I think probably I just go with ELO, much simpler. I think later we will want to have handicaps. Even at 9x9 Mogo is all by itself although I expect other programs to eventually catch up or get close later.
For the Anchor, I think I will take David suggestion and start with AnchorMan. There is nothing that prevents us from experimenting with this later and finding a better Anchor. Now the question is: What initial rating to give AnchorMan? It's rather arbitrary anyway, so I probably stay with 1500.0 An interesting thing we could do is set up each player to have 2 identities, one of them is just the player with a 1 stone handicap. They would be identical in every way except that they would receive separate ratings. They would never be paired against each other and of course a 1 handicap player would never be paired against another 1 handicap player since this is impossible. Just a thought. - Don On Wed, 2006-12-13 at 11:32 +0100, alain Baeckeroot wrote: > Le mercredi 13 décembre 2006 05:53, Don Dailey a écrit : > > > Does a 1 kyu difference mean I can give you 1 stone if I am better and > > expect to come out about even? > yes, 1 handi is 0.5 komi. > > > > Does this all work out in a transitive way? If a 6 kyu can give a 7 > > kyu 1 stone, and the 7 kyu can give an 8 kyu 1 stone, can the 6 kyu > > expect to play even with the 8 kyu player giving 2 stones? > yes, and it works surprisingly well. > > > Would this simple system work: > > > > 1. Start all players out at the same kyu rating. > > > > 2. Pair randomly. > > > > 3. If you win your match, modify kyu rating slightly down. > > > > 4. If you lose your match, slighly change kyu upward. > > Kgs works like this (with more subttle algorithm). > > > > > > All this is applied on top of handicaps of course. > > > > But unless 2 players are an integer kyu apart, a handicap would be > > slighly > > unfair to one side or the other. Is it sufficient to modify the > > ratings in linear proportion to the amount of "unfairness?" > > Less than 1k difference is nothing for weak players. It is only > meaningful for strong players (several dans or pro) > The link below is stats on even games from European Go Federation > http://gemma.ujf.cas.cz/~cieply/GO/statev.html > > As GNU Go is rated 6k on kgs , this should give more than 30% > for a 9k to beat gnugo in even games. > > The traditional way for adjusting handicap needs 3 win in a row (this > is rather difficult) > The fun way is changing handicap after each game (for human > the psychlogical part is very important, one can manage to lose > with many handi due to emotive factor or desire of revenge ...) > > Maybe for computer the handicap could be remembered between 2 oppononents, > and the global rank estimated from this ? > > GNU Go does not eat memory, even at level 10 it is small and rather fast. > At level 0 it is very poor in reading (rated 2k below level10 gnugo on kgs) > but level 8 should be rather good. > On cgos 9X9 i checked the first 100 000 games of GNU Go 3.7.4 and found > less than 10 nearly nearly identical games (against viking) and less than 5 > were rigorously identical. So i bet on 19X19 this will not happen at all. > > my 2 cents. > alain _______________________________________________ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
