Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Help needed for upcoming Chessmaster Custom Personality Championship

Author: Michael Cummings

Date: 02:33:40 06/07/00

Go up one level in this thread


On June 06, 2000 at 20:49:48, Jorge Pichard wrote:

>Michael you probably have done enough CM testing at 40/2 but you have never
>tested CM=10 vs CMQueen+ at any time control.
>
>Pichard

My answer to that is this. I could take your CM Queen setting and maybe change
one parameter by a point. This would make, depending on what I changed very
little difference to your personalitiy strength, and mine would be pretty much
identical.

So no matter what setting you say is the best, there will always be more
settings to test.

I also have little doubt that you have tested all of the other personalities
that have been floating around since the release of this program, and even going
back to CM5000, in which some setting were transfered from and altered.

I have had numerous settings better than the default. And also had good result
against the SS=10, even beaten it on numerous occasions. But in the long run
after much testing I settled on CMSS=10 settings.

I had a CM6Cummings personality which in my limited testing beat all the other
settings floating around at the time. But in the end gave up on it, I had spent
long enough testing.

One setting might beat all other CM settings, but it may loose against another
chess program, where as the setting that lost in the CM tournaments, can usually
beat another chess program.

That said test what you like, give your result, and hopefully at the end you
will feel as though you achieved something.

Unless CM8K makes up for the CM7K disgrace, then no matter what settings you
use, chessmaster will be dead in the serious chess community and no one will
care.

CM6K had a buzz around it. CM7K fizzed out very fast.



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.