Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: killer moves?

Author: Uri Blass

Date: 05:42:49 11/05/03

Go up one level in this thread


On November 05, 2003 at 08:37:17, Steve Maughan wrote:

>Uri,
>
>>>perhaps I do not understand something but my chess program has no best move
>>>in case of fail low because it does not claulcate exact score for moves that
>>>fail low.
>
>With fail soft Alpha-Beta you can get a psuedo 'best' move and score.  Some
>programs store this move in the hash table (e.g. Shredder).

How do you know it?

Shredder is not a program that comes with source code.

>
>>To be more correct I am not going to store a move when I have no idea about a
>>move that gives a value that is bigger than beta because if the first move
>>fail high I also do not know the best move and in theory it is possible that
>>the first move is the worst move but still good enough to be above beta.
>
>So if you don't store fail highs or fail lows this means that you only store PV
>nodes i.e. ones where the score is between alpha and beta?  This is not normal -
>most (all?) programs store moves where the score is >= BETA.
>
>Regards,
>
>Steve

I see that I replaced >= with > in my post.

It should be bigger or equal and not bigger.
I store only killer moves when the score is bigger or equal to beta.

Uri



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.