Geoff Kaniuk's referee exercises

These scenes and questions were posed, with suggested answers, in BGJ editions 137, -8, and -9, of 2006.

Would you like to place yourself in the shoes of a tournament referee?

  1. The draw has just been posted at round 3 and a youngster is tearfully complaining to his mother that he is only 19k, but has to play White against a pensioner who has now reached 16k. She is upset and calls the referee.

  2. A player from one of the top boards went to get a coffee and a fag with a hand held computer in breast pocket. The referee notices the player going for a walk and studying the game record intently. What would you do in this situation?

[Since the development of strong AI for personal use in 2017, the context of this question has changed massively. You might find it most interesting to consider how you’d deal with the situation both today and in 2006, perhaps substituing hand held computer for smartphone. Let’s not let this one take over the thread, though, eh?]

  1. Two players are in Canadian overtime, playing by the book with lids covering the bowls. Black has left the prisoners in the lid. At some point Black absentmindedly tried to take a stone from the bowl and managed to push the lid through the rim. The white prisoners of course fell in, and the players took the time to fish the stones out of the bowl. The game carried on without incident, but it was very close and after counting was done, White had won by half a point. After the analysis, a bystander helped to clear the stones away and discovered a lone white stone in Black’s bowl. He told Black what he had found; Black became upset and called the referee.

  2. Two players have joined the Lightning tournament at the European Congress, which is being run with the Ing speaking clocks.Towards the end of round 1, two players come charging at the referee complaining something is wrong with the clock. Having calmed them down outside the playing area, you ascertain that one of the players has lost on time. He is claiming that he couldn’t hear the clock, and it timed him out. What do you do?

  3. A game between shodans had entered the late yose stage. Black counted very carefully, writing down the size of territories on a piece of paper. Black got up from the table with the paper, walked over to friends who had finished and showed it to them. After some discussion Black returned to finish the game. What would you have done if White had called the referee?

  4. A close game between a 3k and a 4k has just finished. Both players have passed. The 4k then notices some small yose. They consult the referee.

  5. During Canadian overtime a player has been using stones from the bowl instead of the 20 stones counted out, but playing very quickly. Neither player noticed this, but when the flag falls both players agree that more than 20 stones had been played. They decide to consult the referee.

  6. A player who has about 2 seconds left passes, thinking the game is over, but forgets to press the clock. His opponent thinks there is a valid dame point left and as she is about to play notices that the flag has fallen – so says that the first player has lost, and claims the game. The first player is upset, having passed, and they decide to consult the referee.

  7. A close game was in the 1 point in gote stage. White has four stones arranged in the bent four shape as in the diagram with live outside stones.


Both players have passed, White thinking that “bent-four-in-a-corner is dead”. Black is a self-taught rapidly improving player and she passed thinking “I’ve captured four stones on the edge and so I am alive”. Having passed, they now discover that they have differing views on the corner situation. They call the referee.

  1. A 2k and a shodan sit down to a game with 6 komi, and the Shodan playing Black places six stones in White’s bowl as komi. The game is very long, hard fought and close. They count up. Black: “you have 63 points.” White: “you have 64, so with komi, I win by 5.” White: “I didn’t see that!” The referee is called.

  2. A game between a Chinese national playing White and a British 3d at a UK international tournament is counted giving the result that Black wins by 1 point. White is crestfallen, because she had a group with one (false) eye in seki. She calls the referee.

  3. A game between two 5 kyu has achieved the state of having a triple ko on the board, and both players are grimly hanging onto the ko fight. You are watching the game and note that Black is two points ahead even if White wins the ko. They decide to consult you.

  4. One way of resigning is to clearly place two stones together inside your opponent’s territory. Sometimes one wishes that one’s opponent would resign, because the situation is clearly hopeless. Some think that you can signal in this case by placing a stone at the 1-1 point inside your own territory. Two shodans had got into this state, but both thought they were well ahead. Black placed a stone at the 1-1 point. White took this as a resignation, stood up and shook Black’s hand. They cleared the stones away, and they each separately reported a win by resignation to the Draw Master who was sitting at a table collecting results. The referee was called by the Draw Master.

I’d also like to attach two real situations which have happened to me at tournaments.

[I forgot to add a 14., woops.]

“15.” White (me, 9k) has been giving handicap to a low-ranking DDK (16k). After all moves of the game which could have altered the result had been made, Black began to play nonsensical moves inside White’s territory and continued this for some time, during which period an intermediate third party (11k) appeared at the board and told Black that he should resign, to which Black told the third party to go away. Eventually White times out but does not want to pursue a complaint, since he has very little experience with manual overtime; however, what if bystanders had alerted the referee?

