Odd Cases 🤔 in the Japanese Rules

Either counting territory lasker-maas style, or area chinese-style (with button go), as those give identical results. Life and death at that point would be “EVERY stone still on the board is alive” (like Chinese rules / AGA).

I am not sure that I follow. Since this is so exceptional, I am ok in practice (for non computer play) with relying here on long-cycles the same as it is usually handled in Asian rulesets: it is not specified strictly and formally that exactly when the first repetition, or third, or any specific moment in the long cycle occurs then the game is immediately stopped and force no result, but rather “if the infinite cycle is going on and detected and the players do not want to deviate from it”, so in practice a referee should probably warn the players that they either deviate or no result and then apply no result if they do not want to deviate (or the appropriate result in this case, for example in a normal double-ko seki that they keep disturbing both stop positions have identical score so there would be a winner).

Note that the difference with this kind of “long cycle” and the “normal” no result cycle (triple ko, etc) is that in this one there is no cycle within a single “game phase”, but rather the same position is reached again as a stop-position after more than one game-stop.

It should have identical formal consequences with best play if the rule were “once exactly the same stop-position is reached for the second time after a pass-pass game stop, the game is immediately scored with ALL stop positions that have occurred between both repetitions of the stop-position considered and scored independently”. As one can argue that if a player wanted to deviate from that possible “cycle involving game stops”, they should have done so before.

1 Like

Well thank you! for writing that article, because I read it with great interest. I’m still kind of interested in improving on the J1989 rules (what is even the point of playing a territory-scoring game if you have to fill all the dame or else officially be judged in seki and not solid territory? Also, neither OGS nor any other online server that I know of actually enforces this aspect of the J1989 rules), but I got bogged down considering Spight’s and Pauli’s and Jasiek’s numerous detailed proposals. I liked Pauli’s proposal to rule long-cycle situations as seki, but he has so many other unrelated weird innovations in his ruleset that I couldn’t separate them out clearly enough in my mind to get to an actual opinion about it.

I have yet to play a game that got even close to involving a superko situation, but I still think that either winning or losing a game by a superko rule would be extremely unsatisfying in either case.

ETA: also, I am only just beginning to try to understand Ing’s proposals for handling long cycles. At some point I found this effort to explain them in clearer terms, but I haven’t really digested it yet:


1 Like


My point is, avoiding zugzwang is only slightly easier for your suggestion than for mine.

(Mine could also use, ​ ​ ​ “going on and the players do not want to deviate from it” ​ ,
and without that, ​ “they should have done so before” ​ applies to mine too.)

1 Like

Is it not true that under Japanese rules the defending side makes the first move when there is a dispute?

In that case Black would have time to fill up and make two eyes. Though I guess if that were the case, White can claim to be alive.

No, I don’t think so.

In general, if the life and death status depends on who gets to play next, then the position is potentially unsettled. Further, if it is worth enough points to sway the game, then it could become a case of both players lose.

Otherwise, if the unsettled position is not large enough to change the outcome of the game, the game may also end with it scored as an anti-seki, where both sides are considered dead, but those dead stones are not removed and no territory nor captures are counted.

For the original position shared by @antonTobi:

I think this is actually an anti-seki, if both players pass at this position and try to score. White can demonstrate that the black stones are dead (in the hypothetical play with special ko rules, as explained by @antonTobi in the post above), but also Black can demonstrate that the inside White stones are dead. Thus, all of the Black stones and the 11 inside White stones are dead. However, none of these stones are removed from the board, since they don’t reside in anyone’s territory. Thus, no territory nor capture points are scored for the right side of the board, and the only score would come from the 28 points of territory that White has on the left side.

Note that in this position, Black could “safely pass” and not have White capture those stones during the normal alternating play. The Black stones are only at risk during the hypothetical play (with special ko rules) for determining life and death status. Thus, instead of accepting this non-ideal anti-seki outcome, where Black loses by a lot of points, Black could ask to resume the game, which gives White the first move. However, since White cannot accomplish anything under normal alternating play (except to fruitlessly cycle the double-ko), Black can eventually play an extra stone to capture those White stones and settle for a jigo.


I think I’ve come across this a few times and never understood it. Why does black need to take the other ko instead of completing the capture on the left? In hypothetical play aren’t you just trying to demonstrate the status of a particular group?

1 Like

I presume it would reside in the “enable a new stone to be played that the opponent could not capture” condition:

Stones are said to be “alive " if they cannot be captured by the opponent, or if capturing them would enable a new stone to be played that the opponent could not capture.”

Or at least, it may depend on on how you interpret and apply that condition.
Here’s an application of that condition in a different situation:

(these are both the same shape, but just differing in whether it is neutral for black to collapse the seki or strictly bad for black to collapse the seki).

These positions are normally determined to be a seki without white having to throw in at C or D and lose a point. This is despite the fact that black unconditionally can play A and then B to kill the two white stones in the corner! Doing so “enables” white to capture two or three black stones by playing new white stones at C and D that black “cannot capture”. So black is not allowed to determine the white stones as dead even though black can forcibly capture them in hypothetical play.


@hexahedron has already helped to explain about how it could follow from a potential interpretation of the enables clause. Sorry, I should have been more explicit.

However, the Japanese rules commentary implies, through examples discussing moonshine life, that this “enabling rule” should not be interpreted in this fashion.


Thanks for the clarification, I keep forgetting about the “enables” clause. Is the idea maybe that it is just to close a few loopholes and so it should apply only to local situations like the ones @hexahedron posted, not to shapes on the other side of the board?

If the moonshine shape gets more closely intertwined with that ko somehow, does its status become questionable?

1 Like

I don’t think so. Official examples intentionally show a few tricky dependencies, and imo enabling represents the concept that only 100% clean captures of strings with 0% life are granted for free from territory, anything else - exchanges (whether they are profitable or not), tricky things where the capture is only possible at some cost - need to be played out in the game.

The “loss elsewhere” is a global idea (as is hypothetical play itself in J89, using the pass for ko rule) and is an important change from pre-89 Japanese rules (like the Korean rules) where L/D was judged locally. In those rules a local or combined moonshine life is actually alive (see recent Nongshim cup dispute under Korean rules - differing from both Chinese and Japanese).

A plausible explanation of why and how J89 rules - global hypothetical play with global enabling - can work even with a double ko seki present is this. The double ko reinforcement problem (like the Davies’ example) is a different one though that does cost a point in reinforcing.




Regardless, the widespread confusion on this point does not speak well for this ruleset.

1 Like

In the recent weeks, there has been significant discussion about the Japanese across several new threads. I just wanted to make a note of them here.

  1. Is an unfinished Ko worth a point in Japanese scoring? asked about the status of ko left unresolved at the end of the game, in a manner that mirrored historic rules disputes that predated and likely influenced the 1989 version of the Japanese rules.

  2. Interesting double/potential triple ko(/seki ?)! brought up a case of a self-contained moonshine life position occurring in a 9x9 game.

  3. Is white alive under Japanese rules? discussed technical aspects about life and death determination, particularly the concept of the “enables clause”.


Interest thread :slight_smile: Will read through in more details when I get time


I didn’t read everything yet. But my first thought on the first example is: Does the fact, that passing is a legal move in any position mean, it can’t be a mistake? The opener suggests that both players are punished for passing. I have to say, I find that kinda fair. They both failed to settle the position, so you can’t really blame the ruleset to punish them both. Why would you?