[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Bug-gnubg] woolsey.met vs. snowie.met (10.000 matches) finished
From: |
Joern Thyssen |
Subject: |
Re: [Bug-gnubg] woolsey.met vs. snowie.met (10.000 matches) finished |
Date: |
Sat, 10 Jan 2004 10:30:17 +0000 |
User-agent: |
Mutt/1.4.1i |
On Thu, Jan 08, 2004 at 06:12:34PM +0100, Achim Mueller wrote
>
> Final score: 4829
> 5171
>
>
> Comments on this?
51.7% +/- 1%
So we're confident that snowie.met is better than woolsey.met. I think
1.7% +/- 1% is quite a difference. This is compable to the improvements
of Snowie2 to SNowie3 and gnubg 0.13 to gnubg 0.14!
I'm not sure that the experiment is optimal. It might be better to check
each match score, e.g., play a number of games at each match score. A
wrong "entry" at -n,-m where n and m are small could make ripples
throughout the MET.
Maybe some of the other list readers can comment on this?
Jørn
pgpYkBdXCNvbH.pgp
Description: PGP signature