Computer Chess Club Archives


Search

Terms

Messages

Subject: Pro Deo 1.1 testing by SSDF is wrong !

Author: Andreas Stabel

Date: 02:09:03 06/02/05


Hi,

I started to react when Pro Deo lost all its games at the end of the testing
against two programs. I don't remember the programs, but it was those before
the current testing against Shredder and Junior.

If I remember correctly Pro Deo was leading both matches and then suddenly lost
all 3-5 games at then end of both matches. I thought at the moment that it was
peculiar, but statistically possible. I continued watching Pro Deo and saw that
it won a game in the next match and didn't think more of it before now.

It is now clear that the test against Junior is wrong (see posts by Ed Shröder
and others). The match against Shredder also seems to give too bad results.

All this started to happen at a definite moment in the previous matches. I don't
know if they always run Pro Deo on the same machine - if so something must have
happened to it. If it is run on several machines I must admit a can see no
explanation except perhaps a systematic change in the settings that has turned
out to be bad.

I would call for the testing to be halted until this problem can be solved and
then for all Pro Deo matches to be run again.

I don't think Ed Shöder and his very nice program deserves this.

Still - I would like to thank Tony Hedlund and the other testers in SSDF for
the nice work and incredible amount of time they put into this.

Best regards
Andreas Stabel



This page took 0.01 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.