How do people keep track of ko and superko in tournaments?

I would like to revisit a question that came up in this thread:

I understand pro players have practiced memorization and they keep game logs and stuff, but since ko/superko refers to not repeating the board that occurred at any point of the game, how do they keep track?
What if they find out later? Is there any record on this?

It’s rare and not very probable, as I understand it, but still, the question is killing me. :face_with_raised_eyebrow:

1 Like

As I mentioned in that thread, you don’t have to memorize every past position, but rather just recognize whether a repeating pattern is being set up. Once players play something elsewhere (i.e., a ko threat), that essentially resets the position that you need to keep track of.

It is extremely unlikely that anything more than a six move cycle (triple ko) would occur. The shorter cycles are limited to just a handful of patterns:


I have played close to 1000 games of go so far. Superko hasn’t come up so far. :slight_smile:

Normal ko is easy to spot since you always see the last move your opponent played.


When you combine the fact that it will likely occured in particular sequences or within a hand full of moves, and the fact that even strong but non competitive players can replay the entire game off the top of their heads…

I guess the answer is, how do we recognize anything in this game?
Git gud noob.

Honestly, if you’re impressed that high level players can recognize that all recently place stones have been captured and we are back to a previous state… You should watch some pro games. Those dudes reconize much much more complicated stuff then a pattern of 6 stones :stuck_out_tongue:

Having said that, I am sure at least one tournament has missed it.


There is a story, though I’m not sure I remember it correctly. Cho Chikun was playing a big title match game in the 1990s and they were playing a ko. Cho couldn’t remember if it was his turn to take the ko, so he asked the timekeeper or record-taker, who told him that it was his turn to take. Cho took the ko, and it was realised that actually it had been the turn of Cho’s opponent to take, but Cho couldn’t be penalised since he had taken official advice. So the entire game was annulled.


@yebellz, @richyfourtytwo @VicktorVauhn It’s not that I don’t appreciate the input, I do. But my question is not about how unlikely it is or how possible is that pro players are way past that menial problem.
I see all types of “what if/ how about” highly mathematical/ programming questions in forums, so is it that mine is not fancy enough because I’m not an IT person? I feel my question is brushed off.

@bugcat Thank you! Your reply led me to this thread My conclusion is that referee is tasked with keeping up and if they fail this can be cause for annulment of a game.

Sorry if it came across like that, that was not my intention. I’d still say that ‘not at all’ is an honest answer to (the superko part of) your original question as far as lower level players like myself are concerned. For higher level players it is probably not a huge problem for reasons that others have pointed out already.

On the one real live tournament I played in so far we were not even told which ruleset we were supposed to use. Probably that meant Japanese, as that is the most usual here. I would be surprised though if even half the participants knew what the exact superko rules in Japanese rules actually are.

This is a somewhat strange situation. We go players tend to be so proud about our game having such simple rules (when compared to chess for example), but it turns out only few of us are fully aware of all rules! :slight_smile:


What answer WOULD satisfy you here?

What relevance does a highly mathematical discussion have here? You asked a question and received an answer.

You didn’t receive a non technical answer “because your not an IT person” (none of us know if you are an IT person, how would we shape our answer around that?)
You received a non technical answer because you asked a non technical question.

It is a pattern of moves, you asked how the pattern of moves is recognized… the answer is experience.

You say
" But my question is not about how unlikely it is or how possible is that pro players are way past that menial problem."
Yes it is. Or it is at least relevant.

The tournament judges are not going to be pros likely, but they are going to be high level experienced players… So examples of the heights of what is capable with experience and study should certainly be valid for discussion of how practically a lesser task can be handled.

You are honestly brushing off 3 peoples explanations.

1 Like

Please read more carefully. You are so fast to dismiss my question you don’t even understand my question. I did receive an answer from a person who bothered to reply properly, you can compare the answers if it helps. This is all I’m going to answer to this.

I won’t even bother with you trying to tell me what I mean.

3 people who offered 3 variations of “it’s not gonna happen 'cause pros, so don’t ask”.

I don’t think that’s a fair characterization of my reply, which has nothing to do with pros. I offered a resource which I think helps explain how to recognize cycles. Beyond that sort of pattern matching, I don’t think there are any magic tricks for memorizing past positions, except that it will come more naturally as one gains more experience with the game.

I don’t see how it’s productive to further engage in this discussion if you are going to dismiss responses that appear to be offered in good faith.


Maybe my reply concerning yours was a bit strong, sorry.

I think the misunderstanding here is that I’m not asking “how it is done?” in order to replicate it; I’m asking “in case it occurs, how is it supposed to be treated?”. Think of more of a wiki and less of a curriculum answer, that’s what I was looking for.

