Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Null move question

Author: Bo Persson

Date: 01:34:59 07/30/03

Go up one level in this thread


On July 30, 2003 at 02:57:32, Rick Bischoff wrote:

>Hi everyone,
>
>I am trying out null move and am running into a problem-- I am trying the null
>move when beta is still +INF, so what happens is that it calls quies(-INF,
>-INF+1) which will of course fail high.  Is there something else I am missing or
>do I need to avoid null move if beta is +INF?

Yes.

One way to avoid this problem is to not start out with +/-INF as the interval.
Many (most?) programs use a windowed search, where you start out with an
estimated score +/- some margin. This will also save nodes, maybe even more than
a null_move search.

Of course, you will then also have to handle the case where the estimate (or
guess!) is wrong, and all moves fall outside the window. The ususal solution to
*that* problem is to widen the window, maybe several times, until you either
find a move within your window, or get to a +/-INF window eventually.

>
>Thanks,
>Rick


Bo Persson
bop2@telia.com



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.