User-friendly Reading of the Japanese 1989 Rules

[RobertJasiek]

Your attached table:

Classes? What would be the classification?

I have not checked every detail of yout table, but #1 under J2003 you have analysed wrongly. It has one capturable-1 and one capturable-2 string and each string's status applies to ALL its stones.

[Cassandra]

... but #1 under J2003 you have analysed wrongly.

Sorry Robert,

A typo. Of course Black's chain in example #1 is capturable-1. Has been corrected.

... and each string's status applies to ALL its stones.

I'm well aware of the fact that the status given is given uniform for either a single stone or a chain (i.e. "all of its stones") after the end of the evaluation.

But the status found may be found only for one single stone of the complete chain (i.e. "at least one of its primary points").

I think we can use 3 questions to highlight the subtle, but decisive, distinctions between our approaches.

• Should chains be treated as a whole (i.e. "for all") ?
• Should status determination of a chain depend on previous results for others ?
• Should "local" be bordered by "living" chains of each colour ?

I suppose your answers being "NoYesNo", mine are "YesNoYes".

[RobertJasiek]

Firstly I specify for which purposes I design a ruleset. Then I try to design the ruleset for those purposes. Your questions are of comparatively little relevance for that.

Here the major purpose is to design a Japanese style ruleset. Such requires certain characteristica for life, as stated earlier.

If I were to design some territory scoring ruleset with some form of life concept in it, then I would follow a much greater freedom (much greater than yours).

[Cassandra]

Here the major purpose is to design a Japanese style ruleset. Such requires certain characteristica for life, as stated earlier.

As stated earlier, you have chosen the long and hard way when starting with "life".

I had the understanding of your first posting that your aim is the creation of a user-friendly reading of the 1989 Nihon Kiin rule set.

I also had the understanding that you want to use "independently alive" only during Analysis.

Should this be true, you would have to explain, which of the examples on Life & Death you have chosen that must not fulfill your requirements (or elsewhere would make some precedence necessary to fit its Nihon Kiin result).

[Harleqin]

• Should chains be treated as a whole (i.e. "for all")?
• Should status determination of a chain depend on previous results for others?
• Should "local" be bordered by "living" chains of each colour?

I suppose your answers being "NoYesNo", mine are "YesNoYes".

Regarding the first question, are you saying that of a connected string of stones, some may have a different status than the others?

[Cassandra]

• Should chains be treated as a whole (i.e. "for all")?

Regarding the first question, are you saying that of a connected string of stones, some may have a different status than the others?

No, as I have written in my reply to Robert's posting, it is beyond dispute that a (= one and only one) status is given to a chain (= connected string of stones), what means to all of its elements (= stones). This is not the moment when "as a whole" vs. "at least one" will provide a problem.

The problem - in my opinion it is one - arises before, during the course of the evaluation / determination sequence, in the event of the chain in evaluation being captured.

The fate of a single stone that has been captured during Analysis is bound to the fate of its successor. What - im my opinion - must be true with a chain, too. Remember that during "Play" stone and chain are treated in the same manner, they have equal characteristics.

A successor of a chain must mandatory occopy all of its primary points. If any of its primary points is occupied by an opponent's stone, there will be no successor of the chain.

But among chains that can be captured, 1989 Nihon Kiin rules and also Robert's capturable-1 give "life" to

• one permanent stone "under the stone" (= Uttegaeshi)
• one permanent chain "under the chain" (= Nakade)
• one permanent stone "under the chain" (= e.g. Black's 4-stone chain in example #1)

In my judgement, the latter is inconsistent.

Black's 4-stone group in #1 in theory could consist of - let's assume - more than 280 stones (may be even more, I don't know) on the whole board. Of which only one (= 1 !) will have a successor.

Instead of assuming "live" to the whole Black chain (and also to the single White stone that will have no successor at all - that's due to "would enable" respectively "local-2"), resulting in "Seki" for #1, I would like it more to have "removable" for the single White stone and "stable" (can neither be forced to 2-eyed nor to removal as a whole) for Black's group, resulting in "Seki" for #1, too. "not independently alive" for both would give the same result.

[RobertJasiek]

The inconsistency of the kind you describe can occur. The consistency between WAGC-alive and J2003-alive is always given though, and I consider it to be much more relevant.

I am not sure why you ask about "to use 'independently alive' only during Analysis". For Scoring, 'independently alive' is applied. This is used for strategic planning before Analysis. So why do you suggest I might use it only during Analysis?

For 'independently alive' as in the Simplified Japanese Rules, I do not care (much) which rare examples might behave differently.