Usually there is no need for the players to remember. In case a triple ko arises (which is the minimal necessary for the need of a superko rule) that is important enough for both players to not back down, the game ends there and then with a draw.

There’s this page, where several examples of triple ko are given, and this one with quadruple ko.

In any case, I think even an average SDK will probably be able to recognise a repeated board position (or at least, I believe I would recognise it).


I think the misunderstanding here is that I’m not asking “how it is done?” in order to replicate it; I’m asking “in case it occurs, how is it supposed to be treated?”

Well… You did name your thread “How do people keep track”, and proceeded to open with

“I understand pro players have practiced memorization and they keep game logs and stuff, but since ko/superko refers to not repeating the board that occurred at any point of the game, how do they keep track?”

So I think to say we are not reading carefully is a bit unfair.
You need to ask the question you intend to ask, and cut it with this crap
“I won’t even bother with you trying to tell me what I mean.”
Cause we are not mind readers, we can only read your own thread title.

1 Like

Bottom line is the TD will decide. And how will the TD know? Well usually he’ll just know. But if not he could consult the game record since most big games are recorded live. If not… well he would probably just do his best to decide somehow.

BTW pro games they do not make super-ko (triple-ko) illegal, but instead declare the game void. This means they don’t need to know the exact point the repetition begins, so it’s much easier. They just see the cycle (usually 6 moves) will not end and declare no result.


TD is the referee?

I found some older papers of Go associations on the effects of using computers(sadly I didn’t keep the links because I didn’t see much traction) and they noted that with computers there could be a shift in ko recording (the rest I didn’t really get), because even for pro players it wasn’t unheard of to miss a ko, being deep in the game. I’ll see if I can find it again.

And a couple of actual cases I managed to find, no matter if the player or the referee missed it, they declared the game void, like you said.

Generally I found it is the referee’s duty to keep track, but nobody’s perfect.

TD is the Tournament Director, so yes, the referee.

1 Like

In a regular ko, the board position will not repeat because there has to be an intervening ko threat. It is easy to keep track of because it happens immediately, although even in pro games, sometimes people have what is called “chakgak” in Korean, an “illusion”, and will make a mistake and take back the ko without a ko threat. In this case, the game is lost, regardless of whether one is winning, or it’s at the end. This happened once I saw in top pro competition. In a friendly game, I think it’s okay to just say, wait, I just took it, you need to make a ko threat first.

In “unusual” ko positions, imagine something like there are three related kos keeping groups alive. Again, it’s going to be clearer than it sounds in actual play, because the ko threat of taking a ko is going to be pretty clear and happening immediately, that is what is keeping the game in a repeating loop. It is extremely rare. For example, both sides are going to have to keep the kos going, one side can’t just connect a ko and still be on the road to capturing the group. If neither side wants to connect, you can just say it’s dual life, and there are no points there, keeping in mind that we aren’t suspending capturing and basic ko rules, so it could be considered a source of endless ko threats, for example. This all sounds extremely complicated, but you will know, if you know the last move played and the basic rules, what’s legal. Whether a legal move is a good idea, is something you’ll have to figure out, and everyone has trouble with that.

When I was just starting, I used to worry if these complicated rules would come up just when there were a few kos around the board, for example, in the end, when there’s a lot of “half-point” kos that you are filling. Imagine that you have three half-point kos, for example. You take a, the other takes b, you take c, the other takes a, you take b, the other takes c, you take a, etc. etc. Then I realized in order for this to happen, one side has to be able to capture two of the kos from the starting round and the second round, which is impossible. Again, sounds complicated, but in real life it’s immediately visible that you can’t capture a ko that you’ve already captured, so the board position cannot repeat.


I’m not sure which ko rule is used in China, but superko is not used in Japan and Korea. Perhaps superko only used in a few Western countries (New Zealand, AGA?).

I think most players use the basic ko rule (ban to recapture one stone that just captured one stone). In practise, the basic ko rule is good enough in 99.99% of the cases. I have played thousands of games and I cannot remember playing a game where it wasn’t.


All Area Scoring rules I’m aware of are using any superko rule as default.

Japanese rules are using positional superko as well

Article 12. No result
When the same whole-board position is repeated during a game, if the players agree, the game ends without result.

Japanese rules fail to define what happens if one of the players doesn’t agree. I’d bet in this case the TD would violate their own rules and call any result they like.

That’s the reason I like superko rules. They catch the remaining 0.01% without introducing an additional rule to catch rare cases.
Superko = Simple ko + longer cycles.

I can remember 2 cases, where the call moderator function was used, because “game will not end (tripple ko), we need help”. One of which got reported by ~8 users.


I understand it’s a lot to ask, but could you link the game? I’d be interested to see it played out. Probably it’s too much work to find it, but I thought I’d ask.