Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: passed pawn evaluation.

Author: José Carlos

Date: 09:20:52 01/12/04

Go up one level in this thread


On January 12, 2004 at 11:07:19, Daniel Clausen wrote:

>On January 12, 2004 at 10:02:07, Robert Hyatt wrote:
>
>[snip]
>
>>give white two pawns on d2/e2.  Give black two pawns on a2/h2.  Both kings
>>in the center.  Which side do you like?  :)
>
>Black - even if the black pawns wouldn't be so advanced and would still be on
>a7/h7. :)
>
>Problem is, I wouldn't really know what to code here. Of course you can code
>something so the engine won't exchange all pieces so it won't end up in this
>hopeless situation. But maybe it would be better to not let this pawn structure
>occur earlier in the game already, even if there's still pieces on the board.
>Sheesh, hard decision! :)
>
>Sargon

  Set a penalty for this structure in pawn endings. Then scale the eval for it
in any position against material. For simplicity imagine you had material from
10 (all pieces) to 0 (only pawns). Then you could create an array like:

int SCALE[] = {100,90,70,40,10,0,0,0,0,0,0};
eval = STRUCTURE_PENALTY * SCALE[Material] / 100;

  Use this kind of trick for any eval term that relates to material and you'll
see a nice improvement in your engine.

  José C.



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.