Computer Chess Club Archives




Subject: Re: MTD(f)

Author: Tord Romstad

Date: 07:19:33 12/20/05

Go up one level in this thread

On December 19, 2005 at 17:58:53, Dann Corbit wrote:

>Also Tord Romstad has worked extensively with MTD(f), although I do not know
>which versions of his engine Gothmog used one technique or the other.

Gothmog 0.2.10 and earlier used PVS, all later versions used MTD(f).
MTD(f) worked much better.  There is no doubt that the reason was
that my MTD(f) search was less buggy.  This illustrates what in my opinion
is the major practical advantage of MTD(f):  It is by far the easiest search
algorithm to implement efficiently and without bugs (at least for beginners).

>I guess that he may be able to answer specific inquiries also.

I'll do my best, but I fear I won't have much time to read or post here until
after Christmas.

>For sure, that is worthwhile to play with, if just for the fun of it.

I think that experiment with new and exotic search techniques is a
useful exercise, even if you end up not using them.  Using non-standard
algorithms forces you to think more, and improves your understanding
of how the search works.  I use PVS now, but I think experimenting
with MTD(f) earlier has helped me understand the subtleties of PVS


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.