Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Singular extension question

Author: Robert Hyatt

Date: 10:32:36 03/16/03

Go up one level in this thread


On March 16, 2003 at 06:30:19, Martin Bauer wrote:

>Hi,
>
>how can I see, if one move in a position ist much better, then all others? I
>think AlphaBeta gives me only for the best move an exact value. So may be the
>second best value is only slightly under the best value, but in fact is is much
>worse.
>
>So what is the common way to detect if i should do an singular extension? Or did
>I missunderstand something in AlphaBeta?
>
>Regards
>
>Martin


Search the first move with the window alpha,beta.

Search the remainder of the moves with alpha-X, beta-X.  If all moves fail low
there, while the first move did not fail low on the original alpha/beta window,
you know the first move is > alpha, and the rest of the moves are <= alpha-X,
where "X" is the singular margin.

That's the trick.

It isn't quite as simple in reality, as the first move might produce a score
> alpha, and then another move fails high on alpha-X,beta-X, so now you have
two moves that seem to be better than the rest.  You have to re-search both
of them to see if one is "X" better than the other (IE the second move could
be the real best move and it could be singular.)

It gets technically messy...  but it is doable.



This page took 0.01 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.