Question About Japanese Scoring

  Difficulty: Beginner   Keywords: Question, Rules

In this example, White has 16 empty points surrounded, which makes for 16 points under the Japanese system. My question is this: In most games I have played, a formation like this would be left alone, yet as I see it, since you lose points if you play within your own area, would it not do Black good to play:

Scoring Issues (Japanese vs. Chinese)  
Scoring Issues II  

Now, White only has 12 surrounded spaces! Yes, Black lost one stone, but White lost 4 points! It seems counter-intuitive to me why this would normally be left alone until the end of the game. (When I have scored games using ViZiGo, the score is exactly the same in both cases... which doesn't make sense to me.)

Bill Spight: Hmmm. My guess is that ViZiGo actually computes the area score and converts to the Japanese score (territory). With area scoring you do not lose points by playing inside your own territory, because the area remains the same. Area scoring is a easier to program, I believe.

By territory scoring White has lost 4 points of territory, but gained one captured stone, for a net loss of 3 points.

Scoring Issues (Japanese vs. Chinese)  

The key point to remember that you can remove dead stones at the end of the game for free! So, if Black plays B1, White should play W2, as it fences Black in and doesn't decrease the territory. You had 16 points of territory. Now we have 15 because of W2, but we get a prisoner, because we get to remove B1 at the end of the game. So we have the same score.

Bill: As HolIgor points out below, White W2 is not necessary.

Scoring Issues (Japanese vs. Chinese)  

If Black plays B3, White can respond at 5 without altering White's score. If Black continues to play stones in White's territory, White can answer them without altering the score. White can even gain points by ignoring black moves. B7, for example doesn't need to be answered. The black group is unconditionally dead. After removing the black stones at the end of the game, White will have 3 fewer points of territory, but will have 4 more prisoners!

Don't waste arrows on a corpse  

White should not respond to B3 with W4. The marked stone is dead until proven otherwise. If Black adds more stones inside White's territory, only then White should respond with stones inside White's own territory.

-- FCS

Don't waste arrows on a corpse  

Let me add several words to the last diagram. White should not reply at all in this situation; the black stone will remain dead anyway and thus White's score would increase by one point. Imagine how many moves Black should make to live in the corner. This is an example. Only after 4 black moves White makes a move that secures a kill. The other moves may be used to take some profit elsewhere.

It is interesting to note also that even if White ignores 8 in the diagram, Black will have life and will take 16 points from White spending 5 moves for it. This makes about 3 points per move which is pretty low in the middle of the game.

-- HolIgor

Bill: Black should live at a, not 8. If Black 8, White a - Black b, White c makes ko, threatening to kill Black's group by removing its second eye.

By living, Black takes away 14 points and makes 3 for a swing of 17 points.

I think the statement asserted by FCS above is key:

  • "The marked stone is dead until proven otherwise."

The question for me is, whose responsibility is it to prove the liveness or otherwise of a group played in enemy territory?

In the example above (Scoring Issues II), one can imagine a situation in which Black by playing B1 asserts that the Black stone can live. However, Black doesn't think it is beneficial to play any more. White asserts that the stone is dead, but correspondingly doesn't think it is beneficial to play any more. Who has the responsibility to prove their position?

For experienced players, perhaps this is a non-issue. For someone just getting started (like me) this is troublesome.

I really like the Japanese idea of scoring prisoners (JapaneseCounting), but I dislike the (to me) poorly defined end-of-game conditions. As a result I'm tending to the Chinese scoring (ChineseCounting).

I should add that once the responsiblity is cleared up then at least the situation is well defined. For instance, if it's Black's responsibility then Black must keep playing until achieving two eyes. White will wait until the appropriate point and jump in to kill (assuming passes are allowed.)

This involves a significant amount of calculation on both players' parts to determine whether it's advantageous to begin playing in enemy territory. For the beginner this is just too hard and one guesses.

