Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Analysis of Crafty vs. GM van Wely ICC Blitz game

Author: Pete R.

Date: 09:59:11 07/16/99

Go up one level in this thread


On July 16, 1999 at 11:04:56, James Robertson wrote:

>On July 15, 1999 at 23:16:07, Pete R. wrote:
>
>When we work through the game forwards, we have to work through the analysis
>backwards. :)

Hey, whatever it takes. :)  It's actually more like a few steps forward, then a
few more steps back.  Right now I'm trying to find a move for white that refutes
23...Kd7.  Not looking good for white so far.  If that holds then 23...Rg7 is
definitely on the trash heap for black, but it would also mean that we have to
go back another ply and find an antidote to 23...Kd7 for white.  This whole
variation rests on the difficult to analyze concept of "compensation".  So far
White appears to get compensation eventually after 23...Rg7 but not 23...Kd7.

This game is an interesting example of the increasing impact computers will have
on chess.  Consider the usual mathematics vs. this example.  If black plays Rg7
as his 23rd move, white can force him into *one* specific position at his 31st
move.  If we apply the usual mathematics, with this 15 ply gap there should be
millions of potential positions that could be reached.  This is mathematically
true, but 99.999% of them are worthless, and because of transpositions and moves
 forced by virtue of no playable alternative, there is only one worthwhile line.
 So what I think will happen is that as computers become increasing powerful,
many existing variations will be resolved to an advantage for white or black
from the opening through the endgame.  Chess is not a random game, so the
calculations of 10^42 possible positions are highly misleading in terms of how
many playable lines there are.  In this particular subvariation the big question
 (in addition to the usual question of the poisoned pawn) is whether white's
knight sacrifice to open black's kingside gives enough compensation.  Because of
the nature of the game, weakness created in the opening have ripple effects
right out the endgame, particularly where pawn structure is concerned.  Because
of things like this the true number of strong, practical lines is going to be a
tiny number compared to theoretical possibilities. Don't get me wrong, there are
still going to be a *lot* of those lines, but the greater computer power gets
the smaller that number can be resolved to.  Imagine that something like the DB
chips on a card hits the marketplace.  Then imagine what the real opening
theorists whose names start with GM could do with a tool like that.



This page took 0 seconds to execute

Last modified: Thu, 15 Apr 21 08:11:13 -0700

Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.