In this triple ko game on OGS, despite one player having an eye and the other not, the game was adjudicated by a moderator to be a tie.
It was, but the ruling i think, is wrong. Even in the comments it states that white is dead, using the same triple Ko rules that were used in your own game.
That game is also from 2014, and had different moderators taking a look at it.
Then again, they both also agreed to call the game a drawā¦ or at the very least to call the game a no result game. and because of that, then the game can be called a no result game. Which wasnāt the case in your game.
Under the Japanese rules column in the site rules it says:
āSuper-Ko
Repetitions are forbidden
Board repetition is allowed, if neither side is willing to break the loop the game is annulledā
Donāt these rules kind of suggest it was a no resultā¦?
While I obviously canāt make a strong case here, as Iām not a moderator or admin or anything.
What I would say is this.
The game you played had an in-depth discussion about the rules of the game. And considering you were using Japanese rules in the game. @mark5000 went to official rule sights to come up with his ruling in the game.
So, what I would say is that the ruling that was given was both fair, and indeed should stand.
I would also add:
Perhaps the Rules in OGS need to be edited and updated. But, having said thatā¦ I also donāt think it a particularly great use of time, scouring OGS in order to find shreds of examples, to support how you feel about the result of the game. I, of course, do not mean to offend here at all. But, the game was adjudicated, and a result was declared.
I donāt even know if the result of a game can be overturned either. But you would have to ask a mod for that knowledge.
Just seems like the site has rules that cover this exact situation is all.
The previous game was probably judged to be a tie because those are the site rules.
That may well be so. but if we are going to go by rules that were first in place then the Nihon ki-in Rules were established April 1989, and as such have covered the problem of the Super/ triple Ko since then.
Sure, there are other rule sets. And they dissagree over this issue.
That being saidā¦ the case here is that there was a different adjudicator, and they used a far more standardised set of rules than the ones that OGS has.
However, there is a real conversation to be had here over the future of OGS and with this problem arising again in the future. Especially, as OGS has now been the home of Pro games. (Even if the Pro games had their own referees and adjudicators.)
But reguardles of all of that. I would still argue that mark5000ās ruling should stand.
Results cannot be changed, only annulled. Even that; only in the case of the game being both ranked and neither player having resolved a batch process since its completion.
1 Like
Good to know my firend. But what is a āBatch Process?ā I dont believe I understand what that means.
In your game there was no loop. You were obliged to pass at move 99. This put the game into scoring phase.
4 Likes
OGS uses Glicko-2 to handle our ratings. It gets increased accuracy by processing games in batches (either 15 games long or 1 month long, whichever comes first). The rating changes you see after each game are merely an estimate, and during this time a game can be annulled. Once a batch has been processed, all games prior to the current batch can no longer be altered.
3 Likes
Wow! thats actually super interesting, and super good to know. I thank you for taking the time to give a great answer.
@Kosh I had totally forgotten about that move at 99 in reference to the game.
2 Likes
But the game continued. I donāt think one is obliged to agree the game is over just because both players pass.
Play continued and as of move 103 it was in an infinite loop.
The site rules clearly say; āBoard repetition is allowed, if neither side is willing to break the loop the game is annulledā
It seems fairly unambiguous.
Whether the site should be actually using its own rules is a separate issue.
Forgive me, but at this point, we are quite litteraly arguing in circles. Nothing more or new is being added to this ādebateā.
And by the community guidelines of the Forums, such a thing should be avoided.
Now, I say that not as a rebuttle to what you said, but as an example of the maluability of the rules on OGS. Yes OGS does have rules. Butā¦ the important point here, is that by the rules of the actual game, as set out in the Nihon ki-in rules, which were the rules chosen by Mark5000 in his ruling, as well as taking information and advice from other Moderators, and engaging in discussion over your game, a result was found.
Does it strictly go agains the rules? Absolutely not, because the Moderators, and āLeadersā in the forums. They have every right I think to descide a result in either direction.
The direction that was taken was one that is supported by official, if also external, Rules of the game of GO.
Really, any further debate on this topic though is little more than meaningless I think. wE have gone over the rules to such a degree, that even an idiot like me is able to debate the Super/ Triple Ko rule.
