terry mcintyre wrote: > Any estimate of winning probability is only as good as the estimates > of whether particular games are actually won or lost. So what? This is about the sophistication of the approach you are using. Monte Carlo programs use the sophisticated approach of viewing the game globally and to consider what they need to do to win. I cannot believe anyone is still suggesting that they just pick off whatever they can and hope for the best.
Here is an example of why this works so well and why your greedy approach is so wrong: Consider a position where there are 2 groups left that are being fought over. One of these groups is very large and the other is quite small. The computer must win one of these groups or it will lose the game - but if it wins either group it wins the game. The computer estimates that attacking the huge group will result in an impressive win with 80% confidence, but attacking the small group will result in a "tiny" win with 85% confidence. The way you guys are thinking, the big group is the way to go - somehow you must be imagining that it gives you a greater chance of winning the game (because it's a much bigger group.) But if either attack fails, you lose. Clearly, attacking the small group gives you a better chance of winning. But you say, "what if the computer doesn't reliably estimate this and it's wrong?" Well, I'm sure it will make a lot of wrong estimates, but does that change anything? Does that justify it abandoning the strategy of trying to win? Monte Carlo play-outs are where these programs get their scoring estimates. If they are wrong, you can blame the play-outs. If you decide to count stones instead of wins, you are still subject to error. There is a fundamental difference, which apparently isn't obvious to some of us, between playing to win and playing for stones. These are 2 conflicting goals and like anything else in life, if you have conflicting goals both goals suffer. There is only one way around this - if you can give one goal complete priority. Many of us have tried to incorporate both goals and I have not heard of anyone yet having success. If you believe you can have your cake and eat it too, then you are really saying that you don't believe these goals conflict. But we have already proved they conflict. Change your program to count stones instead of wins and see what happens. - Don > > Evidently, even strong programs fail to recognize the impact of > nakade, which will alter the score not by one point, but by ten or > twenty. Their estimate of winning probability is totally wrong. Good > players winnow out losing moves and stick with good moves - the basic > premise of minimax searching. Losing a big group will lead to a win > only if one obtains equivalent compensation elsewhere. Good players > sometimes make sacrifice plays, but failing to recognize that one's > group is lost will totally skew one's estimate of one's winning chances. > > > > ------------------------------------------------------------------------ > Be a better friend, newshound, and know-it-all with Yahoo! Mobile. Try > it now. > <http://us.rd.yahoo.com/evt=51733/*http://mobile.yahoo.com/;_ylt=Ahu06i62sR8HDtDypao8Wcj9tAcJ%20> > > ------------------------------------------------------------------------ > > _______________________________________________ > computer-go mailing list > computer-go@computer-go.org > http://www.computer-go.org/mailman/listinfo/computer-go/ _______________________________________________ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
