Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: For Stuart Taylor: CM8000 - CM6000 3.5-4.5

Author: stuart taylor

Date: 17:05:37 05/20/01

Go up one level in this thread


On May 20, 2001 at 15:45:16, Peter Berger wrote:

>Posted by stuart taylor on May 20, 2001 at 09:45:15:
>
>In Reply to: For Stuart Taylor: CM8000 - CM6000 2.5-1.5 posted by Peter Berger
>on May 20, 2001 at 09:25:34:
>
>
>>Er! Whaaat???? Was I wrong all along????????Can this be believed?
>
>>Well, atleast it's a promissing sample for CM8000, but still very slight and
>>short compared to what I believed I've been seeing to counteract it.
>
>Now that already sounds a little different: "believed I've been seeing" ;-)
>
>As far as I remember there have been a lot of negative fan remarks about CM8000
>while the match against Deep Fritz was running without any further facts given.
>
>Then there was one guy who did a match against CM6000 on a single computer
>without taking care of disabling the CM6000 usage of CPU while it's the
>opponent's turn .
>
>We know very little about the strength of CM6000 - look at the little number of
>games against top competitors in the SSDF list .
>
>I haven't seen a single post so far giving any evidence that CM6000 is stronger
>than CM8000 but I admit I probably missed a few .
>
>Maybe the one who claims that CM8000 is a rip-off and people should get their
>money back should do some tests himself ;-) ??
>
>I don't think many tests with CM6000 will be done as it is quite time-consuming
>to play manual games .
>
>>But thanks!
>
>>S.Taylor (gasp! gasp!)
>
>OK ( as I had to sit next to my computers due to work anyway ) - I continued a
>little :
>
>CM8000 - CM6000 1/2 1/2 1-0 1/2 0-1 0-1 1/2 1/2
>
>So CM6000 fought back ; but let's talk about the match of CM6000 against Junior5
>again : see how you can get anything from 0.0-2.0 to 1.5-0.5 for CM8000 out of
>the above sample easily ?
>
>My impression watching the games was that CM8000 and CM6000 are very similar and
>of compairable strength . Opening books seem to be identical - to know which one
>_really_ is better in a direct match against each other you'll need more than
>100 games IMHO .
>
>And maybe CM8000 would win the next 5 games straight : that's quite conceivable
>.
>
>IMHO the above little result is enough to dismiss the theory that CM8000 SIMPLY
>SUCKS !
>
>Games 5-8: PGN availlable on request .
>
>If it is clear which one is better in a direct match we still won't know which
>one is better against other opponents - you see it is basically hopeless :-) .
>
>I think the Chessmaster fans do a lot of tests and produce a lot of data - yet ,
>the results could be better if people worked together .
>
>Take the personality challenge - several people come up with new personalities
>every other day ; and there are a lot of test tournaments, but I have never seen
>any test result posted on which parameter change has a clear positive or a clear
>negative effect .
>
>But I am not in the CM "business" - I simply appreciate _your_ passionate and
>enthusiastic search for THE BEST OF THEM ALL !
>
>I really hope you will have much fun in this last year you'll devote to
>computerchess .
>
>pete

Thanks for such a helpful post, and all the special testing and for the
encouragement!
S.Taylor
S.Taylor



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.