Le vendredi 22 décembre 2006 14:50, Don Dailey a écrit : [...] > 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.
I strongly disagree with this. One player need handicap because his plays are inferior (overconcentrated, paranoiac defense, unefficient moves...), his understanding of global position and urgency is much weaker, his reading skill is poorer, knowledge (tesuji, life...) smaller and he makes more blunders than opponent (each additionnal blunder can be seen as a need for one more handicap stone). White does not need to overplay to win. Just keep calm, play normal move and punish every mistake is very often enough for winning. I often play handi games with white, and i am always surprised that i can win without overplay (sometimes mistakes but not voluntary overplay). Overplaying (or trying trick moves) is always dangerous and this is not the good way for improving. This may have some teaching virtue, but it's a kind of "stealing" the victory, so not a glorious one ;-) Handicap stones gives global advantage, and help for simplifying position and not being crushed in an difficult corner sequence. (9 handi strategy for black is just "avoid local disaster") And i m pretty sure that MC bots will correctly use the handicap stones, either free placement or ala japanese (on star points) > 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. http://gemma.ujf.cas.cz/~cieply/GO/statev.html gives 20% victory for even games against a 4-stones-stronger opponent (in the range 15k-5k) so this seems closer to 250 ELO. Maybe 5 handi could be considered too, as the additional center stone is very helpfull (breaks many ladders and help escape ...) without changing too much the game on the borders (6 and above is really a different game) According to pro, 4 stones handi (on star points) is equivalent to 40 points komi, but for computers it is probably different. > > 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 > Alain _______________________________________________ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
