Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Attack Tables

Author: Vincent Diepeveen

Date: 06:30:44 06/26/01

Go up one level in this thread


On June 26, 2001 at 05:25:22, Bas Hamstra wrote:

>On June 25, 2001 at 12:56:26, Artem Pyatakov wrote:
>
>>I have decided to keep track of attack tables in my program, because I think
>>they will save computation time in the eval.
>>
>>My question now is this, given the experience of most people in this group, what
>>is the most useful info to keep in the attack table?
>>
>>Should I just keep the number of times that white is attacking the square and
>>black is attacking the square? Or is it very useful to keep track of what actual
>>pieces (and on what position on board) are attacking the square on both sides?
>>
>>Or is it some other kind of attack info?
>>
>>Thanks for all of your responses.
>
>Two possibilities (there are probably more) are:

>Keeping track of which pieces attack a certain square, where you number pieces
>uniquely from 0 to 31. This info fits nicely in 32 bits, since there can be no
>more than 32 pieces. It can give some complications for pawn promotions, but
>that can be handled.
>
>Another possibility is using bitboards. For a certain square, you can read from
>an attacktable the bitboard that contains the set of squares that attack that
>square.

You're only describing a sucking bitboard solution, please checkout
gnuchess 4.0, there was an entire different solution there, and that's
the one where he's looking for.

Note with the gnuchess 4.0 version you can also order moves better.

Best regards,
Vincent

>Best regards,
>Bas.



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.