Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Something's wrong with Crafty?

Author: Robert Hyatt

Date: 06:55:26 07/14/02

Go up one level in this thread


On July 14, 2002 at 02:10:03, Russell Reagan wrote:

>On July 14, 2002 at 02:03:46, Ricardo Gibert wrote:
>
>>The bottom line is, does it play ok. The actual score matters less.
>
>This is very true. I've often thought that it's odd that every chess engine I've
>ever seen evaluates positions in basically the same way, in terms of pawns. I
>have quite a few ideas for my chess program that I'm not sure how I would score.
>My scores might be completely weird. I have heard of evaluation being different
>in other games, where a ratio of scores is computed, and if the score approaches
>0, that's good for one side, and if it approaches infinity, that's good for the
>other side. Never seen anything like that in chess though.
>
>Russell


The idea is doable.  IE compute N terms, and rather than saying
score = sum(scores)/N

do something similar but use the square of the sums.  So that one big
advantage will be more important...



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.