Life In 19x19
http://lifein19x19.com/

A piece of the rules puzzle - the Korean / Japanese dilemma
http://lifein19x19.com/viewtopic.php?f=45&t=18378		 Page 2 of 2
Author:  jann [ Tue Sep 28, 2021 3:30 pm ]
Post subject:  Re: A piece of the rules puzzle - the Korean / Japanese dile
Pio2001 wrote:
Maybe because a sponsor who invests hundred of thousands in a tournament would prefer having secure rules.

I agree with your other points but I doubt a sponsor would mind a juicy scandal / rule dispute in a tournament (for extra exposure). :)

Otherwise stress testing any system involves challenging conditions a bit different to everyday ones. Even in the bot example, the fact that we know (by experience) they don't play perfectly in complicated tsumego (and lightvector even trained a bit on them) shows those are valid and interesting tests of their limits.

For territory scoring some limits are obvious. Strings must be dead or alive as a whole. No combination of statuses could score 3-torazu correctly, thus it needs to be seki and played out in normal play for a small score drift.

But the above four examples are different. Any average player can discuss hypothetical play and score all four correctly - only the official rules can't.
Author:  CDavis7M [ Tue Sep 28, 2021 5:17 pm ]
Post subject:  Re: A piece of the rules puzzle - the Korean / Japanese dile
Pio2001 wrote:
Maybe because go is the only abstract combinatorial game that doesn't have a completely defined rule.

You are missing the point. Go would be a worse board game if it had such a rules.
Author:  Cassandra [ Tue Sep 28, 2021 7:39 pm ]
Post subject:  Re: A piece of the rules puzzle - the Korean / Japanese dile
jann wrote:
But the above four examples are different. Any average player can discuss hypothetical play and score all four correctly - only the official rules can't.

Any amateur Kyu player with a solid knowledge of Igo Hatsuyôron 120's interdependencies can solve that tsume-go, correct play in subvariations included. No AI on this planet can.

The attitude that your artificially designed tsume-go for rulesets have in common is simply a destructive one.
As I have proven here, no one of your tsume-go for rulesets creators is able to solve the reverse tsume-go (show the last seven moves before that position) for the created positions, which makes evident that these to not have the slightest connection to the real world on the Go board.
You all are simply looking for a ruleset that "heals" a very special kind of simple obvious mistakes from "play" during the status confirmation.

You think that you have found a weakness and you use all your energy in search of a "suitable" position. But this weakness does NOT exist IN that ruleset, but only in your comparison of this rulesets with others. One of which YOU assess to be perfect.
But you do not want to play games under that seemingly "perfect" ruleset only.

AI is designed to play real games.
No software developer on this planet would make the requirements for successfully treating Igo Hatsuyôron 120 the foundation of their program.
Author:  Pio2001 [ Wed Sep 29, 2021 11:24 am ]
Post subject:  Re: A piece of the rules puzzle - the Korean / Japanese dile
CDavis7M wrote:
Pio2001 wrote:
Maybe because go is the only abstract combinatorial game that doesn't have a completely defined rule.

You are missing the point. Go would be a worse board game if it had such a rules.


Chinese, AGA, New Zealand, French and British rules are completely defined. And it doesn't seem to me that they play a worse game.
Author:  jann [ Wed Sep 29, 2021 5:18 pm ]
Post subject:  Re: A piece of the rules puzzle - the Korean / Japanese dile
Pio2001 wrote:
Chinese ... rules are completely defined.

Hmm, including their triple ko draws?

I think they COULD be completely defined, but this would need some clarifications:
  • a clear threefold repetition rule (draw/loss/win depending on stones lost in the loop)
  • and clearly stating that superko applies in dispute phases only, after first resumption

I actually think AGA/NZ/etc. would also be better with such "Renaissance rules" approach, to become closer to the Chinese game.
Author:  John Fairbairn [ Thu Sep 30, 2021 1:53 am ]
Post subject:  Re: A piece of the rules puzzle - the Korean / Japanese dile
Quote:
Chinese, AGA, New Zealand, French and British rules are completely defined. And it doesn't seem to me that they play a worse game.


In practice, Chinese pros ignore their own rules and treat triple ko as a draw or replay, ignoring superko.

In practice, most Americans, French and Brits ignore their own official rules (which are really only there to avoid irritating and repetitive questions from beginners) and play what they call Japanese rules.

In practice, virtually nobody plays go in New Zealand.

And as for being completely defined, surely the whole tenor of Rules19 (the forum that replaced L19) is that they are not.
Author:  Pio2001 [ Thu Sep 30, 2021 12:18 pm ]
Post subject:  Re: A piece of the rules puzzle - the Korean / Japanese dile
When I wrote that go was a game which rules were not completely defined (with Japanese rules in mind), I was not talking about superko, which is an easily solved problem. I was talking about the removal of "dead" stones.
What is a dead stone is very difficult to define.
Author:  RobertJasiek [ Thu Sep 30, 2021 3:03 pm ]
Post subject:  Re: A piece of the rules puzzle - the Korean / Japanese dile
John Fairbairn wrote:
play what they call Japanese rules.


Which they also ignore;)
Author:  RobertJasiek [ Thu Sep 30, 2021 3:07 pm ]
Post subject:  Re: A piece of the rules puzzle - the Korean / Japanese dile
Pio2001 wrote:
Chinese, AGA, New Zealand, French and British rules are completely defined.


The truth is: it is easy to completely define them by slight clarifications or (Chinese) simplification.
Author:  RobertJasiek [ Thu Sep 30, 2021 3:11 pm ]
Post subject:  Re: A piece of the rules puzzle - the Korean / Japanese dile
CDavis7M wrote:
Pio2001 wrote:
Maybe because go is the only abstract combinatorial game that doesn't have a completely defined rule.

Go would be a worse board game if it had such a rules.


But... there are completely defined go rules (Olmsted, Berlekamp et al) - yet go is not any worse than before.
Page 2 of 2 All times are UTC - 8 hours [ DST ]
Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group
http://www.phpbb.com/