Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Beating MTD(n,f)

Author: Vincent Diepeveen

Date: 16:42:08 06/05/01

Go up one level in this thread


On June 05, 2001 at 16:03:22, Dan Andersson wrote:

>>
>>I'm not personally convinced that mtd(f) is any better (or worse) in _real_
>>positions where the score fluctuates normally, of course.

I don't see how to beat MTD at material only positions, except
with a better move ordering. Anything having a better move ordering
wins of course from a bad ordered tree. At those depths the
difference between MTD and PVS is a peanut compared to something
that improves branching factor bigtime.

>I agree, it's hard to prove its superior to PVS-negasout. But in calm
>positions/parts of games it has its ups. But it's really easy to implement time
>management code for fail high/low in the driver part of the code.
>
>Regards Dan Andersson

Behaviour of MTD is a horror in normal positions when
score slowly drops. Also with pawn being 1000 units MTD is hard to use,
as well as most commercials which give big scores for some patterns there
MTD is not usable.

However for programs like SOS, Cilkchess and similar programs MTD is a
perhaps very good working.

Without hashtable i hear MTD is impossible to use though because of qsearch
issues.

Best regards,
Vincent




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.