On Tue, 2007-01-30 at 10:16 -0800, David Doshay wrote: > Second, with respect to computer algorithms (and I know that the > primary thrust of your argument was NOT directed towards computer > Go) > I think that SlugGo shows rather well that some algorithms do not > scale well. SlugGo gives GNU Go about 72 times as much thinking, and > while it could be argued that some of our heuristics and evaluation > functions sometimes lead us to make worse moves, in a statistical > sense when SlugGo decides to make a move that GNU Go considered to > have a lower value, it is often correct that it is a better move > (even if rarely the "best" from the view of a good human). SlugGo is > at best 2 stones stronger than GNU Go against a third opponent. > Don's > "curve" does hold much better when SlugGo plays GNU Go, where we > have > seen that we can get SlugGo to beat GNU Go with 7 handicap stones > with enough lookahead time.
But my understanding is that SlugGo is like Botnoid, it improves to a point but isn't truly scalable. A truly selective program cannot be scalable unless it is selective in an admissible way. I think SlugGo is a true selective search program right? Computer chess programs are called selective, but they are really what I call "progressively full width." They prune more aggressively near leaf nodes but as they iterate to deeper levels, those same nodes start looking more like root nodes than leaf nodes and get more and more consideration. If any of these pruned moves turns out to be best, it is guaranteed to be eventually discovered. Thus it's scalable. - Don _______________________________________________ computer-go mailing list [email protected] http://www.computer-go.org/mailman/listinfo/computer-go/
