Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Are CM9000 human personality ratings accurate for slower time controls?

Author: Eye Witness

Date: 16:36:35 09/06/03

Go up one level in this thread


I tested some of the personalities in a beta test for the program the ratings
are based upon USCF game in 30 minutes.

>I just read in the CCC archive that CM9000 personality ratings were established
>at a time control of fischer 5/10.  Is that true for the human games CM9000
>played as well?
>
>Has anyone played CM9000 personalities at slower time controls?  How accurate do
>they seem?
>
>Back in the late 80's when Larry Kaufman was editor of Computer Chess Reports
>the rule of thumb was to take a computer's speed rating and subtract 200 and
>that would approximate it's tournament time control rating.  Should 150 be
>subtracted in the case of CM9000 'human' personalities (fischer 5/10 is not as
>fast as G/5, so subtract less) to better approximate tournament time control
>ratings?
>
>
>Regards,
>
>Kerry



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.