[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Bug-gnubg] Unexpected results Gnu vs Gnu
From: |
Paul Hankin |
Subject: |
Re: [Bug-gnubg] Unexpected results Gnu vs Gnu |
Date: |
Fri, 1 Sep 2006 21:41:49 +0100 |
On 1 Sep 2006, at 21:30, Peter Carlson wrote:
Ran an interesting test. Gnu windows versus Gnu cli. Both use the
same settings file, both I assume, use the same neuro net and
libraries.
Test:
Ran gnu windows. Set player 0 to gnu with same eval as the eval
settings
Before and after each roll, copied matchid and postionid and pasted
into the cli
requested a hint on the doubling and a hint 3 for the next moves
Played the moves given back as a hint by the cli in the windows game.
Outcome:
gnu windows beat the cli 16 to 5 over 10 games.
Question:
Why? Is there something obvious I am missing here? Is there a way
to make this be more even so the results would be closer to 50/50.
This isn't an interesting test; 10 games is far too few to show any
difference in skill between the two versions.
If you want closer results, play more matches.
--
Paul Hankin