such,
take a standard King's Indian Defence,
Pirc
Defence, or Dragon Sicilian. These all in‑
valve a
kingside fan ghetto with ...g6 and
g7. Now
White often lines up his queen and
bishop
with IVd2 and e3; in that case playing
h6 to
exchange bishops on g7 can be good for
any
number of general reasons, e.g. because
Black
has subsequent weakness on his dark
squares,
because White' s centre will be pro‑
tested
in the case of an ...exd4 exchange, or be‑
cause
White's potential gain of space after d5
will be
more effective. But the move h6 may
equally
be bad, e.g., if it allows Black to ex‑
change
off his only bad piece and block the
game, or
if it diverts White' s queen from the
action,
or if it helps Black to play ...c5 and at‑
tack a
weakened centre. The reality is that such
a
decision tends to be very subtle, ultimately
depending
upon a great number of other less
obvious
factors; e.g., how quickly Black can
play
...b5, whether a well-timed ...exd4 fol‑
lowed by
...c5 or ...Ze8 has tactical advantages,
whether
White' s a3 will slow the attack or pro‑
vide a
target, what reorganizations of pieces
are
available to both sides, and so forth. From
many
years of analyzing with top players, I be‑
live
that with the aid of calculation they assess
this
decision more or less instinctively based
upon
experience and judgment, without re‑
sorting
to verbal reasoning. Later, the move can
be
explained to an audience using simplified
generalities.
But going the other way, from
words to
move, is extremely unlikely to pro‑
duce the
right decision. And it is particularly
hard to
act upon raw concepts. For example,
one
might trade off time or development to se‑
cure
some other positional advantage (an out‑
post,
doubled pawn for the opponent). But one
question
of many is: how much time/develop‑
mint can
I give up before, e.g., a counterattack
becomes
a problem? This is solved by some
combination
of analysis and judgment, and
very
seldom by abstract reasoning.
People
commonly interpret what I've said
in terms
of rules and exceptions, i.e., modem
players
know the rules and have discovered the
exceptions.
Or that a player should in fact learn
the
rules first and then learn their exceptions.
But
that's not what I'm saying at all. The point
is not
whether there are X exceptions to every Y
instances
in which the rule is true. Naturally
• • •
factors
like space, outposts, and 'better' minor
pieces
will always correlate, however mildly,
to
winning percentage. Rather, I am asking:
a) Does
a master think in terms of a given
rule
(and exception) at all?
b) Is it
even useful to think in those terms
when
confronted with a specific position?
Granted,
there are cases in which the sim‑
palest
way to make my argument (and I have
occasionally
done so) is to point out that there are
so many
exceptions to a particular traditional
rule
that it becomes meaningless, and so by im‑
placation
the modem player must not be using
it. An
extreme case is one examined in Chap‑
ters 1
and 2, where I'm not sure whether the
rule
'The player with more space should avoid
exchanges
(and vice-versa)' is true even 50% of
the
time! But let's say that figure were 60% or
70%: it
would still be terribly limiting to be
thinking
about a prospective exchange in terms
of such
a rule. Even when the rule expresses a
situation
that is true in many more cases than
not
(e.g., 'a knight on the rim is grim' ), is it con‑
structive
to think in those terms? Or is it better
to
examine the concrete situation in all of its
subtleties,
without prejudice, using calculation
and
pattern recognition as tools? Which does
the strong
player do? In my observation and
from
many conversations and analysis with
them,
grandmasters don't think much in terms
of rules
and exceptions at all. Indeed, in some
cases
where they have done so, some quite
playable
positions have escaped deserved con‑
side
ration because of they 'looked' bad on gen‑
eral
principle. Increasingly, that attitude is being
replaced
by a pragmatic analytical approach.
Other
models have been floated to deal with
these
issues. One such is the idea that rules are
valid,
'other things being equal'. I won't go
into
detail here, but the question in practice is
whether
anyone can decide by explicit means to
what
extent other factors in a typical middle game
are
'equal' . Since the diverse considerations are
interdependent
and also time-dependent, and
since
they can be so unbelievably complex, the
determination
of what weight to give one ab‑
strict
rule or another seems to me impossible in
almost
all cases. Then there is the idea of as‑
sassing
and utilizing explicitly identifiable im‑
balances.
But juggling and weighing the many
and
often subtle imbalances (and anticipated
