Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: How do authors rate against their own programs?

Author: Robert Hyatt

Date: 07:42:51 11/14/98

Go up one level in this thread


On November 13, 1998 at 02:39:10, David Blackman wrote:

>On November 12, 1998 at 19:36:02, Chris Moreton wrote:
>
>>I am the author of Rival chess which has a rating floating around 2000-2100 on
>>the ICC while I feel good if I stay above 1000 for any length of time.  Is their
>>anywhere I can find out how other authors rate against their own programs?
>>
>>-Chris
>
>I don't know of any authoratative source about this, but the short answer is
>that most programmers would be completely smashed by their own programs, unless
>they are quite new to chess programming and still have a very weak program.
>
>One possible exception is Peter McKenzie, author of LambChop, who posts here
>occasionally. Peter is a former New Zealand champion, and although i don't think
>he has a FIDE title, he could not be far below IM strength. Lambchop is a
>reasonably strong program, especially for its positional play, but i think Peter
>would have some chance against it if he played seriously.
>
>My own program, Desperado, is much stronger than i am, and it is definatly not
>one of the strongest programs out there. (My Australian rating is 1504.
>Desperado hasn't competed much against rated players recently but might be about
>2000. An earlier version once managed to fluke a win against an IM.)


I occasionally beat Crafty, at full strength, but it isn't exactly about
"chess" when I do.  I have to (a) avoid a tactical mistake which is hard
against crafty, (b) play into positions where I think it might not be able
to whip up some sort of tactical activity, and (c) I use specific knowledge
about what it does well and what it doesn't do well to help guide me.  I
probably win around 1 of every 20 games I play, when the games are longer (ie
not blitz) so I can try to avoid tactical mistakes.  But this is really helped
because I know the program so well, particularly how it evaluates positions.
It isn't "easy" to exploit it, but every now and then I get away with it.  Of
course, 1 out of 20 is not exactly spectacular...



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.