In his history articles, John Fairbairn mixes a lot of interesting and entertaining information with some salt of doubtful opinions. The following citation from "Rules poser in the Kisei" is a typical example:

"[...] the western hubris that rules discoveries are made here and not in Japan, we may [...] point out that Hayashi's main article was entitled 'Tradition and logic in the Japanese rules of go - discussion of the endgame, dame and local life and death.'"

Which, if anybody, Westerner has said that rules discoveries were made here and not in Japan? Rather some Westerners have said that most and the most important discoveries on modern (Japanese or other) rules were made by Westerners. This can be supported by listing and comparing discoveries, if necessary. Such is realism and not hubris.

If the mentioned article by Hayashi should contain some noteworthy discoveries, then what are they?

It is not like one person would have done all the research. Some very important discoveries were made decades ago. Then it is a matter of opinion which of the important discoveries were the most important. E.g., one might call these candidate topics: first well-defined ruleset, equivalence proof, definition of life, some specific definitions of life, definition of strategy for games in general, definition of strategy for go, definition of ko.

I think it is all a matter of how you look at it.

You can argue that the most importand rules are discovered in the east (how else would we come to play it now in the west?)
or that the impurant rules are made at the west. (New sealand being part of the west )

I use discovery and made on purpose
In the east our game is more seen as something almost god given. and crazy situations are seen as some rare species, where the gods had did not pay attention. Not for nothing there are stories that it is played in the clouds, the hand of god, and that i was given to us by the emperer, a living god long ago.

In the west our game is seen as a human game of logic and intelligence. To be played on cumputers, and we need rules to cover any situation, because that is what computers need. irrespective if the situation can uccur in actual play.

And still it is the same game.

_________________
Promotor and Librarian of Sensei's Library

Of course, Go had been invented millenia ago in the East. This thread is about MODERN rules discoveries though.

In modern times, only few rules discoveries appear to have been made in the East. Presumably Ikeda Area III and Ikeda Territory I Rules are some. They are compromise scoring rulesets though. Concerning regular scoring rulesets, there have been J1949, WAGC, J1989, C1988, K1992, Ing 1975, Ing 1986-97.

***

Which discoveries have been made by them?

J1949: none

WAGC: none

J1989: life can be classified into uncapturable and capturable (correct, good definitions are missing though), a definition of Japanese style territory can be approached via eye-point and in-seki (correct, good definitions are missing though), (the hypothetical ko rule is not an important discovery but nonsense for the purpose of practical applicability)

C1988: none

K1992: none

Ing 1975: none (or was Ing Counting discovered already then, would need to check)

Ing 1986-97: Ing Counting (important in theory but horrible in practice), first rough sketch of a basic ko types classification, (the Ing ko rules are a failure more than a discovery though)

***

All those rulesets failed to discover or describe correctly all the really important basic definitions:

- hypothetical-strategy
- force
- two-eye-formation
- uncapturable
- capturable-1
- capturable-2
- territory
- seki
- dead
- eye
- eye-space

Even worse, J1989 went astray and confused players by avoiding two-eye-formation entirely.

Ing Rules, which presume exact related definitions, failed to define these terms:

- ko (in general)
- ko stone
- cyclically / repeatedly
- invariation

***

Of those important discoveries that should have been made, only J1989 and Ing86+ made only a small fraction and failed terribly at providing correct and complete definitions.

So while the Eastern rulesets and their authors discovered only little, very much has been discovered by Western rules researchers.

***

Game rules were transferred to Europe (the initial place outside Asia) mostly verbally. Early written documents had been mostly insufficient, see Eurogo Vol. 1.

***

Reference to divine relation is a bad excuse. Disoveries that were not made were not made indeed. It is that simple.

The presence of rare exceptions does not justify weak, ugly, flawed discriptions of the standard life and death concepts used in Eastern rules.

I do not see the basis for your distinction of "compromise" and "regular". I regard all the "regular" rulesets you mention as deviations from Ikeda (A III/T I) rules.

_________________
A good system naturally covers all corner cases without further effort.

I may be misunderstanding what you mean, but I'm pretty sure you can define living groups without needing to make reference to either "two-eyes" or "seki."

1) Groups of stones will be removed from the board when they have zero liberties.
2) Players may place no more than one stone per turn.
3) At game-end, all groups which are agreed by both players to have no chance of avoiding capture will be treated as if captured.

This is obviously not a complete ruleset, but I think it is entirely sufficient in itself (when taken within the context of a complete ruleset which defines such things as "groups" "liberties" and "game-end") to be a robust definition of life and death.

And then (for Japanese style rules) you are back at having to have some such definition of life and death, for which J1989 avoids the two eyes concept altogether.

1) Groups of stones will be removed from the board when they have zero liberties.

That is my definition of life and death. Rule #3 above is just a shortcut.

Or do you mean that J1989 doesn't define capture at all? I haven't read it, that's why I say I'm probably just confused about what you're saying.

