This really interesting. The strongest Mogo no longer uses UCT at all. It's tuned instead to do very deep and narrow searches. I've seen that other programs that use UCT are using very small C values to make uct also do very narrow searches.
The strong programs also have very smart playouts. With the patterns they use, I expect that when they make a random choice it is between a very small number of options. Can we even call these programs UCT/MC any more? They are not using the bandit problem UCT search any more. They are more like greedy best-first searchers that keep going deeper in the variation with highest winning probability. David > -----Original Message----- > From: [EMAIL PROTECTED] [mailto:computer-go- > [EMAIL PROTECTED] On Behalf Of Olivier Teytaud > Sent: Sunday, February 10, 2008 9:35 AM > To: computer-go > Subject: Re: [computer-go] Re: computer-go Digest, Vol 43, Issue 8 > > > That translates to mean that MoGo no longer uses upper confidence > > bounds, and only uses means. It also means that MoGo will _never_ > > explore improbable children (after a few sims) unless the RAVE value > > yields an unusually high estimate for it. Is all of that correct? > > > > Precisely: I don't see why you would be wrong, but empirically for 9x9, > we have played games against high-level humans and for the (few :-) ) > games that mogo lost, we tried to see which moves were erroneously > chosen > by mogo; if we restart mogo at the same position with a huge > computation time (30 minutes of a fast octocore) mogo always changed > his > mind and moves to a better move. > > So: > - theoretically, I don't see any reason for mogo to be asymptotically > consistent > - there are long computation times during which mogo focuses on a bad > move > - however, we have not seen a case of bad move for which mogo keeps > this move in case of _very_ long computation times > > ==> if someone beats the release MoGoR3 with > very large computation times (time x nbcores = 4h, 1 to 4 cores) > I'm interested in the sgf file and the analysis > _______________________________________________ > 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/
