Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: how much does eval effect strength?

Author: Tim Foden

Date: 04:58:50 09/18/03

Go up one level in this thread


On September 18, 2003 at 06:29:03, macaroni wrote:

>How much searchdepth is it worth giving up for a stronger evaluation function? I
>often find, that while my program is playing GNUCHESS, or AnMon, both the other
>computers give a reasonable + score in a position in which neither side is up a
>pawn (and neither side will be in the ply 8 searches they are managing). My
>almost entirly tactically based program gives a slightly worse score, and loses
>20 moves later (the position being dead lost the whole way). But lets say, I
>code a much better eval function, and as a result, lose about 1 ply. Is it worth
>it? or does the extra ply really count heaps.

It can be worth it.  It depends on how much evaluation you already have.

I did an experiment once... I made a version of GLC 2.18 that used only material
balance in its evaluation.

It played a 20 game match against GLC 2.13.  During the match 2.18 averaged 2 to
4 ply more than 2.13 (mainly due to extra cut-offs, the NPS search speed was
only about 20% faster).

The result... 2.13 won, by 19.5 points to 0.5 points.  This gives roughly a 640
ELO rating difference.

Normally I would expect 2.18 to be about 50 ELO stronger than 2.13.

So loosing the 'positional' part of the eval function seems to have caused
around a 700 ELO drop in strength.

Conclusion: the evaluation is important.  :)

>
>
>[D]2r5/1b2k1p1/p2pp2p/4n3/1pP1P3/8/P1PNB1PP/2K2R2 b - - 0 22



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.