“16.” White, a 3d, has been giving handicap to Black (me, 7k), who is playing his final game of the tournament, having lost all the ones he’s thus far played. Late in the endgame, White exclaims that it looks like Black has won. Black takes this as a resignation and begins to eagerly clear the stones from the board, happy to have finally won a game. However, as Black is doing this White is startled and protests, upset, that the game is still continuing and that he hadn’t conceded. What if White had called the referee?


I guess these are hard to answer without having the actual tournament guidelines by hand. Some of these really depend on them, I think.

  1. Life isn’t always what you want it to be. If there was a draw to set up the tournament, then this is always going to involve a random component, there’s not much one can do about it except for organising tournaments without draws (e.g. pairing by strength directly)
  2. Probably in this day and age a player from one of the top boards should know better than using their phone for looking at a Go game while still in the middle of playing one. If I were to organise a tournament, I’d specifically want a rule against using smart devices in any form during play to prevent this kind of uncertainty.
  3. I’d have to know the rules more specifically. If the rules specify that all prisoners should be kept in the lid throughout the entire game, then Black, although likely accidentally, did break this rule and thus the game should be declared a win for White by disqualification. If no such explicit rule is stated, then I would consider the game as being counted incorrectly. I’m sure there’s a line in the tournament guide about what the proper action should be with incorrectly scored games.
  4. Unless the clock is indeed malfunctioning, I guess that saying “one couldn’t hear the clock” isn’t really relevant: a time-out is a time-out. If the clock is malfunctioning, I’d say the player has not lost on time due to their own fault and would suggest continuing the game with a new timer (unless there’s something about clock malfunctioning in the guidelines)
  5. As a referee, one has to assume the worst in suspicious circumstances. It’s the showing of the paper that excludes the possibility that the conversation was about something other than the game. Discussing an ongoing game in the vicinity of the players is a no-go.
  6. This depends on the ruleset. Under Japanese rules for example, this would be a void game. Under Chinese rules, the game should resume. Etc.
  7. Again, I’m not sure how the guidelines are stated. See my answer to 3.
  8. Passing is a move, and requires pressing the timer. The player has lost by timeout.
  9. This depends on the rules again. Under most rules, the dispute would be played out. As far as I’m aware “bent-foun-in-a-corner is dead” rings true not because it is a rule, but because that’s the status of the group after actually playing it out. It also is not true if there are unremovable ko threats on the board.
  10. Again, what is the proper way to deal with komi? If the rules declare that komi is added at the end, then what the hell was Black doing? If the rules declare that komi is counted by giving the opponent prisoners, then we have to assume that these prisoners were indeed in white’s bowl. Moreover, depending on the game it may be possible to reconstruct the number of prisoners, unless there was a large ko fight. Aren’t players supposed to keep a record of the game anyways?
  11. I assume the rules of the game state no points in seki? In any case, the Chinese player is supposed to have been aware of the rules they were playing with, thus it’s their own responsibility for any mistakes resulting from not being aware of the rules.
  12. Again, what is the ruleset? In Japanese, the game would be void. In Chinese one of the players must have made a mistake by violating superko first. That Black is ahead when giving up the ko is up to Black to discover. I’d interpret calling the referee as both players refusing to back down from the ko fight, which is a strategical error for Black, but I’m not refereeing to protect people against making strategical errors.
  13. If there’s no record, I’d declare the game no result. Otherwise I’d ask the player to score the game properly based on the record.
  14. is missing?
  15. I’m not sure why White timed out here. In a tournament setting I would not want to interfere with a game, even if an opponent is playing pointless moves. This is not for a referee to decide. A winning player should be able to defend adequately until Black runs out of legal moves.
  16. “Looks like you have won” is not a proper way of resigning, Black shouldn’t have cleared the board.

Geoff’s answers make it clear that these exercises are under Japanese rules. I suppose the BGA still used Japanese rules at the time.

  1. This is not a bent-4 situation (in the sense of the bent-4 is dead as stated in the 1949 version of the Japanese rules). It’s just an unsettled position under any rule set.

Black can just live unconditionally by capturing:

And if instead white starts by atari at E2, black would capture and it would turn into a ko.