On the other hand if it's White's responsibility, Black plays B1 and asserts life and White has to lose a bunch of points to kill it. Then Black plays one more stone and sits back while White loses even more points. (This seems really dumb, I trust this is not the answer.)

-- Dave Finlay

lavalyn: In Japanese scoring, after two passes the players must agree on which groups are dead. If there is disagreement, they settle it in continued fighting, then return the position to the two passes and evaluate knowing for certain whether the disputed stones are alive or not. (The player that wants to save their stones goes first.) This eliminates the really dumb method of needing three stones to kill one.

Dave Finlay: What if they can't agree but neither player thinks it benefits him or her to continue playing? Do the just have to play? (What does that even mean if pass is a valid play?)

lavalyn: Because there is no change in points during the dispute, it is always favorable to continue play. Remember, the final scoring position is that laid out after two passes. There is no cost to continued play.

Oh. Forgot about the concept of ko here. The only valid external ko threat in this phase is passing. It is assumed that all external ko threats have been cleared. Why? Because they should all have been used up arguing over one last capture. If there was a bigger ko in the fight for life, they would have played it out in the main game, not in the dispute phase, and it really would have mattered to play an additional move to eliminate that ko for life.

Lavalyn, thanks for the detailed explanation. It makes sense when it is the responsibility of the player who asserts life to play. If he really believes his stones are alive then it benefits him to play.

However, the player who asserts life may believe his group is finally dead, but he is simply trying to bait the opponent into playing in the opponent's own territory and reducing the final score. It seems to me that the player who asserts life must finally get the stones to a position in which the existence of two eyes is easily shown.

When you shake it all down, it's still a lot of calculation. Tough for beginners. -- Dave Finlay

I think I'm not completely understanding what you are saying here. I thought that after two passes, any disputes about groups have to be settled by the player declaring a group dead to play the first move, each player must play the same number of moves and they play until both pass again.

By playing the same number of moves, the score shouldn't change unless the disputed group is really found to be alive. It also ensures that the disputer does not lose points by playing stones to prove that the disputed group is dead.

Playing an equal number of moves is key here. This way it wouldn't matter if you played Japanese or Chinese counting, I think. If there is no place for the second player to play a forced move (to make the moves equal), he or she can just give one stone as a prisoner, unless the last move wasn't in someone's territory.

If only one player passes and the other player plays on, that player can keep passing, watching the other player play and only interfering when he/she thinks it's necessary.

Anyway, I don't think that the person trying to save his/her stones goes first, because they already declared them alive by passing. So I don't understand lavalyn who says that he should go first. Unless I'm misunderstanding something, but I didn't see someone mentioning the equal number of moves here.

-- El Draco

El Draco, you are right regarding who plays first when trying to agree the life/death status of groups, but there is no requirement for Black and White to make equal numbers of moves.

If after both players have passed White claims a black stone is dead, while Black claims it is alive, then both players would start examining sequences with White playing first to determine the truth of the matter. The question is essentially whether, with best play by both players, White can always capture the disputed black stone.

When this question has been answered to both players satisfaction, both players will agree whether the original black stone is actually alive or dead. Then just as Lavalyn described all the stones are put back to the position after the two passes, and the score is then counted with this newfound wisdom.

Since all the moves played out by Black and White after the two passes are thrown away, they have no effect on the score beyond clarifying for both players the status of the disputed stones. You could say these extra moves are not really played, they are just imagined/discussed by the players - and it makes no difference how many moves of each colour are imagined! :)

When you speak of making equal numbers of moves for Black and White I suspect you are confusing Japanese rules with AGA rules which I believe do have such a restriction. -- Mike Terry

After carefully re-reading once more, finally I get it! Thanks Lavalyn, El Draco and Mike Terry for the help and explanations. Apologies for being so boneheaded. -- Dave Finlay

To Dave: You're welcome :-)

To Mike: I think there are different ways of doing the same thing. I heard about the equal moves way from the go club in my city. You get the same result with equal moves as when you removed them all afterwards. So there is no need to remove them afterwards. But I agree that it is just as valid, however I don't see the equal moves counting as a restriction, more like a convenience ;-).


