On Fri, 2008-08-08 at 14:35 -0300, Mark Boon wrote: > > On 8-aug-08, at 14:16, Don Dailey wrote: > > > Also, it seems silly to me to find super strong players only to > > heavily > > > > handicap them. What's with that? > > > > Actually, that's not so silly. I think a case can be made that super > strong players tend to have a more consistent level than weaker > players. 8-dan pro is also pretty close to the highest level of > humans, so you have a fixed measuring point instead of a moving one > when the program's level goes up.
Yes, except the next match will be some other player at a different level with another arbitrarily decided handicap. > Having said that, I actually thought the pro played some questionable > moves, so there goes the argument about consistency out the window. > I'm also wondering to what extent the pro was maybe deceived by the > poor play in the faster games, maybe he thought he could get away with > some tricks that then back-fired. This of course is the problem with a single data point like this. The pro could be capable of winning 9 out of 10 such matches once learning Mogo's weaknesses, but this single result must serve to tell the whole story. One other thing. Even though I am a firm believer in scalability, I don't believe Mogo truly had better chances in long games. Unless there is some hardware issue that prevents this (which is very possible) then Mogo should have had better winning chances with the short games. It was clear the strong player utilized much more of his time in the fast games and didn't take advantage of his time in the long game (the one he made mistakes in.) - Don > > Lastly I agree that you get a different result from pros once some > money is involved. Offer him $,1000 if he wins and watch what happens. > Most pros can't help themselves once money is at stake :-) > > > Mark > > _______________________________________________ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