So both players were wrong with passing in this position. I think in that case, this corner would be scored as a seki under official Japanese rules (the position is final after passing)

Geoff suggests that resuming the game is the proper referee decision in this case (perhaps taking into account the level of the players). I think that this deviation from the official Japanese rules is common in European amateur tournaments.
Resuming still leaves the problem of who plays first, which in this case would make the difference of black living unconditionally or living in ko, so it could decide the game.


Contrary to chess, go does not have this tradition. Official pro games have recorders, so the players don’t record it themselves. As far as I know, amateurs are not required to record their games.


My 2 cents:

  1. I think AGA rulesets and any other alike to ask players to “save” the prisoners are stupid and only work in elite games. Back in the days, we played in college, the games were like many people around and people talk and more than two play etc. It’s chaos and many times we ran out of stones and took the prisoners to continue to play. Then we were using Chinese rulesets. Japanese ruleset is elegant mathematically, but is just not good for general public.

  2. I learned from a youtuber and thought it brilliant. He, as RD, asks every player taking picture of the game before counting. Fox actually has an app to count and is very accurate(I never used it myself).

  3. On AI cheating, I just give up.

My nephew, who was a competitive chess players, told me this story many years ago. He was much stronger than his opponent in this game and put him in trouble very quickly. His opponent asks as a courtesy if he can use the restroom. When he came back, he played the best possible move to get him out of the trouble. Then the same thing occurred again. The two moves to get him out of trouble were beyond his strength in my nephew’s view. So the third time his opponent was in trouble and asked for restroom break, my newphew said: sure, I actually need to go too. They came back, soon, his opponent resigned.

I thought this was a solution. Only now I realized this only works if one is much stronger than the cheater. I can’t walk away from the table on my own time. Don’t think it is cool to disallow opponent to use the restroom. What if I do?

  1. On the resignation, I think AGA rules state we have to say: I resign. I always do the two stones thing, thought it is cool. Looks like the rule is scrutinized by lawyers to eliminate misunderstandings. :joy:

This would not be an issue in Texas. The mom will just slap the boy’s face: don’t be a baby. :joy:

Years ago, I was watch 10 years old girl soccer game, the best team in dallas. One girl made a misjudgment. the coach yelled at her: Megan, this is exactly what I told you not to do.

1 Like
  1. In the past, I have encountered players who paid komi with prisoners at the start of the game. They were older players who had been playing since the 70s or early 80s, so I’m guessing there was a period when this was somewhat common (perhaps this habit evolved when most players in the West were autodidacts and local go communities were small and isolated).
    I think in most cases it would be easy enough to count stones to verify their claim, but they should be told to not do it again.
  1. As white timed out responding to nonsensical moves, I assume this game had absolute time settings. I think it depends on the country if they have special rules about trying to flag the opponent by playing nonsensical moves after all neutral points are played.
    But whatever the case, white should make a complaint themselves. Bystanders are just bystanders and they should stay out of it.

  2. If white had complained, I think a referee would have granted them a win, because he did not resign and black destroyed the position. Or perhaps the game could be reconstructed (a regular 3d should have little problems doing that) to resume the game.
    But perhaps white meant what he said and black was in fact winning, so he decided to let it pass.

1 Like

I’d grade this answer a D. This is testing your ability as a referee to empathise and explain, so that participants understand the rules and procedures and feel they have been treated fairly, and you’ve basically said “tough titties kiddo, go cry to momma!”. Maybe you aren’t aware but in Britain (*1) most tournaments are McMahon, so the “round 3” was an important clue. What has presumably happened is the 19k kid won his first 2 games, whilst the oldie 16k lost theirs, and the kid has now been drawn up to get a complete pairing. So explain this to the kid and mum, with some “well done for winning your first 2 games, it’s because you are doing so well you are now playing stronger players”. As to being white, confirm with them if they feel this is unfair as usually the stronger player takes white in club games. If so, explain that as this is a tournament the colours are random (*2), but reassure them that as white they get komi to balance going second.

*1 and Europe, but I think bracketed swiss is more common in USA
*2 actually not totally as pairing programs can try to balance your total number of black/white across an event but nvm

P.S. and yes Britain used Japanese rules for tournaments back in 2006, we switched to BGA rules which are a minor variant of AGA a few years later.


In the past, I have encountered players who paid komi with prisoners at the start of the game. They were older players who had been playing since the 70s or early 80s, so I’m guessing there was a period when this was somewhat common

There seems to be a reference to this practice in the comic piece The Masters of Go from BGJ #79, which is set in 2025 and written in 1990. The whole joke, however, is that the characters are both old and incompetent.

