> In a game of self play, if both parties are employing only monte > carlo, ... random simulations... wouldnt it be very weak... > ... and some playing around I am clearly mistaken because its works > quite well.
Yes, it doesn't make sense but it does indeed seem to work :-). > I have seen papers implementing patterns and prior knowledge with > monte carlo (dated 2006) has that become a "standard" now, and when > people refer to monte carlo they dont mean absolutly random? Just using monte carlo playouts is still weak. Monte carlo + UCT is strong. (UCT is a specific algorithm, so people are now referring to the set of similar algorithms as "MCTS" for Monte Carlo Tree Search.) No one is using purely random playouts in any top program, but David Fotland described Many Faces' playouts as "fairly light". I think Crazy Stone (as described in detail in one of Rémi Coulom's papers) or Valkryia perhaps have the heaviest playouts, and Mogo is somewhere between? Too much knowledge in the monte carlo playouts and you use too many CPU cycles and cannot do enough of them. (And the real situation is even more complex than that, as sometimes adding knowledge that seems really fundamental for human players does not seem to help, even before allowing for the extra CPU cycles used.) Darren -- Darren Cook, Software Researcher/Developer http://dcook.org/mlsn/ (English-Japanese-German-Chinese-Arabic open source dictionary/semantic network) http://dcook.org/work/ (About me and my work) http://dcook.org/blogs.html (My blogs and articles) _______________________________________________ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
