Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: MTD(f) and Wikipedia

Author: Djordje Vidanovic

Date: 12:32:56 01/20/06

Go up one level in this thread


On January 19, 2006 at 15:01:04, Joachim Rang wrote:

>Hi,
>
>after a question below regarding the "superiority" of MTD(F) and the article on
>Wikipedia about that I decided to modify the entry and added a sentence in which
>I question the "superiority" of MTD(f) over PVS since it has "practical issues".
>Naturally some guys over in Wikipedia now want to see proof of my statement and
>want to know what kind of practical issues there really are. So I have two
>questions:
>
>1.) Would you support the statement that MTD(F) is _not_ superior to PVS and
>while save in some circumstances a few % of nodes has "practical issues" which
>make it a less desirable choice?
>
>2.) Could you (briefly) name some of the "practical issues"?
>
>regards Joachim
>
>P.S.: If someone feels encourage to reformulate the article on MTD(F) and
>Negascout (PVS) more knowledgeable please feel free to do so. ;-)


One of the practical issues regarding your second question might be that a
program that employs MTD(f) can't include any large eval terms if it cares at
all about the speed or ever 'catching up' with a PVS search program

Djordje





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.