Click Here To Show Diagram Code

[go]$$W
$$ -------------------
$$ | . O X 3 X X O . .
$$ | O O O X 4 X O . .
$$ | X O O O X O O . .
$$ | 1 X O 5 X X O . .
$$ | X 2 X X . X O . .
$$ | X X O X X X O . .
$$ | O O O O O O O . .
$$ | . . . . . . . . .
$$ | . . . . . . . . .[/go]

There is no regular ko ban in this position, only the pass-ko bans. If white can only have a single pass-ko ban at a time then black could take back in one of the kos without first passing. That is the difference.

Oh I see where is the misunderstanding. The position I proposed is not that one above but the following:

Click Here To Show Diagram Code

[go]$$W
$$ -------------------
$$ | . O X . X X O . .
$$ | O O O X . X O . .
$$ | X O O O X O O . .
$$ | . X O O X X O . .
$$ | X X X X . X O . .
$$ | X X O X X X O . .
$$ | O O O O O O O . .
$$ | . . . . . . . . .
$$ | . . . . . . . . .[/go]

Click Here To Show Diagram Code

[go]$$W
$$ -------------------
$$ | . O X 1 X X O . .
$$ | O O O X 2 X O . .
$$ | X O O O X O O . .
$$ | 3 X O O X X O . .
$$ | X X X X . X O . .
$$ | X X O X X X O . .
$$ | O O O O O O O . .
$$ | . . . . . . . . .
$$ | . . . . . . . . .[/go]

That is a different position and white would be able to use the pass-ko and the regular ko ban to demonstrate the black group is dead. The result would then be a seki. This is exactly same as with regular pass-ko rule because you still have two ko bans.


You could create a position like this to get different results.

Click Here To Show Diagram Code

[go]$$W
$$ -------------------
$$ | . O X . X X O . .
$$ | O O O X . X O . .
$$ | O O O X X O O .
$$ | O O X . X X O . .
$$ | O O O X . X O . .
$$ | X O O O X O O . .
$$ | . X O O X X O . .
$$ | X X X X . X O . .
$$ | X X O X X X O . .
$$ | O O O O O O O . .
$$ | . . . . . . . . .
$$ | . . . . . . . . .[/go]

Click Here To Show Diagram Code

[go]$$W
$$ -------------------
$$ | . O X 1 X X O . .
$$ | O O O X 2 X O . .
$$ | O O O X X O O .
$$ | O O X 3 X X O . .
$$ | O O O X 4 X O . .
$$ | X O O O X O O . .
$$ | 5 X O O X X O . .
$$ | X X X X . X O . .
$$ | X X O X X X O . .
$$ | O O O O O O O . .
$$ | . . . . . . . . .
$$ | . . . . . . . . .[/go]

Is it desirable or undesirable to let white claim there are three ko bans now? I'd say it is undesirable, and I find it hard to imagine a position when we want one side to defend SIMULTANIOUSLY in more than two kos. Such things may exist but the examples in j89 don't require it.

I think now your are right Jann. I am not able to find an example in which one player need more than one pass-ko ban. Good point indeed.

Coming back to my proposal with the "no ban in double ko" rule the point is that in some double ko situations even a single pass-ko ban is an issue:

Click Here To Show Diagram Code

[go]$$B
$$ -------------------------------
$$ | O . O . X X . . O . . O X . .
$$ | O O O X O X X O O O O O X . .
$$ | X X X O . O O O X X X X X . .
$$ | . . X O O O X X X . . . . . .
$$ | . . X X X X X . . . . . . . .
$$ | . . . . . . . . . . . . . . .
$$ | . . . . . . . . . . . . . . .[/go]

Click Here To Show Diagram Code

[go]$$B
$$ -------------------------------
$$ | O . O . X X 2 3 O . . O X . .
$$ | O O O X O X X O O O O O X . .
$$ | X X X O 1 O O O X X X X X . .
$$ | . . X O O O X X X . . . . . .
$$ | . . X X X X X . . . . . . . .
$$ | . . . . . . . . . . . . . . .
$$ | . . . . . . . . . . . . . . .[/go]

and white can neither retake the ko at (pass-ko ban) nor the ko at (regular ko).

That was not my quote, I still think the left-or-right examples from the other topic show why one pass-ko ban is not enough - IF Japanese-style results (no ko interaction) are desired.

Yes I agree Jann. In this thread I only consider positions with double ko and in this context and in Japanese-style rule your comment was to me quite interesting.

I don't get what interactions in you mean in that specific position, Jann. I think I treated the position incorrectly before, it is a bit tricky to have a pass-ko and a regular ko ban, now I don't think there is any difference from regular pass-ko in this case because of how black takes the kos first. Because black takes the kos first he will have a regular ko ban in one and a pass-ko ban in the other, allowing black to capture everything.

It is basically the same as your pass-once except that possibility of reusing the pass-ko ban is not discarded AND it solves the issue of allowing use of pass-ko to force multiple moves in a row in at least some positions. Like with any change or interpretation of anything in J89 there will be different results in some positions (that is the intention) but it does appear to work with all of the examples.

Anyway, I don't expect it to work 100%. For one thing J89 still has unclear "enable", a double-ko cycles, and a lofty idea of a "Japanese" understanding of what is the correct result. It is too subjective to really match everyone's idea of what 100% is. I am more interested in understanding the benefits and limitations.

