On Sat, 2008-10-11 at 23:49 +0100, Raymond Wold wrote: > Don Dailey wrote: > > Are you speculating that if enough play-outs are done, the idea might > > prove to be superior? I never actually considered that. So perhaps > > with 5000 playouts using the win/loss score wins, but at 50,000 using > > the margin might be better? This is easy to test with simple MC go > > programs. If I get a chance I will test it for you (since you do not > > have a program of your own) and report the results. I'm not holding my > > breath however ... > > Indeed, that was my /guess/. And one aimed at rationalizing why this > particular idea might somehow work. It might well be that something else > is needed to utilize the score difference after a play-out, or it might > be that it really is worthless... Without a much deeper understanding > (in a mathematical sense, not a player-skill sense*) of go, I don't see > how we can find out which ideas really can't work.
I believe there have been many attempts to make this work. These attempts are based on the intuition that the margin approach should be better even though it is clearly inferior. So in my opinion they start with a wrong premise. And this wrong premise is combined with another wrong premise that the win/loss thing is inherently flawed, but works due to some fluke. So the idea is to somehow combine something that seems to work very well, with something clearly inferior and get something that works better than both! I'm guessing that most approaches based on combining the two scoring systems is like taking great wine and mixing it with cheap wine and hoping to come up with something superb. Also, one must consider if it's even possible to get much more out of the win/loss method anyway. How much more can you gain making it fight when it is dead lost or dead one? That is probably why it has been so difficult to improve, because you have so little to gain. To me it's now a question of finding a way to make it behave more like people do without weakening it too much. I don't have a lot of hope that you will improve it and even if you do, it will be a small improvement. But I think it's possible to improve the behavior we find so ugly without unduly weakening it. - Don > _______________________________________________ > computer-go mailing list > computer-go@computer-go.org > http://www.computer-go.org/mailman/listinfo/computer-go/
signature.asc
Description: This is a digitally signed message part
_______________________________________________ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