For J2003, I did care a lot which examples would behave how:
- In principle, ALL examples from Japanese professional or related sources should fit: rules examples, games, books, journals.
- As an exception, when Japanese professionals changed their judgement over time, then their most recent judgement should fit. (Like 3-points-without-capturing shape or moonshine life.)
- As an exception, where the rules experts among the Japanese professionals provided purely artifical examples AND were inconsistent / incomplete in their analysis, not necessarily all their judgements must fit. (I.e., J1989 official commentary examples II.16-18.)
- Where purely artifical examples were invented by Western rules experts, any somewhat reasonable judgement should be considered fitting.

J2003 Rules meet these requirements. (And do so better and much more exhaustively than what the Japanese professionals had studied as examples.) Slightly different rulesets, which have not been worked out so far though, might have chances as well. (Winfried Borchardt had an idea to compile something like J2003 but with a 2-layer process of life types. I think he has not finished working out that yet.)

[Cassandra]

... WAGC-alive ...

Do you refer to the 1979 rules text ?

I am not sure why you ask about "to use 'independently alive' only during Analysis".

During the stage of Analysis, "independently alive" is the one and only status, your user-friendly rules look for.

For 'independently alive' as in the Simplified Japanese Rules, I do not care (much) which rare examples might behave differently.

I suppose you even care less what the reasoning might be (e.g. within "user-friendly").

- As an exception, when Japanese professionals changed their judgement over time, then their most recent judgement should fit. (Like 3-points-without-capturing shape or moonshine life.)
- As an exception, where the rules experts among the Japanese professionals provided purely artifical examples AND were inconsistent / incomplete in their analysis, not necessarily all their judgements must fit. (I.e., J1989 official commentary examples II.16-18.)
- Where purely artifical examples were invented by Western rules experts, any somewhat reasonable judgement should be considered fitting.

I feel very uncomfortable with your statement above.

Sentence 1 (which in itself provides no problem, because it only says "use state-of-the-art") follows "Example provided by Japanese." >>> "Example evaluated be Japanese."

Sentence 2 follows "Example provided by Japanese." >>> "Characteristics of the example judged by N.N. (probably non-Japanese)." >>> "Characteristics of the Japanese evaluation judged by N.N. (probably non-Japanese)" >>> "Japanese judgement judged by non-Japanese."

Sentence 3 follows "Example provided by non-Japanese." >>> "Characteristics of judgement judged by non-Japanese." >>> "Non-Japanese judgement judged by non-Japanese."

[RobertJasiek]

WAGC-alive: I refer to http://home.snafu.de/jasiek/wagcmod.html

Within my rulesets, I care very much what the reasoning is.

Concerning your uncomfortable feeling about sentences 2 + 3, why are you uncomfortable with them? Because it is well known that the Japanese professionals changed the rulings of 3-points-without-capturing shape or moonshine life over time?! Don't you believe that? Then read, e.g., the rules texts J1949, WAGC, J1989! - And what is your problem with my characterization of Examples 16-18? That I have been able to describe the huge gaps in their analysis, to find move-sequences overlooked by the professionals, to point out their inconsistency in their analysis and to assess the missing relation between terms in commentary to terms in the rules?

[Cassandra]

@ Harleqin: A supplement to my posting above.


Let's further refer to example #1 and assume that the "desired" final result is "Seki" (i.e. none of the 5 stones will be taken off the board as prisoners).

Following the precondition that only "not alive" (used as in 1989 Nihon Kiin rules) chains within opposing "alive" chains will become "dead" (i.e. will be taken off the board as prisoners), we can consider 2 alternatives, both giving the chains identical statuses.

1) Both White's single stone and Black's 4-stone chain are "not alive".
2) Both White's single stone and Black's 4-stone chain are "alive".

In alternative 1) none of the 2 chains in evaluation is situated within opposing "alive" stones, so both of them will remain on the board.
In alternative 2) there is nothing "not alive" that probably could be taken off the board.


Reasoning could be as follows:

1) "not alive"
White's single stone will be captured and its primary point will not constitute part of an "alive" chain.
White's stone is in danger (because"not alive") to be turned into a prisoner, should Black's chain become "alive". But:
Black's 4-stone chain will be captured and its primary points (i.e. "all of") will not constitute part of an "alive" chain".
So Black's chain is "not alive", too.

2) "alive"
Black's 4-stone chain will be captured and one of its primary points (i.e. "at least one") will constitute part of an "alive" chain.
White's stone is in danger (because now being inside opposing "alive" chains") to be turned into a prisoner. So it is mandatory that its status must become "alive", too.
White's single stone will be captured and its primary point will not constitute part of an "alive" chain. It becomes obvious now that the procedure used to get Black's chain to "life" will not work here.
A second procedure is wanted in desperation. Suddenly it is regognized that during the evaluation sequence White had occupied at least one point of the board, where NO White stone had been before in the primary position. Let's turn this into the second procedure, using "would enable ...".
Be aware that there could be examples, where "where no White stone had been before" could go too far away from the battlefield (i.e. beyond "living" groups).


