Many Faces won its match today against James Kerwin 1p, played at 7
stones, by 4.5 points. The program was running on a 32-core cluster
supplied by Microsoft. The match was played live in front of a press
briefing at the 2009 AAAS general meeting.
The game record will be available on KGS shortly, if it is not
already. Note that KGS scored the game incorrectly, so the result in
the game record is wrong.
As David says:
This version of Many Faces plays pure Chinese rules, so it will place
handicaps in nonstandard places. It counts all stones on the board
at the
end of the games as points, including the original handicap stones.
KGS scores differently. In a Chinese handicap game it gives one point
compensation to white for each handicap stone.
We need to play with pure Chinese rules. Keep this in mind when
setting the
handicap. Each handicap stone is worth one point more than a Japanese
handicap stone.
At the end of the game, the KGS score will be incorrect. It will
show white
with one extra point for each handicap stone.
Thanks to both David and Jim for participating. I got a lot of good
questions at the AAAS press conference; hopefully this will add some
popular interest.
Bob Hearn
_______________________________________________
computer-go mailing list
computer-go@computer-go.org
http://www.computer-go.org/mailman/listinfo/computer-go/
