> By the way: a 9x9 CGOS server using japanese rules... I have a dream.. ;) > > > Lars > Hi Lars,
I don't want to get too philosophical here and start another "rules debate" so I'll start by saying that I'm not that interested in rules as such. It's way more interesting to me to focus on playing strength. Japanese rules is a fairly large diversion from this in my opinion. When I designed CGOS I didn't want to get involved with the complexities of Japanese scoring because the real motivation for CGOS was to encourage the development of strong playing Go programs. I wanted it to be as easy as possible to get started in this, not just for the program developers, but also for myself. How do you score Japanese games correctly in an automated way in the face of program disputes? There was a lot of discussion about this at one point - one of them being to let gnugo score the games. But of course it involves more protocol to the server such as dead stone agreement and such. Nothing that isn't possible, but it makes things way more complicated for everybody. For a server like CGOS it's a heavy price to pay. As Nick Wedd observed, it's very difficult to get all the program authors to get the software correct for any kind of protocol. Of course if I were to do it, I would simply set up the rules and the protocol and I would not be forgiving of program error (it would be up to the programmer to get it right or risk losing games.) It's already like that on CGOS. Some programs don't do positions super ko correctly and they sometimes lose games. I get a report every month or two about this, claiming CGOS doesn't know how to handle Ko properly but in every case the complainant was wrong. I can easily imagine what it would be like with Japanese rules. I would be getting hammered with sincere pleads to "check the code" because they are sure it's wrong. I chased down every case of the KO bug, I never ignored the complainant just because I knew they would be wrong but I wanted to give them the satisfaction that I checked it out with an open mind (ok, I admit my mind wasn't open - I knew they were wrong but I still always did the homework to set their mind at ease.) Since you have a dream - I will tell you my dream. I wish there was a single set of standardized rules that everyone in the world honored. These rules would promote the game by being as simple as possible so that beginners would understand it. Japanese rules have almost certainly hurt the game in Western cultures. I believe this because I almost didn't discover the game due to these rules - you simply cannot understand Japanese rules without already being good at the game. With Chinese rules, you can learn the rules quickly. Having said all of that, I'm very interested in Japanese scoring from an engineering point of view. I think it would interesting challenge and a lot of fun making my own program able to handle them properly. Of course this is a requirement for a serious commercial program. I have a un-serious commercial palm program that does not know Japanese and it's a requested feature, although not heavily requested. I acknowledge that Japanese rules are popular and here to stay. I would really get interested if I had a strong program and I was in the "polishing" phase of program development, where you provide a fancy GUI and lots of cool features. Even then, it's a lot of work because there is not such thing as one set of Japanese rules, there are many variations. In fact it would be interesting to enumerate those variations, such as ko, komi, handicap system (should there be "compensation"), suicide, time-control and so on. There is surely a sensei page on this. - Don > DD> AnchorMan uses that in KGS mode - it will pass quite early sometimes and > DD> mark dead stones based on the territory statistics you are talking > DD> about. > > DD> So I assume the play-outs are chinese and the move selection is the same > DD> as our bots except you won't move into an intersection that is owned by > DD> either player? > > DD> How reliable is that? I had to be pretty conservative in AnchorMan > DD> about using that, it would fail to defend territory unless I made the > DD> threshold for ownership pretty high. > > DD> - Don > > > > DD> Lars wrote: > >>> I had build an Monte-Carlo GO-Engine (GOMonCy) wich uses the Japanese >>> scoring system. It reached a win rate against GnuGO 3.6 level 10 of >>> stable 50%-52%. I used territorry-statistics about the Monte-Carlo >>> outcomes. You get a probability for every field telling you who is the >>> owner. It works quite good, but I thougt that nearly everyone is using >>> such statistics, isnt't it? Using a threshold to decide that a field >>> belongs to a player you can also handle seki situations. Of course, if >>> it is losing, the engin will break the seki situation an continue >>> losing.. >>> >>> Am Montag, den 05.11.2007, 16:54 -0800 schrieb [EMAIL PROTECTED]: >>> >>> >>>> Jason House said: >>>> >>>> >>>>> What about seki situations? >>>>> >>>>> On Nov 5, 2007 1:41 PM, <[EMAIL PROTECTED]> wrote: >>>>> >>>>> >>>>> >>>>>> It takes some tricky analysis to work out the Japanese score, due to >>>>>> uncertainty about life/death; likewise it's not easy for a program to >>>>>> recognize when moving is no longer to its advantage. >>>>>> >>>>>> How about bringing in a Monte Carlo routine after both players have >>>>>> passed?--as a scoring referree, set to fill up the board (but avoiding >>>>>> eye-filling >>>>>> >>>>>> >>>> ------------------------------------------------------------------- >>>> -->> and self-atari (except in ko situations) <<-- >>>> ------------------------------------------------------------------- >>>> >>>> >>>>>> until all legal >>>>>> moves are played... >>>>>> >>>>>> >>>> ----------------------------------------- >>>> This email was sent using AIS WebMail. >>>> http://www.americanis.net/ >>>> >>>> >>>> _______________________________________________ >>>> 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/ >>> >>> >>> > DD> _______________________________________________ > DD> computer-go mailing list > DD> computer-go@computer-go.org > DD> 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/ > > _______________________________________________ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
