eric (12/11/82)
Many thanks to cca!mclure for the interesting Prestige
games. That article seems a very good use of this net, because
it picked out games which show an interesting problem in the pro-
gram which might well be soluble. If it is hard to solve, the
discussion of why it is would be very enlightening to me.
I really enjoyed the contrast to be found between the first
and third games. An important principle Prestige followed in the
first game but ignored in the third is not to allow open lines
for opposing rooks. I would be interested to hear conjectures as
to why it followed that principle in P.-Moody (where failing to
do so would probably only cost a half point) and ignored it in
Tingblad-P., where it could, and did, cost a full point.
Against Moody, it locked up the pawns with 36 c4, keeping
Moody's rook passive, and preventing a draw. A piece and a pawn
up versus Tingblad (after 1 e4 e6 2 d4 d5 3 Nc3 Bb4 4 e5 c5 5 a3
Bc3 6 bc3 Ne7 7 Nf3 Nd7!? Very ambitious; aims at complete cen-
tral hegemony with f7-f6-fe5 and cd4. 8 h4?! f6 9 h5?! 0-0 10
Nh4. Consistent at least. 10...fe5 11 Bd3 cd4. Again very
ambitious. Safer is e4, keeping the pawn advantage and forever
keeping the prelate out of h7. 12 Bh7 Kh7 13 h6 dc3?????) it
failed to play the obvious g6. This is the same idea of not
allowing the opposing rook more space by keeping an opposing pawn
in its way; and the principle seems much more crucial in a posi-
tion where temporary mobility of opposing pieces is of the essen-
ce.
Perhaps the move is not obvious to every player. It was to
me because I shifted my thought patterns into piece-ahead-
temporarily-on-the-defense mode. In that mode I pay much atten-
tion to how many pieces are being really active near the center
of the action, and I eschew one *more* pawn when it costs one
more opposing piece in the action, or one less of my own. This
principle seems fairly easily quantifiable, and in my experience,
it is pretty easy to determine when it ought to be operant and
what it demands.
Comments from programmers?
teklabs!reed!eric