|
BGonline.org Forums
New GNU Strength. Some Q&D tests.
Posted By: Michael Depreli In Response To: New GNU Strength. Some Q&D tests. (eXtreme Gammon)
Date: Friday, 5 April 2013, at 8:21 a.m.
I completely agree with Frank,
I will strongly advice to not use long match for benchmarking for any program. AS Frank pointed, there are precision issue, also MET above small value are interpolated and no-one knows really how good these interpolation are. anyway who are playing 64 point match?
I'd suggest using match <11 or unlimited game.
Also if using match, the same MET should be using on the playing and analyzing program. it does not really matter which one as long as they are the same.
I wanted to use 100 unlimited games for the test games but couldn't work out how to stop the program so selected the longest match possible which was 64 games.
I only really wanted a very quick indication which I understand is not ideal and I don't plan on doing any extensive tests.
I'm guessing the GNUBG team can come up with some more significant numbers
Michael
|
BGonline.org Forums is maintained by Stick with WebBBS 5.12.