Such a definition is insufficient for Japanese rules! If you used only such a definition for the status determination for scoring, then before scoring the players would have to remove stones by actually approaching and filling their liberties. You should know that such changes the game dramatically because, under traditional territory scoring, a player would be forced to sacrifice part of his territory for the purpose of filling it with his own approach and liberty filling stones.

Your definition works well only for stone scoring, area scoring and some sorts of pass stone scorings.

And then you forget to appreciate the strategic definition level of life. The liberty-alive definition is very low level. The much higher level definitions WAGC-alive or J2003-alive are useful as strategic concepts under every scoring method, if only one adds a definition module. E.g., when you look onto a position during the middle game, you want to speak of high level life rather than low level life.

http://www.cs.cmu.edu/~wjh/go/rules/Japanese.html

This is a good translation of the official Japanese rules.

That is the point: life is a strategic concept, not one of the rules. The rules just tell you which stones to remove.

Of course, this does not hold under all rule sets.

_________________
A good system naturally covers all corner cases without further effort.

That's why I had my rule #3 listed above as well -- "Groups of stones which both players agree will never be able to escape capture may be treated as captured at game-end."

This is the way it works in practice in amateur games as well as being a concise and clear rule - if the players cannot come to resolution on group status by themselves due to confusion among beginners or whatever then they seek the help of the tournament director (who of course may seek the help of a stronger player if he himself is an improper judge of the situation, but in practice that has never been necessary as far as I am aware.)

Another option is to provide a means (as Japanese rules do) to provisionally play out the life and death situation on a separate board, if necessary, so that the score is not altered by its resolution.



Harlequin has my response to this 100% correct - a definition of "two-eyes" is not necessary from a rules perspective, as they are actually derivations of the rules which are known to players for strategic purposes, not (or they do not need to be) basic rules in themselves.

Seki may be defined without reference to two-eyes as well:

Seki is a situation in which groups of stones are in a position such that if one player would attempt the capture of another group, his own could be captured without killing his target. Territory adjoining the contested liberties is not counted towards either players' score. (Addendum for Japanese rules: Territory solely controlled by one group in a seki situation does not count toward either players' score.)

Even when you add some sort of superko rule, this definition of seki stands on its own. As a superko violation would allow the group which move would violate superko to be captured, when that group must initiate the sequence, the result is still a seki.

And in fact, as a basic rule, this definition is unnecessary except for Japanese rules, because of the special treatment of eyes in seki. If territory has been defined (as it always is to my knowledge) as "intersections on the board which are completely surrounded by a single group of living stones" then the null territory between groups follows already.

From my understanding, at least, when one speaks of rulesets one doesn't speak of encyclopedias that detail strategy and emergent behavior -- instead one is defining an algorithm by which the basic game itself may be played. A "rulebook" may contain tutorials on how best to play the game, but such a chapter is not actually part of the "rules of the game."

Also, my definition shows easily why the Japanese rules which define bent-four are so difficult. They introduce a whole special case for a situation which would most times naturally fall out into the conclusion they desire, but logically would very rarely be seki instead.

Earlier you wrote:



Either you take #3 seriously or you don't but when you want to do both it becomes impossible to reply.



Not in amateur games in general; such a rule is used only for some rulesets. Also it does not work in case of disputes.



Yet more concise as rules is ommission of that rule.



Not in case of a disagreement.



Rules do not become any simpler by hiding their essential contents and putting it elsewhere.



It has been necessary; I have watched that.



The do not prescribe it being done on a separate board.



The question is relevant for Japanese style rules. There some definition of life (or a substitute for it) is necessary. It can be a definition using two-eye-formation or capturability.



We are not that far. I know that my rules models have some readers but more realistically still the vast majority of players has only an informal understanding of life as a strategic concept while life developed in terms of rules definitions is created mostly independently from that.

Whether life is part of the rules or derived from the rules depends on which rules we are speaking of. Japanese style rules use the former while typically for other rules it is the latter.



In case of Japanese style rules, life belongs to the rules themselves; without life they would be incomplete.



There are better and more precise definitions of seki. Actually yours is not a definition; it is too ambiguous.



Your usage of "group" is ambiguous. Earlier you used it in the meaning of string, here you might be using it as set of strings.



As you see in my rulesets, seki is not needed for Japanese style rules, either.



I wrote more precise definitions.



Japanese style rules do not completely forgo that though.

Well, it's a serious shortcut. I don't see why calling a stone a stone makes it less serious.

But anyway, I've screwed up with my objectives for this conversation. I wasn't spoiling for a fight, nor was I (initially) trying to out-rule the rules-meister, who is definitely more practiced than I at stating things without ambiguity when he wants.

Mainly I just was wondering about the statement that J1989 rules require two-eyes but don't define them. It seems to me that it is entirely possible to unambiguously define life without making reference to eyes, and I was wondering if that is in fact what they did, or is it actually a flaw in them as written?

The statement you gave to which I responded initially seemed ambiguous in that it sounded as though you were saying that the flaw was in the lack of definition of "two-eyes," not in the lack of definition for life.

They tried and failed, see my commentaries.



The flaw is in teaching the players as readers of the J1989 rules badly, especially while not knowing yet of Chris Dams's later proof.