Click Here To Show Diagram Code

[go]$$B
$$ +---------------
$$ | X X . P P X . O O Z Z . |
$$ | X P P P X X O O O Z O O |
$$ | . P X X . X O . O Z Z O |
$$ | P P X X X X O O O Z . . |
$$ --------------------
$$[/go]

I thought that: B starts the bent4 ko and takes it first, W throws in at right, B captures there as well. At this point (if there is only one pass-ko ban) W passes for whichever of the two kos B chooses to have protected. Next B resolves one of the kos and W captures in the other, thus saving central stones and even bent4 will be alive (directly or on enabling). Right?


Hmm, I think J89 is actually not bad (if you accept its approach to ko), better than I thought in the past.

This is my current thought of how the position goes when black tries to capture in confirmation.

Click Here To Show Diagram Code

[go]$$B
$$ +---------------
$$ | . X . O O X . O O X X . |
$$ | 1 O O O X X O O O X O O |
$$ | O O X X . X O . O X X O |
$$ | O O X X X X O O O X 2 3 |
$$ --------------------
$$[/go]

Black protects with pass-ko, the other ko will be protected by the regular ko rule.

White passes for because he can take in neither ko shape, one protected by pass-ko and the other by regular ko rule.

finishes the ko that was protected by the regular ko rule. (There may be an ambiguity about if you are allowed to switch the pass-ko ban around but we don't need it here).

White still can't take in the bent-4 because of the pass-ko ban, black takes there also with .

Click Here To Show Diagram Code

[go]$$B
$$ +---------------
$$ | . X 7 O O X . O O X X 5 |
$$ | X O O O X X O O O X O O |
$$ | O O X X . X O . O X X O |
$$ | O O X X X X O O O X . X |
$$ --------------------
$$[/go]

So it appears that black only needs one pass-ko ban to capture everything in this position.



I agree that J89 is not so bad. It is just that I think that we can't talk about perfection or treat its concepts formally all the time. The semantics of J89 are not always clear enough, but I do like how it presents a simple and elegant approach. I still doubt it really works except in practice (when we are really just happy to finish the game).

I'd also add that the faults in some other rule texts are lot more comical than J89, for one thing we tend to just ignore anything not doing with the position on the board. That is really a faulty approach but understandable because it is the position in the game that interests us.

I don't understand this. As I wrote I think W passes for the protected ko (B 1). What does it mean to pass for a ko that is not protected anyway?

Hm, maybe you mean passing for the ko itself is forbidden by the protection? But then how can such protection be eliminated? You need a way for that to prevent anomalies (or a different kind of double ko abuse).

It should just be "white passes" because white has no move, it is not a pass for a ko in any sense. My point was that black takes unprotected ko first.

The idea is that neither player can automatically defend multiple kos by referring to just the pass-ko rule. If white can fight the second ko then the game is either not over or black is asking too much. In this case the game wouldn't be over because if white had a few ko threats he would start the ko (or?), so that is different from pass-ko in J89. The only extra that black gets in status confirmation is the single pass-ko ban in the bent-4, allowing positions that are regarded as end-of-game positions to be handled correctly (hopefully). Or you could say I need two pass-ko bans, or something else.

If the rule really forbids passing itself for the ko, the question is how can such protection be lifted? Which seems essential...

I am a little confused.

Click Here To Show Diagram Code

[go]$$B
$$ +---------------
$$ | . X . O O X . O O X X . |
$$ | 1 O O O X X O O O X O O |
$$ | O O X X . X O . O X X O |
$$ | O O X X X X O O O X 2 3 |
$$ --------------------
$$[/go]

Black protects with pass-ko, the other ko will be protected by the regular ko rule.
The only interesting move white can make with is pass for protected ko (ko )
and then white will be able to save a great part of her stones.

That means that we need to handled two pass-for-ko in order to reach the expected japanese result (all white stones are dead).

BTW has somebody found a position in which the "no ban in double ko" rule associated to the "traditional" pass-for-ko (for each ko) contradicts the expected japanese result?

In that case black would take in the bent-4 shape and protect the other ko using pass-ko so white would have to pass for that one.

I am failing to explain this What I am trying to say is that you can have one pass-ko ban at a time, but you don't have to declare it ahead of time or maintain it for the next move. Basically, you can ban the recapture in one ko using the pass-ko in any position, it does not have to be declared when the ko shape is declared, the ko shape does not become ineligible for a pass-ko ban by not introducing or removing the ban. This might be more flexible then needed, but for simplicity just pick any available j89 pass-ko bans in every position as the single pass-ko ban.

Sorry, I don't understand. The ban can be lifted by passing but then you adjust the play (to end the ko that was passed for) and use your single pass-ko ban in the other/second ko. If the second ko is then passed for you finish that ko before giving any chances there.

I think I understand the idea now. But this essentially allows preventing ko recapture for TWO turns (which is what you did in the left-right example: one turn with normal ko ban, next turn by insta-shifting the protection there, and only then can the opponent pass for it - otherwise W could save himself in that example).

I don't see how such wild idea could survive even a simple approach ko (J89 #10).

You mean to say there is problem that one may not be pass for a ko when it is protected by the regular ko rule? That seems valid but not really what I meant. What I meant or intended was that you could pass to lift any potential ko-ban, it is not conditioned on it being used to protect that ko at that time. That is you don't get to postpone the pass for a ko like you suggest by not protecting the ko using pass-ko.

I open to the suggestion that passing for an unprotected ko may create some weirdness, but maybe a pass for a ko only really happens when the ko is protected by either rule?

Now I'm not sure again if I understand the idea well, how would this prevent the closed double ko loop for example?