Yes, there are lots of problem with this. And some of my bots will automatically pass if doing so gives it the immediate win, so already I know of one program that this will not work. As soon as you pass, the game is over.
Of course CGOS could be modified with a rule not to end the game until each player has made at least one non-pass move, but that solves nothing unless all the bots were aware of this rule. Something that might be similar, would be a rule to play the first N moves randomly. It would be a true handicap, but it would not be a consistent handicap. For instance 3 random moves may turn out to be reasonably good, or quite horrible. Perhaps it would be reasonable to start with N non-corner random edge moves which might have a similar effect to pass moves. Or maybe the first N moves could be specified (to be weak moves) and the bot moves to each in turn, subject to it's availability. - Don On Sat, Jun 6, 2009 at 7:49 AM, Álvaro Begué <alvaro.be...@gmail.com> wrote: > On Sat, Jun 6, 2009 at 3:18 AM, Heikki Levanto<hei...@lsd.dk> wrote: > > On Fri, Jun 05, 2009 at 03:49:11PM -0400, Don Dailey wrote: > >> Handicap games opens a can of worms. > > > > Of course, any program is free to give its opponent any handicap it > wants, > > by passing in the opening (if the opponent didn't pass last). > > > > It is up to the operator of the bot to decide when and how much handicap > to > > give, and how to analyze the results. The handicap-giving program can > play > > under a different name, so that for CGOS it looks like a totally separate > > entry, with its own rating. > > This has a few problems: > * Let's say black plays on D4 and then white passes, to give some > handicap. Now black passes and wins the game by n^2 points, according > to Tromp-Taylor rules. > * Similarly, if black tries to give some handicap and passes on the > first move, white may pass and win the game by komi. > * Even if everyone agrees to not do this and continue playing until > at least both players have one stone on the board, what happens when > two programs that give handicap meet each other? > * Then there is the minor issue that a program might place the > handicap stones differently if it knows how many stones it can place > before the opponent starts playing. > > > Álvaro. > _______________________________________________ > 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/
