Vlad and Stuart, I'm not completely closed on this issue - but there is lot going against it ....
It's easy to adapt monte carlo programs to have the goal of trying to win as much space or territory as possible but many of us have studied this as see that it seriously weakens monte carlo programs. But this is not the real problem. It seems that the handicap system is not reasonable in general for computers. As has been stated a few times on this group - the player with the handicap starts from a losing position and must play a different kind of game in order to win. It seems that playing the best move possible (best in the sense of maximizing your territory gain) is not the best strategy when playing a handicap game. You literally have to play foolishly in order to dupe your opponent into losing. And the sentiment of the group seems to be that they would rather focus on programs that play the best moves. I agree, as playing GO is difficult enough for computers and playing this other "side game" in addition imposes too much of a burden. Having said all of that, if I felt the sentiment of most of the CGOS participants were in favor of handicapping, I would do it but I don't get that feeling. I personally think small handicaps in 19x19 might be reasonable because I think playing good moves is still a dominant factor - at least at the levels our programs can handle. I would be reluctant to go beyond a few stones. I don't know what a good number is, but I'll take a somewhat educated guess and say 4 stones, because it supposedly corresponds to almost 400 ELO points - which I have come to think of as a conceptual "barrier" of superiority. If you are 400 ELO superior your losses will be rare and in the linear version of the ELO formula 400 is considered certain victory (you can't win ELO points from beating someone 400 points weaker or more. The linear formula therefore cuts off at 350, so that you still get a little bit for beating a weak player.) In fact, I am curious about this - and have a question for all the monte carlo authors. What kind of expectancy is reported when your program is handicapped by 4 stones in 19x19 games? - Don On Fri, 2006-12-22 at 09:59 +0100, Vlad Dumitrescu wrote: > Hi, > > On 12/22/06, Stuart A. Yeates <[EMAIL PROTECTED]> wrote: > > On 12/21/06, Jacques Basaldúa <[EMAIL PROTECTED]> wrote: > > > Handicap play is a *different* problem. > > The rules of go include rules for handicapping. > > It seems to me that this implies that a complete solution for the game of go > > must include the ability to play such games. > > Yes, of course. But is it that difficult? The goal would 'just' have > to change from "winning" to "getting the best possible result". Now if > one has already solved the game for the former goal, it should be > trivial to adapt it for the latter, right? > > As a matter of fact, after solving the game for any goal, almost any > computer science related matter would become rather trivial, I think > :-) I.e. if the NP complete problems are solved, only easy ones > remain! > > best regards, > Vlad > _______________________________________________ > 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/
