Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Fidelity versus Rebel et al

Author: Fernando Villegas

Date: 13:38:13 10/05/97

Go up one level in this thread



On October 05, 1997 at 05:59:21, Robert Sullivan wrote:

>I have a Fidelity Chess Challenger, which has a clock speed of 3mHz. If
>one played the newer programs against Fidelity with the maximum ply
>searched held equal, would this result in a comparison of the relative
>strengths of each program, independent of processor speed? In other
>words, would this exercise yield information about which program played
>the "best" chess?
>Robert

Probably the mistake -if there is any- in Robert approach is that he
assumes that equal ply-search- tree means equal  numbers of nodes and so
any difference in the quality of the game could be asociated ONLY with
the  quality of the code. It is not so simple; differences in the
structure of the program, that is, of his  quality, means very  often
-or ever- differences in the number of nodes searched in the same number
of plys. Maybe a program that search with a lot of prunning see less
moves in 5 plys that a brute force program,  that looks  all moves. But,
at  the same time, maybe the speed of the CPU could let the newest
program to see more moves in the same numbers of ply , even if he is
selective, that the old one that try to see everything.  So, at the end,
it seem that there is  not any way to compare programs on the ground  of
external coertions, like time alloted, number of plys, number of nodes
per second, etc. Maybe even the question has no sense at all, as it was
the old discusion about how strong is Deep Blue  IF we put it in a
normal PC instead of the IBM monster. Speed, number of nodes per seconds
and things like that are such an integral part of the program as the
code lines, so I believe..



This page took 0.02 seconds to execute

Last modified: Thu, 07 Jul 11 08:48:38 -0700

Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.