Computer Chess Club Archives




Subject: Re: mate threat extension/null move

Author: Stuart Cracraft

Date: 12:18:22 09/25/04

Go up one level in this thread

On September 25, 2004 at 14:59:07, Zach Wegner wrote:

>On September 25, 2004 at 14:46:06, Peter Fendrich wrote:
>>On September 25, 2004 at 13:51:39, Zach Wegner wrote:
>>>>When you find a mate it first will get x > beta... Right?
>>>>Then you will never reach the following statement.
>>>No, this is when the side to move gets mated, giving a score < -MATE + MAXPLY.
>>Right but I meant that return beta will not show -MATE+MAXPLY one ply up.
>>This nulmove only code, however, and the answer is dependent of what's happening
>>in the alpha/beta stuff later on...
>Sorry, I misunderstood you. But the mate threat code is typically not (x < -MATE
>+ MAXPLY) but (x == -MATE + ply + 2) so that first when the side to move is
>mated it returns -MATE + ply (ply being 2 more than where the nullmove is
>called) and then at ply + 1 it returns MATE - ply (this is a simple
>change-return val instead of beta at a fail high)

But doesn't returning -MATE+ply from search() mean that the program
might postpone making a mate since it is not returning -MATE+depth
and so the ply is increasing and so more credit for deeper mates
whereas depth is decreasing so more credit for more shallow mates?


This page took 0.02 seconds to execute

Last modified: Thu, 07 Jul 11 08:48:38 -0700

Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.