Life In 19x19

HermanHiddema wrote:The decision is not arbitrary, it is based on a good understanding of the spirit of the rules. Most people who themselves have a good understanding of the spirit of the rules are quite capable of predicting such decisions with reasonable accuracy.

I simplified. When disputes in otherwise predictable environments (Edo / Meiji Japan Go) arise then because they focus around the there ambiguous aspects. Such aspects lead to unpredictable decisions.

Javaness2 wrote:I find it nice that Chinese professionals don't like superko

Among Chinese, Korean and Japanese professionals, there are each two factions: liking versus disliking superko. (I do not have enough evidence for Taiwanese professionals.)

Is that in the same way that there are those who believe in Santa, and those who do not believe in Santa?

I have been reading the ing rules. If that is the best translation I am afraid they are way too complex for a beginner or for me to understand. So if "my" rules define about the same game I am proud because they are much simpler. I am confident that the practical issues raised here can be solved once the players agree on the statuses of their groups and after removing and filling following the agreement. For example by the scoring methods mentioned by Robert such as pairing stones or rearranging the stones. The boring end then only exists if players refuse to resign or to agree. But that is also boring under present rules.

I believe in Santa!

cyclops wrote:For example by the scoring methods mentioned by Robert such as pairing stones or rearranging the stones.

Please call counting procedures 'counting procedures'! They are not scoring methods.

HermanHiddema wrote:
Actually, Japanese rules have been understood perfectly for centuries. It is just that that understanding included the option of having a referee decide disputes. The modern approach, where we want an algorithm that can perfectly decide every possible situation, is not inherently better, IMO, it is just more popular under current cultural norms.

I don't think that's true, one is a game for 2 players, and the other is a game for 3.

If we ask, "Which is a better game for 2 players" I think one of the two is disqualified.

shapenaji wrote:

HermanHiddema wrote:
Actually, Japanese rules have been understood perfectly for centuries. It is just that that understanding included the option of having a referee decide disputes. The modern approach, where we want an algorithm that can perfectly decide every possible situation, is not inherently better, IMO, it is just more popular under current cultural norms.

I don't think that's true, one is a game for 2 players, and the other is a game for 3.

If we ask, "Which is a better game for 2 players" I think one of the two is disqualified.

I doubt human nature will ever allow us to get rid of referees altogether.

Certainly, though, the algorithmic approach creates a better game for computers to play.

RobertJasiek wrote:Having a highest authority to make an arbitrary decision is a way of handling an ambiguous scoring system, but it is not a way of allowing easy understanding in the sense of predicting the authority's decisions. A scoring must be easy BEFORE and AFTER possible decisions. This is so if before and after are always the same, i.e., if scoring is independent of arbitrary decision-making.

The referee is there because humans occasionally disagree. That is unavoidable, no matter how clearly defined the rules may be. It might be nice to replace the referee with an algorithm or automated system but even determining the life and death of a group is NP-hard - it's impossible, regardless of the counting rules.

This system seems unnecessary. What's wrong with the current system? Time consumption? This is worse. Accuracy? Not really any room for error no matter what you do. And as soon as you lose a stone you're screwed.

Charlie wrote:determining the life and death of a group is NP-hard - it's impossible, regardless of the counting rules.

Area scoring does not need to rely on life and death at all. At the end, a stone scores for a player by being on the board. This is not NP-hard or harder but it is of constant (immediate) "complexity": O(1). One only needs to identify the stone's colour. Determining all stone colours and all territory intersections is in O(n), i.e., linear complexity.

The difference between scoring and counting:

http://home.snafu.de/jasiek/endrules.html
http://senseis.xmp.net/?Scoring
http://home.snafu.de/jasiek/int.html

Annihilist wrote:This system seems unnecessary. What's wrong with the current system?

Which system are you referring to?

RobertJasiek wrote:

Annihilist wrote:This system seems unnecessary. What's wrong with the current system?

Which system are you referring to?

Whichever one we use. Territory or Area scoring. It doesn't matter. Neither need improvement I don't think.

Annihilist wrote:Territory or Area scoring. [...] Neither need improvement

Area scoring:

It does not need improvement, but rules that mention life and death can be improved by not mentioning life and death.

Territory scoring:

It depends on which form of territory scoring is meant; usually the referred to form is life-and-death-territory-scoring in the Japanese / Korean style. Extremely great improvements are possible for every particular ruleset of that style. E.g., the various exceptions apparent from

http://home.snafu.de/jasiek/wagcflaw.html
http://home.snafu.de/jasiek/j1989c.html
http://home.snafu.de/jasiek/difference_ ... rules.html

could be cut down to a reasonably small amount in

http://home.snafu.de/jasiek/sj.html

Note that the remaining exceptions are major ones. In particular, there are different rules for the alternation and the analysis at all.