Computer Chess Club Archives




Subject: Re: About False Fail Highs, professionals, and MTD searches

Author: Dieter Buerssner

Date: 09:56:13 04/13/02

Go up one level in this thread

On April 13, 2002 at 02:39:29, Tony Werten wrote:

>What happened a while back, was that a move would fail high (on score+1), do the
>"check research" (ie score+200), fail high again, print out "fail high" and then
>fail low (on score+pawn). What happened was that new ab-windows would trigger
>new extensions etc. This can be solved (almost) by making sure that during the
>search for the first move, all extensions are triggered, ie search with
>(-MATE,MATE) ie drop aspiration search.

I think, it is both, pruning and extensions. Essentially anything that depends
on search bounds. For my program, and search trees I debugged, it was more
pruning than extensions, but these two are IMHO conceptionally very similar
things anyway.


This page took 0.04 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.