I hope it has become apparent that the reasoning for 1) is simpler than that for 2), needs only one procedure instead of two, and provides no side-effects on chains not evaluated so far - in contrary to 2).

If there remains some uncertainty, it might be preferable to chose "not alive", because this is the choice that has NO side effect on neighbouring chains, never.


If you go through the results of "my idea" for the 1989 Nihon Kiin rules examples, you will recognize that in the somewhat "mysterious" examples (containing Double- or Triple-Ko) it will make no difference to the final result (excluding their well-defined parts, as Bent-Four), whether the chains have been evaluated as "stable" or "removable". Really important is that they did NOT get the status "2-eyed". What is true also with Roberts "independently alive" vs. "not independently alive", using "user-friedly".

This way - using the minimal possible scope for (independent) life - minimises potential side-effects in situations one had not been aware of before.

[Cassandra]

WAGC-alive: I refer to http://home.snafu.de/jasiek/wagcmod.html

(quotes from that text will be in blue, formatting is done by me)

In a position, a string of a player is two-eye-alive if the opponent cannot force no intersection of the string with a two-eye-formation on.

This is the reason for your "independently alive" of Black's 4-stone chain in example #1.
"for all" would result in "can force at least one intersection of the string with no two-eye-formation on."

J2003-alive is defined like in J2003 as either uncapturable, capturable-1, or capturable-2.

This is the reason for your "independently alive" of White's single stone in example #1.

In a position, a string is WAGC-alive-in-seki if it is J2003-alive and not two-eye-alive.
In a position, a string is WAGC-alive if it is either two-eye-alive or WAGC-alive-in-seki.
In a position, a string is WAGC-dead unless it is WAGC-alive.
...
Chris Dams has proven:
WAGC-alive equals J2003-alive.

Sorry, Robert, but that's trivial, isn't it ?

J2003-alive = uncapturable OR capturable-1 OR capturable-2
uncapturable = uncapturable-two-eyed OR uncapturable-Seki
capturable-1 = capturable-1-two-eyed OR capturable-1-Seki
capturable-2 = capturable-2-two-eyed OR capturable-2-Seki
two-eye-alive = uncapturable-two-eyed OR capturable-1-two-eyed OR capturable-2-two-eyed
WAGC-alive-in-Seki = J2003-alive AND NOT two-eye-alive
WAGC-alive-in-Seki = uncapturable-Seki OR capturable-1-Seki OR capturable-2-Seki

WAGC-alive = two-eye-alive OR WAGC-alive-in-Seki
WAGC-alive = (uncapturable-two-eyed OR capturable-1-two-eyed OR capturable-2-two-eyed) OR (uncapturable-Seki OR capturable-1-Seki OR capturable-2-Seki)
WAGC-alive = (uncapturable-two-eyed OR uncapturable-Seki) OR (capturable-1-two-eyed OR capturable-1-Seki) OR (capturable-2-two-eyed OR capturable-2-Seki)
WAGC-alive = uncapturable OR capturable-1 OR capturable-2
WAGC-alive = J2003-alive

Excel would complain about a circular reference.

Within my rulesets, I care very much what the reasoning is.

Perhaps you should care about "marketing", too ?

Concerning your uncomfortable feeling about sentences 2 + 3, why are you uncomfortable with them? Because it is well known that the Japanese professionals changed the rulings of 3-points-without-capturing shape or moonshine life over time?! Don't you believe that? Then read, e.g., the rules texts J1949, WAGC, J1989! - And what is your problem with my characterization of Examples 16-18? That I have been able to describe the huge gaps in their analysis, to find move-sequences overlooked by the professionals, to point out their inconsistency in their analysis and to assess the missing relation between terms in commentary to terms in the rules?

I suppose the approach we prefer is very different.
There is no problem with "state-of-the-art has changed over time" or "a non-Japanese had found inconsistencies within the rules text or its application".
When I read your sentences then arises an implicit feeling like "There is someone (may be non-Japanese) who decides what to include and what not to include to get the final result he wants." If this is not your intention, you should think about better "marketing", as written above.

My "marketing-friendly" suggestion would be the other way round.

1) Develop a rule set that is well-defined and consistent. (You have done already.)

