Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Efficacy of moveordering()

Author: Robert Hyatt

Date: 08:33:08 09/09/00

Go up one level in this thread


On September 09, 2000 at 02:02:46, TEERAPONG TOVIRAT wrote:

>Hi,
>
>I want to estimate the efficacy of each method
>of moveordering() . Some problems arise.
>
>Should I use branching factor to determine the
>efficacy ?  If so,the next problem is how to obtain
>the factor ? It varies from ply to ply and from
>game to game. In order to approximate it, should
>I use geometric mean or arithmetic mean ?
>
>Or it would be better to use total searching time ?
>How u did it ?
>
>Thanks,
>Teerapong


There are two issues:

1.  size of the tree.

2.  how long to search the tree.

If you implement your new move ordering idea efficiently, then (2) is _always_
the right way to test.  Because any new move ordering code will likely slow you
down a bit for each node, and you then only hope you search fewer nodes so that
the added computation is offset by less searching.

if you do a "quick and dirty" implementation, just to see if ti will work, then
looking at the size of the tree is the best way, so that sloppy programming
doesn't bias the results against the new idea.  Of course, there is still the
problem of doing an efficient implementation so that it is a "net gain".



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.