Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Bitboards and Quick Killer

Author: Brian Richardson

Date: 16:56:37 08/15/00

Go up one level in this thread


On August 15, 2000 at 18:37:44, Larry Griffiths wrote:

>On August 15, 2000 at 17:14:20, Brian Richardson wrote:
>
>>I would think you would want to try the killer moves before even generating
>>captures and non-captures (assuming you already have the killer move, of
>>course).  Then just keep it simple with a MVV/LVA move ordering, which will
>>place captures before non-captures (count pawn promotion as a capture too).
>>Later do the same with hash moves first, then killers, then generate and so on.
>>Each program uses a different set of algorithims, and you will just have to
>>experiment with your mix.
>>
>>Brian
>
>This is more along what I am thinking.  I still must generate the moves for the
>killer piece to see if it is still a valid move.  Pawns moves are generated
>en-mass because of my bitboards so if the killer piece is a pawn, then all pawn
>captures or pawn moves have to be generated in one fell swoop.
>
>Larry.


Rather than "generating" moves for the killers, try a quick is it still legal
routine.

Regarding Tom's post:  I cannot speak for a lot of programs, but at least for
Tinker and Crafty, doing killers before captures seems to work better.

Brian



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.