Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: futility pruning?

Author: Bas Hamstra

Date: 16:09:06 11/09/03

Go up one level in this thread


On November 09, 2003 at 02:13:47, Christophe Theron wrote:

>On November 09, 2003 at 00:36:30, Daniel Shawul wrote:
>
>>   I am confused about futility pruning at frontier nodes(depth==1).
>>well,the theory says if
>>  mat(node at depth==1) + mat(gain of move) + maximum positional score < alpha
>>   prune all non-checking moves.Ok! Well what I don't understand here is
>>if we don't use the pruning technique here the stand pat cut off will do it
>>at the quiescence node.So what we saved here is just making the move,right?
>>E.Heinz says FP At frontier shrinks the tree by 60%.Please help me out here.
>>
>>regards
>>Daniel
>
>
>
>You cannot stand pat here. You have not reached the QSearch yet.
>
>You are at one ply from the horizon (remaining_depth==1). You cannot just stand
>pat. But if you see that the minimum score (alpha) that you have to reach to
>make this branch relevant is well above what you can expect from a particular
>move, just don't even try this move (unless it's a check) because it is clearly
>futile to try it.
>
>
>
>    Christophe

What he means if if you do not futility prune, you make the move and then goto
qsearch. Where you quickly find out Score >= Beta so FP saves only that move (in
an inaccurate way).

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.