Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Chessmaster 60000

Author: John R. Menke, Sr.

Date: 15:58:18 11/11/99

Go up one level in this thread


Hi Alvaro,

This is just personal opinion, of course, but that seems to be what you are
stuck with.  I'm not aware of any rigorous scientific studies proving the case,
one way or the other, re the strength of various user-defined Chessmaster
personalities.  The closest is Shep's tests, which I consider trustworthy, and
you should consult his web site if you haven't:

  http://sccs.8m.com/index2.html#top

My feeling is that the new piece values in CM6666 are an improvement over
CM6000.  They are a close approximation of the piece values as determined in a
generally respected statistical study by Larry Kaufman, published earlier this
year in Chess Life magazine.  The "catch" is that they are based on human chess
games, which might not translate 1:1 to be valid in computer chess evaluation
algorithms, but I can't imagine a better way to do it!?

I am less sure about the other changes, factors such as king safety, etc.  There
may be a good rationale for those changes, but I'm just not privey to that
information.  Probably somebody else can speak more knowledgably in that regard,
somebody who goes back a couple of years on this board and/or at
rec.games.chess.computer.  I'm a relative newcomer in this area, just started a
few months ago.

As for shortcomings in the opening book:  If you use the same opening book for
CM6000 and CM6666, that shouldn't matter as far as determining the relative
strenghths of those two personalities.  But you could probably devise
improvements in the opening book that would make any personality somewhat
stronger, although I don't know how much stronger.  It might not have as big an
effect as is imagined.  It would also depend on the level of play selected.
Openings don't matter as much at lower levels (elo <2200) of play.

--JRM




On November 11, 1999 at 09:18:21, Alvaro Rodriguez wrote:

>Hi,
>      I´ve been thinking about the Chessmaster 6000 personalities. Maybe the
>default cm6000 personality is better then cm6666 but I´m sceptical to the
>opening lines that cm6000 chooses. Sometimes, after only 20 moves, cm6000 is
>allready behind in material or in position, sometimes, the game is allready lost
>after 20 moves.
>
>      Maybe the best personality is to combinate cm6000 settings but with
>another opening book, maybe a GM-book.
>
>I would like to hear comments by others!
>
>Alvaro

worth a look, if you do much web surfing...
http://www.alladvantage.com/go.asp?refid=FTZ387




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.