Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Precision about NegaMax (on B. Moreland's site)

Author: Robert Hyatt

Date: 21:59:22 04/08/04

Go up one level in this thread


On April 07, 2004 at 21:45:41, Mathieu Pagé wrote:

>Hi, I want i confirmation that the algorith on this page of Bruce Moreland:
>
>http://www.seanet.com/~brucemo/topics/hashing.htm (In the middle of the page)
>
>it is named : AlphaBeta but I think it is NegaMaxWithAlphaBeta.
>
>Am I right ?
>
>Mathieu P.


This is a terminology mix-up.

The basic search is called minimax.  A simplified version that is recursive is
usually used, and it is usually framed as a "negamax" implementation.  But there
is no _actual_ difference between the two as far as results go, but the
recursive negamax is generally liked by those that don't choke on recursive
code.

alpha/beta is an enhancement to minimax, not something totally independent.  IE
you can do a minimax alpha-beta or a negamax alphabeta that produce identical
results...

Cray Blitz used a non-negamax search as Fortran-66 had no recursion at all.

Crafty has been negamax from the beginning.



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.