd4
C) Something el se?


Te s t S e v e n

Test 7.1


Following on from Test 5 . 3 , this is

an extrem ely difficult position to try to
assess. I played 19 .l:1a 3 (A) over the
board, after some thought, but found
thi s un satisfactory.
My initial idea was 19 .l:1e1 (C), but
after 19 ... i.. xf3 it looked as though
Black was doing well.
19 tLJgs (B), a suggestion by the
computer, looked good, and on seeing
this, I regretted that I did not look at it
more closely. That said, looking at it
ag ain, it is not so convincing. These are
the three most active options, but the
quieter continuation s might turn out
to be more effective, if the more com
bative moves do not impress. The one
really clear assessm ent i s th at 19
d4?? i.. xf3 is a blunder, losing a piece.
The position is to be regarded, at
least initially, as unclear'. It is unlikely
that anyone is clearly worse, as there

are strength s and weaknesses for both

sides. If the position is genuinely un
clear, the implication i s th at unless ei
ther player m akes a mistake from thi s
position, t h e final result should e n d up
in a draw, with neither player being
able to force a win . This i s much as at
the start of the game. The one way of
losing is to make a mistake. The best
way of avoiding loss i s to avoid mis
Caution, not aggression, is therefore
the watch -word. Let us go through the
position in greater depth, after the ini
tial assessm ent, and consider 19 .i:!.e1
(C) 19 ... i.. xf3 . My notes at the time ex
amined only 20 gxf3, keeping the
queen s on the board. After 20 .. :Yf5 2 1
"iid4 tLJd7, White h a s slightly more ac
tivity th an Black with the pieces, but
his pawn structure is badly dam aged.
" E qual, but no more", I wrote earlier,
and without trying to an alyse in ex
treme depth, this seem s fair enough.
It is interesting th at in my notes at
the time, I did not even mention 20
xf3 ! ?, protecting the pawn structure.


Why we L o s e a t Ch e s s
White is then slightly better after
20 ... 'ilVxf3+?! 21 xf3, moving towards
a favourable endgame. I feel sure th at I
would have assessed thi s position cor
rectly. I h ave no fear of a queen ex
ch ange. Indeed, I would almost cer
tainly h ave decided that 20 ... 'ilVd4+! i s
best, a n d highly unpromising if White
is thinking of playing for an edge. After,
for example, 21 i.. e 3 'ilVxb2+ 22 g l
tte8 2 3 ttab1 'ilVC3 White will later re
cover the pawn on b7, but this will lose
time, and Black's knight will reach good
central squares after ... ttJd7. There is a
clear danger th at White could easily
end up worse, and thi s is exactly wh at I
wanted to avoid.
We move next to the line I actually
played, 19 tta 3?! (A) 19 ttJd7 20 tte1
(but not 20 tte3 ? 'ilVxf4 2 1 l:Ie7 fS, and
Bl ack win s a pawn). Then 20 'ilVb4!
proved uncomfortable for me.

Sometimes it is better to use a com

pletely new perspective. I h ad been as
suming th at I needed to bring the rooks
into pl ay quickly, giving myself the
chance of actively developing the
bishop. I took it for granted th at the
knight was pinned, and th at at some
stage Black would exch ange on f3, if
required. But maybe the knight can
h ave an impact?
It took me a long time to think
about 19 ttJgS!? (B), even after the
gam e, and even with the h elp of the



Black is still keeping pressure on the

b2-square. I did not h andle thi s posi
tion well, and I was fortun ate th at he
opted for a perpetual when short of
time. But that is a stage of Test 7.4.


If 19 ... i.. x d1 ? 20 ttJxe4 i.. c 2

(20 ... i..b 3 ? 21 tta3 wastes time) 2 1
ttJxd6 with a big plus for White. Bl ack
needs to improve, with 19 .. :YWfS . Then
perh aps 20 'ilVd4 (20 1!Vh3 ttc8 is much
slower) 20 ... l:Ie8 2 1 h 3 'ilVc2+ 22 g 3
i..fs 2 3 'ilVf6 'ilVd3+ 24 h 2 'ilVxds 2 s ttd1
ttJd7 ( 2 S ... ttxd1? ? 26 'ilVxf7+) 26 xg6+
hxg6 2 7 l:txdS. Even h ere, 2 7 ... ttJcS i s
slightly better for Black. White's bishop
still remain s undeveloped, and his rook
still obstructed.
This suggests, as an alternative, 19

Tes t S e ve n
1i.e3 (C), but after 19 .. .'Jd7 followed by
... tt:Jf6, the knight suddenly reaches a
good square.
There are various other possibilities
for White, including 19 as, 19 b3 and
19 l:.fl, but none seem s completely
secure. Bl ack quickly plays ... tt:Jd7.
By now, it is easy to see that analysis
can easily run around in circles. After
some thought, I played what seems to
be the natural move, having found
nothing better over the board. It looks
like that there ought to be an improve
ment for White, but even in later analy
sis, I cannot find a clear drawing line, let
alone a win. So perhaps I was worse?
Around the roundabout, White's origi
nal exit, 19 l:.a3, now looks the best, but
I had to find good moves later on.

now do I respond? I n the end, I played

18 iLC2!?, which I described as a "wimp
out" in an earlier set of analysis. Play
continued with 18 :iVf6 19 iLC3 iLd7,
which was critical, and the point at
which White needed to improve. We
sh all examine thi s below, although of
course the con scientious and thorough
reader will already h ave examined this.
Sometimes it i s appropriate to grab
the exch ange, but sometimes it should
be delayed. It i s always a difficult deci
sion to m ake. White certainly needs to
consider 18 1i.xf4! ? exf4 19 h3 "iYf6
(maybe the simple 19 ... tt:Je s, with rea
sonable compen sation for the ex
ch ange), and indeed I con sidered it,
and rejected it.

Test 7.2
Colin Crouch - Andrew Lewis

Kid l i ngton


As we saw in Test 3 .4, Lewis h as of

fered an exchan ge sacrifice, but how

This game was played a few years

back, before my writing on Tal, Stein
and Kasparov, and also my book on wins
by 2 700+ players in Modern Chess: Move
by Move. I would like to think that by
now I would have been more fully
aware of the idea that even in the wild
est positions, the best play by both sides
should end up in a draw, unless of


Why we L o s e a t C h e s s
course one of the players is already bet
ter. What this in practical terms would
mean is that if your position feels okay
before the opponent's attempted sacri
fice, and if you see no clear advantage
for your opponent, then you are at lib
erty to accept the sacrifice, however
dangerous the opposing sacrifice might
appear. 'Fortune favours the bold', and
bold defending, as well as bold attack
ing, should be praised.
Bl ack h as no win after 19 .. :f6 ? ! ,
which sets u p an apparently extremely
dangerous attack, but it turn s out that
Black has no more than a draw after 20
hxg4 al+ 2 1 i..b l i.. x g4 (he h as to
aim for ... idS) 2 2 l:I.dl ii.b2+ 2 3 C2
i.. C 3 24 Cl i..b 2+ 2S Cl. And, second,
th at the sneaky 24 .l::i. c l ! ! gives White an
unexpected advantage after 24 ... iLfS+
2S dl i.. x bl 26 e6+ h 8 2 7 f7.
This line would h ave been extremely
difficult to foresee for either side.
So 18 i.. xf4 is extremely promising,
but 18 iLC2 !? f6 19 iLC3 iLd7 i s al so
fully playable, provided I play accu


I h ave avoided the tactical battle,

but now I h ave to con sider the posi
tional battle. I should not be worse, but
my next move is poor.
I played 20 l:I.dl? (B), but I soon was
forced to appreci ate that the rook was
on an unfortunate square, and that
after 20 ...l:I.f8 21 h3?! (mistimed, but
even h ere 2 1 l:I.hel l:I.xf3 ! ? 2 2 gxf3 xf3
2 3 xf3 l:I.xf3 24 l:I.d3 l:I.xf2 clearly fa
vours Black) 21 ...iLlh6 22 l:I.d2 iLlf5 23
iLlel iLld4 my position was gradually
falling apart. There was not much to be
done after 24 iLxd4 exd4 25 iLld 3 i.. h 6
26 bl l:1.e8 27 dl l:1.fe4 28 f3 l:I.e3 29
Cl i..f 5 30 xa 3 l:1.xf3 3 1 l:I.ddl l:I.g3 32
'iia 7 l:I.xg2 33 'iix b7 .l::i. e e2 34 b8+ f8
3 5 xf8+ xf8 0-1.
During the g ame, I was regretting
th at 20 h3!? (A) would h ave been m ore
accurate, so that after 20 ... iLlh 6, White
h as a wider choice, other th an 2 1
l:I.dl ? ! , in reply. 2 1 l:1.el would certainly
h ave improved, trying to bol ster the e4square, but Black is still better: for ex
ample, 21 ... iLlfS 22 iLld2 iLld4 23 iLxd4
exd4 24 iLle4 'iVe s 2 S i.. d 3 i..fs 26 f3 (26
C2 i.. x e4 2 7 .:t.xe4 .:t.xe4 2 8 i.. x e4 d3
29 i.. x d3 g S+ favours Black) 2 6 ... iLxe4
27 fxe4 (27 i.. x e4? d3 28 i.. x d3 al+ 29
d2 'iix a2+) 27 ... .l::i. af8 28 .:t.hfl l:1.xfl 29
.:t.xfl .:t.xfl+ 30 'iVxfl g S+ and White is
uncomfortable. In such lines, Black has
excellent play on the dark squares, with
opposite-coloured bishops, but White
has no corresponding pl ay on the light
squares, his own pawn s blocking his
bishop and queen.

Te s t S e ve n
So wh at else i s there? Clearly I need
to bring my pieces into pl ay, but in the
game I soon h ad to give protection
with l::t d 2, thereby losing time. This
suggests an improvement, with 20
fl!? (C).

h as a flight square for both knight and

rook, but also that the es-pawn is no
longer pinned. If, for example, 2 1 h3 e4!
22 .txe4 .txc3 2 3 hxg4 l::t e 8 24 tiJg s
l::tfxe4 2 S tiJxe4 'iWf4+, and Black ends up
with a winning advantage. There are
various other possibilities for White, but
while the knight is on g4, and creates
pressure, White faces problems.
The best way for White, it seem s, is
to kick the knight out immediately,
with another 'something else' move,
namely 20 h3!? (in stead of 20 .t:.d1 or 20
.t:.f1), and then 20 ... tiJh 6 21 l::tf1 ! . The
only problem is th at my intention was
to avoid pushing the knight away to a
better central square.

Suddenly White's quiet pl ay with

ii.c2 and .tC3 m akes more sense. If he
has time to regroup with, for example,
20 ... l::tf8 21 h3 tiJh 6 22 .te4 .tfS 2 3
tiJd2, White h a s finally con solidated h i s
centre, with perh aps a slight edge.
However, 20 ...'iVf8 !, a suggestion
from John Emms, is an improvement for

The point is not just that Black now

If then 21 ... tiJfs, White would

probably be quite h appy to break the
central tension with 2 2 g 3 tiJd4 2 3
tiJxd4 ( 2 3 .txd4 l:txf3 24 i.. e 4 l:txb 3 !
favours Black) 2 3 . . . exd4 2 4 gxf4 dXC3.
Thi s might not necessarily be the most
critical position where there is a choice
of several moves, but it is certainly one
of them. Thinking in term s of a player
trying to analyse this position over the


Why we L o s e a t C h e s s
board, the natural inclination would be
to think th at Black is better, despite
being the exch ange down . Black has
two advanced pawn s on the queenside,
whereas White has a weakened pawn
structure there, making it difficult for
White to bring his pieces into play. If
the bishop moves, for example, ... c2
would be a reflex reply.
If the reader is finding it confusing
over the intricate changes in move or
der, and the sudden emergence of tac
tics, then me too. I still feel th at it is
almost a m atter of principle that White
ought to be able to keep some sort of
edge in the King's Indian, with the ad
vanced central pawn structure, but thi s
turn s out t o b e extremely difficult to
prove. Every time that White feel s that
h e has a secure edge, there always
seem s to be some unexpected way for
Bl ack to hit back.
Positional chess is so often ex
tremely delicately bal anced, particu
l arly when it involves sh arp position al
pl ay. It might often be the case th at a
player might h ave four natural moves
with which to carry out a m anoeuvre,
but it is far from clear which precise
move order needs to be pl ayed, given
the case th at the opponent might al so
h ave four possible manoeuvres. Of
course, at any stage one of the pl ayers
might well ch ange the balance of pl ay,
possibly ending up with tactics, or pos
sibly with a slightly varied new strate
gic pl an . A move not only puts a piece
on to a new square, but al so removes a


piece from its previous square.

For the moment, playing ag ainst
complicated King's Indian openings, I
h ave been playing the Torre (1 d4 tDf6 2
tDf3 g 6 3 g 5), before trying again
something more ambitious.

Test 7.3
P Sowray C Crouch



Thi s position ought to be fully play

able for Black, one might think, but
after a few more moves White started
to emerge ahead, and later won . It was
important for me to try to work out
how and wh ere I went wrong . My as
sumption during my initial an alysis
was th at I defended far too cautiously,
and that I should h ave fought fire for
fire. After further scrutiny, I now h ave
come to the view that the next few
moves for Black were correct, ending
up with a slight but tenable disadvan
tage. The attempt for Bl ack to play for
an advantage, in m any lines, would

Te s t S e v e n
have ended up as an advantage for
White, often after tactical play.
I played 1s ... e s ! ?, a move l ater than
I should h ave done, as we saw in Test
4.4. Black's ... bS proved to be unneces
sary. Instead, Bl ack could h ave tried
taking up the ch allenge with lS ... b4 16
liJdS exds 17 exds+ <t>d8 18 liJc6+ ..txc6
19 dxc6.

Such minor piece sacrifices in the

Sicilian are extrem ely scary, and I felt I
wanted to avoid this. Black i s not
threatened with checkm ate yet, so
there i s still some time for flexibility,
and the chance to pl ay 19 .. :as ! . Then
a quick and unexpected repetition with
20 l:td4! liJb6 21 e s ! liJC4! 22 e2
liJb6, difficult to envisage a few moves
in advance. In stead, 20 ..td4? xa2 2 1
i.xf6+ <t>C7 ! 2 1 b3 liJa3 would h ave lost
quickly for White.
A third possibility i s the simple developing approach with lS ... ..te7 ! ? 16
b3, then either 16 ... b4 or 16 ... liJa3, with
tough play, probably ending up around
equal . Black's quiet move would per
haps h ave been the way to play for a

win, but i t would take a con siderable

amount of hindsight to think about
this line.
I felt comfortable with my central
pawn push, but 16 b 3 ! ! took me out of
my comfort zone.

The big problem from Bl ack's point

of view is that he weakens the e-file if
he captures on d4. White then can
open up the e-file, with eS or liJds,
when he is taking the initiative. I did
not expect my opponent's move, but
now wh at can I do?
This was a difficult choice. For a
long time, I concluded that my reply,
16 ... exd4, was a mistake, but it looks
okay, and it was the next move th at
was my mistake.
The computer suggests, after a bit
of prodding, th at 16 ... liJa3 17 liJdS d8
gives an advantage for Bl ack, but this
seem s unlikely. Computer analysis
tends to concentrate on trying to avoid
the loss of a piece, but strong hum an
players will want to keep the attack
going, overriding loss of material . In
deed, 18 l:td2 exd4 19 ..txd4 leaves


Why we L o s e a t Ch e s s
Bl ack uncomfortable, for example after
19 ... 3Le6 20 'it'b2 b4 21 3Lxf6 gxf6 2 2
16 ... 'ilVas is slightly more accurate,

All sorts of pieces were flying

around in my mind, and eventually I
decided th at it was too ri sky for me to
try to keep the extra piece and win . I
was more worried that I would end up
losing. So I blocked the e-file with
17 ..ltJes ! 18 fxes dxes anyway, and
later lost. My l ater reaction after the
game was th at returning the piece was
a mistake, leaving me with a positional
disadvantage. Surely I could h ave
pl ayed more ambitiously, with an un
clear position, and with ch ances of
keeping the extra piece?
Of course, one cannot m ake an as
sessment completely abstractly, and
precise evaluation of pieces and varia
tion s needs to be con sidered. In home
analysis, I examined several lines, both
from thi s position, and in earlier posi
tion s. I decided th at in the end Black
h ad gone wrong on move 17, with the
possibil ity of a complicated perpetual . I
al so decided that even with best play,
Black h as not got an objective advan
tage. I cannot play too ambitiously.
It was only a long tim e later th at I
started to look again at the quiet line of
the g ame, deciding in the end that with
an improvement, Black's play would
h ave led to close equality. In other
words, my assessment in the early part
of the g am e was correct (apart, of
course, from the imprecise 14 ... bS), but
that there was a serious mi stake l ater
on . We shall consider the improvement
soon, but first it i s time to examine the
wilder lines.

in the sense th at after 17 bxc4 bxc4,

White can only choose between 18 .l:!.f3
and 18 .l:!.g 3, and cannot play 18 tte3 ? in
view of 18 ... exd4. This m ay be com
pared with 16 ... exd4 17 3Lxd4 'ilVas 18
bxc4 bxC4, when White h as the choice
between 19 .l:!.e3, 19 .l:!.f3 and 19 .l:!.g 3 .
Even so, 19 .l:!.g 3 3Le7 20 g s hxg s 2 1
.l:!.xg s 'ilVa3+ 2 2 'it'dl still keeps an edge
for White.
Anyway, after my immediate cap
ture on d4, he took back with 17 3Lxd4.


Te s t S e ve n
It is, of course, a bad defen sive mis
take to retreat m eekly with 17 ... liJb6 ?
White crashes through in t h e centre
with 18 e 5 . Black needs to play actively,
either returning the extra piece, or us
ing it in a counterattack ag ain st the
white king . So 17 .. :a5 ! ? needs to be

We m ay quickly reject 18 liJd5 ? as

premature: 18 ... liJxd5 19 exd5+ d8 20
bXc4 lIxC4 is safe for Bl ack, and he has
an edge. White's king is not so safe.
18 e 5 ? ! i s the m ost immediately ag
gressive line, but Black again h as quick
counterpl ay, with 18 .. :a3+ 19 dl
ttJb2+ 20 d2. Then 2o ... liJxd3 ? would
be too greedy, in view of 2 1 exf6+ liJe5
22 .txe 5 dxe 5 23 'iVxe 5+ .te6 24 fxg 7,
with advantage to White. In stead,
2o ... dxe5 21 .txe 5 ! (Black is better after
either 21 xe 5+ iJ.. e 7, or 21 fxe 5 liJxg4
22 liJxd3 23 'iVxd3 iJ.. x g4), and it is be
coming even sharper.
The computer suggests two win
ning moves, but 2 1...liJxd3 ? is in the
end far too entertaining to be wise: 2 2
iJ.. d 6+ iJ.. e 6 2 3 iJ.. x a3 liJxel 24 iJ.. xf8

liJxg 2 2 5 iJ.. x g 7 liJxf4 2 6 f3 lIxC3 2 7

'iVf4 i s g ood for White.
21 ... JJ.. e 6 ! is much better, and good,
but a nightm are for a player trying to
analyse over the board. Or, for the Tal
sch ool of attacking chess, many would
take a positive delight in playing
through such lines, knowing that when
neither player can analyse to a clear
result, artistry in chess prevails over
m ere calculation. Pl ay might continue
with 22 f5 liJxd3 23 'iVxd3 l:td8 24 JJ.. c 6+
e7 2 5 .txf6+ g xf6 2 6 liJd5+ d6 27
g 3+ (27 fxe6 xc6 2 8 e7 i.. x e7 win s
for Black) 2 7 ... xc6 28 'iVC7+ xd5,
which eventually leaves White running
out of checks.

It would take far more th an a quick

glance for Black to convince him self
th at his king is safe in the centre, but
he h as some defen sive ammunition
though h aving an extra rook plus two
extra bishops. Al so, perh aps even more
significantly, White's king is on an
open and exposed square, with checks
and a discovered check by Black's king
cutting down White's option s. If, for


Why we L o s e a t C h e s s
example, 30 c4+ bXc4 3 1 bxc4+ 'it>d4 3 2
f4+? (but 3 2 'iVxd8+ 'iVd6 i s al so a win
for Black) 32 ... 'it>cS+, and yes, thi s is a
counter-check, so White i s not allowed
to play f2+.
The mistake took pl ace much earlier
in this line. Back to 17 ... 'iVas, and now
18 bXC4 ! .

B y recovering the piece quickly,

White slows down his attack, and al so
his opponent's attack. We are now in a
sh arp positional set-up, rath er than all
out tactics. The computer suggests that
the position is equal, but after further
examination, White keeps an irritating
edge. Black is not equal, and the move I
actually played, 17 ... tDes ! , was closer to
full equality.
H ere 18 ... .l::t x c4 does not work, as
White h as a resource with 19 eS J::i. x d4
20 exf6+ 'it>d8 21 .l::t x d4 xC3 22 'iVe3 ! ,
choking Black's counterpl ay. Therefore,
18 ... bxC4 (or 18 ... 'iVa3+ 19 'it>dl bXC4),
but White still keeps an edge after ei
ther 19 .l::t e 3 or 19 .l::t g 3 Ji.e7 20 gs hxg s
2 1 J::i. x g s, a side-line we h ave noted be
fore, after a tran sposition .


After 1 9 J::i. e 3 'iVa3+ 2 0 'it>dl Ji.e7 2 1

eS dxes 2 2 J::i. x es Ji.e6 White stands
well, the most direct way being to sac
rifice the exchange with 23 J::i. x e6 fxe6
24 'iVxe6. Not pleasant for Bl ack.
So we return to the m ain line, and
Bl ack's return sacrifice at move 17.
White regains the knight, then 19 tDd s
tDxd s 20 exd s .

I was understandably concerned that

I had done nothing to develop my king
side pieces since move S, and that my
king was still stuck in the centre. I was
greatly relieved that I could gain a
tempo with 20 .ia3+? (A), and then
castle quickly. This unfortunately
proved to be a serious mistake. After 21
Ji.b2 Ji.xb2+ 22 'it>xb2 0-0 23 d6 'iVcs 24
d2! (a quiet queen move which I h ad
underestimated), White has excellent
central control, and an advanced and
dangerous passed pawn. Black h as none
of the usual Sicilian counterplay on the
queenside, although of course I tried for
some. We return to the finish later.
20 f6! (B) is not so far from being
level . Black keeps his bishop in play,


Tes t S e ve n
making it far m ore difficult for White
to advance his passed pawn .

Perh aps it would be only just for

White to be able to keep a slight edge,
given that h e i s better developed, and
has a good passed pawn . So 2 1 .l:i.C3 !
i.a3+ 2 2 <t>bl "ifd6 2 3 .l:i.xc8+ .txc8 24
.l:i.dl 0-0 25 .tal .td7, and the struggle
is likely to continue after the adjourn
ment. Bl ack h as not yet equalized. A
computer suggestion gives 2 6 h4 l:!.e8
27 g5 f5, and it i s not so clear that
White has broken through after either
28 gxh 6 xh 6, or 28 5 e4.
It is time to return to the m ain line,
after 24 d2.

The next few moves were played at

some speed, with 2 5 d7 l:!.cd8 26 l:!.d6 b4
27 l:!.d1 as 28 .tc6 "ifb6 29 "ife2 "ifcS 30
l:!.ld3 hS 31 gxh s .tfS 32 .te4 .txe4 3 3
xe4 fs.

I h ave m anaged to keep up some

counterplay, hoping that there is going
to be an in accuracy by my opponent
during the time scramble, m aybe due
to some pressure in front of his king,
and above all, pushing the two passed
pawn s in the centre. I did not expect
my position to hold, but White was
clearly nervous of my pawn s.
34 "ifc4+ ! "ifxc4 35 bxc4 slaloms
through with White's second pawn .
Both players h ave two connected
passed pawn s, but it is not difficult to
work out that White's pawn s are fur
ther ahead, closer to promotion . For
example, 35 .. .f4 36 c5 f3 37 l:tdl .l:i.f4 3 8
c6 .l:i.C4 3 9 l:!.el e 4 4 0 .l:i.xe4 l:txe4 4 1 c 7
l:!.ee8 4 2 dxe8"if l:!.xe8 43 .l:i.xd8.
White played instead 34 h4?!,
which gave me a little encouragement,
but once White h as the opportunity of
attacking with h6, Black is likely to run


Why we L o s e a t C h e s s
out of time. Then 34 ...e 4 3 5 l:td1 led to
my final mistake, which is worth re
cording for psychological interest, but
not really worth giving as an exercise.

35 ... a4? 36 bxa4 g ave away a pawn

for nothing. White has to capture of
course, as otherwise ... a3+ is likely to be
winning. There are so m any examples
in chess of giving up a pawn to break
up the opponent's pawn structure (in
cluding 30 ... hS in this g ame), and
sometimes such moves can be played
almost automatically. H ere, all I m an
aged t o achieve was t o weaken m y b4pawn, while giving my opponent a
passed a-pawn .
I was annoyed afterwards with my
misjudgement, but in the end it did not
m ake much difference to the final out
come. The tactics remain the same.
Play continued 36 J!Ve5+ 3 7 '.t>b1 (see
diagra m) 37 ...f4, and 1-0.
A strange position to resign a game,
except for the point that this was an
evening London League game without
a quickplay finish. I resigned at the ad
journment. There is no point in travel -

ling to central London when with the

help of computer an alysi s (and, of
course, my opponent was al so more
th an competent to analyse without
computer help), it is easy enough to
find a win by the opponent.

The liveliest line is 37 ... e3 38 'iVC4+

Wh 8 39 l:te6 "iVg 3 40 h 6 g xh 6 41 "iVd4+
Wg8 42 'Yi'b6, and White wins.

Test 7.4

Kid l i ngton



Following on from Test 7.1, I was

Te s t S e ve n
understandably nervous about thi s
position , m y pieces all being o n the
edge, except for the pinned knight and
the exposed king, whereas Black's
pieces are on good positions or on open
lines. My position is not yet drastically
bad, but I need to be careful, and cer
tainly I am not better.
Someh ow I need to activate my
pieces, or to exchange some m aterial . I
chose the second idea, playing the sim
ple 21 h3?! (A) 21 ii.xf3 22 l::i. xf3 l::i.fe8,
but my position was worse, and I was
fortunate that h e eventually decided to
take a perpetual when he was better,
but short of time. The problem with the
move I tried to defend with i s that it i s
using u p another pawn move, when
really I needed to concentrate on bring
ing my pieces together.
Even 21 rj;;g3!? (C) is a possibility, in
viting Black to exchange on f3, while
also m oving the king to a safer square

can b e useful . If, for example, Black

were to carry on with the line of the
g ame, with 2 1 ... ii.xf3 2 2 l::i. xf3 l::i.fe8,
White now h as 2 3 Itfe 3 ! , and the f
pawn i s safe. 2 1 ... ii.fs 2 2 Itae3 also
keeps White's pieces in play.
21 'iVd4! (B) is a much simpler line,
indeed almost too simple if White is
still hoping to outplay his opponent.


There is no check on the g l-a7 di

agonal, and al so the less obvious point
that setting up an extra defence on f4

The queens come off immediately,

with 21 .. :iVxd4 22 tLlxd4, and then
22 ... tLlf6 23 l::i.b 3 .l::i.fe8 24 .l::i. x b7 .l::i. x el 2 S
rj;; x el tLlxds 2 6 fS ! , and once White's
bishop is in full play, he comfortably
holds the balance.
21 Wid2?! (C) i s worth considering,
especially if Black is not interested in
steady equality. If 2 1 .. :xd2 ? 22 tLlxd2 ! ,
White suddenly jumps t o good squares
with the knight, but 2 1 .. .'iVc s+ 2 2 '.tg 3
ii.xf3 2 3 l::i. xf3 tLlf6 is about equal . There
is a third option . In stead of immedi
ately exch anging, or escaping with the
queen, Bl ack can keep the tension with
21 ... a s ! , when he seem s slightly better.
There are then a couple of dubious
gambit ideas. If 21 il.d2? (C) 21 .. :iVxb2


Why we L o s e a t Ch e s s
2 2 l:tb3 a2, the queen is o n the edge,
but then what next? 2 3 l:txb7 .i.xf3 24
gxf3 tLlC5 win s the exch ange for Black.
Or 21 ae3?! (C) 21 ... xf4 22 e7
f5, and White does not h ave quite
sufficient compen sation for the pawn .
There were clearly quite a few al
ternatives to con sider, but offering to
exch ange the queens with 2 1 d4
looks the best and safest.
We return to the position as di s
cussed on move 22, and then 2 3 xe8+

.i.b2 e1+ 29 f2.

