Wolfgang wrote: > The april edition of "Pour la Science" (which is the french version of > Scientific American) has an article on MonteCarlo methods for go programming. > i thought a little about it, and dont see how this can work.
It does, see also Heikki's reply that didn't include you as receiver but can be found at http://lists.gnu.org/archive/html/gnugo-devel/2007-05/msg00012.html > However they claim to have some success, in particular for 9x9 games. > who knows more about that ? 80% or so of the messages to the computer-go mailing list over the last year have been about Monte-Carlo go, so pretty much anybody doing computer go seriously knows quite a bit about it. > Is this a subject one should look into ? Yes. > what do people think ? It's an enormously big deal and we have only seen the beginning so far. GNU Go has no such code integrated in the distribution yet but there's work going on, see http://trac.gnugo.org/gnugo/ticket/150 and the MonteGNU entry at http://cgos.boardspace.net/9x9/standings.html /Gunnar _______________________________________________ gnugo-devel mailing list [email protected] http://lists.gnu.org/mailman/listinfo/gnugo-devel
