Maybe we need a Japanese speaking go player to explain/describe the rule in question.
I find it difficult to believe that what is meant is entirely different from the traditional Japanese way of determining the status of disputed groups at the end of the game. Refinement, perhaps yes. Making clear what might be ambiguous. But not TOTALLY different.
Japonese counting
-
Mike Novack
- Lives in sente
- Posts: 1045
- Joined: Mon Aug 09, 2010 9:36 am
- GD Posts: 0
- Been thanked: 182 times
-
Cassandra
- Lives in sente
- Posts: 1326
- Joined: Wed Apr 28, 2010 11:33 am
- Rank: German 1 Kyu
- GD Posts: 0
- Has thanked: 14 times
- Been thanked: 153 times
Re: Japonese counting
Gérard,Gérard TAILLE wrote:Let's take the following position: what is the status of?
According to the loop you highlited in your last postDo you really believe a japonese referee will declare this stone alive?
There are a lot of flaws in the rule. Here the way to handle a loop in the confirmation phase is simply not defined and common sense must apply.
You do not really understand.
in your example is NOT attached to the double ko.Therefore, it is dead. It is completely irrelevant for the status of
that White could capture Black's stones with a. Just because no group in the upper left is under consideration now.There are various examples in the J89 text, which explain this principle.
Probably it would help your understanding, if you studied life-and-death problems 19 to 21.
There, the White bent-four-in-the-corner is attached to a Black group that does not have two eyes. The status of White's group depends on the number of liberties of that attached Black group.
Quite the contrary, an isolated bent-four-in-the-corner, which is surrounded by a two-eyed group of the opponent, is ALWAYS dead, no matter the rest of the board.
+ + + + + + + + + + +
"Ad infinitum" in the confirmation phase ALWAYS results in the status "alive" for the group under consideration, if that group had NOT been captured before.
If the group had been captured before, "ad infinitum" ALWAYS results in the status "dead" for the group under consideration, as no new stone can be placed on the board that will remain there forever.
The really most difficult Go problem ever: https://igohatsuyoron120.de/index.htm
Igo Hatsuyōron #120 (really solved by KataGo)
Igo Hatsuyōron #120 (really solved by KataGo)
-
Cassandra
- Lives in sente
- Posts: 1326
- Joined: Wed Apr 28, 2010 11:33 am
- Rank: German 1 Kyu
- GD Posts: 0
- Has thanked: 14 times
- Been thanked: 153 times
Re: Japonese counting
Status assessment under J89 has NOTHING to do with "common sense".Gérard TAILLE wrote: BTW I am also surprised to see you conclude here it is a seki while the common sense tells us black is dead! Remember that, in normal play, black is dead even if she plays first.
And again, it is COMPLETELY irrelevant what might have happened before the game stopped. Otherwise, there would be NO special ko rule for the status assessment.
However, this special ko rule is unsuitable for achieving the desired purpose in all cases.
The really most difficult Go problem ever: https://igohatsuyoron120.de/index.htm
Igo Hatsuyōron #120 (really solved by KataGo)
Igo Hatsuyōron #120 (really solved by KataGo)
-
Cassandra
- Lives in sente
- Posts: 1326
- Joined: Wed Apr 28, 2010 11:33 am
- Rank: German 1 Kyu
- GD Posts: 0
- Has thanked: 14 times
- Been thanked: 153 times
Re: Japonese counting
I doubt that you have an understanding of what is "behind the written rule".Gérard TAILLE wrote:Oops I am very surprised to see you act as these western players who do not really understand what is behind the written rule.
You simply exploit what is strickly written even if the result looks stupid.
Many of your comments are evidence that you haven't even studied the comment and the life-and-death examples extensively.
The authors of the rules tried to declare several types of triple-ko "dead" with the crowbar, but apparently knew that this was against their own rules.
Otherwise, there would have been no need to resort to a construct such as declaring a group dead by "dissolving of the seki", for which there is no equivalent in the legal text.