Both king s look to be in trouble, but

a critical attacking piece in either side
is pinned.
My opponent thought until the last
minute. As the earlier minutes ticked
by, I started to feel more relaxed. If he
could find a win in stantly, then so be it.
If he h as to take a few extra minutes,
he would not be able to find a more
complicated win . Eventually he decided
it was time to take the draw with
29 e3+?! 30 f3 e1+ 31 f2 1/1-1/1.
I still h ad the feeling that I should
probably h ave been losing, and in later
analysis I was able to find something,
with the h elp of an endgame
zugzwan g . But who would be able to
find the line, the first part of the proc
ess, and then spend th e extra tim e to
confirm th at the critical line is indeed a
zugzwan g ?
The critical l i n e is 2 9 . . .f1 ! (or, after
repetition, 3 1 ... l:tfl) 30 .i.xf6+ xf6 3 1
8+ e7 3 2 d4, when the king and
pawn endgame is equal if the queens
and rooks are exchanged.

White is 'almost' close to being

equal, but it turn s out th at he i s not
quite there. I h ad misjudged thi s posi
tion, and none of my pieces are on a
good square. The only hope, I felt, was
to try to take control of the long diago
nal , with 24 b3, then .i.b2, when given
the chance. Black played 24 tLlf6
(24 ... tLlC5 i s also promising, and if 2 5
.i.b2 tLlxa4), and then came 2 5 g3 ! ?
O n other m oves, he will soon win a
pawn . I was h oping that Black would
not be able to find a winning plan after
25 J1e1 26 C2 h1 27 c8+ Wg7 28



Tes t S e ve n

However, it turns out that Black can

still win with 32 ... h S ! ! .
If then 3 3 d2 ??, Black win s with a
cross-pin after 3 3 ... h4+ 34 'it>h 2 xf2.
If in stead, 3 3 as h4+ 34 Wf3 d1+
35 'iWxd1 1:Ixd1 36 Wg4 1:IxdS 37 'it>xh4
.l:!.xas, and Black will eventually grind
away for a win in the pawn-up rook

ending. Even the inoffen sive a-pawn

ends up in a zug zwang trap.
Or if 33 h4 as, and White's pawn s
are now blocked, and h e will lose to
zug zwang in a king and pawn ending
after 34 'iVd2 xf2+ 35 xf2 lhf2 36
Wxf2 Wf6 3 7 Wf3 WfS 3 8 g 3 f6 3 9 'it>e3
gS 40 fxg s fxg s 41 hxg s wxg s 42 Wf3
WfS, and a pawn drops. A long varia
tion, but al so elementary.
Thi s leaves 3 3 'it>f3 . Instead of look
ing for complications with another
zugzwang, there is a simple line with
33 ... 'iVd1+! 34 1:Ixd1 1:Ixd1, when a
pawn drops, and Black eventually win s
the rook endgame.
Thi s would h ave been an excellent
win for my opponent, except th at he
missed it.


Test E i g h t
8.1 Black to play

8.3 Black to play

A) 2 1 ... gS
B) 21 .. .fS
C) Something el se?

A) 22 ...'it>d7
B) 2 2 ...WfS
C) Something else?

8.2 Black to play

8.4 White to play

A) 2 1 ... as
B) 21 .. . a7
C) Something el se?

A) 2 2 ..ib4
B) 2 2 'iVg 3
C) Something else?


Tes t E ig h t

Test 8.1

British League (4NCL) 2006

Continuing from Test 2.1, White h as

just castled, and Black now castled
him self, 10 ...0-0. Both m oves seem a bit
too automatic, and indeed lazy. As dis
cussed earlier in this game, White
should try d5, or even dxC5, if given th e
chance, and Black should h ave avoided
this by exchanging on d4. I h ad the
higher Elo rating, and so in theory it i s
more likely that he would m ake a slight
mistake at some stage than I would.
Thi s edge is, however, slight if the posi
tion i s drawish, with little chance of
m aking mistakes. The stronger player
therefore h as the incentive to try to
keep the position lively, but it is easy to
overpress. Wh at h appen s, for example,
if the opponent does not m ake mis
takes? Quite often he will keep a slight
edge, when with best pl ay the position
will remain equal at best.
White him self makes another minor

in accuracy, with 11 e4. Clearly thi s does

not m ake his position worse, but the
point i s th at White still h ad prospects
of keeping an edge. 11 .l:Id1 ! ? seem s
promisin g . If then 11 ... cxd4 12 'iVxd4!,
and White has gained a clear tempo
when compared with ttd1 instead of
e4, the pawn push being of little rele
Bl ack, of course, has altern atives,
but after, for example, 11 ... d6, White is
now better placed for 12 N, and if
12 ... cxd4 13 tt'lxd4 tt'ld7 ? ! there follows
14 tt'lc6 and White wins a pawn . Black
h as to find something else, but he is
not equal .
In the game, I decided it was high
tim e to exch ange, with 11 cxd4 12
tt'lxd4 tt'lc6. Black is h appy now after 13
tt'lxc6 dxc6 14 'iVe2 'iVC7, and could
think of aiming for an edge, so White
develops with 13 iLe3. Then 13 .. J:tcB,
showing Black's intention is to focus on
White's slightly weakened 'English
pawn ' on c4.

After 14 .l:Iac1, I h ad several reasonable altern atives, many of these ending


Why we L o s e a t C h e s s
up a s equality. I tried 1 4 'ilVC7, telling
myself th at there was no reason to
move the knight just yet.

came lively, continuing with 15 ... CZJxd4

16 'ilVxd4 (maybe 16 iLxd4! ?, probably
equal) 16 ... .i.f6 17 'ilVd2. Now I saw the
glint of an initiative with 17 :e5, si
multaneously attacking two pawn s.

Even so, the immediate knight ex

ch ange, 14 .. .'Jxd4, and only then ...'ilVC7,
is a safer way of equalizin g .
The computer al so suggests 14 .. .f5,
which I admit I h ad not seriously con
sidered. After 1 5 exf5 CZJxd4 16 iLxd4
.l:i.xf5, Black h as too m any 'pawn is
l ands', and White is slightly better. So
m aybe I was correct in not analysing
deeper on this over the board.
My opponent too developed quietly,
with 15 l:rfdl, with equal pl ay. Perh aps
though he should h ave tried for a slight
edge with 1 5 CZJb5 'ifl>8. Then not 16
'ilVxd7 ?? l:rfd8, and the queen is
trapped, but 16 .l:i.fdl or 16 f4 might
still give White a slight push, provided
he can keep his pawns secure.
By now I h ad weighed my opponent
up, and decided th at he would h ave
been more th an h appy with a secure
draw. This gave me slightly more op
portunities to push for a slight break of
the symmetry. Before long, play be-


Then 18 i.d4 'ii'x e4 (if 18 ... 'iVxd4 19

'ilVxd4 iLxd4 20 l:rxd4 with a slight edge
for White, but 18 ... 'ilVg 5 can be con sid
ered) 19 i.xf6 gxf6. I felt reasonably
confident h ere, reasoning that it was
now up to White to prove that he could
keep the bal ance. He thought for a long
time, and I saw nothing m ore fearful
th an 20 'ilVxd7 l:rxC4 2 1 'iUxa7, and then
I saw 21 ... 'iUd4!, with a likely draw after
22 .l:i.bl 'iVc 5 2 3 b4 'ilVC7 24 'iUxC7 l:rxC7.
I h ave to admit that I h ad no sense
of real dang er, but 20 l:rC3 ! aimed for
checkmate, and forced m e to think
h ard. If, for example, 20 ... l:rxc4?? 2 1
l:rg 3+ 'it'h 8 2 2 'ilVh 6, which is the e n d of
g ame. 20 ... .l:i.c5 2 1 .l:i.g 3+ l:rg 5 2 2 .l:i.xg 5+
fxg 5 2 3 'ilVxg 5 + 'it'h 8 24 'ilVf6+ 'it'g 8 25 h3
'ilVxC4 26 l:rd4 'ilVcl+ 2 7 'it'h 2 'ilVC7+ 2 8 g3
al so win s for White.
I was briefly h orrified until I realized
th at the king could slip away with

Te s t E ig h t
20 Wh8!, avoiding the rook check. A
close escape.

White tried 21 xd7, and I was still

hoping to try to pl ay for a win .
In stead 2 1 lIg 3 .l:!.g 8 (the reason why
Black must m ove his king first) 2 2
xd7 could easily tran spose into 2 1
xd7 (see l ater), with, for example,
22 ... e2 23 !ttl .l:i.xg 3 24 h xg 3 .l:i.f8 2 5
xa7 xb2. Black's position i s slightly
uncomfortable, but it seem s to hold.
2 1 6 g 6 2 2 'iYxg 6 fxg 6 2 3 !txd7
.l:i.fd8 24 !txd8+ .l:i.xd8 25 Wf1 is a
straightforward draw if both players
pl ay sensibly.
In the g ame, Bl ack now jumped too
quickly for the attack. After 21 fS? (B),
there was more a sen se of relief, th at I
was not getting checkm ated, rather
th an the pleasure of wild attack. I h ad
not really thought about dropping
pawn s, m aybe on the basis that ' I was a
pawn up, everything goes up and eve
rything goes down, life goes on .' But 2 2
'Yi'xa7! i s n o w quite simply good.
White is simply a pawn ahead, and
a second i s about to drop.

There i s not much that needs to be

noted, except that 22 ... .l:i.xc4?? 23 e7
leads to a quick checkmate, and that
2 2 ... .l:!.fd8 2 3 .l:i.f1 leaves White m ore
than comfortable. If pawn s are not
necessarily the soul of chess, chewing
up your opponent's pawns provides
good nutrition .
Buckley instead tried 22 'Yi'e7?! es,
and the crisis h ad passed. Play was
sh arp, and I later h ad a strong attack,
without either player handling the po
sition particularly well. We resume th e
an alysis later in Test 11.3.
Earlier, Black should h ave tried
21. .l:!.g8! (A)



Why we L o s e a t C h e s s
2 2 :tg 3 ( 2 2 g 3 llgd8 ! ? 2 3 'iYxd8+
llxd8 24 llxd8+ 'itg7 25 :tcl e2 26 b4
'iVb2 27 llcdl 'iYxa3 should draw)
22 ... llxg 3 23 hxg 3 , with a position we
have discussed above. Then 23 ... llf8 24
xa7 'iYe2 25 llfl 'iYxb2 26 'iYe7 'itg 7
keeps the balance.

that it i s not all th at effective after ac

curate pl ay by Bl ack. Pl ay continued
16 ... axb4 17 exd6 .1i.xd6 18 llel b8 19
b5 .1i.xg3 20 hxg3 'iYb7.

Test 8.2

B u ry St Ed m u nds


Some quiet pl ay early on, with 1 d4

tbf6 2 tbf3 b6 3 e4 .1i.b7 4 tbe3 e6 5 a3
tbe4 6 tbxe4 .1i.xe4 7 .1i.f4 .1i.e7 8 e3 0-0 9
.1i.d3 .1i.xd 3 10 'iYxd 3 d6 11 0-0 tbd7 12
lladl e5 13 .1i.g3 Wie7, then 14 b4!? I
h ad expected 14 d5, and this seem s to
give a slight edge for White, after ei
ther 14 ... e5 1 5 e4, or 14 ... exd5 1 5 Wixd5 .
Pl ay continued with 1 4. . .exb4 15
axb4 a 5 .

Now Pert played 1 6 c 5 , which I was

impressed with at the time, and which
I awarded an excl amation m ark in my
notes just afterwards. The trouble i s


50 White recovers th e pawn, and

Black is stuck with an isol ated pawn . It

is difficult, however, to take the second
pawn, and if Bl ack can keep it, and pro
tect it with care, White will h ave to
watch out for the pawn being ad
vanced, and/or one of the open lines
for the black rooks being a threat.
White played 2 1 llbl. Now I took the
wrong plan , with 2 1 ... lla5?! (A) 22
'iYxb4 e4?!. The problem is th at Bl ack
is moving his pieces far away from the
weak isolated pawn, and he is going to
h ave to work h ard to cover the pawn .
Black cannot attack very h ard on the
king side, and so White soon keeps the
There are two basic methods of de
fending again st positional pressure.
One is to set up counterplay, where the
opponent is weak. The other i s to con
solidate, as firmly as possible ('overpro
tection'), so that the defender covers all

Te s t E ig h t
weaknesses on his own part of the
board, with later thoughts of his own
counterplay. On thi s occasion, since
White has no genuine weaknesses on
the kingside, Black should h ave over
protected on the queen side.
Therefore, (B) 21 ::071 (and indeed
there m ay be others, such as 21 .. JUc8
or 2 1.. . .l:!.a2) 2 2 .l:!.xb4 .l::tfa8, and Black's
pieces are tightly compacted, but with
ch ances of springing out again.

The knight protects the isol ated

pawn, the queen and rook cover th e
knight, and either the queen or rook
could bounce back into play, and none
of Bl ack's pieces is confined to passiv
Carrying on, with for example 2 3
cl al 24 .l:!.bc4 .l:!.8a5 2 5 c6 .l:!.xcl+
26 xcl xc6 2 7 xc6 al+ 2 8 'it>h 2
'it>f8, Bl ack can find simplification, and
even the better pawn structure. After
all, if Black can keep his passed b-pawn,
and White's king is a long way from the
queen side, it i s Black who h as the bet
ter ch ances. In addition, White's dou
bled pawns are often a slight weak-

ness, and m aybe something more sig

nificant. It would be rare in an end
g ame for pawn s on f2, g2 and h2 (or
h 3 ) to be less useful than pawn s on f2,
g 2 and g 3 .
Back i n the m ain line, and pl ay con
tinued with 2 3 e7 d 5 24 al tiJf6.

In my contemporary notes, I sug

gested that " Black has played actively,
and the position should probably be a
draw." Thi s now seems wildly optimis
tic. Compare this diagram and the test
position, and it is clear that Bl ack has
gone downhill. I now m ade several
slight mistakes before the time control,
and these all added up to a losing posi
tion . Then unexpectedly h e missed
something, in the quickpl ay finish, and
suddenly I h ad the chance of holding
the draw, but I was too short of time to
see this. So there are still some more
question s to follow in Test 11.4.
After pl aying through several lines,
it seem s th at Black is slightly worse,
and will h ave to work h ard and accu
rately to hold the balance.
After 2 5 tiJe5, I played the weak


Why we L o s e a t C h e s s
2 5 ...I;!.a5?!, and soon fell into difficul
ties after 26 I;!.xa 5 bxa 5 27 I;!. a 1, when
my pawn soon fell, and I h ad to scrab
ble around for any sort of compensa
tion . Possibly I could h ave tried to re
engage my pieces with 2S ... I;!.dd8, but
even here White is standing well after
26 I;!.fc1 4:JdS . Then after 27 'iVb7 ! White
has taken control of the seventh rank,
and this is why Black should h ave tried
21 ... I;!.a7 much earlier.

Test 8.3

B riti s h League (4NC L)


We h ave seen thi s position before in

Test 2 . 3 . Both players h ave pl ayed
slightly in accurately, but it is now
about equal .
I continued to develop with 14... iLd7,
which at first might look a slightly pas
sive set-up for the bishops, but White's
bishops h ave even less development
opportunity. Black needs to get his rooks
working together as soon as possible.


White continued his plan of attack

ing on the kingside, with 15 h4. I h ave
needed to con serve my energy in term s
of brain power. Translation : I pl ayed
l azily, because of my health at the time.
If there is a n atural and obvious move, I
would play quickly. Often thi s i s by no
means a bad idea, for any pl ayer, but it
i s important to h ave a sixth sense of
feeling that on certain occasion s the
pl ayer must slow down, whether it is a
sen se of danger, or whether, as here,
there is an advantage to be g ained
from finding the less obvious move.
Here I pl ayed 15 ... h6, pushing the at
tackers quickly away, and gaining some
time to try to develop on good squares.
I suspect th at in my younger days, I
would h ave preferred l S .. .fS ! ? 16 exf6
iLxf6, quickly opening up the centre,
and counterattacking . Both king s are
still in the centre, but the important
difference, in term s of trying to take
the initiative, is that Bl ack is able to
castle quickly, bringing the rooks into
the centre, whereas White's bishops
are blocking the rooks and king s .

Te s t E ig h t
Here 17 ttJxe6? d6 18 fS gxfs 19
S+ e7 favours Black, but 17 ttJe4
iLe7 18 hS leads to sh arp pl ay. Black
could continue either with h and-to
hand pawn fighting with 18 ... e S , or
consolidate his king side with 18 ... g 8.
Either seem s reasonable.
My notes at the time stated that:
"My regret i s not so much th at I did not
play this, m ore th at I did n ot even con
sider it." Probably the alternative move
is interesting, but not necessarily bet
The g ame continued with quiet po
sitional play, with 16 ttJe4 ..tc6. I sensed
that if I was trying to overprotect my
self by defending the pawn on c4, I
would lose m omentum with my pieces.
If now 17 iLxC4 ttJb6 18 iLd3 ttJxa4,
and Black m ay h ave a slight edge.
So in stead 17 'ife2, and now it is
time to protect the c-pawn with
17 b5. White simplified with 18 axb5
axb5 19 l::t x a8+ ..txa8 20 g3 h 5 .

th at for the time being, the emph asis

will be on the pieces. The knights on
both sides are on good squares, but I
can only take advantage of thi s after
careless play by the opponent. I was
more interested in term s of taking the
initiative, with the queen and bishops,
and m aking use of the a8-h l and a7-g 1
diagonals. What I did not like, though,
was the positioning of my kin g . If Black
were to be castled king side, White
could possibly regenerate a pawn at
tack, with g4. If my king did not move,
then obviously I could not bring the
rook into play. Thi s leaves the possibil
ity of bringing the king to the queen
side, probably safe enough, but al so
losing time.
First, both sides continue their de
velopment, with 2 1 iLg2 ..tc6 22 0-0,
reaching the quiz position, and now it
was time for me to decide what to do
with th e king .

2 2 0-0 ( C } keeps the king away

from the open queenside, but after 2 3
g4 hxg4 2 4 xg4 g7, followed by
... l::th 8, Black h as partially covered
. . .

All the pawn s o n both sides are well

covered, and there is unlikely to be a
pawn rush on either side. This suggests


Why we L o s e a t C h e s s
White's attack. White can, of course, try
to continue the attack with 2 5 h 5 .l:th 8
2 6 hxg6 l:th4 2 7 'iWg4 fxg 6 28 lbg 5, and
h e i s better. This is the sort of position
that I was worried about, and decided
to avoid.
So it is a simple king move, but
22 ri;j8! (B) is best, and then de
fending with ... ri;g 7.

pl ayed 2 3 ... l:ta8, still with equality, but

in stead continued wastin g time with
the king, with 23 ri;c8?, and then 24
'it'h2 'it'b7?! (m aybe 24 ... .l:!.d8 ! ?, although
White is now better). I h ad completely
missed the force of 2 5 b 3 ! cxb3.

Black h as now gained a full tempo

for manoeuvre when compared with
... 0-0, ... ri;g 7 and ... .l:th 8, and is safe, with
ch ances of an edge. If, for example, 2 3
l:tdl ri;g 7 24 i. e 3 lbxe3 2 5 "iVxe3 .l:td8,
and Bl ack is h appy, although at the
moment it i s too early to claim an edge.
My text move, 22 ...d7?! (A), i s bad,
and it gets even worse when Black con
tinues with the king m anoeuvre to b7.
It i s not so much that the king's final
square is necessarily worse, but rather
th at Bl ack has foregone two decent
m oves with other pieces. At the end of
his plan, White ends up with a signifi
cant advantage.
After 2 3 l:tdl, Black could h ave


Now h e tried 26 C4?!, and it looked

highly effective, but the quieter 26
2 ! ?, with a positional grind, seem s
better in the longe term, with 2 6 ... .l:i.a8
27 "iVxb3 .l:!.a4 (otherwise c4 for White)
28 lbg 5 i.xg 5 29 hxg 5, and Black will
h ave to work h ard to hold the bal ance.
Play continued with 26 ... bxC4 27
xc4, and now Black's king i s exposed
to the pieces.

Tes t E ig h t
I h ad to defend the b-pawn with
27 JWb6 for as long as possible, other
wise my position would collapse. I was
hoping th at m aybe I h ad chances of
scrambling for a draw. It was now
White, rather than Black, who started
to take the initiative on the long di
agonal, with 28 ttJC3 .l:.d8 (if 28 ... b2?,
then simply 29 xb2).
Thi s i s a typical sort of frustration
position for the attacker. There are
many obvious choices to play for a win,
but nothing quite seem s to work, with
the position fizzling out to a draw on
best play. Wall goes for the apparently
deci sive fin al effort - and loses!

White must remember h e i s a pawn

down, and so he must move quickly. 29
J::!. d 3 ! ? leads to tactical play, but soon
ends up in a drawn endgame after
29 ... b2 30 xb2 'ilVxb2 3 1 ttJxdS exdS 3 2
J::!.b 3+ W C 7 3 3 'ilVxc6+ Wxc6 34 .l:.xb2.
Black cannot afford to lose the isol ated
d-pawn, but White i s unable to force
the pieces away. An opposite-col oured
bishop endg ame with rooks.
There i s no obvious improvement,

so thi s position seem s at least pl ayable

for Black. Now 29 ttJxd S exd s led to a
critical position.

White can try 30 .l:.xdS, and if

30 ... b2?? 31 .l:.b S ! . Entertaining vi sual
chess. Instead, 30 ... xdS 3 1 i.. x dS+
J::i. x dS 3 2 'ilVxdS+ Wc8 33 'ilVxf7 b2 34
'ilVC4+ 'it>b7 3S 'ilVe4+ Wb8 3 6 xb2+
'ilVxb2 3 7 'it>h 3, and various assorted
checking lines, leaves White under
pressure, with only two pawn s versus
the bishop. Maybe White can still hold
on, or m aybe not, but it is difficult.
The best line is the ultra-positional
30 'ilVd3 ! ! , which probably neither
player would h ave con sidered over the
board, ignoring Bl ack's two isolated
passed pawns. White's pieces are now
far more centralized than Black's, and
his extra kingside pawn is much more
effective th an Black's pawn s, even
though he is a pawn down. Black would
be unable to keep both isolated pawn s
together, and when one drops, ' White
would still keep the positional advan
tage. Black would suffer the problem of
th e isolated pawn, while White would


Why we L o s e a t C h e s s
still b e able t o take advantage of his
extra king side pawn, with e6 and/or fS
pawn breaks.
In stead, Wall blundered with 30
ii.xd 5??, and now 30 .. J:txd 5 ! 3 1 1::tx d 5
b2, winning for Black (and indeed
31 ... 'ii'f2+ 32 c,t>h 3 'ii'f3 is even quicker).
The most likely explan ation for White's
blunder, perh aps, was that he saw both
of his captures on dS as leading to a
likely advantage for him after a double
exch ange (indeed a triple-exch ange),
but missed that in playing the wrong
move order, his opponent h ad the op
portunity of an intermediate m ove. A
common blunder, which we h ave seen
many times in this volume. Here Wall
loses the g ame.
After 3 2 ii.xb2 'ii'x b2+ 3 3 c,t>h3 'ii' b l
34 c,t>h2 'ii'e l, White's rook cannot
move, and Bl ack can capture the ex
ch ange at leisure, staying the exch ange
up. On 3 5 'ii' b 3+ c,t>C7 3 6 'ii'd l 'ii'f 2+ 3 7
c,t>h3 ii.xd 5 38 'ii'x d 5 'ii'g l I knew that I
would be given a long string of checks,
but I felt I would survive the perpetual
and win .

So it proved, although naturally it

took a long time for Black to take com
plete contro1. Play eventually finished
with 39 'ii'c4+ c,t>b6 40 'ii' b 3+ c,t>a5
(40 ... c,t>cS i s more efficient) 41 'ii'C 3+
c,t>b5 42 'ii'd 3+ c,t>a5 43 'ii'd 2+ ii.b4 44
'ii'd 8+ c,t>a4 45 'ii'd 7+ c,t>a3 46 'ii'd 3+ c,t>b2
47 'ii'e 2+ c,t>c3 48 'iWf3+ c,t>d4 49 'iWc6
'ii'fl+ 50 'it'h2 'ii'e 2+ 51 c,t>h3 'ii'g 4+ 52
c,t>h2 'ii'e 6 5 3 'ii'c 2 'ii'C4 54 'ii'f 2+ d 5 5 5
'ii'f 3+ 'it'e6 56 'ii' b 7 ii.a 5 57 'ii' b l 'it'e7 58
'ii' b 7+ 'ii'C 7 59 'ii'a 6 ii.b6 60 'ii'a 8 'ii'C 5 61
c,t>h3 'ii'g l (now Black's king finds
breathing space on g7) 62 'iWf3 c,t>f8 63
g4 hxg4+ 64 'ii'x g4 'ii'f l+ 6 5 c,t>h2 ii.e3

Test 8.4

Thames Va l ley League


Following on from Test 3 . 1, this

tim e the puzzle is to find the winning
idea for my opponent. I h ave pl ayed
dreadfully so far, and he should be able
to find good ch ances of a substanti al


Te s t E ig h t
advantage just with good and logical
pl ay. There might well be a forced win
somewhere, though, without any
chance of Black wriggling out.
Nurmoh amed immediately pl ayed
for a direct attack again st my king,
with 2 2 .tb4?! (A), and it looks ex
tremely threatening, but Black was
able to squeeze out with 22 :a4! 2 3
ttJd6+ f8, and now White h as no time
for a winning discovered attack, as
White's bishop is threatened.

We return to thi s later i n Test 10. 2.

White i s clearly not worse, but as pl ay
continued, any slight mi stake by White
could easily turn m atters around.
Perh aps the easiest way of demon
strating how Bl ack h as gone wrong
over the l ast few moves i s by showing
wh at i s h appening after the best move,
22 g3! (B).
The queen and bishop are no longer
tied down, allowing them to maintain
maximum flexibility. Black's king i s
stuck in the centre, whereas o n g 8
rather th an e8 the king i s much safer. It
i s not just solely the question of the

After 24 .tC5 'Yi'c6 2 5 'iVe3 g8 Black

has covered the immediate danger,
and was even able to win, after some
ten se pl ay.
Wh at would finally decide after 2 2
'Yi'g 3 i s that Black cannot protect his g 7 pawn . If White invades there, the rest
of Black's position crumbles. 22 ... g6 2 3
ttJxb2 axb2 24 e s, for example, takes
complete control for White on the b2h 8 and b4-f8 diagonals, with the bl ack
king being stuck in the middle. Or
22 ... l:1.g8 23 eS ttJdS 24 .txh 7, and
White joins in on a different diagonal.
Let us return to the main line, and


Why we L o s e a t C h e s s
the diagram after 2 S .. :>t>g8.
Then he played 26 es?, natural
enough, but I had expected 26 .i.bS !
'Va8 and now 2 7 eS, which i s more ac
curate timing. White then doesn 't h ave
to worry about the bishop on c S .

Bl ack has a coupl e o f altern atives

h ere, but neither holds the balance.
If 2 7 . ..t2Jds, the move I was thinking
of, then 2 8 'Vf3 fS 29 exf6 .i.xf6, but
now 30 J::t C 1 ! , and Black is effectively in
zugzwan g . Or 27 ... tLJg4 28 'Vg s h S 29
'Ve7 ! (29 tLJxf7? ! i s good for entertain-


m ent, but after 29 ... .i.a6 30 tLJxh 8 xbS

3 1 tLJg6 .i.xfl 3 2 tLJe7+ ..t>h 8 3 3 'Vxh S+
tLJh 6 34 'iit xfl e4 Black holds)
29 ... tLJxes 30 f4 'VdS 31 fxe s 'VxcS+ 3 2
'iith 1, and White h a s a clearly winning
It could be argued that since White
was winning anyway after either 22
.i.b4? ! or 22 g 3, why should the anno
tator be criticizing either m ove? In the
long run, Bl ack i s dead. The question is
really one of practical aesthetics. The
pl ayer with a winning advantage needs
to convert it into a win as quickly and
clearly as possible. If the attacker
messes around, Black can mess around
himself, muddying the waters. Then
there might be chances of counterplay.
In practical term s, Black would lose
very quickly after 22 g 3, and would
h ave slight practical chances after 26
.i.b s, if somehow White miscalculates,
but in the g ame, with 26 e s ?, the posi
tion was suddenly messy.

