Don Dailey wrote:
I'm not sure I used the program correctly - it's rather complicated and
I'm not that great with statistics. If anyone is interested in the
settings I used I can provide that.
Hi,
The only subtlety here, is that bayeselo is tuned for chess, and assumes
that draws are possible, and White has a small advantage over Black. In
order to make it work maybe a little better with Go results, you could
do this (assuming games.pgn contains game results):
readpgn games.pgn
elo
advantage 0 ;no advantage for playing White
drawelo 0.01 ;draws are extremely unlikely
mm
exactdist
ratings
No draw at all would be "drawelo 0", but it generates some numerical
problems in the algorithm for computing confidence intervals.
I don't expect changing these parameters would change the final ranking
much, anyway. Maybe they would be a little more compressed, because a
win in chess means more in terms of strength difference than a win in Go.
I would be very interested if you could lower your threshold to 197
games, and include december results, which would include Crazy Stone in
the list ;-)
Rémi
_______________________________________________
computer-go mailing list
computer-go@computer-go.org
http://www.computer-go.org/mailman/listinfo/computer-go/