Should the declared results for the examples 16 to 18 be the intended ones, the authors of the rules should have pored over the wording of Article 7, 2. longer.
Again, study the deciding difference to examples 19 to 21 (bent-four instead of double-ko), which can be solved (i.e. giving the declared result) without any contrary to Article 7, 2.
+ + + + + + + + + + + + + + + +
This position does NOT have the same properties as ...
... this one, right?
By what reason should it be necessary to enforce that the marked White groups had the same status in both cases? Only because a double-ko is involved?
+ + + + + + + + + + + + + + + +
In examples 19 to 21, the status of the (White) bent-four-in-the-corner depends on the number of liberties of an attached (Black) group.
-- If Black's group has too few liberties, the bent-four is alive.
-- If Black's group has a larger supply of liberties, the bent four is dead.
By what reason should it be necessary to enforce a different behaviour for the status of a double-ko?
The really most difficult Go problem ever: https://igohatsuyoron120.de/index.htm
Igo Hatsuyōron #120 (really solved by KataGo)
Igo Hatsuyōron #120 (really solved by KataGo)
-
RobertJasiek
- Judan
- Posts: 6273
- Joined: Tue Apr 27, 2010 8:54 pm
- GD Posts: 0
- Been thanked: 797 times
- Contact:
Re: Japonese counting
Common sense is not needed. My J2003 interpretation of J1989 treats it correctly:Cassandra wrote::wt: in your example is NOT attached to the double ko.Gérard TAILLE wrote: what is the status of :wt: ? [...] common sense must apply.
The big white string is uncapturable due to the infinite alternating sequence B[kkPP]*, where k is ko capture and P is ko-pass. White triangle is neither uncapturable nor capturable-1. Its local-2 is:
White triangle is not capturable-2. Therefore, it is dead.
Re: Japonese counting
As I wrote earlier you seem to misunderstand what it means to "enable" a new live stone. If you cannot carry out a capture without enabling something (that wasn't originally possible) elsewhere, that is not a valid capture in J89. Spatial locality means nothing - that's why pass-for-ko (with global enabling) was invented for.Cassandra wrote:It is completely irrelevant for the status of
The problem with double kos and ko passing is one of the oldest well known defects of the text. But because it would break almost anything it is safe to assume that such perpetual ko passing loop is not allowed - even if the rules authors overlooked the problem.
Other than that, there are several beasts where the ko pass rules change local outcomes and require protecting moves that would not be necessary in normal go. Those are, I think, accepted consequences of J89.
-
RobertJasiek
- Judan
- Posts: 6273
- Joined: Tue Apr 27, 2010 8:54 pm
- GD Posts: 0
- Been thanked: 797 times
- Contact:
Re: Japonese counting
Infinite ko pass cycles have been intended by the J89 authors, see official commentary II.25b. http://www.cs.cmu.edu/%7Ewjh/go/rules/Japanese.html
Re: Japonese counting
I meant using a remote double ko (with enabling rule) to prolong analysis. Those moonshine-like problems are the main reason for special rules in confirmation (trying to achieve locality without explicit localization). They clearly aren't allowed - even if the text fails to handle them as intended.
-
Gérard TAILLE
- Gosei
- Posts: 1346
- Joined: Sun Aug 23, 2020 2:47 am
- Rank: 1d
- GD Posts: 0
- Has thanked: 21 times
- Been thanked: 57 times
Re: Japonese counting
You tell me to study the examples given in the rule but at the same time you say that examples 16 to 18 are mistaken. How can I trust the results given in all these examples?Cassandra wrote: I doubt that you have an understanding of what is "behind the written rule".
Many of your comments are evidence that you haven't even studied the comment and the life-and-death examples extensively.
The authors of the rules tried to declare several types of triple-ko "dead" with the crowbar, but apparently knew that this was against their own rules.
Otherwise, there would have been no need to resort to a construct such as declaring a group dead by "dissolving of the seki", for which there is no equivalent in the legal text.
Should the declared results for the examples 16 to 18 be the intended ones, the authors of the rules should have pored over the wording of Article 7, 2. longer.