Test N i n e
9.1 White to play

9 .3 White to play

A) 24 llb 5
8) 24 'iWe2
C) Something el se?

A) 2 8 tiJd3
8) 28 tiJc4
C) Something el se?

9 .2 White to play

9 .4 Black to play

A) 2 5 e 5
8 ) 2 5 .l:!.f1
C) Something else?

A) 2 5 ... ..txh l
8) 2 5 ... C2
C) Something el se?

10 1

Why we L o s e a t C h e s s

Test 9.1

London League 2006

There is plenty of sh arp and un
usual play in this opening, and inevita
bly both pl ayers make mi stakes, espe
cially given that we were pl aying an
evening game in the London League,
where time limits are quick. Quite pos
sibly a deep theoretical an alysis m ay
pinpoint a suggested improvement in
the opening part of the g ame. There
was always a suspicion in thi s g ame,
after all, th at in a complicated line
White was quite often in danger of
ending up worse, and this should not
happen . It i s Black, not White, who has
to work h ard to equalize.
I do not wish to concentrate on
Nimzo-Indian opening theory with
much depth, trying to assess sub
variation s in move A12 1, or wh atever,
on move 8 or 11 or 14. Thi s book is con
centrating far more on searching for
any clear mistakes th at players should
be capable of avoiding over the board.
The more interesting question, from
our point of view, is not opening the
ory, but rather, what mistakes are
made over the board, and h ow such
mistakes can be eradicated.
We start with a Nimzo-Indian, with
1 d4 tiJf6 2 c4 e6 3 tiJC3 .i.b4 4 e3 c5 5
.i.d3 .i.xC3+ 6 bXc3 d6 7 tiJe2 e5 8 0-0 e4
9 .i.c2 .i.e6.
Cox was playing quickly, so I was

10 2

sure th at he h ad experience in this po

sition. In contrast, this was new to me.
Clearly h e was deliberately provoking
my dS push, and I decided th at here I
should be provoked, in return setting
up counterpl ay on e4.

So, 10 d5 .i.f5, and h ere I immedi

ately played 11 f3 !? My opponent sug
gested afterwards th at White could
effectively g ain a tempo with 11 tiJg 3 ! ?
.i.g6 12 f3 exf3 1 3 gxf3 . This would cer
tainly be so if Bl ack were to exch ange
on c2, but if h e were to wait, h e can try
som ething different, maybe 13 .. :iVaS ! ?
I n the g ame, Bl ack played 1 1... exf3.
My original intention was 12 .i.xfS fxe2
1 3 xe2, with a good bishop-pair except th at when I looked more closely,
I started to wonder wh at I was going to
do with the dark-squared bishop, and
al so how White could prevent Bl ack
from moving the knight to a strong
square on e S . So I ch anged my mind,
preferring to keep the pawn structure
more flexible with 12 gxf3 .i.xc2 13
xc2 tiJbd7.
We h ave very close to an equal posi-

Te s t N i n e
tion . I like the extra pawn in the centre,
but I am worried about the doubled
isolated pawns.

I am not too worried about the

slight exposure of White's king, but I
am not quite sure wh at to do with the
bishop. I played 14 tiJg3, which might
or might not be fully accurate. I do not
know. The computer suggests th at 14
e4, 14 l:!.bl and 14 3 are al so equal .
Any of these might al so be equally
good, very slightly better, or very
slightly worse. Probably, as we sh all
soon see, 14 e4 would lead to a tran s
position .
For myself, it h ardly seem s worth
while examining extremely subtle dif
ferences when analysing afterwards.
Most players will tend to g ain more
from looking for more basic mistakes,
and trying if possible to cut down on
simil ar types of error.
If I were to reach thi s position again,
I suspect that I might well be thinking
of 14 l:!.bl ! ? in stead, forcing Bl ack to
con sider wh at to do with his b-pawn .
Thi s i s a reasonable possibility, not

necessarily a clear improvement.

Black tried 14 ... g6, to cover any tiJfS
attack. Then 15 e4 e7. Of course there
could be a tran sposition with 14 e4
e7 15 tiJg 3 g 6, and thi s in m any ways
looks the more n atural way of playing
it, but it would end up the same.
White now h as an obvious way of
taking the initiative with 16 f4. This
seemed so obvious and natural a way
of playing such a move, that I did not
con sider any alternatives. The 4V2
pawn push in the centre looked good,
provided I do not create too m any extra
weaknesses for my king .
A reality check suggests that, ac
cording to the computer, there are at
least a dozen alternatives, probably at
least equal, and in some cases even
better. 16 l:!.bl ! is suggested as a slight
improvement, forcing Black to decide
wh at to do with his b-pawn . If 16 ... tiJb6,
then 17 a4! , and if 17 ... tiJxC4?! 18 3
tiJb6 19 as, with advantage to White.
Maybe at about the fourth attempt of
an alysing this position, this seem s the
best, with ch ances of keeping an edge.
Backtracking a couple of moves, maybe
Bl ack could h ave developed more sol
idly with ... C7, rather th an ... g 6 and
... e7. Who knows?
16 . . 0-0-0!? There are several other
choices, maybe castling on the oth er
side, with 16 ... 0-0.
Now there is an obvious possibility
of the centre blowing up with a pawn
sacrifice on e S . Or maybe White can


Why we L o s e a t C h e s s
hold the balance, perh aps only tempo
rarily if he has a better occasion to
break with e S .

I did not h ave t o think about thi s

position too h ard. 17 e s ! is a highly
thematic central pawn sacrifice, in
such lines as the Modern Benoni and
the Four Pawn s Attack in the King's
Indian . White gives up a pawn, win
n i n g Black's d-pawn, but losing his e
and f-pawn s, quickly bringing h i s
bishop and rook into active play, and
setting up a strong passed d-pawn .
Quite often there is a question of tim
i n g i n such pawn explosion s. I f you
don 't sacrifice or g ambit immediately,
it is often best not to do so at all . I did
not like the idea of Black restraining
White's e-pawn with ... J::i. d eS or .. .lUeS
or ... ctJg4; and earlier, the only sen sible
continuation to my f4 push was an e S
push . So i t h a s t o b e done immediately.
Several months later, I wondered
whether I could have improved my
game, by developing with 17 l1b1. The
idea is that when Black too plays a quiet
move, White's rook is now on an active


file, and he can play eS with an extra

attacking piece in play. This seems to
make sense until one realizes that if
Black can prevent or reduce the impact
of eS, White would then be under pres
sure. Therefore 17 ... ctJh S ! , a useful move
in the game as well. Whether White ex
changes on h S , allowing Black the open
g-file, or whether he allows the ex
change on g3, White is under positional
pressure. If he cannot find any good
gambit play on eS, his e4-pawn is weak,
once the knights have gone.
So again, White h as to gambit im
mediately, or not at all .
Back t o the sacrifice, and 17 ... dxes
18 fxes.

Now the captures on e S seem far

too dangerous, and I expected a knight
move. Indeed, lS .. :iWxes ? gives White
far too big an attack after 19 ..tf4 fie7
20 fia4. We see an example of the '4V2
pawn attack', the standard four pawn
attack, but with a fifth doubled pawn
on c3. Thi s pawn covers the d4-square
again st the queen, m aking fewer es
cape routes from e S .

Te s t N i n e
Here I was expecting 18 ... lDg4, to
which I planned to play 19 lDe4 lDgxe s
(19 ... lDdxeS 20 g s f6 2 1 xf6 lDxf6 2 2
.l:txf6 tran sposes) 20 g s f6 2 1 xf6
lDxf6 2 2 .l:txf6. I assumed this was go
ing to be slightly better for White, al
though now it seem s th at it i s only
equal . For example, 22 ... .l:thf8 23 .l:Iafl
l::t xf6 24 .l:txf6 .l:tf8 2 S d6 d8 26 l::t xf8
'iYxf8 2 7 'iYa4 lDf3+ 28 cj;>g 2 lDh4+ 29
cj;>g l, and Bl ack would be advised to
take the draw. 19 f4! ? is also to be
In stead, Cox went for counterattack
with 1S ... lDh S ! ?

Now 19 lDxh S gxh s 20 i.f4 lDxes 2 1

fS+ lDd7 2 2 .l:Iae1 i s probably equal,
for example, with 2 2 ... l::th g8+ 2 3 cj;>h 1
"iVh4 24 .l:te4 'if6 2 S 'ixf6 lDxf6 2 6 l::t e 2.
I chose the m ore aggressive idea, 19
White keeps a slight edge after
19 ... lDxe s 20 g s f6 21 .l:txf6. Black
tried in stead 19 ... xes?!, when he ad
mitted afterwards th at he h ad mi ssed
my next move, 20 lDgS !.
Thi s i s the high point of my attack. I

win the exch ange because of the un

stoppable knight threat on f7. Al so,
take note of the position of the bl ack
queen, on an apparently open board.

The only way the queen can move i s

backwards, th anks in part t o White's
doubled pawn s. Black would, of course,
welcome the chance of a queen check
from g4, but there i s no way to attack
th at square, even in two or more
moves. White's pieces and pawn s are
well coordinated.
Cox thought for a long time, and
decided th at while the rook h as to go,
he needs to keep the extra pawn, with
20 ... fS 2 1 lDf7 f6 22 lDxhS .l:IxhS.


Why we L o s e a t Ch e s s
W e have just gone through a phase
of sacrifice and counter-sacrifice, with
the initiative swinging from one side to
th e other. White has sacrificed a pawn,
in return obtaining good piece activity;
Black has returned the exch ange,
eliminating White's more attacking
pieces. Now Bl ack has the more effec
tive piece and pawn form ation, even
though he is down in m aterial . White is
not under immediate attack on the
kingside, but h e h as to be careful about
wh at might h appen in the future.
I played the natural developing
move, 2 3 J::i. b l!?, trying to create pres
sure on the queenside. I underesti
mated his reply, 2 3 :iVa6!, when Black
has covered all his queen side pieces,
and threaten s to take the critical C4pawn, and with all likelihood, White's
open lines in the centre and king side.

At the time, I felt no real option

other th an to block the c4-a6 diagonal
with 24 J::i. b S (A), which at least
squeezes Black's queen to the edge, but
it looked far from convincing.
If at all possible, the defender needs


to centralize, covering king side, queen

side and centre. Indeed, 24 'ifie2! (B) is
best. Black's queen i s on the edge, so
White can create counterpl ay on the
rest of the board. I was worried that
Black was about to g ain time by attack
ing the queen, with 24 ... 4:Jhf6, but 2 5
.l:i.e1 ! covers the immediate threat. If
25 ... J::i. e 8??, then obviously 26 "iVxe8+.
White is fighting the pressure, not
merely defending.
Instead, if 2 5 ... 4:Je4 2 6 J::i.b 3 (there
are alternatives, probably ending up
level) 2 6 .. .tt:Jb6 2 7 J::i. a 3 "ivxc4 28 "iVxC4
4:JxC4 29 J::i. x a7 b8 30 .l:i.a4 b5 31 ..if4+
b7 3 2 .l:tb1 4:Jb6 3 3 .l:i.xe4 fxe4 34 .l:txb5
J::i. c 8 35 J::i.b 3, with a likely draw.
Should we be disappointed that
White cannot find an edg e ? Probably
not. The position is starting off as finely
bal anced, and if both players keep the
delicate bal ance, then the proper result
should be a draw. It is not too difficult
to appreciate that in my an alysi s later, I
h ave tried several possible alternatives,
either to keep an edge for White, or if
not, then clear equality. This is the best
line I h ave tried, and it keeps play sim
Probably we should add that a move
earlier (omitting 23 J::i.b 1 "iVa6), 2 3
"iVe2 ! ? i s certainly a choice for White.
Possibly Bl ack could try 23 ... 4:Je s 24
.lte3 b 6 ! ?, followed by a l ater .. .f4, with
reasonable compen sation for the ex
change sacrifice.
We now return to the m ain line,
and 24 J::i. b S?! J::i. e S.

Te s t N i n e

It i s starting to be difficult for White

to coordinate his pieces, especially as
Bl ack is intending to play .. J:!.e4. I de
cided it was time to set up a counter
exch ange sacrifice with 25 iYf2, and a
25 ... ttJhf6?!, which is natural and logi
cal enough, centralizing a knight on
the edge.

Even so, inaccuracies are starting to

creep in on both sides, in the l ast ten
moves before the tim e control i s
2S ... .l:i.e4! would h ave been highly ef
fective, taking control of the fifth rank,
with an attack of the pawn on c4, a

possibly nasty check on g4, and also an

effective defence again st i..f4. The
rook's remit goes further after 2 6
.l:i.xcS+ ttJxcS 2 7 'iYxcS+ d7 2 8 iYbS+?
(28 i.a3 iYb6 is better, but Black i s bet
ter) 28 ... 'iYxbs 29 cxb S .l:i.a4, cutting
open White's pawn s.
Then 26 i..f4!? was a difficult deci
sion, and, of course, tim e was getting
short. The bishop i s on a good diagonal,
but it can al so be threatened and
kicked out.
The alternative is an immediate ex
ch ange sacrifice, with 26 .l:i.xcS+ ttJxcS
27 xcS+ 'it>d7 ! (rather than 27 ... b8
28 i..f4+ 'it>a8 29 C7, when White is at
least equal, and quite likely somewh at

If now 28 i..f4, we see a good reason

why Black has del ayed ... .l:i.e4. He plays
28 ... .l:i.c8, knocking out White's queen
and bishop attack, winning a doubled
pawn . White can try in stead 28 i.. a 3,
but he still has not got a genuine attack, and after 28 ... ttJg4 Black is better,
with thoughts of ... ttJe3 or ... ttJe s . Before too long, White will notice the lack


Why we L o s e a t C h e s s
of the g-pawn, whether i n a middle
game, or, with the passed f-pawn com
ing through, the endgame.
After the text move, Black blocked
off White's queen side attack by 26 ... b6,
avoiding any sacrifice on c 5 . 26 ... tDe4?
would h ave been too combative, and
would lose after 27 Itel g 5 ? 28 ':xe4
Itxe4 29 .i:!.xC5+ tDxC5 30 'iVxc5+ <t>d7 3 1
'iVC7+ e8 3 2 'iVc8+ f7 3 3 'iVxf5+.
White pl ayed 27 .tg3, with the
none-too-subtle thought of placing the
queen in front on the diagonal, and
trying to give m ate.
I did not like the idea of giving up a
few pawns with 2 7 Itel Itxel+ 2 8 'iVxel
'iVxa2 29 'iVe6 'ivxc4 30 'iVc6+ <t>d8.
There might be a coupl e more checks,
but then wh at? If, for example, 3 1
.tC7+ e8 3 2 'iVd6+ f8 3 3 .td6+ g 7
34 'iVe7+ h 6, and White run s out of
checks. Al so, if 3 1 d6 e8, and again
the king escapes.

In the g ame, Black forces the re

moval of White's queen, with 27 tDe4,
but where should the queen move to?
Thi s is Test 1 2 . 1


Test 9.2

H i l l i ngdon League 2006

As we saw in Test 4.2, my opponent

h ad just missed a winning opportunity,
which I only bel atedly saw myself. We
are now back into norm al pl ay, with 17
.te2, and I was slightly better, in term s
of technical chess analysi s. The psy
chology of the g ame was highly against
me, though, as I was ultra-nervous
throughout the rest of the g ame, wor
ried that every move I h ad to think
about was a blunder, and th at every
move, good or bad, was on an eggshell .
I was able to convince myself, though,
th at the first move was fine.
He played norm ally with 17 ... h6,
then came a few reasonable moves,
with 18 g4 .tg6 19 g5 hxg5 20 tDxg5
tDh7 2 1 .txe4 'iVf6. Clearly we are now
in a ph ase of tactics. I h ave won a
pawn, but it will take tim e to con soli
date, given the openness of the centre,
and the ten sion on the bl-h7 diagon al .

Te s t N i n e

In my notes just after the g ame, I

was critical of 2 2 f3, missing the idea of
his next move. 22 i.xg 6 ! xg 6+ 2 3 ttJe4
is certainly easier to pl ay for White.
Whether it is 'objectively' better, in the
sense that it i s the best outcom e after
the best m oves by both sides, is not so
G ait sacrificed the exch ange with
22 ...l::i.x e4! 23 fxe4 i.e7, creating confu
sion in the position . I am m aterial
ahead, and I h ave the bigger pawn cen
tre, but I now h ave problems with my
minor pieces, with my knight being
precariously pinned. It is possible too
th at one of White's central pawns will
drop, allowing Black to continue the
counterpl ay el sewhere.
The position is extremely difficult to
h andle, and I recognized this at the
time. My next move, 24 'it'al, seem s
correct, escaping any pin s and checks
on the light-squared diagonal, but my
follow-up move was a mistake.
In stead, 24 l::i.fl looks promising, but
after 24 .. :iYxd4 2 5 i.b4 i.xg 5 26 xh 7+
i.xh 7 2 7 l::i. x d4, Black keeps in touch

with an unexpected combination of

fork and skewer, with 2 7 ... C 5 ! 2 8 i.xc5
i.e3, keeping the m aterial bal ance.
Not, of course, 29 l:td5 ?, allowing a fork
with the other bishop with 29 ... i.xe4+.
There is al so another possible bishop
fork, with 24 .l:!.gl 'ilVxd4 25 i.b4 i.xg 5
26 'ilVxh 7+ 'it'xh 7 2 7 l:txd4 .ie3 ! , level .
Clearly it is not a straightforward win
after Black h as given up his rook.
After 24 ... ttJxg5, we h ave reached
the test position. The next move ended
up in the difference between an advan
tage and a quick loss.

I pl ayed 2 5 e5? (A), overlooking th at

after 2 5 ...'ilVf5, the geometry is again in
favour of Black after 2 6 l::i.fl 'ilVxh 3 ! 27
i.xg 5 'ilVxh4 28 'ilVxh4 i.xh4. Nominally,
Bl ack's m aterial advantage is slight
(two bishops versus rook and pawn),
but the two bishops are powerful, with
no good way of combating them, and
no doubt the passed g-pawn will even
tually push through .
S o I changed mid-stream with 2 6
i.xg5 i.xg5, when I a m worse, but with
still a chance of making a battl e. My

10 9

Why we L o s e a t C h e s s
next move was a blunder, though,
which forced me to resign quickly, as
we'll see in Test lOA.
Earlier, 25 '1J./l! (B) is much better,
with a clear advantage for White after
2 S .. :xd4 26 xg s xg s 27 "iiix g s
"iiix C4 28 e s ! (better th an the obvious
28 '1J.d8+). Black h as only a pawn in re
turn for the exch ange, and his bishop is
ineffective, while White's m ajor pieces,
plus a possible advance with the h
pawn, create great problems.
Again, after my earlier big mistake, I
got nervous, and made further mis
takes and blunders.

Test 93

Britis h League (4NCL)


Continuing from Test S.2, there fol

lowed some obvious development
plans with the next couple of moves,
with 20 '1J.c1 c6 21 '1J.hd1 .ixb2 22 'it'xb2.
His next move, 22 ... 'it'b8, i s again obvi
ous, but perh aps not quite accurate,

1 10

involving a slight loss of time. White's

knight on dS looks strong, but does not
immediately threaten anything, and so
there is no need to kick it out just now.
22 ... b 6 ! improves, with the idea of
... 'it'b7 (a more active square), and
... 4JcS . If White were to try 2 3 4Jb4, hit
ting the c6-pawn, then after 23 ... 4JcS,
Black h as clearly gained time in com
parison with the line in the game. This
pawn move may at first look anti
positional, weakening the queenside
pawn structure (especially the pawn on
c6), but sometimes pawn pushes can
create more th an compen sating
strength s. Black's king and two knights
work well together, with the h elp of the
two pawn s covering squares on the
fourth rank.
In the game, White h ad to move the
knight, 2 3 4Jb4, and Black played the
useful but less aggressive 2 3 ...4Jb6.
Looking on the computer afterwards, I
felt slightly mystified th at White is
given an edge h ere, whereas earlier on
Black is given the edge. Maybe the posi
tion i s in the end about equal, although
over the board I was startin g to get
worried, as his knights are moving to
good and dangerous squares, notably
with ... 4Ja4+. Black still, of course, has
his passed pawn, a worry if we reach
an endgame. I became over-cautious
over the next few moves, and thi s af
fected a critical choice a few moves on .
We go quickly towards the test posi
tion, with 24 'it'c2 'it'C7 25 4Jd 3 4JbC4 26
a4 4Jb6 and now 27 4Jb2, so that the

Tes t N i n e
of the g ame. The intention for White
here should h ave been to exch ange a
pair of knights.

knight covers both a4 and c4.

Then 27 lbd S . I was perh aps too

highly concerned with the idea of Bl ack
being able to reach b4 with a knight, so
I tried to cover with 28 lbd 3? (A), with
the h ope perh aps that he would finish
with a draw, with 2 8 ... lbb6 29 lbb2.
He was more than h appy, though,
to try to play for a win with 28 b6!,
with his knights and pawn s coordinat
ing well for both attack and defence.
We shall examine the rest of the game
later in Test 1 3 . 3 .
2 8 lbC4! (B) would h ave been much
better, even giving chances of a slight
edge for White. It i s well known th at a
pair of bishops will often work better
together than bishop and knight. In the
game, for example, White would h ave
enjoyed h aving the bishop-pair, cover
ing squares of both colours (imagine a
bishop on b2, for example, rather than
a knight). What is often less fully exam
ined is the knight-pair. Sometimes it
can be useful for the knights th at be
tween them th ey can press again st
squares of both colours, as in the rest


If now, for example, 28 ...lbxC4 29

xc4, the bishop h as good play,
whereas Black's knight accomplishes
little. Black might try in stead 28 ... lbb4+
29 'it'b2 b6 30 lbxas bxas 31 ii.C4 ttxd1
32 ttxd1 tte7, but he has lost coordina
tion with his queenside pawn s, and
White has the better minor piece.
White in fact h as a slight edge, because
of Bl ack's loss of time on move 2 2 .
There are, of course, possible alter
natives for Black, but it is not easy to
find quick equality.
At the time, I asked myself, "Why
did I reject my simple and direct move?
Partly because his knight on as was so
much on a limb that it looked crazy to
try to exch ange it. But his knight is use
ful in th at White always needs to cover
the b3- and c4-squares. Exch ange th at
knight, use the light-squared bishop
again, and White can start becoming
active. Or at the very least he is not
hoping for repetition ."


Why we L o s e a t C h e s s

Test 9.4

London League 2006

was still feeling very unwell,
slightly less th an two years after my
stroke. I needed both physical exercise
and mental stimulation to try to bring
myself back to life.
For the London League m atches th at
year, I generally walked three miles to
Edgware to catch some exercise, took a
bus into town and then walked a fur
ther mile or so to the Briti sh Museum .
On the bus, I would catch up on some
Sudoku puzzles. The bus journey from
Edgware to Baker Street was about the
right length of time to complete the
exercise, although occasionally I would
make a mistake, based on an incorrect
assumption . This, of course, was frus
On th at day, I noticed I could not
concentrate on the puzzle, suggesting
perh aps that I was unlikely to be on
form in my chess that evening. I am
reason ably h appy with my early moves,
but sure enough I made a couple of
ridiculous blunders later on .
Pl ay started with 1 e4 c5 2 ct:Jf3 e6 3
d4 cxd4 4 ct:Jxd4 ct:Jf6 5 ct:JC3 d6 6 g4 h6 7
h4 a6 8 g2 g6 9 g5 hxg5 10 xg5 e7
11 d2 ct:Jbd7 12 0-0-0 e5 13 ct:Jb3 b5 14
f3 b7 15 ..t>bl b4 16 ct:Je2 as 17 h3
l:!.a6, when Bl ack was doing fine,
though I h ave yet to prove an advan


White pl ayed 18 C3?, a common po

sitional error - and I am sure I h ave
done thi s myself m any times. Black is
intending to rush his queen side pawns
quickly, and th at can indeed be fright
ening. White has to remember that
Black is not thinking of promoting his
pawn s, but rather he i s aiming for
pawn-versus-pawn contact in front of
the king, aiming to open up lines. Un
fortunately, White's pawn push accel
erates the battl e, in favour of Black.
18 xd7+! would h ave been better,

and if 18 .. :Vxd7 ? 19 xf6 xf6 20

ct:JcS, winning the exchange, but
18 ... ct:Jxd7 19 l:!.hg l is about equal .

Te s t N i n e
I h ad the luxury of setting up a mi
nor piece sacrifice in order to break up
White's centre pawn s, with 18 ... a4 19
lbbc1 lbxe4 20 fxe4 iLxe4+ 21 'it>a1

My original intention was naturally

to capture the rook, 2S i..xhl (A), with
thoughts of a significant plus. But my
eyes were dim , and I could only see
with h alf of one eye, and occasion ally
do not 'see' things in my mind. Proba
bly the biggest single problem for me
when starting to resume playing chess
after my stroke was that I was unable
to see the whole board, both visually
and metaphorically. Occasionally I saw
ph antom s, thinking m aybe th at one or
two of my pieces or pawn s were on
different squares, rather than where
they were actually at. I could not see
the whole board, and because of re
striction s of my vision, I could not see
the bottom right-h and square. I h ad to
rely on memory.
My biggest blind spot was on the
bottom right-h and corner, and here of
course my queen on a8 was absolutely
critical to my calculation s. In earlier
years, I would h ave been able to see the
position as a whole. In 2006, though, all
I could do was remember th at my
queen was on a8, simultaneously cov..

Black's attack i s massive. 22 cxb4 i s

t h e only move. In stead, 2 2 .i.xe7 lbb3+
2 3 lbxb3 axb3 24 a3 'iVa8 would h ave
forced m ate on the a-file. The problem
of White's c3 push is not just th at Black
has opened up the ... b4xC3 pawn ex
ch ange, although in some cases it
might h ave proved useful . The much
bigger problem is th at White has badly
weakened his b3-square, allowing po
tenti ally winning tacti cs with ... lbb3+.
So I continued with 22 lbb3+ 23
lbxb3 axb3 24 a 3 'iVa8, with a m ating
threat on a3, and al so the threat of
winning the rook on h i . Then White
has to play 25 'iVC1.
It was di sconcerting that Black was
not already winning, and th at I h ad
given up a piece for a couple of pawn s
and an attack. For complete security, I
would want to give checkm ate, or at
least to gain material .

1 13

Why we L o s e a t C h e s s
ering both the h l-square and the a-file.
It was difficult at the time to cover both
the long diagonal plus a-file, as well as
what was going on in the rest of the
board. Indeed, I was un able to visualize
the whole board. I could see part of the
board, and such visualization is, of
course, extremely important for chess
pl ayers. I was unable as yet to find
ways to visualize and see different
parts of the board, but later I was able
to learn how to improve on this.
What I 'saw' was 2 S ... ..txh l 2 6 'iWcS+
'iWxcs 2 7 l:!.xcS+ ..tdS 2S .l:i.xdS, m ate.
Except this is a complete illusion . He
cannot move his rook to cS. He can
move his queen, 'iWcl-cS, but he cannot
move the rook as well, l:tCl-CS, in a sin
gle m ove. The queen and rook are not
doubled on the c-file. Logically this i s
absurd, since the queen i s firmly an
chored on cl, to protect the sacrifice on
a3, and there i s no rook coverage on
the c-file, just behind the queen .
Maybe for some readers thi s might
at first be of relevance only to partially
sighted players. Try though to think of
a complicated combination, when you
h ave to think of possibilities h alf a
dozen moves in advance, when som e of
the pieces on either side h ave m oved
on, and other pieces and pawn s h ave
disappeared. The player would not be
able to see the position in advance, as
the pieces are not yet there. What i s

1 14

important i s to try to 'vi sualize', to

imagine with cl arity wh at the position
is a few moves along the line.
My problem with my eyesight was
certainly a significant problem in this
g ame, but ultim ately the deci sive prob
lem is visualization, and thi s was some
thing I clearly h ad to work h ard on .
I played 2 S ..tc2?? (8), inexplicable,
until you know the full thought proc
ess. I felt I h ad to cover th e c-file, but
this was irrelevant and wasted pre
cious time.

In the g ame, White played 26 l:!.d2,

with advantage, and my bishop on c2
was not even very effective. Al as, there
were blunders on both sides before
White later won . We return to the
g am e later in Test 13.1.
The n atural 2 S ... ..txh l, as noted, is
good, while preparing this with 25 /6
(C} 2 6 ..te3 ..txh l 2 7 l:!.xh l dS might
even be fractionally better (avoiding
2S ... i.xh l 2 6 i.xe7 ! ?).

