Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Branching factor too high -- need ideas

Author: Dann Corbit

Date: 22:45:21 09/16/04

Go up one level in this thread


On September 17, 2004 at 00:56:23, Jeremiah Penery wrote:

>On September 16, 2004 at 17:37:48, Stuart Cracraft wrote:
>
>>Hi -- my program's branching factor on an old IBM notebook
>>for Win-at-Chess at 1 second per move on is 4.28 solving245/300 positions.
>>
>>This is too high a branching factor.
>>
>>(It was calculated by averaging all the timeperply/timeperpreviousply for all
>>iterations on all problems in the suite.)
>>
>>What do people do when their branching factor is too high. What kinds of things
>>best improve the branching factor?
>>
>>Thanks -- Stuart
>
>What is the average depth of search?  I would guess that a very short search in
>the WAC positions will produce an abnormally large EBF, because of the nature of
>the positions and the general instability of very short searches.  It would be
>much better to test some relatively deep searches on different types of
>positions.

A branching factor of 4.28 for all of the WAC positions is very high, even for
short searches of a second or two.

The very best programs will be under 3 (on average).



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.