Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: COMPUTER CHESS LEAGUE (CCL) introduction

Author: Peter Fendrich

Date: 13:03:19 05/19/98

Go up one level in this thread


On May 19, 1998 at 14:07:46, Ed Schröder wrote:

>>Posted by Peter Fendrich on May 19, 1998 at 11:16:59:
>
>>In Reply to: COMPUTER CHESS LEAGUE (CCL) introduction posted by Ed Schröder
>>on May 19, 1998 at 09:53:48:
>
>>Ed,
>>I think a league is a good idea but I don't think you should compute
>>ratings out of thousands of games based only of these programs. It's too
>>few and it doesn't help if you play more games between the same
>>programs.
>
>More participants are planned, this is just a warming up.
>
>
>>A league with more programs involved and maybee somewhat changed
>>conditions is my advice...
>
>What are these changed conditions?
>Any proposals?

What I had in mind was different kind of tournaments instead of only
one. Like different time controls (the blitz league, fisher time etc),
different themes (like sicilian or queens gambit etc, with and without
opening lib's and other types of conditions. That would give some more
added value than just another "1000 of games rating" with probably about
the same results as the SSDF tests.

What I'm thinking of is a variety of results under a variety number of
conditions.
Together with the SSDF testings on a variety of different hardwares,
which I don't regard as 'unfair' (not arguing about Fritz here, just the
concept of different hw's) in principle as long as Rebel9 as well as
others on P90 is regarded as completely different players than the same
programs on P300, we will get an ocean of data to argue about...

Tha last one is probably the longest scentence I've ever written!

//Peter




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.