Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Search algorithms

Author: Robert Hyatt

Date: 07:43:25 11/13/03

Go up one level in this thread


On November 12, 2003 at 21:02:50, Dave Gomboc wrote:

>
>No, he isn't.  MT-SSS* isn't mtd(f).  mtd(f) is the new algorithm that is better
>than MT-SSS (a.k.a. mtd(-oo)).
>
>Dave


OK.  But mtd(anything) is depth-first.  There is no way in a normal
search a depth-first search strategy will search the same tree, in the
same order, as a best-first search strategy like SSS*, except for
pathological cases.  I suppose you could corrupt the idea of depth-first
to say that it searches the same tree in certain cases, and make that
true.  But not for the general case.

IE the concept of the "open list" simply doesn't exist in depth-first,
by definition.  I'm constrained to search the tree as it is generated,
contrary to best-first.

However, it is all moot.  I've only seen one SSS* program ever compete in
a computer chess event, that being a version of HiTech.  And they turned
SSS* off and went back to PVS once they reached some "endgame threshold".



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.