> Wow! That's quite impressive. In a review on the BGA site ( > http://www.britgo.org/reviews/mf11rev.html ), the author (who claims to > also be 1-kyu) says he is able to beat Many Faces v11, even giving it a > nine-stone handicap. Is the version of MF which played the exhibition > game that much stronger than version 11? Or is the 13-year old 1-kyu ten > stones weaker than the BGA 1-kyu? > > I don't mean to be disrespectful, I just am wondering how the same > program can beat one 1-kyu and lose a 9-stone game against another.
I agree that it's impressive, and I agree that it's surprising but there are some things that you must bear in mind. Go programs may play at a Dan level in positions they understand, which tend to be tactical ones. Where they make blunders is in strategic situations, or in tactical situations where they misss the point for some reason. If the game developed in a certain way, the human opponent might be surprised by a sudden show of strength. This is of course just speculation since we haven't seen the game. Dan _______________________________________________ gnugo-devel mailing list gnugo-devel@gnu.org http://lists.gnu.org/mailman/listinfo/gnugo-devel