I would kindly ask you, in this case, to drop this, and respect the ruling. Going on about it isnt going to change anything in a meaningful way anyways.
Similarly you passed on move 103 and you had no choice but to do so. It is not an infinite loop if it has to include passes. The consecutive passes trigger article 9.
2 Likes
For the purposes of clarity.
Article 9 (final)
-
- When one party abandons the start and then the other party also abandons, the game is āstoppedā.
-
- After the game is stopped, the game ends when both sides confirm the death and death of the stone and the ground, and agree. This is called āthe endā.
-
- If one side requests to resume the game after the game is stopped, the other party has the right to arrive first and must respond accordingly.
even more important for this specific case:
Article 7. Life and death
1. 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. Stones which are not alive are said to be " dead ."
2. In the confirmation of life and death after the game stops in Article 9, recapturing in the same ko is prohibited. A player whose stone has been captured in a ko may, however, capture in that ko again after passing once for that particular ko capture.
This, as far as I can tell, makes all the moves up to 103 completely legal. however, this appears to mean that if w continues with 104-107 wJ6, bF9, wC9, b cannot take back either at C8 (normal ko rule), or at J6 (because, since the game stopped at move 100 as per article 9, now article 7.2 is active), b must pass (or give up his group of the corner, but that only delays the inevitable) at move 108, leaving w capable of capturing at E9 at move 109.
therefore the upper right black group is dead.
Edit: continuation
actually in this game, it never entered the stopped phase as per Article 9, as neither player has passed, therefore rule 7.2 is not active, and neither side appeared to be willing to break the cycle.
however, if b were willing, they could pass at move 69, and then 70-72 wJ3, bG2, and then w must pass move 73 as any other move is self-atari. b can then pass move 74 to enter the stopped phase, then if the game resumes under w, they may start with moves 74-75 wG3, bJ2, now w must pass 76 or commit self-atari, where now b may continue 77-79 bG1, wJ2, bD3, and now w cannot capture either D3 (normal ko rule) or G3 (article 7.2 is active), meaning they must pass move 80, leaving b to capture at move 81 with J2.
So it would be dead by proper ruling if that path were taken.
Of course, that would be slightly early, so it would be better for b to finish endgame first.
3 Likes
7.2 does sounds like itās worded a bit ambiguously. āRecapturing in the same ko is prohibitedā just sounds like itās enforcing the normal ko rule even though the game is in a confirmation phase. Like it sounds as though if you had no ko threats then you could pass and recapture assuming for some reason that happens.
I think to really tie this up nicely Iād like to see another example explained using these rules, whereby a group is alive in double ko. I believe this situation comes up in the large avalanche joseki for instance. So it would be reasonable to say it can be fairly common. In fact the one Iām thinking of is mentioned here https://senseis.xmp.net/?DoubleKo
If you wanted you could for instance combine this with say a ten thousand year ko on the board somewhere else to get a triple ko, of which thereās a standard example on Senseiās library here
https://senseis.xmp.net/?MannenKo
Iām picking this example because for instance in this triple ko I would assume since if one looks at the two positions separately, one person is alive in double ko and the ten thousand year ko I supposed to resolve to a seki, that thatās how they should remain if we follow the above articles but maybe itās the case if you fight the kos it plays out differently? Especially with the funny no recapturing sounding rule.
1 Like
ok, letās simplify this and call the two kos in double ko A and B (b having A and w having B), and the Thousand Year Ko, T.
Also, letās say itās the first example on the senseiās page Ten Thousand Year Ko, while also the first double ko on the senseiās page, with neither being particularly bigger than the other, and itās blackās turn. the most recent turn was the last other endgame (gote dame even), there are no ko threats (other than localized to the two L&D situations), and nobody has passed yet.
My first pattern to search is ācan b attempt to kill the thousand year ko, assuming thereās nothing left on the board?ā (assuming heās capable, of course)
so he starts at T, where w then takes A, and if that is big enough b takes B, and w takes T (and, of course, b takes the white group thatās dead in double ko). if A was not big enough, b makes the ko direct, and then w will take T or capture at B (whichever is bigger). If w takes T, b takes A, but w gains nothing from taking B in response (as he is dead in double ko anyway), so w just wins the Thousand Year Ko (unless w believes it is smaller and lets w win the double ko).
so b cannot win the Thousand Year Ko.
