Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Question about evaluation and branch factor

Author: Russell Reagan

Date: 10:51:28 11/20/03

Go up one level in this thread


On November 20, 2003 at 13:18:55, Uri Blass wrote:

>I do not agree.
>
>Maybe if you compare only material with something more complex you are right but
>I see no reason for it to be the case with piece square table relative to more
>complex evaluation.
>
>More evaluation does not mean that you are slower in test suites and it means
>that you often solve the test faster because the evaluation can tell you that
>the move is good move even when you do not see that it wins material.

Take for instance if you only used material evaluation. If the score is +1, then
you're going to get a lot of '=' beta cutoffs, as he said. If you have an
evaluation that is very fine grained (meaning a pawn might be 10000 instead of
100). Consider if you had a score of +1 pawn. In the fine grained version with a
more complex evaluation, you might have a score of 10001, which wouldn't get you
a cutoff (but there is almost no difference between 10000 and 10001).

I think I recall that Tim's engine used three decimal places for its evaluation
(meaning a pawn is 1.000), and in his new engine that he is working on, he chose
to make a pawn only worth 1.00. Given the other info about his posts about '='
beta cutoffs, I guess he hopes to get more cutoffs that way.



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.