T e st Te n
10. 1 Black to play

10.3 Black to play

A) 2 S ... g 6
B) 2 S ... d7
C) Something el se?

A) 26 . . .4:Jg S
B) 2 6 ...d7
C) Something el se?

10.2 Black to play

10.4 White to play

A) 26 ... 4:Jg4
B) 26 ... 4:JdS
C) Something el se?

A) 2 7 g4
B) 2 7 g 3
C) Something else?

1 15

Why we L o s e a t C h e s s

Test 10.1

B ritis h League (4NCL)


As we observed in Test 6.1, White i s

clearly better, Black h aving queenside
weaknesses. He tried 2 3 Ite3?!. This
keeps an edge but I was more anxious
at the time about 23 e6 ! ?: for example,
23 .. J:te7 24 exf7+ Itfxf7 25 h4 g6 26
.l:txe7 .l:i.xe7 27 .l:tc3. It is not quite as
impressive as it earlier looked, though,
and White only h as a slight edge.
Maybe there i s another improve
ment? 23 tLlh4! looks better.

It might seem anti-positional at

first, placing both knights on the edges,
but the knights both threaten to move
into the centre. If Black decides later
that he has to pl ay ... b5, then the two
knights on c5 and f5 will be mighty.
Black can try 23 ... g6, but after 24 f4,
White i s starting a king side pawn
roller. White is much better.
I was still under pressure in the
game, and eventually hit on the plan of
2 3 .. Je7! 24 tLlh4 ..ic8!, covering the f5square. Bl ack would not, of course,
h ave h ad the tim e to defend in this way
if White h ad tried 2 3 tLlh4 a move ear
White pl ayed 25 f4 (not 25 .l:i.xc6 ??
..id7), and on my next move, I quickly
undid my previous defensive play.

2s ... g6? (A) was a serious conces

sion, and should only h ave been pl ayed
if absolutely necessary. Before too long,
I was being ground down, after 26 Itb3
bs 27 tLlcs tLlC7 28 Itf3 Wh7 29 fs gxfs
30 tLlb3. There were a couple of twists
to follow, though, which we sh all see
later in Test 14. 1.

1 16

Tes t Te n
Black should earlier h ave played
25 .i.d7! (B), in stead.

Thi s is a difficult line and Black still

has to defend carefully, but the alter
native is the danger of a position al loss.

Test 10.2
S Nurmohamed-C.Crouch

Tha mes Va l ley League

Somehow he h as unravelled his

pieces after an extremely unpromising
start. Just an extra h alf-tempo, a logical
but slightly l azy move by the opponent,
can sometimes make a big difference.
Play might continue with 26 f5. Then
Black can try active counterplay with
26 ... c5 27 tLlC3 tLlC7 (27 ... cxd4 28 tLlxd5)
28 tLle2 C4 (28 ... cxd4 29 tLlxd4 tLlb5 30
tLlhf3 with a slight edge for White) 29
J:::tg 3, and White's kingside pawn push is
more effective than Black's queenside
pawn push. Black's king is in danger
after a sacrifice on g 5 after, for example,
29 ... .l:!.ee8 30 f6 g5 31 tLlf3.
2 6 ... J:::t c 8! i s a tighter defence, with
the hope th at if White's king side initia
tive can eventually be slowed down,
Black can at some stage set up coun
terplay with ... c 5 . Black is obviously not
better, but it is not so clear that he is
substantially worse. There i s a possible
draw by repetition after 27 tLlc3 tLlC7 28
tLla4 tLla8, but White could still try for


The good news from Test 8.4 is that

now I h ave a reasonable position, hav
ing been in severe trouble. Over the
board, I felt I had reasonable chances,
especially if he made any further slips
in a complicated position . The com
puter analysis suggests that of the two
main lines for Black (26 ... tLld5 and
26 ... tLlg4, both hitting the queen on e3),
one is equal after best, and compli
cated. play, while the other is al so
compli cated but there is a win for
White. In a quick time limit, I would not
be expected to calculate the critical
positions in depth . It would always be a
case of positional sense, in stinct, and
about as much as you could an alyse
given the context of the ticking clock.

1 17

Why we L o s e a t C h e s s
S o which is better? I concentrated
on thinking of where the best counter
attacking and defensive move for the
knight is. I felt at the time that
26 ... ti'lg4?! (A) looked best, with the aim
of taking White's bishop on c S .
The alternative was 26. . .ti'ld5! (B),
aiming to take the knight to one of the
four central squares a move earlier.
This seem s to be a draw. I was almost
certainly worried about 27 g s g6 28
..Ite4!, and it looks like Black's king will
be swallowed after the central knight
has gone.

There are two saving resources for

Black, ending up as a draw. Either
28 ... h6 2 9 ..Itxds hxg s 30 iLxc6 iLxe s ! ,
and Bl ack h as time t o give a perpetual
with ... ..Itxh 2+. Or 2 8 .. :WxcS 2 9 ..ItxdS,
seemingly terminal , but Black h olds
with the help of a couple of pins, with
29 .. .f6 ! 30 xf6 xd6 31 .l:tc1 f8 3 2
'tWxf8+ xf8 3 3 .l:txc8+ g 7 34 l:Ixh 8
xh 8 3 5 Ji.b7 ..Itxes 3 6 ..Itc8 g7 3 7
..Itxd7 f6, ending up with a quiet
draw in an opposite-coloured bishop
ending .

1 18

2 7 .. .f6 is another possibility, with

less bite for Bl ack than in the similar
set-up in the main line with ... ti'lg4.
There is a quick draw after 28 'iVh s g6
29 ..Itxg 6 xcS 30 iLe8 g 7 ! , forcing a
perpetual after 3 1 f7+ h 6 3 2 'iVh S+.
White could al so try 30 'iVh 6 ! ? (threat
ening ..Itf7 m ate), and then 30 ... xd6
3 1 exd6 hxg6 32 xg 6+ f8 with un
clear play, with queen and pawn versus
three minor pieces. My in stinct is to
favour Black slightly, as White does not
h ave much ammunition to favour the
queen . The perpetual would seem
preferabl e.
2 6 ... ti'ldS therefore gives three ways
of m aintaining the bal ance, which
would clearly h ave been progress for
me after my poor opening.
We now con sider the game's
26 ... ti'lg4?!, ambitious but not quite as
steady. Then there is forced play with
27 g5 f6.

Now Nurmoh amed aimed for a tac

tical flourish with 28 h 5? g6 29 ..Itxg6,
but I h ad seen thi s before playing
... ti'lg4.

Te s t Te n
Sometimes, especially in time trou
ble, it is easier to envisage tactics on
either side, rather th an to con sider
quiet positional chess. Here 28 'iVxg4
'ifxC5 29 liJe8 ! gives White a clear ad
vantage. Then 29 ... 'iWe7 30 liJxf6+ 'iitf8
3 1 liJxh 7+ 'iit g 8, with several ways of
keeping the upper h and, for example,
3 2 e4.
I m anaged to escape, with
29 . . liJxes, covering the f7-square.

White could still force a draw with

30 i.e4 xc5 31 e8+ 'it>g 7 32 'iVe7+
'it>g 8 33 'ife8+, but he was still aiming to
force a win. There is also a more imagi
native draw after 30 i.xh7+!? 'iit g 7 !
(30. . .l:txh 7 ? 3 1 'iVe8+ 'it> g 7 3 2 liJf5+! exf5
33 i.f8+ 'it>g8 34 i.h 6 mate) 31 liJe8+
l:txe8 32 'i'xe8 'i'xC5 33 'iVg 8+ 'it>h 6 34
8 liJf7 (34 .. .f5 ! ?) 35 g8 liJe5, which
would have been on just about the edge
of my thoug ht processes.
In stead he tried 30 i.f7+ 'iit g 7 3 1
i.e3, playable but more complicated. I t
is now White who h as t o find a way of
holding the draw after 31 ... h61, as seen
well in advance.

All Black's pieces around his king

are suddenly extremely well covered.
And yes, I played much better in the
second h alf of the game th an in the
first h alf.
White lost co-ordination after 3 2
liJe8+? 'iitf8 3 3 i.g6 'iit e 7 34 liJg7 i.b7,
adjourned but White resigned before
the resumption, (0-1).
In stead, he could still have kept a
draw with 3 2 J:td1 ! (protection for the
knight) 3 2 ... i.b7 3 3 i.xh 6+ l:Ixh 6 34
liJe8+ 'iith 7 35 liJxf6+ 'iit g 7 3 6 liJe8+ h 7
3 7 liJf6+.

Test 10.3

B ritish Lega ue (4NCL)


As soon as I felt I h ad made myself

comfortably equal, as we saw towards
the end of Test 6.4, I made a comfort
able but l azy move, and found myself
worse. Sometimes it is difficult to read
just when a player h as moved from
being level to being worse. It is difficult

1 19

Why we L o s e a t Ch e s s
t o work out whether a player should
then aim for careful equality, or
whether it is possible to play for more.
Here I got it wrong .

Perh aps the tightest pl an for both

sides is a quick repetition of moves
with 26 :Wid7 (B) 27 iLh 3 c6 28 iLg 2
d7. This, over the board, would h ave
been an unexpected finish, but it is
al so logical enough. White needs to
keep the bishop on the line of the
queen, while Black wants to escape
from the bishop.
There are other examples of to-ing
and fro-ing in thi s type of closed posi
tion, and another repetition would be
27 ct:Je3 ct:Jg s 28 ct:Jf4 ct:Je4 29 iVC1 (29
iLxe4? ! .l:Ixe4 favours Bl ack) 29 ... iLc6 30
'it>f1 fS 31 iLh 3 ct:Jg s 3 2 g2 ( 3 2 xfS ?
.l::t x e3 win s material for Bl ack) 3 2 ... ct:Jg S.
I wanted to block out Black's bishop,
with 26 . ct:JgS? (A), but in the geometry
of chess, a knight cannot simultane
ously defend both the light squares
and the dark squares, and White was
able to take over with 27 f4!. If now
the queen s get exchanged, Bl ack's dS-

pawn will soon collapse, for example,

with 27 .. :xf4 28 ct:Jxf4 .l:Ixe1+ 29 .l:Ixe1
ct:Je4 30 ct:Je3 ct:Jf7 31 .l::t e 2 ! , but not 31
ct:JfxdS ? ? xdS 32 ct:JxdS ct:Jd2+.
So I m ade a queen retreat, 27 .. :d7,
and 28 ct:Je3 left White with a signifi
cant plus. I h ave, of course, lost time.


With 28 ct:Je4 I still felt reason ably

comfortable, but can I even hold it?
Certainly I needed to activate my
knight in the centre. If I tried to play
slowly with, for example, 28 ... ct:Jhf7 or
28 ... c6, White would take control in
the centre with 29 fS, quite probably
eventually winning the ds-pawn .

After 29 ct:JfS, there was a quick

Te s t Te n
simplification with 29 ttJd2+ 30 'it'xd2
'it'xfs 3 1 ttJf4, and the prospect of fur
ther simplification . The trouble is that
the backward pawn on d5 will continue
to be weak, before simplification, and
perh aps even more dangerously, after

.l:i.xe1+ 3 3 .l:i.xe1 'iYxa4 34 1i.e6+ ttJf7 3 5

..txd5 1i.xd5 3 6 ttJxd5 ttJg 5 3 7 ttJe7+
'it>h 7 3 8 'it'e2 i s not so clear) 32 ... .l:i.xe7
33 .l:!.e1 .l:i.xe1+ 34 'ilVxe1 ttJf7 35 1i.h 3 f5
36 ttJg6, and White is h eading for a
Fin ally, we find 3 1.. . .l:!.xe1+ 3 2 .l:!.xe1
.l:txe1+ 33 'ilVxe1 1i.c6 34 ..th 3 'ilVe4 3 5
xe4 dxe4 3 6 1i.e6+ ttJf7, and, of
course, White wins, with a possible
lively checkmate after 3 7 d5 1i.xa4 3 8
d 6 'it>f8 3 9 d 7 ttJd8 ? ! (39 . . . 1i.xd7 40
1i.xd7 lasts longer) 40 ttJg6 m ate.
I cannot remember h ow deeply I
tried to analyse over the board, but cer
tainly I saw enough to appreciate that
quiet play does not work, and so I broke
up the position with 31 gS ! ? At last
my knight was back in play, or at the
very least I was able to exch ange my
passive knight for her more active
knight with 32 hxg6 ttJxg6 33 ttJxg6
'it'xg6, then a further exchange with 34
.l:!.xe7 .l:!.xe7.

I belatedly appreciated that Black

was in serious trouble. Quiet play, I de
cided, was losing . If, for example,
3 1 ...ttJf7 32 h 3 'it'g 5 (or 32 .. J::t x e1+ 3 3
l:txe1 l:txe1+ 34 'it'xe1 'it'e4 3 5 'it'xe4
dXe4 36 1i.e6+ 'it>f8 37 1i.xC4 ttJd6, and
Black is a clear pawn down) 3 3 1i.d7
l:txe1+ 34 l:txe1 l:te4 35 'iYd1 f5 36 1i.e6
i.c6, when Black already has problems
on the light squares, and White can
speed up his simplifying attack with 3 7
i.xd5 1i.xd5 3 8 ttJxd5 .l:!.xe1+ 3 9 'it'xe1
"Yi'xh 5 40 'it'e8+ 'it>h 7 41 'it>g 2 ! , with a
nasty little zugzwang (41...'it'g 6 42
Or 31 ... ..tc6 3 2 h 3 'it'g 5 33 .l:i.xe7
l:txe7 34 .l:i.e1, which again leaves White
clearly on top. The knight in the corner
has still not developed.
Likewise, 31 ... 'ilVd7 32 .l:!.xe 7 ! ( 3 2 ..th 3

My pawn structure was even worse

than before, but at least my pieces
were reasonably active, and I was hop-


Why we L o s e a t C h e s s
i n g th at I could squeeze a draw. Thi s
indeed happened after 3 5 l:te1? l:txe1+
36 "iix e1 Wf7 (now Bl ack can hold) 37
iLf3 "iif5 38 "iie 2 'i!Ve6 39 "iid 2 "iif 5 40
iVe2 iLc6 (maybe 40 ... iVe6 and offer a
draw?) 41 iLg4 iVb1+ 42 Wg2 iVe4+ 43
'i!Vxe4 dxe4 44 iLe2 iLd5 45 Wh3 f5 46
Wh4 wg6 47 iLh5+ Wf6 48 iLe8 We7 49
iLg6 'it'f6 50 iLe8 We7 51 iLg6 Wf6
Yz-Yz. White h ad simplified a little too
In stead, 3 5 "iif4! is winning for
White, as my opponent noted after the

41 iLd1 l:rd2 42 e1 l:td3 43 iLe2. Lau

terbach should certainly h ave won this
g ame.
Chess i s an absorbing game, and
while I was feeling niggling pain after
falling down a long set of stairs at the
hotel, I was still able to pl ay a reason
able game of chess, not h aving, for ex
ample, a losing position after the first
dozen moves. White, h owever, played
better than me.

Test 10.4

H i l l i ngdon League

The queen threaten s to invade with

"iid 6, chewing up the weak b6-pawn. If
then 3S .. .fS, the timing is right for 3 6
l:te1 l:txe1+ 3 7 xe1 'i!Ve6+ 3 8 d2, and
White h as a classic good bishop versus
bad bishop win .
3 S ... "iic 2 would b e a n attempt to
counterattack, but White can, of
course, grab the pawn with 36 "iixh 6.
Bl ack runs out of play after 3 6 ... f7
( 3 6 ... "iix b2 3 7 "iig 6+) 3 7 'i!Vc1 l:te2 3 8
"iix c2 l:txc2 39 l:tb1, and Black's rook
gets trapped after 39 ... iLc6 40 iLf3 e6



h ave mercifully m ade very few

one-move blunders, where I m ake a
move with a piece and h e just takes it,
or I miss th at he was attacking my
piece. Perh aps one of my more memo
rable, or forgettable of my g ames, was
in a quickplay at Richmond a few years
back, when pl aying again st Rasa Noe
inkeviciute. I moved my king to an ad
joining file, and immediately she
stopped the clocks, and she went up to
the arbiter. I simply could not begin to
understand what was going on. It was
a few minutes l ater, when the arbiter
h ad arrived, th at I saw wh at the prob
lem was. I h ad moved my king into
check, along a file, and I could not see
the g ap between her rook and my king.
Thi s was a bizarre finish.
It was not just a question of my very
poor eyesight at the time, but rather

Te s t Te n
th at my brain was clearly not in full
focus on chess. Young children m ay
occasion ally try to move a bishop or
queen on a diagonal , and m ay try to
switch from a light square to a dark
square, or vice versa. I was making a
similar mistake with the rook, seeing
her rook moving from the c-file to the
The end result was a draw, after a
time penalty for my illegal move. I was
amazed th at I h ad got away with it,
thinking th at if I m ake a mistake, she
should be allowed to take my piece in a
quickpl ay finish. Apparently though
the arbiter was correct. Thi s was a
quickpl ay finish, not an actual blitz
Back to the G ait game. Not a one
move blunder, but a two-move blun

I tried 27 'iVg4? (A), seeing my oppo

nent's move, but missing that after the
exch ange h e has a win th e following
move, with 27 'iVxg4 28 hxg4 .te2,
trapping the rook. White resigned (O-i).
White could h ave fought longer
with 2 7 flig3 (B) 2 7 ... .th S 28 .l::.b l .th 6,
but one can feel certain th at Black
would eventually win . Bl ack's bi shop
pair is extremely strong .


Test E l eve n
11. 1 White to play

113 Black to play

A) 2 7 ..txh 6
B) 2 7 .l:f.xa8
C ) Something el se?

A) 2 7 ... e5
B) 2 7 . ..1:1g 5
C) Something el se?

11.2 White to play

11.4 Black to play

A) 2 7 fxg 6
B) 2 7 lIxh 6
C ) Something else?

A) 2 7 .. Yiilc 2
B) 2 7 ... h 5
C) Something el se?


Te s t E l e v e n

Test 11.1


London Open


It's difficult to get much worse than

this, for any player of reason able
strength . I do not pretend th at I was
pl aying particul arly well in the early
part of the g am e, but then he made
mi stakes too. Out of kindness to both
pl ayers, we do not cover the first 26
moves. My position looks extrem ely
promising just at the mom ent, but h e
could h ave equalized comfortably i n
the opening. B y now, I was clearly in a
winning position, with a powerful rook
on the seventh and a protected passed
pawn just behind. I could even take his
own passed pawn without too much
In this exercise, I deliberately put in
an obvious blunder to set up a false
trail for the reader, with 2 7 iah6? (A)
27 ... J::t e 1+ 28 'iVfl J::t xf1+ 29 'it>xfl, but
even this is winning for White; his
passed pawn dominates.

There are several other ways of

pl aying the position for a comfortable
win for White, but not the casual 2 7
ClJxc6? ( C ) 2 7 ... .i.xc6 28 .i.xc6 l::t x e3 2 9
fxe3 g 3, for example, allowing Black
too much counterplay. After 30 J::t e 7,
there is a perpetual with 30 ... e1+ 31
'it>h 2 ttJg4+ 3 2 hxg4 'iYh4+ 33 'it>g l 'iVe1+
34 'it>h 2.
The other capture on c6 i s al so un
clear after 2 7 .i.xc6 (C) 2 7 ... l::t x e3 2 8
fxe3 .i.xc6 29 ttJxc6 g 3 . It i s possibl e
th at one or two lines for White might
still keep some sort of edge after deli
cate m anoeuvring, but White should
be aiming for something simpler and
All thi s suggests 27 'f1C7! (C), blocking Bl ack's diagonal to the kingside.

Bl ack's rook and knight h ave some

activity, but Black's queen and bishop
now do nothing. With care, White
should win comfortably.
The next question is why I blun
dered with 27 l::tx a8?? (B). Part of thi s
was that I would h ave been frustrated
by not finding anything to set up a win


Why we L o s e a t C h e s s
with an immediate capture o n c6, and
decided th at it was time to eliminate
the bishop on a8. Partly too I would
have lost concentration . I h ave a win
ning position, my opponent is clearly
on the lower h alf of the board, so why
should I use up energy in such an easy
position ? The trouble i s that an easily
winning position plus lack of concen
tration adds up to the loss of grading
I pl ayed my sacrifice, and then after
27 .. J1Vxa8 I suddenly noticed that not
only h ave I given up the exch ange, but I
h ave allowed Black to control the a-file.

All thi s broke my equanimity. I set

up a couple of passed pawn s with 28
ii.xc6 .l:txe3 29 'iWxe3 ii.xe3 30 ii.xa8
i.. x d4, but Black too h ad strong passed
pawn s, and after another bad move,
which we'll see in Test 12.4, I was los
In stead, if I could h ave found a way
of regaining my composure, I would
still h ave h ad chances of keeping a
slight edge with 28 'iWxC 3 ! .
I h ad still clearly not recovered from


my stroke, and was still m aking many

serious blunders, m ainly through poor
vision, both metaphorically and liter
ally. Here I forgot th at Black could at
tack on the a-file once White's rook was
no longer on a7.

Test 11.2


Thames Va l l ey League 2007

White to play. It is not difficult to

work out that this was a Nimzo-Indian,
with the well-known doubled c-pawns.
Those pawn s can at times end up as
weak, but h ere it i s not so clear that
Bl ack can take any advantage of the
slight positional defect. On the con
trary, White is better, with more space
for his pieces, and has excellent
ch ances of attack on the kin g side,
Bl ack's pieces being con strained on the
l ast two ranks.
The next few moves run smoothly,
beginning 20 ii.gs f6 21 ii.e3, forcing a
pawn to advance in front of the king,

Te s t E l e v e n
setting up a new weakness. The prob
lem for Black is not so much the weak
ness of g6 itself, but more the threat of
a sacrifice on h6, forcing Black's king
into extremely open play.
He tried, m aybe with great anxiety,
21 ... g5, covering the g 6-square, but
also an open invitation for White to
sacrifice, or a dangerous pawn ex
change with h4. The computer explores
the idea of defending with 21 ... .l:tc8, but
after 22 .l:tf3 .l:tfd8 2 3 l:tg 3 'iif7 24 'iih 4
f8, there i s a sacrifice o n a different
square with 25 l:txg 7+ xg 7 26 ii.xh 6.
White ties thing s up after, for example,
26 .. :iVg 6 27 ii.e3+ 'it>g8 28 l:th 8+ ct;f7 29
l:th 6 g 8 30 l::rh 7+ ct;f8 3 1 l::rh 8 .
For the next few moves, White qui
etly undermines Black's king side with
22 h4 l:tg8 23 l:tf3 l:tg6 24 l:th3 ct;g7.

Now I played 2 5 ii.e2. This was some

thing of a waiting move, aiming for a
better diagonal for the bishop when the
storm arrives. There was a possibility of
combinative play with 2 5 hxg 5 hxg 5 26
l:txh6 l:txh 6 27 xg 5 l:tg6, but I could
not find any further breakthrough. With

rooks o n the board ( 2 8 i1.. x e7 l:txg4 2 9

ii.xd6 l:tg6), o r with the rooks off (28
l::rh 7+ 'it>g 8 29 l:txe7 .l:txe7 30 i1.. x e7 .l:txg4
31 .1Lxd6 l:tg6 32 .1LC7 .l:ta6, and Black is
better), Black can at least hold. White's
bishop-pair does not cooperate. This
was a disappointment for me, as I had
pressed my attack as far as I could.
I pl ayed the bishop move, aiming
for a later .1Lg4, but now my e-pawn
was slightly weakened. Maybe instead
25 l:th 1 ! ?, as suggested by the com
puter. The computer then suggests
holding moves, such as 'it>h 2 or l::r h 2 ,
waitin g for t h e opponent to create a
weakness. A more constructive pl an
would perh aps be hxg 5 followed by
3 , when at least we can see what
White i s aiming for.
In the g ame, Black should h ave tried
25 ... l:th 8 ! , avoiding any unnecessary
concession s. White can try for an edge
with 2 6 e6 xe6 2 7 dxe6, but there is
no great chance of certainty for an
edge in over-the-board play. A steadier
line might be 26 l:th 1 d8 27 ii.d3, still
perh aps keeping an edge, but White is
circling around with his pieces.
Over the l ast few moves, White has
never quite been able to m ake his
breakthrough . Wh at is surprising is
th at it is Black who open s up the posi
tion, with 2 5 ...f5 26 exf5 tLlf6.
There are tactics for both sides, with
White's queen and rook being threat
ened, and Black's rook and two of the
pawn s in front of his king are about to
be stripped away. It is now time for


Why we L o s e a t C h e s s
White t o calculate with clarity. Unfor
tunately I did not do this well.

27 h6! (B) i s straightforward,

m aking sure th at White can get at the
black king without any pawn obstruc
tions. 27 ... liJxg4 2 8 l:ixg 6+ f7 29 .ltxg4
drops White's queen, but h e has a rook,
bishop and two pawns in return, more
than sufficient compen sation, espe
cially as White keeps a serious attack
against the king . Indeed, 29 .. .1:[g 8 30
hxg 5 xg 6 31 h 7+ l:ig 7 3 2 xg 7+
xg 7 33 f6+ "i!Vxf6 34 gxf6+ 'It>xf6 soon
win s for White, with his extra piece.
In stead I m ade a bad stumble, with
27 fxg6? (A) 27 ... liJxg4 28 xg4.


Time trouble was approaching, this

being an evening m atch, and my op
ponent tried the n atural 28 ... xg6?,
but I was soon able to gain a winning
advantage with 29 hxg5 liJf7 30 gxh6
'It>h7 31 .ltf5+ 'It>h8 32 l:ig3 l:ig8 33
l:ixg8+ 'It>xg8 34 .ltf2 e4 35 .lte6 'It>h7 36
.lth4 'ilC7 3 7 .ltf5+ g8 38 h7+ g7 39
h8'iV+ liJxh8 40 l:ih7+ 1-0.
Instead, 28 ... liJxC4! keeps play com
plicated. Maybe White should eventu
ally end up winning, but it would be
more due to luck than judgement.
White h ad a much clearer option .

My notes at the time suggested that

I was probably thinking of 2 9 .ltxg 5 ? ! ,
but then after 2 9 . . .h xg 5 30 l:ih 7+ xg 6
3 1 l:ixe7 l:ixe7 Black h as recovered
much of his ground.
In stead, 29 l:ixh 6 'YWd8 ! ? 30 l:ih 7+
'It>xg 6 31 .ltxg 5 "i!Vxd5 32 .l:!.d7 'iig 8
( 3 2 ... 'iie 4? ! 3 3 .l:!.f3 "iVxg4 34 l:if6+ leads
to an attractive checkmate, with
34 ... 'It>xh 5 3 5 l:ih 7 m ate) 3 3 .l:Ig 3 iVh 8 !
34 i. c l ! "iVxh4 3 5 i.h 5+ f5 3 6 l:if3+
e6 3 7 l:ih 7 'YWel+ 3 8 l:ifl "i!Vxfl+ 39
'It>xfl eventually leads to a win for

Tes t E l e v e n
White, but who can doubt th at White,
and indeed Bl ack, could h ave pl ayed it
so accurately?
This was one of m any club games I
played against David Okike over a few
years, playing again st Kin g s Head and
Hayes. I won most of them, but h e h as
scored a win and a draw against me.
The strangest encounter was a few
month s after thi s one.
I walked through Oxford Circus be
fore the start of a London League
m atch, then realized th at I was starting
slightly l ate, so I got up the stairs at the
venue, quickly m ade my way through
the dimly lit room, m ade my first move
as Black, ... c7-c5 on the dark squares,
pressed the clock, sat down, then sud
denly noticed th at I h ad played closer
to th e edge, light square to light
square, 1 e2-e4 b7-b 5 ? ?, in stead of the
intended 1 ... c 5 .
M y opponent briefly wondered what
I had planned with my unexpected
opening move, when it suddenly be
came clear from my shock that I h ad
blundered on move 1, a mixture of poor
eyesight (I am only partially sighted)
and being distracted by having to rush
to the venue. I did not ask for the game
to be restarted at the beginning, and I
suspect that it would have been im
proper for me to do so. My opponent
sportingly suggested that we could start
the game at the beginning anyway, be
ing aware that I was barely able to see,
and that he would rather have an inter
esting game rather than a hollow win.

