Computer Chess Club Archives




Subject: Re: The Limits of Positional Knowledge

Author: Stephen A. Boak

Date: 19:10:28 11/15/99

Go up one level in this thread

On November 15, 1999 at 17:36:28, Bella Freud wrote:


>Chess programs are playing "negative" chess. They avoid obvious loss by ply N.
>For them perfection is in loss avoidance (the not totally obvious opposite of
>win seeking). Therefore no style, no plan, no class. Just dumb obstructionism.

Not quite.  You miss the twin behavior--chess programs seek to maximize overall
score (material, positional, etc)--which is a form of 'win seeking', even if
slower than some people's taste, and even if not the best way of selecting among
available moves in some situations.  They seek obvious score increase by ply N.
This is still not what I would like to call 'a plan', but is is far from mere
'dumb obstructionism' (your 'loss avoidance').  It is indeed some sort of plan,
within the move horizon (plan is to maximize score, within foreseeable future).

Also, not all programs have the same style, especially depending on the variable
settings available in some.  What some of them sometimes do to upset the
material/positional balance in a game would surprise you.  I call that 'trying
to win', just like many humans seek to outplay you by confusing the situation,
creating dynamic chances resting on inequality of positions, imbalances of
material, etc.  It is stereotyping to lump all programs, under all possible
positions, in the same bag.

To say that all chess programs merely move simply to 'avoid loss' is incorrect.


This page took 0.01 seconds to execute

Last modified: Thu, 07 Jul 11 08:48:38 -0700

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