Computer Chess Club Archives




Subject: Re: PVS and MTD(f) branching factor

Author: martin fierz

Date: 01:24:00 06/18/03

Go up one level in this thread

On June 17, 2003 at 20:11:12, Thomas Mayer wrote:

>Hi Russell,
>>>NegaScout/PVS improves alpha-beta by about 10%. For some people MTD(f) works
>>>better, but it has some delicate problems associated with it.
>>But that is a constant speed improvement, and not a reduction in the branching
>>factor, correct?
>correct - like any pure alpha beta has a branching factor of around 6 when we
>say that there are approx. 40 possible moves... -> usually the formula
>5*SQRT(Nodes) is near to practice... any better move ordering would bring a
>constant speed improvement that would have no influence to the branching
>Greets, Thomas

better move ordering reduces the branching factor! you can easily see that from
best/worst case which is sqrt(N)(best) and N(worst) as branching factor for
alpha beta. that's not a constant speed improvement then...


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.