Much l ater, when looking though

my games for thi s book, I m ade the
second big visual blunder in thi s g ame,
pressing the wrong button on my
computer (no, it wasn't a '(' button too long to expl ain, but I'm sure th at
the memoirs of Grandmaster James
Plaskett will give the full story). Any
way, what h appened is th at I m anaged
to delete my game. Almost certainly it
would be possible to retrieve it, but I
h ave not got round to this. I was able to
remind myself that I m anaged to get a
good position in the 1 ... C5 opening, or a
Sicilian, then mispl ayed the queenside,
then he mispl ayed it, and then finally a
win for me.
A long shaggy dog story.
There are, of course, m any more an
ecdotes from events for blind and par
tially sighted players. Alas, such things
are inevitable if half the players cannot
see, and the other h alf can barely see,
and can easily bump into obstructions.
The cl assic question is wh at h appen s
when a blind player plays a deaf pl ayer.
My own nightmare is that my brain
dam age not only makes me only par
tially sighted, but al so th at I suffer
from a degree of aphasia. I can usually
understand wh at is being said, but it
can be very difficult to bring my words
together, and I am terrified of saying
the wrong move, and making a big
mistake. To make a move over the
board, .l::!.fe1, is easy enough, and it is
easy enough also to write it on the
scoresheet, but then to have to say


Why we L o s e a t C h e s s
' Rook H ector Eva One' takes more time,
and the intern ation al Germ an Braille
version, 'Turm Hector Eva Eins', takes
several seconds to work out. By the end
of all this, I cannot remember what I
have written down on my scoresheet.
Fortun ately, Braille pl ayers know
their own allowances. After all, every
player has his or her own disability.

Test 11.3


B ritis h league (4NCL)


pawn push weakens his overall king

side pawn structure.
If he was worried about the back
row, a much simpler option was 2 3 f4! ?
C5+ 24 'iyxc5, and then after either
recapture, White would try 25 d7.
White has a clear edge in the double
rook and pawn endgame, Black's pawn
structure being weakened. If White
was worried about counterattacks,
then surely forcing an advantageous
endgame was the safest way of not
losing the game in a complicated mid
2 3 xa7 ! ? can still be tried. Then
23 ... cd8 24 f1 f4 2 5 xb6 b8 26
c6 xb2 27 d3 will give no m ating
attack for Black, just a little pressure,
whereas White still h as two threaten
ing pawn s.
After thi s reprieve, I threatened to
set up an attack with 23 f4, also keep
ing the queens on the board. Maybe,
after 23 ... a5, I would h ave been worried
about 24 d5, but after 24 ... g 7 2 5
ttd7 f4, Bl ack h a s reasonable chances of
h olding the balance.

Continuing from Test 8.1, Bl ack's

king side attack is threatening, but far
from decisive, and if White is able to
keep an extra pawn, and perh aps an
extra passed pawn l ater, then the onus
for Black i s to keep the bal ance.
Here White pl ayed slightly nerv
ously with 23 h3?!, with no discernible
intention other th an to create an es
cape square for the king on the back
row. Black i s not threatening anything
on the end rank though, and White's


Te s t E l e v e n
He now tried 24 ltd 7, doubling on
the seventh . It i s possible th at h e h ad
previously considered trying 24 ltdS
'iYg 7 25 ltg s 'iYh 6, but thi s seem s
messy. If 2 6 ltbS 'iYg 6, White's best op
tion is 27 ltg s 'ii'h 6, with repetition., as
27 'iYxa7 ? ltg 8 28 'iYb7 ltcd8 ! i s ex
tremely dangerous, with threats of ...
'iYbl+ and ...ltdl, and m ate on h l . If 2 9
lt c l lt d S 30 g4 (otherwi se m ate o n g 2 )
30 .. .fxg 3
checkm ating. 50 White definitely h as to
pl ay carefully.
The other option is an immediate
capture with 24 'iYxa7, and thi s looks
good, even though White h as lost time.
24 ... 'iYe2 2 5 ltcd3 'iYxb2 2 6 'iYe7 ! h as
drawn Black slightly out of pl ay with
his pieces. Then 26 ... 'iYg 7 27 ltd8 ltg 8
28 ltxg 8+ (not, of course, 2 8 ltxc8??
'iYxg 2 m ate) 28 ... xg 8 29 'iYb7 is decep
tive. It might look as though the pawn
on g2 is seriously weak, but as long as
Bl ack can do nothing to add extra pressure (for example, 29 ... 'iYg 6 when 30
ltd7 e S 31 lte7 prevents ... e4), he can do
nothing with his pieces, and White can
gradually improve.
After the move actually played, the
computer suggests several option s for
Black, including keeping the a-pawn
safe with 24 ... a6 or 24 ... aS. The aggres
sive counterplay, with 24 ltce8 2 5
'iYh4 'iVf5, looked t o me far more ap
pealing, not least because my king was
now safe. I felt that m aybe I was better,
and that it was up to him to prove oth

H i s next move 26 b4?!, was unex

pected, a quiet move on the queenside,
before the king side storm s.
2 6 ltxa7 is the more direct way of
setting up a queen side pawn advan
tage. Then after 2 6 ... ltg 8 2 7 ltf3 'iYbl+
28 h 2 ltxg 2+ 29 xg 2 ltg8+ 30 ltg 3
'iVe4+ 3 1 g l ( 3 1 fl 'iYb1+ al so leads
to perpetual) 3 1...ltxg 3+ 32 fxg 3 'iVel+,
with a draw by perpetual . Here the
computer suggestion, 2 7 h 2?, is
deeply un convincing after 2 7 ... ltg S ! 28
ltd7 lth S 29 g4 fxg 3+ 30 'iVxg 3 ltg8.
I carried on with the attack with
26 ltg8, when probably the only sen
sible reply for White is 2 7 g4! fxg 3 28
xg 3, clearly with at least a draw for
Black after, for example, 28 ... ltxg 3+ 2 9
fxg 3 'iYbl+, but with no obvious win
ning plan . In stead, 2 7 ltxa7 ? 'iYb1+ 28
h 2 'iVfl 29 'iVf6+ ltg 7 30 g4 'iVxf2+ 3 1
h l e1+ 3 2 g 2 d2+ 3 3 fl ltd8 i s
not safe. White's king i s under pres
sure, and Black h as no need to repeat.
White missed a tactic h ere with the
move 27 h2?, and then I missed a
good reply.

13 1

Why we L o s e a t C h e s s
ing the tactic, White kept an edge with
2S lIcd 3 lIe6 29 lIdS lIeg6 30 lIxgS+
'it>xgS 31 'iVdS+ 'it>g7, and should h ave
carried on with 3 2 'iVd5 ! in stead of 32
lId6?!, equal, but l ater a loss. Yes, it has
to be admitted that the play in this
g ame has been for the most part ex
tremely unexciting, with too many
mistakes on both sides. Unfortunately
it doesn 't get better, as we'll see in Test
27 ...e S ? (A) is solid enough, but as
Keith Arkell pointed out, 27..JJ.g5! (B) i s
very strong, with a winning queen sac
rifice after 28 1:txf7 J::t e g 8 ! ! 29 J::t xf5
lIxg 2+ 30 h 1 lIg1+ 3 1 h 2 J::t 8 g2
mate. Sometimes the spectator has the
best view. And I think th at neither
player was seeing the g ame well any
White can, of course, avoid the im
mediate checkmate, playing defen
sively with 2 8 f3 lIeg 8 29 'iYf2 lIxg 2+ 30
'iVxg 2 lIxg 2+ 3 1 xg 2. Then 3 1 ...'ilVb 1 !
seem s the best way t o prevent White's
rooks from coordinating properly, and
White still has defen sive problems.
One would not expect White to be
able to h old the draw after 28 g4 fxg 3+
29 fxg 3 'iVf2+ 30 h 1 1:teg 8, but h e can
struggle on in a rook and pawn ending,
with 3 1 'iVd4+ 'iVxd4 3 2 J::t x d4 1:txg 3 33
J::t x g 3 J::t x g 3 34 J::t d 7 J::t xh 3+ 35 g2 J::t x a3
3 6 1:txf7. White i s going to be a pawn
short of safety, though, once the
queen side pawns are eventually ex
ch anged.
After my less accurate move, miss-

13 2

Test 11.4


B u ry St Ed m u nd s


From the end of Test 8.2, White h as

a 5-4 pawn advantage on the kin g side,
while Black h as an i solated pawn on
the queen side. The danger is, of course,
th at Black is going to lose the a-pawn .
N aturally Black will not be able to win,
but h e should h ave good ch ances of
h olding the g ame. White's extra pawn
on the king side is a doubled pawn, and
sometimes the pawn s can get in the

Te s t E l e v e n
way of each other. Visualize a position
where White takes the a-pawn with his
rook, Black pl ays ... 'iVdl+, then h 2
s+ and a perpetual , and the point i s
easily m ade.
I played very poorly over the next
few moves, quite simply losing concen
tration . 27 ... h S ? (A) was bad. I noted at
the time th at " I see from my scoresheet
that I pl ayed thi s more or less immedi
ately, with 29 minutes to go until the
time control on move 36. No time
shortage therefore, but I was con scious
th at there was going to be a quickpl ay
finish coming up. I should h ave
thought much, much h arder, and m ade
White work h ard to try to win the
pawn ."
U sually I am reason ably good at
time pressure, m aking it a rule for my
self that I try to keep at least five min
utes for the last few m oves, to en sure
that I h ave some time, if necessary, to
think about the last couple of moves.
Sometimes, as here, I speed up too
early before the time control . There is
no general rule to deal with thi s type of
question, as different people genuinely
thin k differently, and there are many
players who can think superbly with
the fl ag about to fall, while others pre
fer to take their time, and can get flus
tered by h aving to move in stantly.
On thi s move, I got 'pre-flustered',
just m aking a move, hoping to set up
some sort attack against White's king
side, and vaguely thinking th at at least
I am opening up a flight square for my

king . I t loses a tempo, and m y oppo

nent m akes much better use of it.
I should h ave thought and planned
at thi s moment, deciding h ow to h an
dle the rest of the g ame. My pawn
moves are almost irrelevant. I cannot
improve anything with my pawns, so I
need to concentrate on my pieces.
Clearly my king and rook are stuck to
defence, so that leaves moving my
queen and knight into better pl ay, and
preferably m aking them work together
rather than separately.
Thi s immediately suggests 27 'iVC2!

so that if 28 ttxas??, Bl ack win s im

mediately with 28 ... 4:Je4 ! . The queen on
c2 is superbly placed, putting pressure
on f2, and the second rank as a whole,
and reminding White that he needs to
cover his back rank. Once Black has
seen thi s, all th at White can h ope for is
to keep equality.
In stead, 28 4:Jd7 ? ! 4:Jxd7 29 'iVxd7
'iVC3 keeps Bl ack's a-pawn safe, and it
would be up to White to hold the posi

13 3

Why we L o s e a t Ch e s s
The correct 2 8 'iVC5 i s safe, and then
28 ... 'ivxc 5 29 dXc5 tra8 30 tiJC4 Wf8 3 1
Itxa5 (m aybe 3 1 tiJxa5 ! ?, keeping the
rooks, but Black should still hold)
3 1...trxa5 3 2 tiJxa5 e8 3 3 f3 d7 34
f2 C7 35 tiJC4 tiJd7 3 6 tiJd6 f5 37
tiJe8+ c,t>d8 3 8 tiJxg 7 tiJxC5.

Black i s a pawn down, but it i s diffi

cult to envisage any realistic winning
chances for White, given th at Black's
pieces and pawn s are well coordinated,
and th at White's doubled pawns give
littl e chance of setting up a passed
pawn .
Almost certainly I would h ave been
suspicious of allowing the queen ex
ch ange on c5, but small-scale m anoeu
vring with the king and rook on the
same part of the board m akes it diffi
cult for White to create any significant
advantage, even with the extra pawn .
The simplest plan, 2 7 tiJg4?! (C) 2 8
tiJxg4 'iVxg4, is a n altern ative possibil
ity, and then 2 9 'iVa7 Itc8 30 iYxa5 h 6 .
Bl ack would h ave good chances o f h old
ing the position if the queens are off
the board, but if White were to keep

the queen in active play, the chances

are that he would be able to find a win.
2 7 tiJds (C)

is another try, trying to keep the

queen and knight working together,
but 28 'ih> 7 ! keeps White's pieces the
more active.
With my own misguided move, I
was thinking in term s of exch anging
pawn s with ... h4, with a 4-3 defence.
However, exchanging a good pawn of
my own, for a doubled pawn, is not a
good option, especi ally as I was losing

13 4

Play continued 28 'iVC7! (stopping

...'iVc2) 28 ... h4 29 gxh4 'ii'x h4 30 tiJf3 'iVe4

Te s t E l e v e n
31 xa 5 lbd5 32 a7. My play continued
to fall apart after 32 ... g6?! (weakening
my pawn structure ) 33 d7 'it>g7?! (in
credibly, I was thinking of opening up the
h-file for queen and rook ) 34 d6 lbf6 35
e5, and when I started to recognize
that the queen exchange would lose, if
not necessarily quickly, I retreated with
the queen, without much hope, tryi ng

3 5 ...C2?!. Then after 3 6 lbg5 f5, the

last move before the time control, I de
cided that I had to undo the retreat, as
otherwise the pin on the knight on f6
would lose quickly.
A disastrous finish before the time
control, and really I would expect to
lose this quickly. There is, h owever, a
sequel, as we will see in Test 1 5 .4.

13 5

Test Twe l ve
12. 1 White to play

12.3 Black to play

A) 28 'iWf4
8) 2 8 'iWc2
C) Something el se?

A) 30 .l::!. a c8
8) 30 .. .'!i'lc6
C) Something el se?

12.2 White to play

12.4 White to play

A) 28 fS
8) 28 'iWxC7
C) Something else?

A) 3 1 b7
8) 3 1 ct'lb3
C) Something el se?

13 6


Te s t Twe l ve

Test 12.1


London League


Continuing from Test 9.1, I played

28 'iYf4? (A), pointl ess, as there will be
no checkmate on 0. Indeed, Bl ack was
able to obtain an edge.
Somehow I need to create some sort
of edge on the queenside, and with
files rather than diagon als. A few
month s after thi s g ame, I gave a doubl e
exclamation m ark for 28 'iYc2! (B),
mainly because of the paradoxical ef
fect of playing '1Wc2-f2 and then unex
pectedly 'iWf2-C2. Maybe one exclama
tion m ark.
The critical line would seem to be
28 .. .'jxg 3 29 hxg 3 4Jf6 30 a4! (White
needs to attack quickly) 30 ... .l:!.e 3 31 as
4Jxds ! ? ( 3 1 ... 4Je4 32 axb6 .l::i. x g 3+ 3 3
h 2 axb6 34 "iYb2 favours White; Black
cannot find a way to shift the queen
from a6 to the h-file) 32 l:txfs ! ( 3 2
'iWg 2 ? ! allows Bl ack t o wriggle out with
advantage after 32 ... 4J0 33 axb6 axb6

34 .l:!.b3 'iWxC4 3 5 .l:!.xb6 'iWd3 ; note

Black's central control in thi s end posi
tion) 32 ... gxf6 33 'iWxfs+ 'iitb 8 34 'iYxds,
and White holds the draw. There i s the
obvious practical objection th at no
player could be able to calculate this
over the board. Maybe so, but it is use
ful for players to sh arpen their posi
tional in stincts in complicated posi
tion s. Who knows, if you can find the
best move, the rest might follow.
I was momentarily impressed with
28 'iWb2?! (C), but 28 ... 'iWa4! allows the
queen to squeeze out, or rather to ex
ch ange. After 29 "iYb3 (otherwi se
... 'iWxc4) 29 ... 'iWxb3 30 .l::i. xb3 (30 axb 3 ?
4JX(3) 30 ... 4Jd2 the knight forks two
rooks, and al so an important pawn ( 3 1
l:t a 1 4Jxc4! ?). So White does not want
to allow Black to invade the a4-square.
In the game, after 28 'iWf4 ...

1 was critical in my notes about

28 4Jxg3: "Not a bad move, but there is
a better one", but this seem s harsh. If
your flag is hanging, you want to play
good moves quickly, and not waste time
searching for exotic geometry. That said,


13 7

Why we L o s e a t C h e s s
28 ... ttJes ! 29 l:I.e1 ttJxg 3 30 l:I.xes ttJe2+!
(seen after the game!) 3 1 l:txe2 l:I.xe2 3 2
'iYd6 'iYxa2 is good for Black, i f you have
time to think about it.

It would take a lot of time for a

pl ayer to convince him self that Bl ack
could demon strate th at he eventually
can win after, for example, 33 'ilVc6+
'it>d8 34 'ilVd6+ 'it>e8 3S 'ilVc6+ 'it>f8, and
meanwhile the tim e is ticking. Re
member too th at this is five moves
down a sh arp line. In practical term s,
there was no reason for Cox to want to
touch thi s line.
In the g ame, after Cox exch anged, I
recaptured with 29 'ilVxg3, covering the
squares next to the king. 29 hxg 3 'ilVxa2
30 l:I.d1 l:1.e2 31 'ilVf1 ttJes would in stead
put White under pressure.
Bl ack collected his second pawn,
with 29 'ilVxa2, with every chance of
winning a third pawn . I decided th at
my only chance was to play actively,
with 30 'ilVd6. I felt th at my queen was
m oving through the trap door, but
what else could I h ave done? The trap
was closed with 30 ttJe5.


13 8

I played 3 1 l:I.bb1!. Thi s i s a critical

defen sive plan for White, win, draw or
lose. It was also difficult to find over
the board, although it looks obvious
afterwards. Shifting the rook to a1 is
the only chance. The desperate sacrifice
with 3 1 ktfb1 'ilVxC4 2 2 l:1.xb6 axb6 is not
worth thinking about. After either re
capture on b6, Bl ack sets up a swarm of
winning checks with ... 'ilVg4+.
Then 31 'ilvxc4 was natural and
good. In stead, 31 ... ttJf3+? 32 'it>h 1 l:I.e2
33 'ilVc6+ 'it>d8 34 'ilVa8+ 'it>d7 (34 ... 'it>e7 ??
3S d6+ allows White a win) 3S 'iYc6+ is
only a draw.
3 2 l:1.a1! is again the only chance. Of
course, Black could give a perpetual
immediately, and, of course, he could
find a possible improvement to such a
sequence, but for White, a degree of
discomfort i s preferable to an immedi
ate loss. There is a simple trap th at
Black could fall for, with 3 2 ... 'it>b7 ?? 3 3
l:I.xa7+ 'it>xa7 34 'ilVC7+ 'it> a 6 3 S kta1+
'it>b s 36 l:I.b1+ a4 37 'ilVa7+, but Cox
would h ave seen that without any dif

Tes t Twe l v e
In stead, h e started with checks, with
32 ... g4+ 33 hl e4+ 34 gl, and
quietly improved his pieces with
34 ... b7. Thi s time, White's sacrifice
gives nothing, after 35 .l::[ x a7+ xa7 36
C7+ a6 37 .l::[ a l+ b 5, and the king
escapes to c4.
I fought on with 35 I:tfel, and the
casual 3 5 ... ct:Jf3+? 36 f2 even gives
White an edge after 36 Wf2 C2+ 3 7
xf3 xc3+ 3 8 f2 "iVd4+ 39 f1.
So 3 5 ...g4+ 3 6 Whl "iVf3+ 3 7 Wgl,
and to my con siderable surprise, h e
offered a draw.

Here 37 ... xc3 would have been a

win for Black, after the adjournment, to
be played at some later date. Few play
ers enjoy a second session in evening
games, with time being precious, al
though there are others who like a
quickplay finish even less, with every
thing decided in a time scramble close
to 10pm after a long working day. For
myself, I felt I had to avoid a quickplay
finish, which was my right in the league,
not least through respect to my team
colleagues. I did not want to lose for the

team in what could have been a critical

match, as a result of tiredness and diz
ziness. My opponent clearly did not
want to have a second session, and I
apologized before the start that I could
not play a quickplay finish. At the end of
this particular session, he decided he
wanted to force a draw on the evening.
Had he sealed a move, I felt sure that I
would have gone home and analysed
the position, and decided that it would
not h ave been worth playing on, and
would have resigned.
An expl anation then for an unusual
result: an unusual situation . I do not
know wh at i s h appening in other coun
tries in evening league matches. In
England there i s often con siderable
debate about quickplay versus slow
The late Lev Polugaevsky noted, well
before computers were good at chess
analysis, th at adjournments in interna
tional events would at some stage be
come obsolete, because players with a
good database or engine would h ave
an unfair advantage compared to the
player with less computer help. It is
now generally agreed that games
should be finished in one session, for
morning or afternoon events.
In evening games, computer analy
sis often cuts down second sessions,
because one of the players would see
th at it is not worth playing one, either
because it is a clear draw, or because
one of the players is losing . The few
adjournments th at do carry on,

13 9

Why we L o s e a t C h e s s
though, are often of great interest, and
enjoyable for the pl ayers.

Test 12.2
C Crouch P Roberson

B ritish League (4NCL)


From a Modern Benoni. White has

clearly been able to push his central
pawn s safely, and can create a passed
pawn, but his position i s not quite as
overwhelming as it looks. He still h as to
think; h e cannot pl ay fully autom ati
After some thought, I decided to ex
change the pawn s on d6, keeping an
isolated but advanced passed pawn on
dS, helped by rook and bishop. I played
24 tiJe4.
24 e6 'fie7 i s the obvious altern a
tive, and who could resist a well
protected passed pawn on the sixth ? It
i s unclear, though, how White can
m ake progress in such a blocked posi
tion . After, for example, 25 'fid2 tiJg 7
26 i.. a 2 tiJfs 2 7 i..f2 b S 28 c1 c4 29 b4


l:tac8 30 axb s axb s 3 1 i..b 1 tiJg 7 3 2

i.. e 3 fe8 3 3 i.. c 2 l:ta8, i t is still pleas
ant for White, but what next? Black too
has a protected passed pawn .
In the game, Black defended on f6,
with 24 i.. g 7. 24 ... i.. xf4 2 5 i.. xf6 looks
much too dangerous for his king, so he
must try to hold his position .

I now continued my planned pawn

exch ange, 25 exd6. It is again possible,
but slightly illogical, to try 25 e6. Black
can sn atch a pawn with 2 S ... 'fixa4, now
th at White's knight h as moved away.
26 tiJC3 'iYb4 27 tiJa2 'fia4 is a drawing
possibility, although Black could try for
more with 2 7 ... 'fias .
In stead, 2 5 as ! is seriously t o be
con sidered, and might well be best, but
leads to complication s and counter
complication s. I would h ave con sidered
it, but would probably h ave decided
that the move I played was simpler,
and gives White a clear edge. I would
h ave been h appy with my move avoid
ing any possible sh arp play.
After 25 as fxe s 26 fxe s bS 27 exd6
cxd4 28 tiJxcS 'fibs 29 tiJxb7 'fixb7 30

Te s t Twe l v e
dxC7 xc7 White has a comfortable
extra passed pawn . It i s possible that I
would h ave been concerned about
whether I would h ave a genuine ad
vantage after the more natural 27 e6
c8 2 8 a2 c4, and decided in retro
spect that m aybe 29 iLe7 l:tfS 30 ttJxd6
ttJxd6 3 1 xd6 ttJxdS did not give
White much of an edge.
In the end, I went for the simpler
option . This was by no m ean s bad, but
was still sub-optim al, and it turn s out
th at I soon m ade a mi stake even in the
'simple' line.
Play continued with 2 5 ...ttJxd6 26
ttJxd6 xd6 2 7 e7. Here Black could
try 2 7 ... xf4 28 .l:i.e4 fS, but White is
doing well after 2 9 d6+ 'iith 8 30 .l:i.e3
'ii'f4 3 1 dXc7 'ii'x C4 3 2 .l:i.d7 g 8 3 3 'ii' d 6.
However, 29 g4? ! would be more
speculative, and after 29 ... xf3 30 .l:i.d3
xd3 31 xd3 l:tf7 32 d6 ttJxdS Black
is better after, for example, 3 3 C4 iLf8
34 'ii' e 6 .l:i.d8 3 5 xf6 ? .l:i.d6.
Bl ack in stead pl ayed 27 ... .l:i.ad8.

Now l over-egged my pl ay with 28

f5? (A), hoping to g ain a tempo by

bringing the bishop quickly into play

with iLg 3 .
Instead, 2 8 WiXC7 (B) 28 . . .xC7 29 d6+
would h ave been simpler and more
natural. Then 29 ...Wif7 30 xf7+ (30
ttJes ? fxes 3 1 xf7+ 'iitxf7 3 2 xd8
l:txd8 33 fxes is over-elaborate) 30 ... l:txf7
3 1 ttJes ! . Black cannot take the knight,
as the pawn is pinned, 31 .. .fxes ? 3 2
xd8. 5 0 after 3 1 . . .l:tff8 3 2 ttJC4 bS 3 3
axbs axbs 3 4 ttJas White has finally ob
tained that safe positional edge that he
h as been trying to achieve for a long
time. The next stage of development is
to bring the bishop to f2, and then aim
for the queening square on d8, via b6.
In stead, I lost a tempo after
28 .. :iVxe7 29 .l:i.xe7 (29 d6+? Wif7 30
xf7+ .l:i.xf7 3 1 dXc7 .l:i.xd1 3 2 .l:i.xd1
.l:i.xc7 leaves Bl ack a pawn up)
29 ... xd 5 .

I no longer had time for .l:!.xC7. I n

deed, I was a pawn down, and h ad to
readjust myself to staying a pawn
down, but with good bishop activity,
after 30 e2 .l:i.fe8 31 .l:i.xe8+ .l:i.xe8 3 2
'iitf2 gxf5 3 3 g 3 . I even later h ad

14 1

Why we L o s e a t C h e s s
slightly the better of a draw in the endgame, but after a mistake by either
side just before the time control, I
missed a chance of winning. We re
sume the position later in Test 14. 3 .
Much earlier, m y superficial original
intention had been 28 'iixd6 (C)
28 ... xd6 29 e7, but after 29 ... f7
Black holds, and may be better. My tac
tical vision was clearly not very good yet.

Test 12.3

Kid l i ngton


We TUsh through the early part of

the game, and zoom to my blunder. As
readers will h ave appreciated by now,
my tourn ament at Kidlin gton was a
treasure-house of my own mistakes.
Two years after my stroke, I was con
centrating on trying to regain my fit
ness as quickly as possible, doing lots of
walking, and playing lots of chess. I was
playing confidently, but unfortun ately
at that stage my brain was not working
properly, and I could not quickly an a
lyse the difference between good
moves and mistakes. I g ave my oppo
nents plenty of opportuniti es.
The game was a Nimzo Indian, 1 d4
l2Jf6 2 C4 e6 3 l2JC3 b4 4 f3 d 5 5 a3
xc3+ 6 bxc3 0-0 7 e3 e8 8 cxd 5 exd 5
9 d3 c5 10 l2Je2 b6 11 0-0 a6 12 l2Jg3
'iic 8 (all known so far) 13 f5 !? 'iiC 7 14
e1 l2Jc6 15 b2 l2Ja 5 16 e4 l2JC4 17
C1 g6 18 h3 Jtc8 19 g5 xh 3 20

14 2

xf6 'iif4 21 e5.

Play is level, in the sense th at it is

unlikely that either player will h ave a
clear advantage. The position is, how
ever, dynamic. A slip by either side can
end in a quick collapse. My biggest
problem is that if White can get his
queen to h6, then that i s the end. My
hope for salvation i s that White h as
several weaknesses in the centre.
I retreated my bishop, 21 ... Jte6. The
computer initially suggests that this
was a mistake, and that 2 1 ... l2Je3 would
h ave been good for Black, but retracts
when it becomes clear that 22 l2Je2
ih 6 23 'iic 1 would h ave been good for
White, as if 2 3 ... xg 2 24 l2Jg 3, winning
a piece. I feel sure th at I would h ave
seen this.
Pl ay continued with 2 2 a2 ec8 2 3
l2Je2 'iie 3+ 24 h1.
Both sides now h ave apparently set
up some dangerous threats, but noth
ing i s going on at the moment. If either
side breaks open the bal ance, it is the
aspiring attacker who loses the back
fire. E ssentially the only forcing m ove is

Te s t Twe l v e
Cl for White, with an exchange of
queens, but thi s is a simplification
rather than a m ating attack.

24 ... h5 was played, gIVing a flight

square for the king, but al so perh aps
weakening the pawn structure. There
are several alternatives over the next
few moves. Then came 2 5 h4 cxd4 26
cxd4. If Bl ack were to try for checkm ate
after 26 ... f2 ? 27 Cl ct:Je3 (27 ... i.h 3
28 ct:Jc3 .txg 2+ 29 'it'h 2 win s a piece for
White) 2 8 ct:JC3 xa2 29 xe 3 ! , then it
is actually White who gives mate.
So it was still time for quiet ma
noeuvring with 28 ... ct:Ja 5 27 ct:Jg3 'iVh6
28 .tg5 f8 29 d 3 .i:!.c4 30 f4.

