[computer-go] Re: thoughts on 100,000 cgos games

Dave Dyer ddyer at real-me.net
Mon May 8 11:06:45 PDT 2006


 I think the best that can be hoped for at this state (both of
machine learning and Go programming) is to identify a subset
of atypical games for manual inspection.  For example, games
where a large group dies, or games lost against a weaker
opponent.



More information about the computer-go mailing list