Don Dailey wrote:
I can't think of any really clever way to handle this. Here is about
the best I can think of right now:
Keep statistics on each program - how often it loses due to time or
illegal moves. Use those statistics to reduce the impact of the rating
formula. I could arrange it so that even if 10% of your games are
lost that way, your rating becomes highly suspicious and given very
little weight in the rating formula. This basically amounts to
treating the games of such players as if they were un-rated games.
- Don
Wouldn't that drastically reduce the utility of CGOS for experimental
bots? I guess it depends on what you want the server to be, whether you
want as many as possible the chance to get a ranking, or whether to get
as accurate a ranking as possible. And can rankings really be accurate
in any case, with just bots playing bots?
It brings me to another topic I've been considering. If I use mirror go
as an opening strategy for a bot, will that be looked upon unfavorably?
It seems that when most of the opponents a bot faces will be stronger
opponents, mirror go, with some sensible logic on when to break out of
it, would be the optimal opening strategy. But somehow I got the feeling
that a lot of people would be yelling at me if I used that in any kind
of ranked or competition setting, no matter how informal.
_______________________________________________
computer-go mailing list
computer-go@computer-go.org
http://www.computer-go.org/mailman/listinfo/computer-go/
