Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: draw by repetition behaviour

Author: Daniel Clausen

Date: 06:14:58 05/31/01

Go up one level in this thread


Hi

On May 31, 2001 at 09:04:04, Gian-Carlo Pascutto wrote:

>On May 31, 2001 at 08:55:44, JW de Kort wrote:
>
>>Can anybody explain why this is happening?
>
>a) a bug  (very likely)
>
>b) a knowledge hole. Your evaluation function does not understand
>the positions which are needed to convert the win, for example by
>scoring them too low. Hence everything but the current position
>looks worse. Your program will not want to lower its score and
>hence plays shuffle chess.

My guess is that (b) is more likely. At least it happens to my engine quite
often. For example if you don't tell your engine that it's a good idea to push
some pawns in the endgame, the engine will happily shuffle kings until
draw-repetition-recognition sets in and your engine pushes a pawn in order to
prevent the draw.

Regards,

Sargon



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.