Here, short o f time, I tried to

squeeze too much with 30 ... .i:!.ac8? (A),
threatening a skewer with ... .l:i.C3.
The more thematic move, 30 ... ct:Jc6!
(B), gives Black a substanti al edge. H e
does not lose time with a n unnecessary
developing move with the rook, but
immediately puts pressure on the cen
In the game, Cutmore tried to de
fend the third row with 31 .i:!.e3, but
Black h ad won time with 31 ... ct:Jc6. Of
course, both pl ayers were by now short
of time, and White h ad clear counter
play with 32 f5, as we'll see in Test 14.2.

Test 12.4

London Open


Continuing from Test 11.1, I pushed

the passed pawn, with 31 b7? (A), but,
of course, the bish op gets stuck in the
I asked myself, rhetorically, "when
was the l ast game I played this badly?"

14 3

Why we L o s e a t C h e s s
Certainly like all of us, I have lost some
games incredibly quickly, but, beyond
beginner and minor level, thi s would
generally be the case of one ridicu
lously bad move, completely misunder
standing wh at was going on, and per
h aps a couple of slight slips, based on a
misunderstanding of the big error.
Alas, after my illness I m anaged to lose
quite a few of these miniatures, some
of which h ave been published in this
book. This game, though, i s different,
and arguably worse. There must h ave
been well over h alf a dozen serious
mistakes in thi s game, and I am only
giving a few sample exercises, to avoid
m aking the reader feel bored. "Yes, I
see the point. Now give up the pawn on
b7, and your bishop will escape. You do
not h ave to make a meal out of this."
Statistically, there are likely to be
more blunders in a poor game of 50
moves th an in a g ame of 1 5 moves.
I was stunned by my previous blun
der in this g ame, giving away an ex
ch ange for nothing when I h ad a win
n i n g position. I was disorientated, but
then felt a sense of relief that I still h ad
the two united passed pawn s. If I h ad
thought more closely, I would h ave
seen that the pawn s were useless, and
that it is time to simplify and aim for
Indeed, 31 tiJb3! (B) was best. If
3 1 ... c2, there would h ave been a temp
tation to bring White's bishop into
proper pl ay, with 32 3l.c6 ? ! , but as


Emms pointed out much later,

3 2 ... tiJe4! h appily picks up the extra f
pawn . If, for example, 3 3 b7 3l.xf2+ 34
Wfl .i.g 3 35 tiJcl tiJxC5 36 .i.xd5 tiJd7,
followed by ... .i.f4. Surprisingly, my po
sition is not so dreadful after 32 Wfl,
and in some lines Bl ack has to be care
ful not to be worse.
In stead, after the dreadful 31 b7??
.i.e5, my bishop was stuck in the cor
ner, my knight was tied to the oppo
nent's passed pawn, and my two proud
passed pawn s could not break the
blockade on c7 and d6. A mess.

3 2 'it>f1 c2 3 3 tiJb3 tiJd7 is h eading

for a win for Bl ack. 33 ...tiJe4 is al so
good, and if 34 c6? tiJd2+ win s a queen .
In stead, 34 'it>f2 h olds the balance to
some extent, it not even being impor
tant whether White's c-pawn stays or
Then 34 c6 tiJC5 35 tiJc1 set up
passed pawn s for each pl ayer, and de
fended again st queening threats. We
continue the discussion next tim e in
Test 14.4.

T e s t T h i rt e e n
13. 1 Black to play

13.3 White to play

A) 3 1 ... 'iVxds
8) 3 1 ... f7
C) Something el se?

A) 34 e4
8) 34 ttJel
C) Something el se?

13.2 Black to play

13.4 Black to play

A) 3 2 .. .fxg 3
8) 3 2 .. .f3
C) Something el se?

A) 34 ... h 8
8) 34 .. JlxdS
C) Something el se?

14 5

Why we L o s e a t Ch e s s

Test 13.1

London League 2006

We h ave seen thi s position before in

Test 9.4. Now there was a whole cluster
of blunders. I was feeling groggy, and
my opponent was no doubt surprised
and nervous when suddenly reaching a
clear and good position after a blunder
by his 1 M opponent. In time troubl e,
both players made errors, and mine
was the most ludicrous of them .
I tried to regain the initiative with
26 ... f6 27 i.e3 'iYe4,


understandably enough, since I am

a piece down, and if I pl ay quietly, I will
lose. Even so, m aybe 27 .. .fS would h ave
been better, keeping my pawn s active.
He replied 28 ct:JC3 ! 'iYxb4 29 ct:Jd S,
making my pawn sn atch look ex
tremely inconsequential . His knight is
now on the best square on the board.

Ali i could do was to set up a pin on

the long diagonal, with 29 . . 'iYb7, with,
one would expect, my best chance be
ing th at h e would lose on time by
thinking too h ard again st very minoT
H ere the simplest way is 30 i.g 2,
g aining a tempo, in view of the threat
ened discovered check with ct:Jxf6+. If
30 ... e4, the straightforward plan is 3 1
l:txc2 bxc2 3 2 'iyxc2, wh en White i s
slightly ahead o n m aterial (bishop and
knight versus rook and pawn), and his
position i s totally secure, whereas
32 ... xdS ? 33 'iYc8+, and White will be
a full piece ahead.
It is difficult to know wh at was go
ing on in his mind some years later, but

Tes t Th irte e n
wh atever h appened, the line h e tried
was not a m assive blunder, contrary to
wh at I thought afterwards. He played:
30 1:[xc2 !? 30 bxc2 31 xC2!? ( 3 1
.ig2 o r 3 1 iLh 3 also gives a n edge).

now played extremely quickly in

time trouble, and a few seconds later, I
noticed that I could h ave taken the
knight: 31. ... xd5!? (A) could h ave
given a m assive shock to my opponent.
Or, on the oth er h and, White might
h ave analysed thi s right through, but
h ad missed that after 32 c8+ ct1f7 3 3
xa6 Black was no longer i n check, and
so h ad tim e to win the rook, staying a
bishop up, with 3 3 ... xh 1+.
It was years before it was pointed
out to me that, in fact, my opponent
could h ave forced a win with 32 xg 6+
f7 33 .id7+ ct1f8 34 .ih 6+ 1:[xh 6 3 5
xh 6+ g 7 3 6 C1, and White h as a
winning attack after 3 6 .. J::t a 8 3 7 1:[g l.
Or, since everyone loves a queen
sacrifice, we h ave 32 ... ct1d8 33 1:[c1 7
34 e4! a8 (34 ... xe4 35 1:[c8 m ate)
35 1:[c8+ li'xc8 36 .ixc8 ct1xc8 37 C4+
ct1b7 38 5+, winning the rook. Here

3 3 . . .li'a8 sets up a counterattack with

the threat of ... 1:[xa3+, but White de
flects the rook with 34 .ib6+! I:txb6 3 5
li'f5, with a catastrophe on c 8 or d 7 .
After the most spirited play, White
h as to find a coupl e of little combina
tion s, and a player might miss these in
time trouble. The computer did not
find the win in stantly, although it did
not think for all that lon g . The clear
implication is th at I was so shocked
that I h ad earlier given away a piece for
nothing, th at even a couple of years
later I do not know how deeply my op
ponent h ad examined all this. It would
h ave been more entertaining for the
spectators if I h ad seen th at I could
'safely' take the knight. And, of course,
there would h ave been a slight chance
for me to escape.
Play finished with 31 .. .'>itf7 (B) 32
1:[gl f5 33 .ixf5 gxf5 34 !iJxe7 li'e4 3 5
li'C7, wh en I resigned (1-0), perh aps a
touch prematurely ( 3 5 ... ct1e6 forces him
to show the win), but ultimately fair
enough. For the record, 32 ... g5 would
h ave been best, so th at White cannot
open the g-file, but Black would be
unlikely to h old for too long.

Test 13.2

B u ry St Ed m u nd s 2006
Not a particularly h appy day for a
significant birthday celebration .
Pl ay started steadily, with 1 e4 c5 2

14 7

Why we L o s e a t C h e s s
tiJf3 e 6 3 d3 d S 4 tiJbd2 tiJc6 S g 3 .id6 6
.ig2 tiJge7 7 0-0 0-0 8 l:te1 .iC7 9 'ii'e 2
b6 10 tiJf1 a s 11 .id 2 .ia6, equal, with
perh aps Black h aving slightly the better
chances. White's play is not very ag
gressive. Black's idea in the early part
of the opening was, with ... .id6 and
... tiJge7, to prevent White from gaining
a tempo with es, by avoiding ... tiJf6.
White readjusts, playing the stan
dard king side attacking plan with h4h s , and tiJf3-h2-g4, but without es. He
pl ayed 12 exd s exd s 13 h4 h6 14 tiJ1h2
'ii'd 7 lS h S , and Black might well start
to think th at White h as overpressed.
Then ls ...l:tae8 16 11ff1.

I played 16 ... fS, which I am not sure

th at I like now, and judging by my ear
lier notes, I did not really like at the
time. I am gaining space, with pawns
and pieces, but the pawn push al so
creates weaknesses for my opponent to
exploit, and these are arguably unnec
essary. I would prefer nowadays
16 ... .ic8 ! . keeping the pieces highly
coordinated, and stopping White even
thinking about .ih 3 .


My idea in the game was th at after

17 .ih3?! d4, as played, White has been
lured into a now redundant .ih 3, and
Black's ... d4 push i s all the better, now
th at the bishop i s not on g 2 .
In stead. 17 lie2 ! , doubling o n the e
file, looks uncomfortable, with Black
h aving pressure on the a6-fl diagonal.
No doubt it i s on thi s basis th at neither
pl ayer would h ave examined this
closely. Th ere is, however, the tacti cal
point that after 17 ... c4 18 l:.ael cxd3 19
cxd3 .ixd 3 ? (19 ... tiJc8 i s equal) 20 l:.xe7
White is suddenly ahead on m aterial .
17 .. .f4 18 g4! ag ain might look ugly for
White, but h e i s still slightly better.
Back to the gam e, with 18 tiJh4.
Now 18 ... 'iVds, trying to win White's h
pawn, m ay be slightly prem ature after
19 .ig2 'iff7 20 .ixc 6 ! tiJxc6 21 tiJg 6
l:.xel 2 2 l:txel .l:tb8. and White h as kept
equality. 50 I brought the bishop into
play with 18 ... .ieS ! ? 19 tiJ2f3 .if6 20
.if4, and now it was time to try to win
the h -pawn with 20 ...'ii'd s 2 1 .ig2 (if 2 1
.ie?, one possibility is t o call the oppo
nent's bluff with 21 ... l:tc8 ! ?) 21 ... 1Iff7 !.

Tes t Th irte e n
Thi s win s the pawn, but as so often
the h ard work is yet to come, as the
opponent has ch ances of gaining time.
H e tried 22 ct:\g6 ct:\xg6 2 3 hxg6 'Yi'xg6,
keeping his pieces better coordin ated.
Then 24 .id6 .ie7 2S .iC7.

When pl aying through this position

much earlier, I was highly critical of my
next move, 2S ... bS, as the game ended
up much l ater on with a weakened
queenside pawn structure. I preferred
in stead 2 S ... .id8, trying to keep the
pawn structure tight. My comments at
the time now seem over-elaborate, and
the move I played was fine. It is only
later that I started to m ake mistakes.
2S ... .id8 is playabl e, but after 26
.ixd8 l:txd8 n ct:\h4 'Yi'd6 28 'Yi'e2 White
has well-placed pieces, and it is far
from easy for Black to break down
White's pawn structure. Bl ack's miss
ing pawn i s only an h -pawn on the
edge, and it cannot contribute directly
to the central struggle, nor is it about
to help start a pawn advance. Black can
try .. .f4 in some lines, but g4 would be a
standard respon se, taking control of

the king side light squares.

I was also critical of my next few
moves, after 26 l:te2 .id8 27 .ixd8
l:txd8 28 l:tae1 l:td6, but I still keep an
edge, and surely there i s nothing
wrong with that. It is only if I am losing,
equal, or with a substantially smaller
edge th an before th at I h ave gone
wrong . Here I should be reasonably
h appy. White still h as active pieces, but
he is not breaking through, and Bl ack
still has the extra pawn . It was later
that I started to make mistakes.
White started by keeping control on
the e-file, with a knight exchange, 29
ct:\es ct:\xes 30 l:txes.

Looking at thi s diagram position, it

can become understandable why I be
came highly critical of my ... bS move,
and indeed if White were still to have a
pawn on h 2, he would have a clear
edge. But the position is as it is. White
has weakened his kingside, and so
30 ...f4 seem s highly sensible. Then 3 1
.ie4 'Yi'g4 (3 1...'Yi'f7 ! ? is al so worth con
sidering) 3 2 g2, and we h ave reached
the test position .

14 9

Why we L o s e a t C h e s s
passed pawn, not yet immediately im
pressive, but a threat later on . Al so
White h as to try to keep his isolated a
pawn safe.

3 2 ...f3? ( B ) was poorly judged, and,

as I admitted at the time, this was
"really the start of where I was about to
go wrong". In other words, other slips
are at best minor, possibly nonexistent.
" For som e reason ", I noted, "I felt I
could just batten down the f-file, but
now of course Bl ack has no real king
side advantage." I h ad blocked up the f
file, when in reality it was essential to
keep it open . Then 33 hl covered all
Black's threats, and now White h ad
lines for his own pieces, as we'll see
further in Test 1 3 .4.
32 ...jxg3 (A) 33 fxg 3 c8 ! ? in stead
leads to ten se pl ay. I am not convinced,
though, about the computer's sugges
tion that Black has a clear advantage.

Test 13.3


B ritis h League (4NCL)


As we saw in Test 9.3, Black h as,

very reason ably, just declined a draw
by repetition . He h as a supported


I decided it was about time to make

use of the long diagonal , with 29 .i.f3 . I
was surprised by his reply, 29 .. .fS?!,
and indeed with his ... g5 pawn push
later. There are certainly legitimate
minority attacks in the late middle
game, usually to create a weakness in
the m ajority pawn s, but thi s i s not one
of them . Before too long, White has a
good and safe passed pawn, although,
of course, Bl ack keeps his passed pawn
as well. Maybe equal ? I h ad expected
in stead 29 ... c5, followed by .. .'jf6, and I
would feel slightly uncomfortable as
White. At the very least, it should be
White, not Black, who h as to work h ard,
and spend time, to m ake use of his
king side pawn majority.
White continues his piece develop
ment, and Bl ack advances his pawn s,
with 30 b2 gS. I cannot keep the h alf
open files with 3 1 a3 ?, in view of
3 1 ... t2Jxe3 3 2 fxe 3 .l::t x e3 3 3 l::!. c 3 g4.

Te s t Th i r t e e n
Therefore I needed to add an extra
pawn defence on the kin g side, with 3 1
h3, and then came a counterattack
with 31 J:td6 32 g4! fxg4 33 hxg4.

White should now be at least equal,

but, of course, Bl ack h ad no need to
allow White a passed e-pawn .
We were now reaching time pres
sure, and few players can then play with
complete accuracy in complicated mid
dlegame position s. It is easy to see that
some of the next few moves are slightly
inaccurate, but it is far more difficult to
find the very best moves over the board.
Both players have to assess whether to
play a move instantly, or to think for 30
seconds, or whether to decide to think
longer at a critical position, at the risk of
hurrying later. It is much easier, of
course, to determine the best move if
the reader or annotator h as a computer,
a fresh cup of coffee, and, above all,
some time to think.
33 Wb7?! is recognisably a time
pressure move. Black sees a couple of
ideas, then argues that he will have no
time to assess or an alyse the next few

moves, and s o h e quickly m akes an un

obtrusive move, which he h opes h olds
the bal an ce, or m aybe more if the op
ponent does nothing more than hold
the bal ance.
33 ... 4Jf6 ! i s a more constructive
move. If White tries to attack, with 34
4Jb4?, a natural enough move, Black
hits back h ard with 34 ... ttxdl 3 5 ttxdl
4JC4+ 36 Wb3 4Je 5 .
White h as t o improve earlier. The
computer sugg ests 34 l::t c 2 l::t e d8 3 5
Wc3, not easy t o think of while being
short of time, as Black's last couple of
moves look almost random . The com
puter implies that the position is close
to deadlocked, and equal, since White
cannot break Black's pressure on the d
file, while Bl ack cannot create extra
pressure, and he will not want to re
lease the pressure him self. Black might
still create a slight edge, though, with
3 5 ... c5 36 l:tcdl 4Jc6 ! , and the knight is
back in play. If 3 7 l:tc2, hoping for a
repetition, Black keeps an edge with
3 7 ... .l::i. e 8.
Back to the game; it i s White's turn .


15 1

Why we L o s e a t Ch e s s
I sensed that Gran at's play was in
accurate, and felt I could even play for
a win. I played perh aps too h astily to
take the initiative, overlooking my op
ponent's reply. 34 e4 (A) looks good,
but thi s proved to be deceptive, as after
34 ... .l::i.f6! it was Black who was starting
to create pressure.
This I found to be an unexpectedly
difficult middlegame to an alyse, and
therefore an interesting test position, I
hope, for the reader. Of course, the test
is one of positional sense, rather th an
To cut down a couple of pages of
analysis, I give only a brief outline.
34 e4 was not a positional mistake.
It is equal . My mistake was only on
move 40, just before the time control .
In stead, 34 0.b4!? (C) 34 ... .l::i.f6 ! 3 5
i.xd5 J::t xf2+ 3 6 J::t c 2 i s equal .
And 34 0.el!? (B) 34 .. J:tf6 ! ? 3 5 J::t c 2
0.b4 3 6 110 0.d5 3 7 .l::i. c 2 0.b4 is a repe
tition . I was fascinated by the thought
that here White could h ave tried 3 7
l1Cl ! ?, a rem arkable zugzwang idea.

We h ave seen thi s position before,

15 2

but with Black to move, rather th an

White. The computer suggests that it is
better for White if it were Black to play
here rather th an White to play. If so,
this would h ave been a rem arkably
delicate zugzwang, with nine pieces
and pawn s on either side.
Every one of Bl ack's moves, except
only for 37 ... l1ef8 ! , leads to a worsening
of Black's play. The fin al result should
end up as a draw, after 3 8 i.xd3 ttxf2+
39 'it>a3 cxd5 40 J::t x d5 .l::i. e 8 41 .l::i. d 7+
'it>b8 42 ttC3, and if 42 ... .l::i. x e3 43 .l::i. d 8+
'it>b7 44 J::t d 7+ b8, with a draw by
repetition. 44 ... 'it>a6? would h ave been
far too ambitious, as White has a big
m ate threat after 45 .l::i. x e3 0.C4+ 46
'it>b3 0.xe3 47 0.d3 . But h ow m any
players would h ave seen all thi s in ad
Back to the g am e, and I delete a few
earlier criticisms on my next few
moves, with 3 5 i.g2 0.f4 3 6 0.xf4 .l::i. xf4
3 7 f3 'it>b7 38 'it>C3 0.b7.

Sensible so far, but the next move

was not good, as we'll soon see in Test

Te s t Th i r t e e n

Test 13.4


B u ry St Ed m u nd s 2006

In Test 1 3 . 2 we h ave seen a psycho

logically dangerous part of the game,
when a player feels h e has been in con
trol for most of the time, and then
something unexpected h appen s, close
to the time control. There is little
chance to readjust emotionally, no
time to wander around or take a tea,
no time just to chill out for five min
utes. The pl ayer has been h appily play
ing for a win, then a few seconds l ater,
he does not know whether his position
is better or worse, or whether he
should attack or look for a draw. Quite
often, a mistake in time trouble can
easily lead to a second, m aybe worse,
slip the next move or the second move.
This h appen s shortly in thi s game.
3 3 ... c4 was reason able. I can do
nothing immediately threatening, and
so I bring my more active pawns into
pl ay, and hope to open something up

for the bishop.

Then he gave a check, 34 .i.d S + , with
three reason able replies. There are two
opposing weaknesses when playing in
time shortage. Either the pl ayer some
times plays like a jelly, just dribbling
through to an endgame, and hoping to
regain composure after the time con
trol . Or, alternatively, it i s possible to
play the sharpest move, with the hope
that it i s best, and in any case one can
h ope that the opponent will m ake a
mistake under pressure.

But h ere, which plan is better?

There are two choices, sacrificing the
exch ange, and quietly moving the king .
This takes time t o reflect, and I got it
wrong .
Almost without thinking, I sacri
ficed the exch ange with 34 ...l:txd S??
(B). I felt sure that there was something
there for me, but there wasn't. After 3S
l:txd S cxd3 36 cxd 3, if 3 6 ....i.b7, then
White h as 37 'YWh S ! , easy enough to
miss, and he win s. The desperate sacri
fice 3 7 .. :xg 3+ 3 8 fxg 3 f2+ 39 f1 does
not work.


Why we L o s e a t C h e s s
I struggled o n with 36 ... b 4 3 7 h4
'iVxh4 38 gxh4 .i.xd3 39 l:txd4 .i.e2 40
l:tc1 l:tf6 41 l:tg4 g6 42 l:tc5, and played
a few more moves, but he did not lose
on time, and so he won .
Instead, 34 . .'JithB (A) should be a
draw, although perh aps slightly the
worse of the draw. For example, 35 l:te8
l:td8 36 l:txd8 l:txd8 37 .i.xf3 g s 3 8
'ifh s 'iVxh s 39 .i.xh s cxd3 4 0 cxd3 l:tc8
41 .i.g6 .i.b7, a small edge to White as
the bishop on g6 i s uncomfortable.


Alternatively, 35 l:t1e4 'iVg 6 36 'iVxf3

l:txf3 3 7 l:te8+ h 7 3 8 .i.g8+ Wh 8 39
.i.f7+ with perpetual . Or in this line,
3 s ... c8 3 6 dxc4 bxc4 3 7 l:txd4 c3 3 8 b3
IUS 39 l:txfs xfs 40 .i.e4 l:txd4 41
.i.xfs l:td1+ 42 'it>h 2 l:txh 1+ 43 'it>xh 1 gs,
and the position i s blocked after, for
example, 44 .i.e4 .i.e2, probably ending
up as an unusual draw.
There is still some life in the game, so
even the quiet and uninspiring move
can blossom into interesting play.

Te st F o u rt e e n
14. 1 Black to play

143 White to play

A) 3 S ... .l:i.g8
8) 3s .. .'!iJe4
C) Something el se?

A) 3 7 tLJg s
8) 3 7 l:txdS
C} Something el se?

14.2 Black to play

14.4 White to play

A) 3 S .. J:tg4
8) 3 S .l:i.dl+
C} Something el se?

A) 3 8 'it'xc2
8) 3 8 b8'i\i'
C} Something else?



Why we L o s e a t C h e s s
was fortunate th at I had an opportu
nity to keep some counterplay.

Test 14.1

British League (4NCL)


As we saw in Test 10.1, I am in trou

ble h ere, with several pawn weak
nesses, and some holes for my oppo
nent's pieces after my pawns h ave
been forced to advance. After some
thought, I found a way to keep my po
sition alive, giving away a couple of
pawn s, while creating some much
needed pawn activity.
I tried 30 ... a4 31 tLla s ( 3 1 tLlcS is
solid enough, and keeps an edge, but
does not attack any of the pawn s), and
now the pawn s went over the top with
3 1 .. . CS!.
It is an important defensive princi
pl e th at if your pawn s are in poor
health anyway, and some will drop be
fore too long, it is a good idea to en sure
th at when the opponent starts picking
up the pawns, you can create piece
complication s while the opponent
spends time chewing up the pawns. I


Obviously 32 dxcS ttxes allows

Black to recover the pawn, and Black's
pieces are now active, with the rook
being an annoyance. Therefore my op
ponent took the other way, with 32
ttxcS?! tLle6, but now Black suddenly
h as tactical resources.

'Therefore' is a dangerous word in

chess, and White should probably h ave
played 3 2 tLlc6 ! tte6 3 3 ttxcs instead.
Black no longer has the ... tLle6 option.
After 33 ... tLla6 34 .i:f.cl .ll d 7 Black is a
pawn down, while White's pieces are
solid. Perhaps there is no quick win for

Te s t Fo u rt e e n
White, but he should eventually prevail.
In stead, after the text move, the po
sition i s still alive.

If now 33 l:txb S ? tLlxd4, forking the

two rooks, and Black is ahead.
There is, it h as to be admitted, an
element of bluff in attempting to es
cape from such defen sive break-outs,
and to some extent I was relying on the
thought th at 33 l:txdS looks risky for
White, when previously he h ad such a
safe edge. Who wants to take risks
when you are already ahead? I could
not try to pretend th at I h ad calcul ated
through to a satisfactory conclusion,
but then there i s a reasonable probabil
ity that most players would be un able
to find a totally clear advantage either.
Here 33 lIxdS lIC7 (also maybe
33 ... lIa7) 34 tLlxfs lIcl+ 3 5 Wf2 lIc2+ 3 6
We3 lIxb2 would undoubtedly add t o a
touch of fear for White to think about,
and if 37 lIh 3 tLlC7 38 lIxh6+ g8 39
lIcS i.xfs 40 l:txc7 l:txa2, when White,
although maybe better, has to deal with
the implications of two connected
passed pawns for the opponent. In prac-

tical terms, considering Black's earlier

dreadful position, this is a success.
So White tried instead 33 tLlxfS lIa7
34 lIxbS tLlgs, when arguably White's
position i s even more critical .

White has two extra pawn s, cer

tainly, but both his rooks and his
knights are under attack, or potentially
under attack. I was in with a chance,
whereas a few moves previously, I h ad
felt my position was close to resign
Then in the h otel, the alarm bell
went. Thi s for m any years has been
almost a tradition al entertainment on
the l ast days of various tournaments in
England, m aybe through m alice. There
h ave been several other occasion s in
recent years in the 4NCL, but these, as
far as I know, h ave h ad innocent expl a
nation s. In a big hotel, with so m any
people wandering around, especially
after most of the g ames are finish ed,
accidents h appen. The story on this
occasion i s th at one of the players
propped his elbow on the bar on one of
the counter ledges, which h ad been


Why we L o s e a t C h e s s
shut, and the alarm was set off. I n
other words, a complete accident.
There was not much chance to con
centrate with a noisy fire alarm. When
all the players had to escape from a
crowded room I am sure that many
players would have been extremely dis
rupted during the last few minutes of
the time control. But what can be done?
Upon resumption, I completely lost
my thread of thought, and lost even
before move 40, starting with the
blunder 3 5 1'!.f4 ttJe4?? (B), overlooking
that h e could take the central pawn
after 36 ttJd6 (or 3 6 1'!.xdS immediately)
36 ... a6? (a quick loss, but 36 ... ttJxd6
37 exd6 ..te6 38 ttJb7 only slightly de
l ays the result) 37 1'!.xd 5 ..td3?? (another
blunder) 38 ttJxe4 1-0.
This was sickening. I could h ave h ad
no complaints if I h ad lost as a result of
my poor earlier play, but losing as a
result of a fire alarm, when I was back
in the game after much effort ... Well,
this was tough.
After the bell, 35.. Jlg8! (A) was, of
course, much better.

Then 3 6 ttJg 3 (Black was threaten

ing ... ttJh 3+) 36 ... ttJe6 37 .l:th4 .l:tC7, and I
am reasonably certain that I h ad con
sidered this position in advance, but
the chaos of evacuation disrupted me,
and my confidence evaporated. A pos
sible and likely drawing line would be
38 .l:txdS .l:tc1+ 39 'it>f2 .l:tc2+ 40 'it>e3
1'!.xb2 41 ttJfS 1'!.gxg 2 42 1'!.xh 6+ '.t>g 8 43
ttJe7+ 'it>g7 44 ttJfS+, repeatin g . The
computer suggests 43 ttJg 3, and if
43 ... 1'!.xa2, White h as a good escape
route with 44 1'!.d6 ! , then 'it>e4-ds, but
43 ... .l:!.cc2 ! leads to a draw.
There is much to be explored in
analysis. In practical term s, White
would h ave h ad to find good moves in
a complicated position up to move 40.

