J2003's global ko pass (unblocking all kos for both players at once) were mentioned recently. I'm not sure if even J2003's explicit localization can make that work. The theoretical defect of such rule is it may allow passing for two remote ko fights in one move, for non-Japanese results:

Click Here To Show Diagram Code

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

I understand the idea Jann but this example does not seem to work.

In confirmation phase all white stones are dead by simply beginning by

Click Here To Show Diagram Code

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

and black wins the capturing race.

Click Here To Show Diagram Code

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

This position above is better but black seems still to be one move ahead:

Click Here To Show Diagram Code

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

Click Here To Show Diagram Code

[go]$$B
$$ +-------------------+
$$ | X X 6 O X O X X X |
$$ | X O O O X O X . . |
$$ | 5 O X X X O X . 4 |
$$ | O O O X . O X X . |
$$ | O O X X . O X X X |
$$ ---------------------
$$[/go]

Click Here To Show Diagram Code

[go]$$Bm7
$$ +-------------------+
$$ | 3 5 O O X O X X X |
$$ | 1 O O O X O X 6 4 |
$$ | 8 O X X X O X 2 O |
$$ | O O O X . O X X 7 |
$$ | O O X X . O X X X |
$$ ---------------------
$$[/go]

Click Here To Show Diagram Code

[go]$$Bm15
$$ +-------------------+
$$ | 1 5 O O X O X X X |
$$ | 3 O O O X O X 6 4 |
$$ | O O X X X O X 7 2 |
$$ | O O O X . O X X X |
$$ | O O X X . O X X X |
$$ ---------------------
$$[/go]

Click Here To Show Diagram Code

[go]$$Bm21
$$ +-------------------+
$$ | X X . . X O X X X |
$$ | X . . . X O X . 2 |
$$ | . . X X X O X X . |
$$ | . . . X 3 O X X X |
$$ | . . X X . O X X X |
$$ ---------------------
$$[/go]

Ok, then what if the race is longer?

Click Here To Show Diagram Code

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

Yes OC Jann it works now.
Good point!

What is the problem? In which sense is there a defect?

In Japanese rules bent4 is dead (OC) due to the pass-for-ko rule, then the seki collapses. In confirmation the only allowed ko threat is a pass, so B can start capturing from left. This is a case of line #2 of my earlier rules classification table (bent4 with nearby unremovable threat).

How do you evaluate the statuses (for middle and right W stones for example) under J2003?

1) Referring only to bent-4 does not do this position justice. Bent-4 is not a nuclear bomb overriding the rest of the board.

2) For at least months, I lack time to apply J2003 to non-trivial positions.

Why not taking simply the well known L&D exemple 7.2. (to avoid an unknown non-trivial position).
If you replace the move pass-for-ko by a move ko-pass the result is completly changed.

I explained how Japanese rules with pass-for-ko work here, including the rest of the board.

I'm not sure how your rules work, but it seems when B starts the bent4 ko, W can respond by starting another ko on the right, then pass for both simultaneously (which is a nonsense from a Japanese viewpoint), thereby keeping some W stones uncapturable with a left-or-right tactic.

Gérard: the official example is different, these rules use local enable, that's why I wrote they may or may not get away with the theoretical defect of this kind of ko pass (apparently not).

It is not how J rules work but how J1949 and WAGC worked for only the precedent positions.

Why not this simplier position:

Click Here To Show Diagram Code

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

J2003 does NOT have GLOBAL "enable".
Thus, both corners are L&D judged seperately, and independent of each other. The result will be the same.

_________________
The really most difficult Go problem ever: http://igohatsuyoron120.de/index.htm
Igo Hatsuyoron #120 (still unresolved by professionals, maybe solved by four amateurs, really solved by KataGo)

Yes J2003 tries to use local enable, which is probably why this flaw was not discovered until now. But local enable <> local capturability. Btw, bent4 + mannenko is slightly different beast anyway: the bent4 player can also flip the thousand year ko before starting the bent4 ko, as threat removal to delay/make harder the start of the losing ko there.


Looks nicer and still seems to work. Shrinked and color-balanced:

Click Here To Show Diagram Code

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

This right side ko shape really deserves its own name, it seems valuable component. The original (used in my first post above) was from Matthew Macfadyen apparently, and that corner shape gives W the additional ability to make temporary seki and recreate the ko later. This simpler one is more direct but nicely similar to bent4, even in behavior almost like an "inverse bent4" - one side can wait and start a losing ko anytime.

Yes Jann I agree. Sorry for having made a bad mixure with J89 and J2003 (too quick answer).

Remember Jann, this right side ko shape was already presented by le_4TC in the post https://lifein19x19.com/viewtopic.php?p=266251#p266251

Click Here To Show Diagram Code

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

What is the result of this game in J2003? It is not quite clear for me if it is a complete seki or if the four black stones in the middle are dead (it seems it is the only group of stones capturable).

The shape itself is well known (the original/complex one even has a name), but the inverse-bent4 seems unnamed so far.

This doesn't seem a finished position, with two indirect kos W should probably play on in normal game.

Yes Jann I know. But my question is here only to understand how is determined territories according to the rule. Here it is not quite clear for me.

These rules seem to bound local enable by the same player's uncapturable or reestablishable strings. So in your example W can freely throw away the edges, thus middle B string looks capturable without enabling.

I think a local enable rule is almost as big a theoretical flaw as a global ko pass. You definitely want to detect ANY kind of problem with a capture, never allow throwing away large parts without consequence (remember your early rules: global play without global enable).