2) Apply this rule set on the "known" examples (e.g. your #0000, Nihon Kiin Life & Death (You have done already.)). The results will be well-defined and consistent, because the rule set is. As a matter of course some of the results may not be identical to those provided with another rule set (e.g. Nihon Kiin) that has been identified as not well-defined or inconsistent or even both.

3) Develop an idea where the different results come from (we can exclude inconsistent application of the rules themselves here) and try to shape your ideas into a kind of "classes" (some elements of what I mean can be found in the "motivation" parts of your texts). This will make discussion easier, because you can discuss about principles to apply (or may be not) and have no need to worry about single "mysterious" board positions. And what is important also: Discussion about "principles" will not be mixed up with discussion about inconsistent rules' application.

The result of 3) could be something like the following:

I think we can use 3 questions to highlight the subtle, but decisive, distinctions between our approaches.

• Should chains be treated as a whole (i.e. "for all") ?
• Should status determination of a chain depend on previous results for others ?
• Should "local" be bordered by "living" chains of each colour ?

The answers can be only "Yes" / "No", never "Right" / "Wrong".

Questions like "What kind of principle do you like more ?" should be absolutely preferred, questions like "What result in this special position do you like more ?" should be avoided in this context.

When these "principles" are identified, it will be possible to "transform" one rule set into the other. You will know which rules how to change or which precedence to include.

[RobertJasiek]

Cassandra, your discussion on #1:

As far as my personal opinion is concernced, I do not care for #1. I would happily accept one of the following: 1) both alive, 2) both dead, 3) Black alive, White dead, 4) White alive, Black dead. The #1 shape is locally non-terminal anyway.

Since the Japanese professionals chose (1) as their latest judgement, that is what I realize in J2003.

If one does not care for the judgement of Japanese professionals, then one is pretty free to choose (1), (2), (3) or (4). In that case, one might try to go for some sort of relative simplicity. (I do not think though that your general approach is simple.)

Can you please explain how general your statement "chose 'not alive', because this is the choice that has NO side effect on neighbouring chains, never" is? Is it meant to hold for ALL examples in ALL positions? If so, prove your statement!

[RobertJasiek]

Your attempt to repeat Chris Dams's proof: You make a mistake already in the first line. It must be:

J2003-alive = uncapturable XOR capturable-1 XOR capturable-2

I have not read the rest of your proof yet. Please check first whether your mistake has impacts on the rest.

I do not consider Chris proof trivial. (That you start with a mistake shows that it is not indeed ;) )

Marketing? Nah, I have done enough of that for my rulesets:)

"There is someone [...] who decides what to include and what not to include to get the final result he wants.": The initial intention had been to include also #16-#18 exactly as the official commentary suggested. But when I realized that that did not make sense, then 1) somebody - and the somebody had to be me since I was doing the study - had to make a related decision and 2) I modified the intention as slightly as apparently possible and necessary to create J2003 as an otherwise general model. It is not like "I chose what I wanted" but rather like "I chose what makes the greatest sense within the given context".

More or less I used your step 1-3 approach for creating J2003, except that I had to iterate about 35 - 50 versions as likely candidates for a step 1 ruleset.

You are right to ask for the step 3 principles. I cannot provide them fully so far because they are too deeply hidden in my research files. So at the moment, for #16-18, I can tell you only from my memory part of those principles:

- "collapse of the seki" in the official commentary has to be criticised as not providing explanation because that concept is not a concept of J1989.
- A pure text interpretation of J1989 should be tried if anyhow possible while consistent with also all other examples. [I found out though that such a pure text interpretation is inconsistent over the set of all known (relevant) examples.]
- As far as possible, the officially stated statuses should be recreated with a consistent method that works also for all other known (relevant) examples and that approaches a pure text interpretation of J1989 as closely as otherwise reasonably possible. [I would have to look up my files to check what is recreated. No time for that now.]

Your sample set of principles is insufficient because it ignores professional judgement entirely.

[Cassandra]

Your attempt to repeat Chris Dams's proof: You make a mistake already in the first line. It must be:

J2003-alive = uncapturable XOR capturable-1 XOR capturable-2

I have not read the rest of your proof yet. Please check first whether your mistake has impacts on the rest.

There is no mistake, Robert.

XOR is unnecessary, OR is fully enough, because the "variables" have nothing in common. They are "exclusive" in themselves.

XOR would be necessary if a chain would be possible, which is uncapturable AND capturable-1, for example, and in this case would NOT be J2003-alive. I think you know that this will not happen.

[Cassandra]

Robert, I will answer to your previous 2 postings tomorrow in the evening.

I just had an extensive posting ready, clicked on "Submit" and was taken by surprise by "You have to login to ...".

As a matter of course the written text had disappeared.