Komi was the customary 11 points, a figure that had been reached on a steady drift starting with 5 in the 1960’s and 70’s and 6 in the 80’s and 90’s.

Tobin was to play Black and his old eyes steadied as he focussed on the empty board. The time keeper started his clock. Several minutes passed.

“Are you going to move?” asked Timmins.

“Shut up! I’m counting!” replied Tobin cantankerously. … Then, after a pause and another quick scan across the board, “I make it equal so far.”

Timmins stirred his komi noisily with his bony middle finger.

“I’m eleven ahead.”

Tobin slowly raised his eyes to glare at his adversary, his head still bent over the board.

“I have sente,” he said.

In 15. the time control was, iirc, progressive overtime – 10 stones in 5m, then 20 stones in 5m, then 30 stones in 5m etc.

I see progressive overtime as being a sort of pseudo-absolute time since it’s not infinitely playable.

1 Like

I think this is in line with the Japanese rules as long as the resumption is hypothetical, which implies that no outside ko threats can be played, and that the end board is scored on the pre-hypothetical position, but with the life/death status determined by the hypothetical playout. Hence, under Japanese rules, the Black group counts as alive and the white enclosing group as dead (actually, I’m not sure, the board is being cut off before it’s visible if the outside white stones are alive), and thus should be scored as such.

The hypothetical-ness being necessary to prevent zugzwang, where capturing a dead group costs you more points than it is going to return.

1 Like

I should prelude my answer with the fact that I wasn’t aware. :slight_smile:
However, I don’t think this changes my hypothetical verdict.

Actually, now that I’m aware that the draw is done in McMahon style, I believe that the question is in fact misleading: McMahon does not need to have a random component, and the reason that the 19k - 16k happens, is not due to randomness in the draw, but due to how the McMahon system works. (Incidentally, this makes it much easier to explain as well, since as a referee I don’t have to appeal to randomness, but to the logic behind the tournament system)

I would disagree. I’m not saying “tough titties kiddo” to the kid, I’m saying that there’s a procedure to set up games in the tournament, which involves drawing (i.e. assigning opponents with others with a certain degree of randomness), and that such a procedure inherently has to result in matches that aren’t fair when comparing it straight with ranking. If the result of the draw is straight from the 19k performing well and the 16k performing badly, then of course I’d explain that, but this information is not clear from the stated situation.

I’d be on the kid’s side that it is indeed unfair, but I would put the blame on the procedure, not on the opponent, or the draw itself. If this is how things are managed, then a referee is just there to state “let’s look at this… Yes, this seems to have been managed according to procedure.” A referee is not there to decide the procedure of the tournament, but to verify that things are going according to procedure.

In other words, I believe the real complaint should be towards the tournament director, not towards the referee.

I’m not sure, but as far as I know a referee’s job is not to empathise. That’s not to say that a referee should act unfriendly, but it is to say that the job of a referee is to make the moral choice that is closest to the guidelines as given for the tournament. So far, in any sports I’ve witnessed, as well as in law courts (where the jury / judge are referees), acting formal and distant seems to be the norm for referees, instead of acting empathic.

As for how to bring the message to the mum / kid, I’d do that more diplomatically than what I wrote above.

I still believe that the lesson that life isn’t all peaches should be included there, especially for the mum.


Japanese rules are a bit awkward when dealing with unsettled positions. However, I don’t think it should be considered a seki.

If the question of who resumes would decide the game, then it would actually be a case of both players lose.

The surrounding text says to assume that the outside white stones are alive.

1 Like

I guess, based on the answers (which I’ve now found), the exercise are more tuned towards how the referee should interact with the players, instead of which decision the referee should make. I’d have to look over my answers again to incorporate that, since it wasn’t clear to me that the communication part was where the real exercise lay.

1 Like

Phew, this thread is going to be chaotic. I can’t even keep track of it all. The OP is essentially 15 separate prompts for discussion packed into one.

1 Like

I commented in the excerpts thread that it was this exact argument over the status of that bent four in the article that, it seems, provoked the BGA to very quickly shift to the AGA rules.

It’s interesting to see that the question’s retained its heat.

1 Like

Judging from those exercises, none truly technical. Seems like British go players are very much like England football fans, trouble makers. :joy:

I don’t see how switching to AGA rules helps. The game is not over, because the status of black’s group is unsettled.


If the Aga rules were used, then the players wouldn’t have thought that the game was finished.

1 Like

Why? Does the status of groups depend on the scoring method?

1 Like