In this game my opponent and I managed to create an eternal or 10.000 year ko. As you can see, we can go endlessly catching a stone. Normally this means an undecided result.
Here is the final situation, where white resigned.
If a mod is free, it is possible for us to “end game as tie” for situations like this, but frankly it’s probably easier to just resign and ask us to annul the game.
I thought OGS used situational superko for Japanese rules… maybe that changed or I’m misremembering? Or there is a bug in detection? Because it looks to me like move 52 should not be allowed after move 46, for example…
@anoek, what do you think? (I can’t investigate myself right now…)
(Or maybe I’m not interpreting superko correctly…)
I feel like these settings indicate OGS should give the game a “noresult”, but actually I don’t think the value is used elsewhere in goban. In practice it just means no superko detection is occurring at all.
No, a mannenko (10 000 year ko) does not prevent the game to reach scoring, it has significantly different behavior.
Your position is not exactly an eternal ko either, seems an interesting triple ko variant. But the behavior is similar: you can just play it for a few cycles, and if neither you nor your opponent plans to deviate from the cycle, it should be a draw or no result.
About the “offer draw” button: iirc some Asian servers have this, and only occasionally lead to trolling offers. Another option would be the server automatically concluding games at the Nth repetition (without prisoner balance change).
Also, Japanese and Korean rules do not use superko, and Chinese rules only use it if the prisoner balance changes during the cycle (or if the game is in L/D playout).
No, a 10,000 year ko is a situation it is disadvantageous for either player to turn the situation into a direct ko (instead of letting it end as a seki). It’s not eternal. The players are just hesisant to start it so they may delay fighting the ko to a later stage of the game, hence the name 10,000 year ko.
An eternal ko situation is similar to eternal life. It’s a long cycle where both players keep sacrificing 2 stones again and again. When both players persist in the cycle and no superko rule forbids it, the game ends in no result.
In your game it’s a triple ko. It’s a cycle of single stone captures. When both players persist in the cycle and no superko rule forbids it, the game ends in no result.
Yes, especially because they were losing before move 35 (on move 31, 33 or 35 black should have just lived with B1 or C1 and white would just be dead).
In my view, that’s funny only because of how unlikely it is that
there will be another should-be-no-result game between them:
That’s otherwise an eminently reasonable request.
My feeling is that it’s super unlikely in 19x19, but much much more likely, judging by the frequency we see on OGS on 9x9.
It could be possible that it occurs more often due to some misplays or something, but even still, it’s more likely to be decisive say on 9x9, probably because some group dying is like most of the board
To prevent trolling you could block seeing the tie offer in zen mode or make it a side button that coles up when the offer is pressed.
Could also make it so that both players need to press the draw button with no confirmation beyond that. This way it can’t be used for trolling.
Maybe the draw button just changes color when a draw is offered.
Offering and accepting a draw is not really the correct way to handle endless cycles under the Japanese rules. Technically, it should be considered a “no result”, so if we were to go to the trouble of adding a new feature to support these situations better (than the current work-around of calling a mod to manually end the game with an annulled tie), we should add official support for “no result” outcomes, which should not factor in ratings adjustments (as it is a non-result). We’d also have to make sure that “no result” tournament games are reasonably handled. Officially, it should require replaying the game from the start.
It still could be considered as both win (like hey so nice to make a triple ko) or both lose (like none could win). It’s more a cultural (philosophical?) thing.
I’m not sure it is so obvious that these games must be excluded from ratings. For example, it is common tournament practice to replay no result games, but even that is not strictly required by the rules - afaik J89 itself says nothing about this.
There is also a bit of linguistics involved. Iirc Japanese refer to these games as “mushoubu”, something like “no contest” or the common “no result”. But Korean rules use (which I assume is) the Korean version of this same word for BOTH repetition and score ties, which is translated simply as “draw” in both cases.
So written rules seem to simply assign a name to this outcome, just like they assign a name to a score tie which is a different outcome. The tournament environment may have some freedom in choosing how to treat those various outcomes (such as whether to include them in ratings).