Again, study the deciding difference to examples 19 to 21 (bent-four instead of double-ko), which can be solved (i.e. giving the declared result) without any contrary to Article 7, 2.
How can we reach the same understanding? Can we concentrate our attention only on the loop you proposed with the double ko?
In the rule I see four examples with double ko : examples 11, 16, 17, 18. In each of these examples the white group is declared dead by the authors. If I understand correctly your previous posts your declare white alive in the example 16, 17, 18.
Let's try to progress slowly. What is for you the result of example 11?
-
Cassandra
- Lives in sente
- Posts: 1326
- Joined: Wed Apr 28, 2010 11:33 am
- Rank: German 1 Kyu
- GD Posts: 0
- Has thanked: 14 times
- Been thanked: 153 times
Re: Japonese counting
"Pass-for-ko" is NOT a global issue, but a LOCAL one!jann wrote:As I wrote earlier you seem to misunderstand what it means to "enable" a new live stone. If you cannot carry out a capture without enabling something (that wasn't originally possible) elsewhere, that is not a valid capture in J89. Spatial locality means nothing - that's why pass-for-ko (with global enabling) was invented for.Cassandra wrote:It is completely irrelevant for the status of
The main challenge of territory rules, like J89, is the iterative process of determining the life-and-death status of groups.
The very first step (which is identical in all the different rules that I know) is to determine whether a group has the status "TWO-EYED life" (which is "unconditionally alive", "uncapturable", etc.).
These "TWO-EYED life" groups enclose and / or seperate areas of the board, which contain groups of still undetermined life-and-death status.
Each of these areas is one "locality", and has to be considered on its own.
NOTHING beyond the border of (a) "TWO-EYED life" group(s) (seen from the inside of any of these areas) affects the determination of life-and-death inside.
The remaining problem is how to handle "ko".
When trying to resolve this problem, you have to be very well aware of potential hidded side-effects.
As we all know, this awareness has been different in the course of history.
+ + + + + + + + + + + + + + + + +
By the way:
I am with you with your understanding of "enable". But this is NOT covered by the legal text.
I hope that we can agree that the primary concerns of the rule were "snap-back" and "nakade".
As I already mentioned before, an addition would have been helpful that the "new stone" had to be established on a board point that had been occupied by the (then captured) group under status consideration earlier.
You will realise that the group under consideration is "unconditionally alive" (i.e. part of a group with two eyes), if ALL previously occupied board points can be occupied again (i.e. a "complete rebirth" of that group).
If only at least one of these pre-occupied points can be occupied again, the group under consideration will be part of a seki.
Last edited by Cassandra on Sat Aug 14, 2021 3:32 am, edited 2 times in total.
The really most difficult Go problem ever: https://igohatsuyoron120.de/index.htm
Igo Hatsuyōron #120 (really solved by KataGo)
Igo Hatsuyōron #120 (really solved by KataGo)
-
Cassandra
- Lives in sente
- Posts: 1326
- Joined: Wed Apr 28, 2010 11:33 am
- Rank: German 1 Kyu
- GD Posts: 0
- Has thanked: 14 times
- Been thanked: 153 times
Re: Japonese counting
See my posting above.Gérard TAILLE wrote:Let's try to progress slowly. What is for you the result of example 11?
The White group under consideration is surrounded by a "TWO-EYED life" (at least this can be assumed for sure) Black group.
Thus, the rest of the board can be considered to be completely empty.
Nothing outside Black's unconditionally alive wall can stop Black from capturing White's stones inside.
Therefore, White's group under consideration is "dead".