Then, itās ācan w, with infinite ko threats, win the Thousand Year Ko?ā
so they start by making the ko direct, and then b takes T, and then w plays A as a ko threat. If the ko threat is not big enough, b will fill to kill and w will kill at B; otherwise b will take B and now w plays T, and b has no ko threats (as A makes no sense as a ko threat), and w will capture the Thousand Year Ko.
All this having concluded, b can attempt to take A, but w will win the Thousand Year Ko (as per his goal)
Then we might wanna get fancy and think, ādoes w have a chance at triple ko to make a thousand year ko?ā
to try this, w must start at A, and then b takes B. Now w cannot take B so he turns and makes the ko direct, followed by b taking T. So now w plays a ko threat at B, and if it is big enough, B will respond and w will take the ko at T, letting them win the Thousand Year Ko, but not the double ko. If the threat was not big enough, b will make a direct ko, and w can either play B or T. If w plays B, b will either play A or capture the Thousand Year Ko (whichever is bigger). If he plays A then w wins the thousand year ko. if he captures the Thousand Year Ko, w will capture and win double ko. If w instead elects for T, b will have no ko threats and w will win the thousand year ko.
So w can win the double ko, but only if doing so is smaller (either in reality or blackās perception) than dying in Thousand Year Ko.
Now can b still maintain a seki for Thousand Year Ko as the normal result?
If so b must be the one to fill, as otherwise w can use his infinite ko threats to win the ko.
So b starts at T, where w makes a ko threat at A, where b will either fill for seki at T or take at B to prevent recapture (whichever is bigger). If he takes at B, w can take T, b takes A, w makes a direct ko, b takes T, and w doesnāt have a ko threat (but he takes B anyway for infinite threats so b can never kill in the Thousand Year Ko), so b will be able to fill T and force a seki while still remaining the winner of the Double Ko (he cannot win the Thousand Year Ko for the same reasoning listed above)
In conclusion, none of these situations seem to run into Article 7.2 as nobody is inclined to pass under normal circumstances.
There are situations where w can reverse the outcome of the double ko, but it depends on b evaluating the capture of bās group being bigger than the relevant move in the Thousand Year Ko, but no triple ko (or passing in triple ko) should result if that evaluation is consistent.
2 Likes
Lord spoon, we are not arguing in circles, you just donāt seem to understand what I am saying.
I do not doubt your interpretation of the Nihon ki-in rules or whatever you are talking about, but that is basically irrelevant.
OGS has a document which explains the rules we play by here in some detail and those are the rules of the site. If moderators are free to just choose which rule set they use to make decisions then what is the point of having an official OGS set of rulesā¦? The site document clearly states that āBoard repetition is allowedā, this means exactly what it says, that board repetition in allowed. So it WAS an infinite loop, no-one was obliged to pass for āarticle 7ā reasons as we were playing under Japanese rules as laid out clearly in the site document.
The decision was wrong, ACCORDING TO SITE RULES. I do not need you to agree with me, Iām just stating the facts.
As for your request that I ādrop itā, I politely urge you to simply remove yourself from the discussion if it is troubling you. Iām grateful for your input but I donāt think your opinions are adding much to the debate.
On this site, repetitions are forbidden under AGA, Chinese, and New Zealand rules, NOT under Japanese rules. This is clear, it WAS an infinite loop under OGS rules and was therefore a no-result, as neither party was willing to ābreak the loopā.
Also, article 7.2 is never active, as under the site Japanese rule set, āboard repetition is allowedā.
and: Under article 9; āIf a player requests resumption of a stopped game, his opponent must oblige and has the right to play firstā
If we are (as I believe we are) playing under the OGS Japanese rules, this game was a clear and unambiguous no-result.
I can only assume that Mark was unaware of the existence of the OGS rule set, otherwise he would have judged the game as a no result I expect.
p.sā¦ Thank you everyone who has contributed this discussion, Iām grateful that anyone has taken the time to read all this!
1 Like