Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Static Exchange Evaluation Question

Author: Frank Schneider

Date: 23:48:15 01/24/99

Go up one level in this thread


On January 24, 1999 at 19:04:24, Will Singleton wrote:

>On January 24, 1999 at 08:02:16, Frank Schneider wrote:
>
>>I once added recognition of pins (to the king), but it was slower and did
>>influence the searchresults only a little.
>>
>
>Frank,
>
>I get burned by pins more than I like.  Do you have any specific pin code at
>all?
Yes, pins are penalized by the evaluationfunction.

Part of Gromits incrementally copied und updated datastructures are
attacktables FIXED16 catab[64], oatab[64]; // (for computer and opponent).
If bit i is 1 in catab[p] then p is attacked by piece #i of the computer.
The problem with attacktables is that calculating and copying them is expensive,
therefore you have to write a lot of chess-specific code that uses them to
get a return for the invested time.

Having attacktables it is quite easy to
a) detect pins
b) calculate a penalty depending on the number of pieces attacking and defending
   the pinned piece

c) do lots of other things (kingsafety, finding pieces that attack a square
   in SEE, ...)

>
>Will

Frank



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.