Test 14.2

Kid l i ngton


From Test 12.3, clearly the position

is about to open up, on both sides. Play
continued 3 2 ... 1'!.xd4 3 3 bl xf5 34


Te s t Fo u rt e e n
40 .l:i.e4 .l:i.h 1+ 4 1 'it'g 2 ltcg 1+ i s a draw.
After my panic, Cutmore played the
simple and natural 36 .i.f6!, and I was
in trouble. 36 ... .l:i.g6 37 e6 was unentic
ing for Black: for instance, 37 ... tiJe7 3 8
exf7+ xf7 39 .l:i.xe7 ltc1+ 4 0 h 2 xf6
41 xdS+ 'it'h 8 42 Vixh S+ 'it'g 8 43 VidS+
'it'h 8 44 Vid8+ l:tg8 4S l:th 7+, picking up
the queen .
So, 3 6 tiJd4 3 7 Vixh S .l:i.g7.

tiJxfS gxfs 3 S VixfS.


I pl ayed 3 S .l:i.g4?? (A). In my notes, I

said that, "as so often, a panicky reac
tion, short of time, m akes things worse.
My last few moves appear to h ave been
correct, but my king i s suddenly ex
posed. Oh dear, h elp, bring another
piece closer to the king, and hope for
the best - but thi s is not good chess
thinking ! I h ave played attacking de
fence, and should carry on with this,
forcing White to go on the defensive.
That way I can keep the balance and
even more."
I should h ave played 35 1:.dl+ (B)
3 6 'it'h 2 tiJd4, g aining a tempo to bring
extra pieces into play, the knight, and
the rook on c8. It still looks fragile, but
White's queen must move, and cannot
find a good way to attack the rook on
d1. Then 3 7 Vif2 .l:i.cc1 3 8 .l:i.e1 .l:i.xe1 39
.i.xc1 l:txc1 40 Vixd4 Vie7 41 g 3 gives
an advantage to Black, but it will be
diffi cult to force the extra pawn to a
win .
Instead, 3 5 tiJe7 ( C ) 3 6 .i.xe7 l:tC1+
3 7 'it'h 2 Vixe7 3 8 g3 .l:i.dd1 3 9 xh S d4

My position was close to resignable,

but my opponent made two mistakes
in a row, in time trouble, starting with
38 .l:i.g3?, overlooking my reply, or at the
very least overlooking my follow-up.
38 .i.xg 7 ! was simple enough, then
38 .. .'iVxg 7 39 'it'h 2 .l:i.c6 40 l:tg 3 .l:i.g6 41
l:txg 6 fxg 6 42 Vig s, keeping an extra
exch ange, with White's queen and rook
being highly active. Straightforward
chess, but unspectacul ar, and my op
ponent was presum ably wanting a
quick checkmate, or a win of the queen .
I squirmed on with 38 tiJfS!, some
h ow holding on with sticking tape. If 39
VixfS .l:i.xg 3 40 'iVh s .l:i.g 7, level. With
care, White can still keep a highly sig...

15 9

Why we L o s e a t C h e s s
nificant edge, though, with, for exam
ple, 3 9 .l:i.f3 lbd4 40 .l:i.d3 lbe6 41 ..txg 7
'iYxg 7 42 .l:i.f2 .
Instead, however dire my position
was, I suddenly had counterplay after
39 .l:i.gS?? .l:i.C1+ 40 'iit h 2 "iVCS.

46 .l:i.g 2 is a draw by perpetual . There is

nothing more after 46 ... 'iWf5+? 47 xf5
lbxf5 48 .l:i.f2.

Test 14.3

C Crouch P Roberson

B ritis h League (4NCL)

I h ave at least a draw! N aturally, I

could not expect m ore. However, after
41 .l:i.xg7+ lbxg7 42 gs "iVg1+ 43 'iit h 3
.l:i.c3+, there was the biggest oversight
of all, with

44 'iit g 4?? 'iWd1+ 45 'iitf4 d4 mate

An embarrassing result, even for the
In stead, 44 g3 'iVh 1+ 45 .l:i.h 2 f1+



By the end of play in Test 12.2 Bl ack

was two pawn s up, but he h ad isolated
doubled pawn s on the f-file, and it
looked more than likely th at at least
one of his queenside pawn s could fall,
with his minor pieces being under
threat, and therefore being un able to
cover his pawn s. I felt relieved that I
h ad recovered from my earlier mistake,
and I was hoping that I could try to play
for an edge .

Black immediately returned the

doubled pawn, with 33 ...f4 34 i.xf4
lbe6. Thi s is not the only line, but
seem s to be the most sensible. The
computer suggests 3 3 .. J::t x e2+ 34 Wxe2
i.C4+ 35 'iitf2 lbd5, 'equal ', but if White
can regroup his pieces to more effec-

Te s t Fo u rt e e n
tive squares, Black would h ave to
careful after 3 6 as f4 37 h 2 b S
tDd2 ! .
I n the m ain line, White played
.i.e3. Black's bishop i s now under
tack again.


In my contemporary notes, I implied

that Black, with his extra pawn, should
be comfortably equal, and that h e
should h ave chances o f being better.
Therefore I g ave 35 J:td8 a query, but
defending the bishop is a natural
enough idea, with chances of h olding
the draw. In stead, any bishop move
would allow a quick l::t d 6, with chances
of chewing up Black's queenside
pawn s.
It turns out th at Black i s slightly
worse in other lines, except for one, so
perh aps .. J:td8 is a ' ? ! ' move rather than
a '?' move.
The best line for Black i s 3S ... tDC7 ! ,
and then i f 3 6 .if4 tDe6, repeatin g .
White h as choices, for example, 3 6 as
bxas 3 7 .ixcS fS ! , but no real advan
In stead, 3S ... xf3 ? ! 3 6 gxf3 (best)

3 6 . . . as 3 7 lId6 tDg s 3 8 C4+ 'it>f8 3 9

f1 ! gives White a clear edge. The
point of the interpolated check is soon
seen after 39 .. J:tb8 ? ! 40 l::t x b6 l::t x b6 41
xcS+, and Black's queenside col
l apses.
Another attempted spirited line is
3S ... .i.b3 36 l::t d 6 tDC7 37 .l::!. x b6 .l::!. x e3 38
.l::!.b 8+ (38 'it>xe 3 ? ? tDdS+) 3 8 ... lIe8 39
l::t x b3, but Black's four isol ated pawn s
will be far less effective than White's
two connected sets of pawn pairs. In
this line, 3 6 ... .i.xa4 3 7 .i.C4 Wf7 3 8 .l::!. x b6
bS 39 .i.xbs axbs 40 .l::!. xb S fS 41 b3 i s
al so difficult t o defend for Black.
Back in the g ame, I tried 36 xa6,
h aving noted that 3 6 lhd S ? lhdS 3 7
.i.C4 lId6 3 8 .if4 l:!.c6 3 9 .i.ds <j;f7 40
.i.xc6 tDxf4 gives no advantage for
White, and leaves him a pawn down .
So thi s sacrifice needs to be shelved.


Now it i s time for a double-blunder,

close, of course, to the time control . He
was very short of time, a few moves to
m ake in the last minute, while I h ad a
few minutes to spare. I never particu
larly enjoy time scrambles, seeing this

16 1

Why we L o s e a t C h e s s
a s the most stressful part of the game.
These days, I h ave the fear of my eye
sight suddenly deteriorating, or possi
ble dizziness when short of time, and
so I felt I needed to pace my play. How
ever, if you h ave five minutes left, it is
sometimes useful to use part of that
time to find the best move.
Here he m ade a tactical blunder
with 36 ...f5?? Even with plenty of time
on the clock, it is difficult to find the
most accurate defen sive move, as if he
moves his bishop, the rook exch ange is
favourable for White, but if he does not
move the bishop, his pieces remain
tied up.
The computer suggestion, 36 ... h S ! ,
looks irrelevant at first, but compare
3 6 ... 'i.t>g7 37 g4!, and the point is quickly
m ade. Black needs to avoid being
pressed down on his king side pawn s.
After 3 6 ... hS, the computer gives the
line 3 7 b4 ii.f8 3 8 as 'i.t>f7 3 9 ii.d3 cxb4
40 axb6, with perhaps a slight edge for
What, however, i s the problem with
the other pawn move, in the game? The
answer i s that Black has m ade a slight
weakness in his pawn structure, the f
pawn no longer defending the e S - or
g s -squares. I h ad looked at ttxds on
move 36, a move earlier. I did not at
tempt to analyse the position a move
later. I h ad already spent time analys
ing it the previous move, and did not
want to eat up extra time on the clock.
As a result, I missed a chance of win

16 2

I h ad the chance of playing 3 7 d5!

(B) 3 7 ...ttxds 3 8 ii.C4 ttd6 ( 3 8 .. .tbc7 39
ii.f4!) 3 9 .tf4 .1:.c6 40 tbg S ! , which i s
now winning.
Maybe the moral i s th at you should
not completely abandon the sacrifice
you h ave rejected. Who knows, a little
pawn tweak, or a small, quiet piece
move, could alter one's perspective
37 tbg5?? (A) was an awful move
anyway. Black h as now found good
squares for his pieces, with 37 ... tbd4!.
After 3 8 ii.f4 h6 39 tbf3, I suddenly no
ticed that 3 9 ... tta8 ! was strong for
Fortunately, a second or two l ater,
another piece l anded, with 39 ... tbxf3? I
finally m ade the exch ange sacrifice
with 40 ttxd 5 ! ttxd 5 41 ii.C4 tbe5. Then
42 jLxd 5+ 'i.t>f8 43 'i.t>e3, and I h ad h opes
of keeping a slight endg am e edge with
two bishops versus bishop and knight.
It was not to be, and he pl ayed accu
rately with 43 ... 'i.t>e7 44 b3 'i.t>d6 45 ii.g8
'i.t>e7 46 ii.g3 tbc6 47 ii.e1 'i.t>d6 48 ii.g3+
'i.t>d7 49 .tf4 tbe5 50 ii.h2 tbc6 5 1 ii.C4

Te s t Fo u rt e e n
d4+ 52 e2 e5 5 3 gl d4 54 e3
xe3 55 xe3 e7 56 g8 f6 57
d3 tbe7 58 c4 e5 59 a6 d6 60
b7 tbg6 61 C4 tbf4 62 b5 C7 63
f3 tbe6 64 C4 d6 65 b7 tbf4 66
b4 cxb4 67 xb4 tbd 3+ 68 b5 C7 69
d5 tbf4 70 f3 tbe6 71 a8 tbf4 72
g4 fxg4 7 3 hxg4 tbg6 74 e4 74 ... tbe5
75 f5 tbf7 76 e6 tbe5 77 b4 d6
78 c8 tbd 3+ 79 WC4 tbC5 80 b5 e5
81 as bxa 5 82 xc5 f4 Yz-Yz.
A frightful number of years ago, I
won the British Under-18 Ch ampion
ship with a string of minor endgames,
solely on the basi s that I was able to
understand such endgames better
than my opponents, who m ade mis
takes. Nowadays the young defenders
seem to pl ay more accurately!

. . .e7 would tran spose i f he tries . . .C7

then ... e6. There was, however, th e
more worrying prospect of 3 5 ... d4 fol
lowed by ... d3, and a severe danger of
being zugzwanged.
Therefore I was relieved that h e
played quietly, a n d I saw the opportu
nity of grabbing Bl ack's advanced c
pawn, with 38 e2 f8 37 d2 e7.
Excellent, I thought, take the pawn, 'job

Test 14.4

London Open


Following on from Test 12.4, Black

played 35 ... C7. 3 5 ... f8 followed by

Except th at the highly natural 38

XC2?? (A) was a serious mistake. After
38 ... d6!,

Black is not only sitting next to the

passed c-pawn (he cannot, of course,


Why w e L o s e a t C h e s s
take i t yet, a s White would reply with
b8+), but more significantly, the king
i s now protecting the ds-pawn, so that
any breakout sacrifice with b8 fol
lowed by C7 no longer allows White to
win Black's ds-pawn .
My last legitimate chance to save
the game was 38 b8+! (B) 38 ... xb8
39 C7 lLle4+ 40 'iii x c2 xC7 41 xdS
lLlxf2, and although Black i s a pawn up,
his position i s totally unwinnable, with
no passed pawn s, no pawn weak
nesses, and with opposite-col oured
bishops, leaving no chan ces of a
zugzwang .
In the game, after Black played
38 ... d6, White somehow convinced
himself that he was a pawn up, and
therefore had no reason to worry about
losing. A big mistake. It was time to
grovel by sacrificing the passed pawn on
the seventh, bringing the bishop back
into play, although he still h as to work
hard to try to set up a draw.


In stead, 39 lLla2? lLla6 40 lLlC3 lLlb8

was by now a clear win for Black, even
though White has two advanced

I tried 41 lLlbS+ xc6 42 lLlxC7 'iiix C7,

with the expectation th at I was going
to resign in a few m oves. After all,
White's bishop and pawn are immo
bile, and Black's knight h as no incen
tive to move, so in effect Bl ack is a
pawn up in a king and pawn endg ame.
We'll see wh at h appened l ater in
Test l S . 3 , with a surprising result.

Te st F i ft e e n

15. 1 White to play

15.3 Black to play

A) 3 9 'it>d4
B) 39 l:.h l
C) Something el se?

A) 4B .. .'JbB
B) 4B .. .fxg4
C) Something else?

15.2 Black to play

15.4 Black to play

A) 46 ... e4
B) 46 ... 'it>fB
C) Something el se?

A) SO . .. .l:td2+
B) SO .. .lDf2+
C) Something el se?


Why we L o s e a t C h e s s

Test 3.5.3.


Britis h League (4NC L)


Following on from Test 13.3, I blun

dered h ere, losing a pawn with 39
'it>d4? (A). I was con scious that my other
pieces were not very active, and so I
wanted to bring my king into play,
overlooking th at I would lose a pawn
with 39 ... .i:!.a8 40 .i:!.a1 ttJc5. Black
threaten s a knight fork on b 3 .
Clearly i f I h ad had the chance of
thinking before the time control, I
would h ave seen this tactic in advance.
It would h ave taken much more time to
find a wholly convincing way to equal
ize, though, Black's pieces being better
pl aced than White's.
39 1ihl! (B) 39 ... .i:!.f6 40 l:i.h 2 ! i s best,
putting the rooks into wh at seem s to
be a blind corner on the h-file, but
White is al so doubling on the only ef
fective file. If, for example, 40 ...ttJC5 41
'uch 1 'uee6 42 l:i.a1, and Bl ack can m ake
no real progress. Bl ack could try in stead


41 ... ttJxa4+ 42 'it>b4 ttJC5 43 'uxh 6 'uxh 6

44 .i:!.xh 6 ttJd3+ 45 'it>C3 ttJf4 46 ..ttl, but
White should be able to hold.
All this was quite a surprise for me
when I resumed my analysis while writ
ing this book. I h ad assumed that while I
knew that my 40th move was a losing
mistake, I was in a bad shape already,
but it turns out that I was equal .
In stead, after the pawn dropped,
Bl ack won with 41 'it>e3 ttJxa4 42 ..tf1?!
(a lack of confidence, following my er
ror; the only chance was to attack, with
42 ,Uh 1 .i:!.f6 43 e5 .i:!.e6 44 f4 gxf4+ 45
'it>xf4 'uf8+ 46 'it>g 3 b5 47 'uh 5, when
Bl ack should be winning, but he would
h ave h ad to work h ard) 42 ... b5 43 ..te2
.i:!.af8 44 .i:!.h1 .i:!.8f6 45 .i:!.h3 .i:!.e6 46 ..td1
ttJb2 47 ..tb3 .i:!.d6 48 'it>e2 .i:!.d 3 49 ..te6
.i:!.xe4+ 50 fxe4 .i:!.xh 3 51 ..tf5 ttJc4 52
.i:!.a7+ 'it>b6 0-1.

Test 1 5.2

B ritish League (4NCL)


Te s t Fift e e n
White is still better after mediocre
play by both sides. I h ad missed a clear
win a few moves ago, as we saw in Test
11.3, and now he is on top. He has a
clearly better queen and rook ending. It
is easy enough for him to create a well
guarded passed pawn o n the queen
side, whereas Black's extra pawn on the
kingside m ay be vaguely useful, but is
unlikely to do anything con structive.
My opponent immediately threw
away h alf his advantage, by offering
the exchange of rooks, with 3 2 1:td6?,
in stead of trying to strang Ie the queen
side with 3 2 'iWdS ! .
Indeed, White soon lost the whole of
his advantage, as Black now fully
equalized with 32 ...f3 33 1:txg6+ hxg6
34 'iWd S.

We were approaching the time con

trol, and I sensed th at the psychological
initiative was now slightly in my fa
vour, and so tried 34 ... 'iWf4+, instead of
the simple drawing line 34 .. .fxg 2 3 S
'it>xg 2 'iWg S+ 3 6 'it>f1 'iWC1+.
Then came 35 g3 'iWd4, on which I
commented: "The best move, but at the

time I thought i t was a blunder... 3 6

'ii'xf3 ... a s I h ad missed this." Still . it
wasn't too bad as I could immediately
regain the pawn with 36 .. :i!VXC4. It was
disconcerting, all the same, th at I h ad
not seen his simple capture.

The position i s now equal, but I was

still h oping to squeeze an edge, out of
sheer pride. Or perh aps punch-drunk
chess. We quickly reached the time
control with 3 7 'it>g2 a s 38 bxa s bxa s
39 'iWe3 'iWd S+ 40 'it>h2 'it>f6. As I noted
earlier, "A wrong change of direction, ...
and eventually I h ad to find a way to
return the king to g 7 . I h ad vaguely
thought that I could pressurize White
by trying to exch ange the queens, but
White's queen was too active to allow
Black to attack."
There i s the further danger that
White has the possible threat of the
outside passed pawn, if given the
chance of h4, g4, h S , and then Black's
king would be seriously displaced if
White could create a passed pawn .
Pl ay continued 41 'iWb6+ 'it>e7?! (still
the wrong direction) 42 h4.


Why we L o s e a t C h e s s
king stays in the centre a little longer)
46 g4, and I still h ave to play accurately.

There is still some life in the g ame.

If 42 .. .'i!t'd4 43 'it'xd4 exd4 44 'iit> g 2 'iit> e 6
45 g4 f5, and there are several ways, as
far as I could see, to draw, one being 46
g5 f4 47 f3 (but not 47 'iit>f3 ? ? d3, and
White is in zugzwang) 47 .. .'it>d5 48 f2
a4 49 e2 'it>e6 50 d3 d5 5 1 e2,
when neither side can m ake progress.
I also felt vaguely uncomfortable af
ter 42 ... 'iit>f8 43 'it'f6, but maybe this was
a safe level draw.
My 42 ... d2 was al so possible,

but now it was his turn to think

about playing for m ore, with 43 'it'b7+
'iit> e 6 44 c6+ 'it>e7 45 'iit> g 2 'it'el
(45 ... 'it>f8 ? 46 iVC5+ loses a pawn, so the


The position is, of course, equal. The

only likely way of losing such a posi
tion, with either colour, would either
be to overpress, trying for a win, or by a
blunder. Somehow, though, both play
ers managed to go wrong .
I played t h e unnecessary 4 6. . .e4?
(A), which vaguely m akes possible
some dangerous checks on f3, but also
m akes weaknesses on Black's own
It was much more precise to pl ay
46 ... f8 (B), covering the opponent's
potential passed pawn . If 47 hS g xh s
48 gxh s 'iit> g 7 49 h6+ 'iit>h 7 S O 'it'f6, and
the position might look scary, but Black
is fine after 50 .. :e4+ 51 f1, and then
a perpetual after 51 ... 'iVb1+. Here 5 1
h 2 keeps the play going, but the
pawn ending is a draw after 51 .. .'Vf4+
5 2 'it'xf4 exf4 53 'iit>h 3 .
Buckley's queen move, 4 7 iVC4?!, was
a square too far. He missed his big
chance to create some pressure, with 47
"1i"C5+!, which forces the king to an un-

Te s t Fift e e n
desirable square. There would be some
uncomfortable zugzwangs later on.
Black's queen looks good at first, and I
was relying on this, but while the queen
is on an advanced square, it h as little
impact further behind. There is only one
square, on e1, to protect both the e4and a5-pawns, and this is the reason
why there could be zugzwang problems.

Here 47 ... e8 ? ! 48 g5 d7 ? ! 49 "iVf8

e6 50 'ilVe8+ f5 5 1 'iVxf7+ g4 5 2
"iVxg 6 'iVe2 5 3 "iVe6+ xh4 54 'iVf5 will
sooner or later queen a pawn for
White. Therefore Black cannot move
the king again h ere, and so has to rely
on a queen m ove, 48 .. :iVe2 ! . Then 49
'iVxa5 'ilVg4+ 50 f1 'iVxh4. Maybe it
shoul d eventually end up as a draw
(though there is an element of doubt),
but Black h as to work h ard. There
would be much m anoeuvring by White
to try to squeeze a slight edge after 5 1
'iVe 5+ d7 5 2 'ilVf6 3 + 5 3 e 1 'ib3
54 'iVd4+ c7 5 5 a5 'iVc2. Of course, if I
h ad played precisely on move 46, it
would h ave been a very simple draw.
Could Black play better with

4 7. . .f6, with a draw i f White allows

the king to the corner? I was worried
about 48 'iVf8 'iVe2 ! 49 g 5+ e6 50
'iVe8+ d6 51 'iVd8+ c6 ( 5 1 ... e6 ?? 52
"iVf6+, winning a pawn with g ain of
tempo) 52 'iVxa5 'iVg4+ 53 f1 'iVxh4,
but it is likely still to be a draw.
In the game, I played 47 ... f8 with a
sigh of relief. He tried 48 'iVC5+, a move
too late. Then 48 ... g8 49 'ilVe3 .

I n o w saw his idea, th at with the

obvious queen exch ange, White ends
up winning after 49 .. :i:Yxe 3 ? ? 50 fxe3 f6
5 1 g 5 fxg 5 5 2 hxg 5 f7 5 3 g 3 e6 54
f4 d5 5 5 a4. But one of the most
basic position al rules of the endgame is
never to be absolutely certain of what
is going on in a pawn endgame, as
there i s so often such a small gap be
tween winning, drawing and losing.
There i s more latitude in piece end
g ames, and so I kept the queen s on,
49 . "iiYa 1. I doubt whether I could h ave
concealed th at I was starting to feel
tired, and so instead of offering a draw,
he made a last winning try with 50 h 5

16 9

Why we L o s e a t C h e s s
His sister, Melanie, said th at David ever
so often completely forgot about his
time, even in the last few minutes.

Test 1 5.3

London Open

And now he was the one t o blunder,

with 51 h6?? ( 5 1 hxg 6 fxg 6 is, of course,
a draw) 51 ... g5, and Bl ack soon pock
eted a pawn . After 5 2 f3 exf3+ 53 xf3
e6 54 h3 h7, there was an even
bigger surpri se. I was about to be a
pawn up, but it was not clear th at I
genuinely h ad a winning advantage.
My opponent was thinking hard, while
I was trying to think of ways of making
progress, and gearing up to h aving to
play ag ainst quick moves from my op
ponent, short of time. But then he just
thought and thought, and lost on time.
I was confused. H ad h e m ade move
60 or not? Had I got the time control
wrong ? H ad I misread the digital clock?
Quite possible when I could see with
only h alf an eye. Or, most unlikely of
all, did he not notice that he h ad lost on
time? After a brief gap, I pointed to the
loss on time, still slightly confused.
Here, the traditional fl ag-fall would
h ave been clearer, but with a digital
clock, I h ad so m any bits and pieces to
look at. To my relief, David agreed that
this was, indeed, a loss on time, so 0-1.



Following on from Test 14.4, all I

could do was to move my king, plus a
few weakening pawn moves. It i s a
simple win for Bl ack, the sort of exer
cise th at you would use to teach chil
dren . All that Black needs to do is to
concentrate on the king and passed
pawn, and White's king will gradually
h ave to give way, because of zugzwang .
Then after a few more zugzwan g s, the
d-pawn will be promoted, with a win
for Black.
I played a couple of token moves,
waiting for the opponent to demon
strate th at he showed h e knew h ow to
do it. I did not expect this to last for
lon g . I centralized with 43 d3 d6 44
d4. There was still a glimmer of hope

Tes t Fift e e n
when he pl ayed 44 ... tt'lc6+?!, still win
n i n g , but not so direct, since Black will
h ave to play the knight back to b8 at
some stage. The most direct plan for
Black would h ave been 44 .. .f6. The
computer still initially suggests that
White is better, but thi s is clearly mis
taken. After, for example, 45 'It>d3 h5 46
'It>d4 'It>c6 47 f4 h4 48 'It>e3 'It>C5 49 'It>d3
d4 50 'It>e4 'It>C4 51 'it>f3 'It>c3 it i s easy to
see that the pawn will queen .
After 45 'It>e3 f5, I could h ave con
tinued pl aying rather longer with 46
b8+ ! ? tt'lxb8 47 'It>f4 tt'ld7 48 'It>g 5, al
though Bl ack eventually win s with
... 'it>e 5 and ... tt'lf6. I pl ayed the opponent
h ere, rather than the board. I reasoned
th at if h e wanted to 'activate' the
knight, rather than squash the passed
pawn, then let him. Amazingly, it
worked. After 46 f4 'It>C5 47 g3 'It>C4 48
g4, Bl ack blundered.

( B ) 4 9 hxg4, and White will soon h ave

two passed pawn s, with enough coun
terplay to equalize. Play suddenly fin
ished 4 9. . .'lt>c5 5 0 f 5 gxf5 51 gxf5 'It>d6
5 2 f6 'It>e6 5 3 b8 tt'lxb8 54 f7 'It>xf7 5 5
i.. x d 5+ Yz-Yz.
There h ave been m any dreadful
g ames in thi s collection, and I am not
hiding losses, even when I did not h ave
the physical strength to be able to find
good m oves. My pl ay was dreadful that
evening, and if he did not win it, then
there is criticism of his own play. If thi s
seem s ungenerous t o m y opponent,
who achieved a draw again st a much
higher-graded opponent, the point is
th at he had the chance of beating an
1 M opponent on his pl ate, and he
missed it.

Test 1 5 .4

B u ry St Ed m u nds

Most reasonably sen sible moves

win, the most thematic being 48 tt'lb8!
(A), in effect a simple king and pawn
Astonishingly, he tried 48...fxg4??



We h ave just reached the time con

trol, and I had played dreadfully, as we

17 1

Why we L o s e a t C h e s s
saw i n Test 11.4. My position i s , of
course, losing . At the time, I noted that
"maybe I did not quite h ave the
strength to aim convincingly for the
draw". To which we can add that in a
weekend tournament, with five long
games, sometimes even six, in two
and-a-half days, even the young and
physically stronger players can become
tired, and make mistakes. I used to play
m any weekenders, from the late 1970s
onwards. Nowadays, it becomes more
What i s interesting i s that for most
of the finish, my opponent, a grand
m aster and the younger player, actu
ally played worse than me in the next
part of the game. Thi s can be proved,
and is not a matter of opinion. Just
now, he has a clearly winning position,
but at a l ater stage I had the chance of
forcing a draw. The trouble is that I h ad
only a couple of minutes before the
flag-fall, and I missed the opportunity.
Pl ay continued 3 7 :ta7?! (the imme
diate 3 7 xf5 gxf5 3 8 ,U,a7 is perh aps
more accurate)

17 2

3 7 ...g8 ( 3 7 .. .'ib1+ 3 8 'it'h 2 'it'g8 39

f3 ! should nevertheless eventually win
for White; h ere 39 'iixf6?? '*Ifb8+ is
Black's trap) 38 "i\Vxf5 gxf5 39 'it'h2 tLlh7
40 tLlf3 'iit g 7 41 tLle5 tLlg5 42 f3 :th8+ 43
'it'g1 :tf8 (back again).

Now 44 'it'f1?! started to move in the

wrong direction. The king moves into
open play, and Black h as ch ances of
giving a check with the rook.
In quickplay finishes, the subtleties
of a delicate zugzwang play m ay be
missed. Correct was 44 :tb7 ! , and if
44...:ta8, then 45 f4 tLle4 46 'u'xf7+ g 8
47 :te7 .l:!.a1+ 48 'it'h 2 ! (which explain s
why White avoids f1) 4 8 ....l:!. e 1 49
:txe6 .l::!. x e3 50 d5, and White h as two
extra pawn s, while his king is safe.
Pert's choice gives Black much better
chances of h olding.
Then came 44...g8 45 e2 :tb8.
Now White's king i s exposed, and Black
h as drawing chances. He played 46 f4
to my surprise, giving Bl ack's knight an
active square, but how else is White to
make progress? Maybe starting again
with 46 f2.