The really most difficult Go problem ever: https://igohatsuyoron120.de/index.htm
Igo Hatsuyōron #120 (really solved by KataGo)
Igo Hatsuyōron #120 (really solved by KataGo)
-
Gérard TAILLE
- Gosei
- Posts: 1346
- Joined: Sun Aug 23, 2020 2:47 am
- Rank: 1d
- GD Posts: 0
- Has thanked: 21 times
- Been thanked: 57 times
Re: Japonese counting
I agree with you Jann, spatial locality means nothing in J89. J2003 tries to define spatial locality but it is another issue.jann wrote:As I wrote earlier you seem to misunderstand what it means to "enable" a new live stone. If you cannot carry out a capture without enabling something (that wasn't originally possible) elsewhere, that is not a valid capture in J89. Spatial locality means nothing - that's why pass-for-ko (with global enabling) was invented for.Cassandra wrote:It is completely irrelevant for the status of
The problem with double kos and ko passing is one of the oldest well known defects of the text. But because it would break almost anything it is safe to assume that such perpetual ko passing loop is not allowed - even if the rules authors overlooked the problem.
Other than that, there are several beasts where the ko pass rules change local outcomes and require protecting moves that would not be necessary in normal go. Those are, I think, accepted consequences of J89.
Concerning the famous "enable a new live stone" I have real difficulty to understand the meaning.
Let's take the well known first example given in the rule, to try and get a better understanding What is the status of the marked stones?
In the rule I see and it is said "Black is now able to play 2, a stone which cannot be captured".
Where is my point? Surely black
cannot be capture but in any case (I mean even if white does not try to capture black), black is able to prove she can occupy this point by the sequence:
That means that you are not enabling something that wasn't originally possible are you?If it is true that means that the four marked stones in the first diagram are dead.
By the way the white stone in the corner is also dead isn't it?
The final conclusion is the same (seki or anti-seki) but the intermediate results are completly the opposite.
Note : surely the wording of 7.2 is not clear but it is not my point.
My point is : what do we really want to say?
-
Cassandra
- Lives in sente
- Posts: 1326
- Joined: Wed Apr 28, 2010 11:33 am
- Rank: German 1 Kyu
- GD Posts: 0
- Has thanked: 14 times
- Been thanked: 153 times
Re: Japonese counting
Then Black has to play this sequence BEFORE the game stopped.Gérard TAILLE wrote:Where is my point? Surely black
During status confirmation, it does not have any meaning that Black played the move
at a point where a Black stones was present before, because this move is part of a sequence for confirming the status of a °WHITE" group.The really most difficult Go problem ever: https://igohatsuyoron120.de/index.htm
Igo Hatsuyōron #120 (really solved by KataGo)
Igo Hatsuyōron #120 (really solved by KataGo)
Re: Japonese counting
Designing rules like this may be possible. J89 chose a different way: ko passes (hoping for a similar effect). You can have either explicit localization or passing for kos, but not both.Cassandra wrote:These "TWO-EYED life" groups enclose and / or seperate areas of the board, which contain groups of still undetermined life-and-death status.
Each of these areas is one "locality", and has to be considered on its own.
The very example (ex 4) you quoted earlier shows this is not correct. The new stone can appear elsewhere, where no capture happened from. What matters is logical connection - made possible by the capture.I am with you with your understanding of "enable". But this is NOT covered by the legal text.
I hope that we can agree that the primary concerns of the rule were "snap-back" and "nakade".
As I already mentioned before, an addition would have been helpful that the "new stone" had to be established on a board point that had been occupied by the (then captured) group under status consideration earlier.
-
Gérard TAILLE
- Gosei
- Posts: 1346
- Joined: Sun Aug 23, 2020 2:47 am
- Rank: 1d
- GD Posts: 0
- Has thanked: 21 times
- Been thanked: 57 times
Re: Japonese counting
Robert, it is not quite clear to me what capturable-2 means. Can you help me through the following examples by just give me the status of the white stones? what is the status of white marked stones?RobertJasiek wrote:Common sense is not needed. My J2003 interpretation of J1989 treats it correctly:Cassandra wrote:Gérard TAILLE wrote: what is the status of
The big white string is uncapturable due to the infinite alternating sequence B[kkPP]*, where k is ko capture and P is ko-pass. White triangle is neither uncapturable nor capturable-1. Its local-2 is:
White triangle is not capturable-2. Therefore, it is dead.
and now what is the status of white marked stones?