Tes t Fift e e n
I counterattacked, or at least aimed
for a perpetual, with 46 tiJe4 47 I1xf7
I1b2+ 48 'it>el 11bl+ 49 'it>e2 11b2+ 50 d3.

Now I missed a draw, by playing

50 11d2+? (A).
There was a draw with 5o tiJf2+!!
(B) 5 1 'it>C4 ( 5 1 'it>C3 tiJd1+ 52 'it>d3 tiJf2+)
51 ... tiJg4 52 e7 tiJxe3+ 53 'it>C5 tiJxg 2
54 xe6 tiJxf4, which i s quite a mes
merizing knight jump with less th an
ten minutes to play in the rest of the
game. At the very worst, Black would be


able t o sacrifice the knight for White's

fin al pawn, and Black can draw the
rook and knight against rook endgame
with care - except that in 1997 I could
not quite h old thi s against Mickey Ad
ams, as my flag was about to fall in a
quickp1ay. So Black still has to be care
ful, and who knows, I could h ave lost
the g ame anyway, with little time left.
After my inferior move, Pert played
51 'it>e4 tiJd6+ 52 'it>e3 I1xg2 (or
52 ... tiJe4+ 53 'it>b4 xg 2 54 e7 b2+
55 a5 I1b3 5 6 .l:!.xe6 xe3 5 7 e8+
'it>g7 5 8 d5 with broadly similar types of
position) 53 l1e7 tiJe4+ 54 'it>e4 11e2+ 5 5
b5 ':e3, and then there was a time
scramble. I cannot remember the finish
in detail, but a plausible finish might
h ave been something like 5 6 xe6
.l:!.xe3 57 'it>c6 h 3 58 e8+ 'it>g 7 59
l1e7+ g 8 60 'it>d5 I1h4 61 tiJg6 g4 6 2
'it> e 5 I1xg 6 6 3 'it>xf5, and White win s,
which he certainly did in the game: 1-0.


Why we L o s e a t C h e s s

Yo u r M ove

N o attempt i s given to suggest wh at sort of percentage scores should be expected

for pl ayers of particular grades or ratings. This would be slightly artificial . By defi
nition, every move I made was, in the end, a mistake, with a 0% score. I am sure
th at any reader would h ave made a better percentage, even a reader who has not
played chess before. By sticking a pin on th e board, the pl ayer cannot stick far be
low 30%, which my own score seem s by contrast puny.
The reader, aiming to improve his or her game, should try to think about th e
exerci se games, and should try to analyse and learn . A reader who scores more
th an 70%, con siderably in advance of the random result, is doing well, but there
are possibilities of improvement on even this. After all , even a 70% score would
suggest th at on each positi on, there i s a 30% chance of a mi stake on these critical
positions, and a score of 30% error would be a worry for most strong players.
How to Audit Your Resu lts

Players want to improve their grading scores. An extra point per ten games over
the year would mean 10 extra points in term s of the English grading system (this
is very easy to calculate ! ) , or approxim ately 7 2 extra points in Elo term s. All this
would mean a substanti al increase in grading term s.
I h ave, of course, calculated my results of the 200617 season, as analysed in this
book, and we shall turn to thi s later. In th e meantime, I h ave gone through the
games pl ayed in the few month s since I wrote thi s book, to see whether I h ave im
proved my accuracy.
Between May 2009 and September 2009, I pl ayed 18 tourn ament or match


Yo u r M o ve
g ames, but not in international events. I wanted to concentrate on this book !
There were 18 games, m ainly but not entirely against con siderably lower rated
My results were: +12 4 -2, which m akes 78%.
The scores were inflated by six easy win s in the British Braille Champion ship.
My seventh opponent, Chris Ross, was a much stronger opponent, and drew
again st me.
The interesting question i s whether I could h ave scored con siderably more if I
h ad avoided mistakes. Clearly if I h ad avoided my two losses, I would h ave scored
at least an extra point. In fact, in one of these games I could h ave won with accu
rate play. Of the four draws, I could h ave found a win against Graham Morrison .
What should h ave h appened then would h ave been : +14 4 -0, which m akes
89%. Thus just avoiding three slips could h ave created an 11% increase in per
Maybe this is the one part of the g ame, the avoidance of errors, th at m akes the
biggest difference in improving one's perform ance. My challenge for the season is
to h alve this 11% slippag e.
It is perh aps important to recognize that you should avoid double-counting. If
you m ake a mistake, and reach what should h ave been a losing position, but later
won, then congratulations, but you are lucky. Wh at is more important, in term s of
results, is to remember the games you h ave lost, or only drawn after h aving h ad a
winning position.

Using the Com puter

In thi s book, I h ave used the computer in analysing my own games, and the rec
omm endation for the reader is, of course, to take advantage of the computer. Do
not overestimate th e computer though, but use it as a tool .
I was relatively l ate in making use of the computer, partly being suspicious of
computer assessments. I h ave been suspicious of reading computer-based as
sessments which h ave clearly been wrong, and h ave preferred my own an alysis,
with board and pieces. The computer i s important in generating ideas quickly, but
even h ere th e hum an reader should treat everything with suspicion.
It is still important to take advantage of making brain -power in analysis. After
all, in a proper g ame, the player is not allowed to use the computer, and the hu
man needs to practice in thinking about the game.
And Final ly ...

Except, of course, chess is not final . Even after the last melting of the pol ar icecaps,


Why we L o s e a t C h e s s
there will still b e chess being played, and new Noah s will still b e enthralled by the
game. It is, of course, quite possible th at theory has gone so deeply that almost
everything has been studied in depth, and that new version s of chess, maybe
based on western chess, or eastern chess, could be tried, m aybe a fusion of rules
with ideas from western and eastern chess. Thi s would be a long way ahead. For
the next h alf-millennium, we still h ave the current chess of Greco and Ruy Lopez.
For individual pl ayers though, any chesspl ayer is mortal . A pl ayer might h ave
learnt the game, become fascinated, learnt so m any ideas of chess strategy and
tactics, maybe becoming a strong pl ayer, m aybe hoping to become an even
stronger player. Then perh aps before the final push, at the peak of the pl ayer's
strength, ch ess ability starts to decline. It is now reasonably well established that
for a professional player, the peak strength will be around the mid-thirties, and
will start to decline slowly. For me, I really only seriously started studying chess at
the time th at theoretically I should h ave been starting to decline. I h ave played a
lot of chess, but did not h ave time to study h ard. Perh aps I can still add to the
g ame, but not through my playing strength . I h ave continued to study the g ame,
but can contribute more to the theory of the game, rather than through pl aying
over-the-board chess. As the old saying goes, I h ave forgotten more about chess
th an most players h ave even learnt.
A S u m m ary of Games, in Chronological Order

Note: See the list of exercises at the end of the introduction for all the test g ames.
This list is used partly as a research tool , to h elp to indicate when and why bad
moves h ave been played. There being rel atively few games in thi s book, all pl ayed
by myself, that it h ardly seem s worthwhile to add a second alph abetical list.
All games were pl ayed in England.
September 2006
Buckley-Crouch, 0-1 (win)
October 2006
Hebden -Crouch, 1-0 (loss)
Crouch -Cox, '/2-V2 (draw)
Gregory-Crouch, 1-0 (loss)
Crouch -Hutchin son, 1-0 (win)
Pert-Crouch, 1-0 (loss)


Yo u r M o ve
November 2006
Nurmoh amed-Crouch, 0-1 (win)
Morris-Crouch, 1-0 (loss)
Randall-Crouch, 1-0 (loss)
Crouch -G ait, 0-1 (loss)
December 2006
Sen-Crouch, 1-0 (loss)
McKenn a-Crouch, %-% (draw)
Crouch-Oryakh al, 1-0 (win)
January 2007
Crouch-Radovanovic, %-V2 (draw)
Crouch-Okike, 1-0 (win)
Crouch- Roberson, V2-V2 (draw) [see al so Crouch -Povah, V2-V2, November 2006,
given in the notes]
Lauterbach -Crouch, V2-% (draw)
February 2007
Crouch- Peacock, V2-V2 (draw)
Cutmore-Crouch, 0-1 (win)
Crouch - Rose, 0-1 (loss)
Crouch -Lewi s 0-1 (loss)
March 2007
Crouch -Gran at, 0-1 (loss)
Wall-Crouch, 0-1 (win)
Sowray-Crouch, 1-0 (loss)
A substanti al m ajority of the remaining games in this period were wins, al
though it i s possible th at a small number of games may have been misl aid, a prob
lem of being partially sighted.
24 g ames h ave been analysed in this collection, of which 7 were wins, 6 were
draws, and 11 were losses, a score of 45%.
Once you h ave gone through your games, the next stage, if you want to im
prove your playing strength, i s to go through your statistics. Altogether in the
main part of the season up to late March, I h ave recorded 5 6 games. A few of the
games might h ave dropped out of the system, sometimes because with poor eye-


Why we L o s e a t C h e s s
sight, I could not read the scoresheet. Also, I should perh aps h ave added the final
game of the season, a London League loss against 1 M John-Paul Wall ace, when I
should probably have been better. I beat him the next time around, to level the
scores, sealing a match win against the formidable Wood Green team, previously
unbeaten for close to ten years.
I have no doubt that in the 5 6 games I h ave recorded, there will be unrecorded
mistakes, often slight inaccuracies. In 24 of my g ames, I found identifi able in accu
racies. As an approximation, I won, or occasionally played a solid draw, in h alf my
games, and I h ave no reason to an alyse these in much detail. The other h alf is of
much more interest. These are the g ames I need to focus on to cut out mistakes,
and to increase my percentage scores.
So: 5 6 games (100%) were played, of which 3 2 g ames (5 7%) seem to be 'clean ',
including two win s against I Ms, and indeed a win against a G M .
Many o f m y opponents were o f con siderably weaker strength, and therefore
did not force me to h ave to play under pressure, so it would be unrealistic to claim
th at I necessarily played well.
The rem aining 43% (24 games) is of more interest. Thi s included: 7 win s
( 1 2 . 5 %), 6 draws (10.7%), and 11 losses (19.6%).
Clearly at the very minimum, I would h ave gained 10.7 percentage points by
avoiding losses, if I had turned losses into draws. The theoretical maximum of
turning bad moves into win s would by 3 1. 1 percentage points, but this, of course,
would be unachievable, in th at we cannot claim th at my opponent would h ave
made mistakes if I h ad pl ayed the best moves.
So we need to examine, move by move, wh at the most likely result would h ave
been h ad I avoided the mistakes, and found the best m oves.
Starting with the Losses

The obvious starting point is cutting down the losses. If there are pl ayers of
roughly equal strength, but not players of top grandmaster strength (where draws
are more likely as there are few mistakes), a typical score might be +40% = 20% 40%.
As an illustration, let us assume that a player improves, and can cut out a
quarter of his losses. Thi s might turn out next season with a result of +50% =20% 30%, an end percentage score of 60% as opposed to 50%. For simplicity, we call
thi s an increase of 10 English (ECF) grading points.
It is, of course, unlikely that a pl ayer will turn all his or her losses into wins. A
m ore likely assumption will be th at h alf the reduction of losses will turn into
draws, and the other h alf will turn into wins. The previous draws would, after a


Yo u r M o ve
year of improvement, turn into wins.
10 grading points (an approximate equivalent of 7 2 Elo points) is quite a big
jump. Even h ere, the pl ayer is far from cutting out losses, and thi s is far from per
fect play. It is, however, progress. A fast improving junior, absorbing the ideas of
chess rapidly, may often show an increase of 2 5 English grading points, with a
score of +65% = 20% - 1 5 %, and perh aps the implication that it would be time to
play stronger and more testing opposition .
For more established players, it i s difficult to gain 2 5 grading points a year, or
even a decade, or ever. All you can do is to chip away at the margins. If your mind
is befuddled when trying to calculate complications, then sadly you will not be
able to calculate ten moves deep, and, of course, there are other physical limita
tion s.
Therefore one can not h ave unusually high expectations. If you can cut out
even a tenth of your losses, thi s is an improvement, and maybe if you start to feel
encouraged, you can try again next tim e around. Thi s is perhaps a more pessimis
tic, but certainly realistic, view th an in the m ain argument of the book. After all,
one of the m ain thrusts of the argument i s that if the reader can learn to cut out a
few basic and not so esoteric weaknesses, you can improve your chess substan
tially. Just cut out a few silly blunders, and there can be a substantial gain.
It i s time to con sider the perspective of the Elo rating system, rather than the
English rating system , a ladder system rather than one of finding th e average. Two
good and useful rating system s, but with highly different perspectives. In the Elo
system, we start off with the rating of the two players, and calculate in an individ
ual game or tourn ament how m any points may be gained or lost in a particular
game. If two pl ayers m ay be of roughly similar strength, then we can imagine th at
one of the pl ayers might gain 20 Elo points over 20 games without gross blunders,
through better understanding of the game. This is a slow progress. Under the cur
rent system of the Elo international rating system, a player can suddenly drop in a
move from a gain of 10 points to a loss of 10 Elo points, a total drop of 20 points.
The D u m best Moves

Maybe it is time to consider the worst moves, and kick them out of the way
quickly. I should perh aps note th at in the previous few years, the effects of brain
dam age h as h ad rel atively little damage to my understanding of the game, but a
much greater impact in my speed of thought, both in term s of speed of calcula
tion, and lack of clarity of eyesight. Al so dizziness and tiredness have tended at
times to make it difficult for m e to think clearly, and I have at times been unable
to concentrate on pl aying anything other than the most superficial an alysi s.

17 9

Why we L o s e a t C h e s s
There have i n fact been much worse periods of pl ay in chess terms, particularly
at times when I have been actively working on writing books. A compl aint for
many other chess writers, but probably worse for me, since my physical ability
over the board has led to a deterioration in play. Just after my stroke, I gloomily
wrote that I h ad probably gone down to 1 7 5 (about 2000 Elo) in playing strength,
but fortun ately I h ave consistently reached 200 ECF, and stayed well over 2 3 50 in
Elo terms. My target ambition is to get back to 2400, but any loss will tend to go
down the greasy l adder, so I need to plan exactly which tournaments to play.
The general conclusion is that I am still quite likely to m ake serious and unex
pected mistakes, and my hope i s that I can cut down the blunders to a rating I
would be h appier to reach .
Time now to look through the most h orrendous blunders; the single moves
that with even the smallest of thought I could h ave recovered my score.
The Very Worst Games

1. Crouch-McKenna
Drawn . I should h ave won easily, but I displ ayed a lack of concentration, then
m ade several bad moves, and I should h ave lost. Half a point lost.
When I think of all the bad games in th at year, thi s is the one which I remem
ber the most readily. I feel confident I could h ave won it easily with better health .
2. Randall-Crouch
A loss, which should h ave been a win, with a strong attack for me. I h ad the chance
of taking a rook, but I worked out that if I h ad taken this, he h ad th e chance of
pushing two m ajor pieces to a winning back row check. In fact there was only one
check, and I would h ave h ad everything covered. To m ake it worse, I h ad the sim
ple opportunity of taking a rook, after h e too h ad m ade mistakes, but I rejected it,
not noticing that after he won a rook in reply, I could then take the rook with
check. A full point lost.
Again, with my mind functioning properly, I should h ave won easily.
3. Nurmohamed-Crouch
By common consent, the venue at thi s club was not the most attractive place to
pl ay. We were playing in a hut, in the winter, with in adequate heating. I played
the opening embarrassingly badly, losing concentration after quickly g aining an
edge, and before long I h ad a losing position, but recovered to some extent, and
even later won after my king somehow wriggled out of an attack.


Yo u r M o ve
4. Buckley-Crouch
Thi s was one of those nightmare games where both players exchange mistakes,
neither player i s able to find a knockout, and eventually my opponent found an
auto-blunder, when even I could find an easy way to win a pawn . Even worse, h e
ran out o f t i m e after a lot o f thought. It was a dreadful g ame, although i f o n e of
the players h ad played even very slightly worse, the opponent would h ave won
quickly. No points lostfor me, as I won.
5. Wall-Crouch
An awful game, with serious mistakes on both sides. My king move, when I at
tempted to bring it towards safety, was quite simply bizarre, bringing it into the
open . My opponent l ater made a serious miscal culation.
One-move Shockers

In each of these cases, I feel I pl ayed well, or reasonably, in the opening, then lost
concentration, pl aying a superficial move quickly, not calculating as far as I should
h ave done. I h ave played sh arply with good effect, but suddenly when the game
became critical, I played limply.
1. Sen-Crouch
Good opening play, and with the chance to sn atch a pawn as Bl ack. Then, at a later
stage of the opening, I needed to think of the one move for my queen to escape to
a good square, admittedly difficult to find a few moves ahead, and instead pl ayed
a 'simpler' line, which soon ended in collapse. I lost, and instead should have been
2. Hebden-Crouch
I trusted him when h e offered a pawn, to keep the initiative. I shouldn 't h ave. It
would h ave been better to grab the pawn, with the aim of breaking open the
queenside. I lost, and with good play it should have been a reasonably comfortable
dra w.
3. Crouch-Rose
An experimental opening to avoid m ainline theory. The opening itself looks play
able, if not especially dynamic, but I missed a big tacti c, which I should have seen
in advance. I lost, although the opening was about equal.

18 1

Why we L o s e a t C h e s s
4 . Crouch-Lewis
I wanted to prove th at the Crouch-Rose game was not really so bad, and so over
night I aimed to find improvements. Maybe the opening was not so bad in analyti
cal term s, but I was under the psychological pressure of proving that my idea was
good, and lurking behind is the thought that if anything h ad gone wrong again, I
would have been extremely anxious about my pl ay. Thi s was the wrong way to
start a game.
I thought I was doing reasonably well in the opening, but a ' Petrosian ex
change sacrifice' knocked me back. Analysis showed that, slightly beyond my game
horizon, I could h ave satisfactorily found a draw after taking the exch ange, but
the position was scary, and I tried to pl ay around the sacrifice, rather than accept
it. f lost, when f had probably a clear dra w.
Four losses, then, when by pl aying better moves on four critical moves, my ex
pected score would h ave been 2'1214.
Over- p laying the Opening

The last four games indicate a lack of confidence, and I lost heavily. Quite often,
though, there is a tendency in my game for over-confidence in the opening, and
there are a few examples to be con sidered. Of course, the best way is to find the
correct level, but it is extremely difficult to find it. Of 'my 60 forgettable mistakes'
in thi s book, a quarter are uncovered on move 16 or earlier. The next quarter move
us only to just beyond move 20. The opening is the most difficult part of the game,
because the vast possibilities of chess still remain, and on every m ove the player
has to assess not only what sort of position is likely to arise (attack, positional play,
complications, or simplifications), but al so wh at possibilities need to be gradually
eliminated. It is only several moves later th at it becomes clearer as we move past
th e opening.
In the next few g am es, I pressed the position much too far, aiming for attack,
but with no real chance of finding a way to safety. It is do or die, but before too
long I would be highly relieved if I could escape to a draw.
When I played at Kidlington early in 2007, I clearly decided that I wanted to
play imaginative chess, with the confidence that spring was about to come. There
were too many bad games as a result.
1. Crouch-Peacock
I played far too imaginatively, and I cannot reali stically claim th at, even after some
m odest improvements, I h ad anything more than the advantage of the first move.

18 2

Yo u r M o ve
My play went out of control, and I was fortunate th at my opponent went for a
perpetual, when he coul d h ave gone for a winning endgame. A lucky draw.
2. Crouch-Gait
Without thinking in great length, I developed my bishop with gain of play, by
threatening a big check with the queen and bishop. H ad I thought for longer, I
woul d h ave appreciated that h e could defend the threat with counterplay, and
that a move later I could have been much worse. H e missed it, but my equilibrium
was disturbed, and I m ade mistakes l ater. A loss, but I was slightly better out of the
3. Crouch-Hutchinson
I could not claim much of an edge, and it should h ave been very close to equal as
White. I found a complicated way of pl aying for an advantage, but this was an il
lusion, and with accurate tactical play by the opponent, I would h ave been much
worse. He missed it, and ended up losing. A win, but I overpressed. I should have
4. Crouch-Radovanovic
I was still out of touch with my opening theory, and I improvised. My opponent
missed a well -known opportunity to give Black a slight edge, and then later I m ade
a bad knight versus bishop exch ange on g6, allowing Bl ack a semi-open file lead
ing towards h 2 . I h ad to work hard for a draw, although there were improvements
for him. Afortunate dra w.
Sim ple Tactical Slips in the Middlegame

Everything seem s to be going smoothly. I am heading for, I h ope, a winning edge,

then suddenly a miscalcul ation, or more probably, a slip in my thought process,
wh en I assume something, but tactically it doesn 't work.
1. Gregory-Crouch
Thi s seemed smooth enough during the game, but while I won a pawn, it was dif
ficult to convert it into a serious edge, as his pieces were active. Becoming short of
time, I immediately sacrificed the exchange, to eliminate a dangerous bishop, but
I did not have the compen sation th at I was expecting. I should in stead have qui
etly moved the king away. A loss, rather tha n a dra w.


Why we L o s e a t C h e s s
2. Crouch-Roberson
I was able to keep a stable positional edge, with a passed d-pawn against a Be
noni. I was then lazy in my calculation s. A dra w, rather tha n a win.
3. Crouch-Okike
I had to give up a queen for assorted play, but I did it the wrong way round, and he
could have improved. I still won, but I could have played better.
4. Crouch-Cut more
A defen sive mistake, rather th an a tactical mistake. I was better, in a sh arp posi
tion, but I soon m ade the big mi stake of playing too slowly in a sharp position of
attack and counterattack. Before too long, I was clearly losing, but then he h an
dled the attack badly, perhaps surpri sed that he was suddenly in a winning posi
tion again st an 1M. He allowed m e a sudden checkm ate. I won, but I very m uch de
served to lose.
Getting Gro u nd Down

It h appen s, and players hate it. With very best play, one cannot, of course, lose a
g ame from the beginning, but sometimes even the slightest mistake, alm ost in
visible, will decide the game. This in fact very rarely h appen s except at the highest
level - in my previous book, I am thinking, for example, of Kramnik-Leko, Dort
mund 2006. U sually what will h appen, if the players play well but not totally accu
rately, is th at there will be a series of minor slips, the defender gradually slipping
away, or the attacker losing his advantage.
Th e examples I give tend in the end not to h ave been rel atively subtle errors,
and there is therefore the chance of being able to learn more from the errors.
1. Lauterbach-Crouch
Even the less experienced pl ayer would quickly h ave seen th at my knight was bad
on h 8 . After a few slips, I should h ave lost, but she avoided the best m ove, and I
was able to hold for a draw. Of course, I did not particularly want to move the
knight to such an uncomfortable square, but my mistake came earlier. To kick out
White's bishop on g 5, I should h ave hit the bishop with .. .f6, rather than ... h 6 . I
needed my pawn on h 7 . A fortunate dra w.

2. Morris-Crouch
I again m anaged to find one of the knights in the corner, this time on a8 rather
th an on h 8 . Thi s was a result of placing my pawn s carelessly, creating weaknesses


Yo u r M o ve
on the queen side. I wanted to advance my pawn to as, to prevent my opponent
pushing forward with b4 to attack the knight on c5, but in the end my knight h ad
to retreat to a8 much l ater. There was the strange incident of the fire alarm just
before the time control, but even so, I deserved to be losing anyway. A loss, even
though I seem to have equalized earlier on.
Messing up the Endgame

There are surprisingly few examples of losing in the endgame, or throwing away a
win . My biggest fear after returning to playing chess was that I would be so tired,
and my eyes dizzy, that I would completely lose the thread of things in the end
g ame. Thi s in fact has not h appened so often, m ainly because if I was not feeling
in good health that day, I would probably h ave gone wrong much earlier, in the
opening or the middlegame. Nevertheless, in earlier years I often m an aged to lose
concentration in the endgame in the fifth hour, m aking ridiculous errors from
good endgames, and then losing. Such disasters stick in the mind, but, of course,
often it was my opponent who m ade the silly losses, and usually I h ave forgotten
about those games.
In general, I do not particul arly fear the endgame as such, but I do fear being
short of time, and I do fear tiredness.
I h ave already noted the poor play again st both Buckley and McKenna. In both
cases I ought to h ave been holding the draw, rather than under the illusion of try
ing to squeeze an endgame advantage, but both my opponents m anaged to play
even worse.
1. Pert-Crouch
I got tired in round 5 in a critical weekender, and as I drifted from a late middle
g am e into an ending my position became worse and worse. My opponent did not
play particularly accurately, and to my regret I managed to miss a chance of a
draw. A loss, but I could have recovered half a point.
2. Crouch-Granat
A difficult queenless middlegame and then endgame, where I started with the
bishops and h e started with the knights. Neither pl ayer h andled thi s with com
plete positional confidence, but in th e end I h andled things worse th an he did. A
loss, but with good chances perhaps of a slight edge at various stages.


Why we L o s e a t C h e s s
Two Difficult Games to I nterpret

Sometimes it is difficult to understand wh at is going on during a game, and al so

difficult to understand at the post-mortem, or afterwards with the computer. If
the game is played at relatively quick time limits, there are almost certain to be
mistakes, quite possibly on both sides. The good news, for both players and spec
tators, is that at least the games are interesting.
In the two London League games I am covering, I m an aged a draw and a loss,
but I could easily h ave ended up with two wins, or two losses. In my earlier analy
sis, I must h ave questioned over half a dozen of my moves, but l ater I h ave decided
that most of my moves seem to h ave been okay, and only a relatively small num
ber eventually seemed questionable. So both players played reasonable chess in
difficult positions.
1. Sowray-Crouch
I probably h ad a slight edge as Black in a Sicilian, but I wasted a move with ... bS,
normal enough, but h ere a waste of time. It i s difficult to believe that Black was
worse, and even significantly worse, but almost everything I tried in my l ater
analysi s ended up as bad. The Sicilian is so often on a knife-edge for both players.
Eventually I found th at the 'natural' check with gain of tempo (with ... .ltf8-a3+ fol
lowed by castling) was a serious mistake, and that the 'unn atural ' .. .f6, defen ding
the e-pawn, appears to equalize. At least a dra w, probably more, from the opening,
but I lost my way later.
2. Crouch-Cox
Even more complicated, and a draw, but with opportunities for a win for either
side. He h ad the more direct winning ch ances just before the time control . I would
h ave needed to h ave found much more complicated options, and I very much sus
pect that I would not h ave found the best option s even if (the standard excuse) my
health was better, with a clearer mind. The difficulties for both sides were based
very much on the question of finding the right balance between positional advan
tage and m aterial advantage, with both sides at times h aving to con sider sacrifice
and counter-sacrifice. A complicated dra w.
And a Little C uriosity

My concentration l apsed in an unorthodox opening, and I was already worse after
my fifth m ove. Before too long, I recovered my edge again. A win.


Yo u r M o ve

My calculation s suggest th at I dropped ten points as a result of serious mistakes,

m ainly through dropped h alf-points, rather than losing from a winning position.
This necessarily involves some degree of approximation, as it i s difficult to estab
lish early on what the result would h ave been if I h ad pl ayed the game more accu
rately. I h ave tried to err to the side of the draw.
In term s of statistics, the most meaningful assessment in practical term s
would be to say th at with accurate pl ay, I could h ave g ained 10 points out of 5 2
games, purely through active pl ay, and avoiding serious mistakes. Thi s is a score of
19.2% th at I h ave lost.
If I h ad been able to play more accurately, my ECF grading would h ave jumped
by 19 points. It is difficult to m ake a reliable Elo calcul ation because of the various
methods by which Elo ratin g s are calculated, depending on the K-factor and the
number of g ames (as well as the average rating) being played. We are, however,
dealing with a ball-park figure of around 140 extra Elo points.
N aturally no player can cut out their weaknesses in stantly, but even h alf an
elimination of clear weaknesses would, on this basis, give about 10 extra ECF
points. Some of the errors in this book m ay well seem extremely obscure and
complicated to many readers, while others might be reg arded as simple mistakes,
which most reasonably stron g pl ayers should be able to avoid much of the time.
The basic question for the reader is to think whether you are capable of avoid
ing at least some of the errors given in the exerci ses, and if so, whether you will be
able to improve in your games. Good luck.
I am going through my new g ames to see whether I can improve on my play.
Maybe at some future date we can compare and contrast our efforts?
The computer these days allows the player to indicate where possible im
provements are to be found. Go through the lines and work out what is going on.
Don 't bother so much about trying to learn opening theory. Just learn, paying
equal attention to the different parts of the g ame: opening, middlegame and
endgame; attack and defence; strategy and tactics. And above all, make fewer mis