The idea of Black and White playing the same number of moves to resolve life and death disputes after both players have passed is tempting, but unfortunately it doesn't always work! That is it does not always lead to the correct game score defined by Japanese rules. Here is the sort of position that gives problems:


Here there is a Seki between the surrounded black and white groups, so none of the points a,b,c,d are territory for either player. But following your procedure, after both players pass Black could insist on playing further moves to resolve some dispute. Black's first extra move will be at c above (which has no cost), and now White would be forced to give Black a free point by playing a matching stone elsewhere on the board. This will make the final score count incorrect by one point!

Playing matched Black and White moves to resolve a dispute works on most occasions, but as it does not always work I would discourage using this method...

-- Mike Terry

Is the info on AGA Rules incorrect? It says area scoring so for White to play elsewhere would cost nothing in this example. However, the page also says that pass stones are used but this would make no sense in area scoring, would it? I thought captured (including pass) stones go back in the bowl under area scoring. -- Dave Sigaty

SAS: The pass stones are so that you can count the score as if you were using territory scoring -- but that doesn't affect the fact that it's really area scoring. If the players agree to use normal methods for counting the area score then the pass stones (and other prisoners) are indeed completely irrelevant. The score is the same however you count it, and it's the area score.
El Draco: Mike, if Black played in no one's territory, as in your seki example, I already said earlier that White wouldn't have to play a move. Of course, White could play on a dame by choice, but that was probably not what the dispute was about.

To El Draco - Sorry, I'd missed your earlier line concerning a player not having to respond if the previous move was a dame point. I can see how your process will work fine in practice, between well meaning opponents. And there's no harm in all this as long as the players do not confuse the process with the actual Japanese Rules, and so are willing to revert to a more direct interpretation of the rules if their opponent requests it. (I.e. as described by Lavelyn)

From a strictly logical perspective, however, there is a problem with your extra proviso - whether a point is dame or territory is only defined in Japanese Rules after life and death issues are settled, and it is precisely the purpose of your procedure to do this! :)

For example, suppose my diagram above were being played out in order to settle the disputed life/death of the white stones? It is not altogether clear whether or not White must respond to Black's move c (and exactly why?). Whether c is territory or dame depends on whether Black finally succeeds in the attempt to capture the white stones, and this is not known when c is played.

-- Mike Terry

I thought about this problem too, but I think it's only a problem in theory. For most games, it simplifies the process, especially if it's just a friendly game. You'll probably not be allowed to use it on a tournament, especially if they use Japanese rules.

What I'm still wondering, is where the equal moves counting comes from. AGA Rules do say that the number of moves must be equal, but that is for the whole game, not only the dispute phase.

-- El Draco

I live in Japan and started playing Go a few months ago using the Japanese rules (they're the only ones I know).

In the first diagram, all black stones are treated as dead and are left until the end of the game when they are removed, UNLESS they can create two eyes. All white has to do is to ensure that black cannot create two eyes. Usually black won't play in the white area unless he or she is confident of making two eyes. Otherwise black is just increasing the number of whites prisoners.

In El Draco's example, the stones are not in seki. If white plays at "b", black "a" kills all the stones. White only has one eye and is therefore considered dead.

Erm, it is seki. The same argument can be used about black's stones. :-) /Vlad

You're right. I've just spoken with my colleague who has been teaching me go. I was confused because black is completely surrounding the white stones and mistakenly believed black had effectively already "won". It is seki. I stand (or sit) humbly corrected. My colleague told me that in the diagram all eyes and false eyes inside black's perimeter would be ignored when counting territory. You learn something new every day :-)

When the game ends, the players continue to take turns to fill in any "dame" spaces by using unplayed stones (situations involving seki are left alone and ignored). When all the "dame" spaces are filled, each player removes his dead from his partners territory and gives them to his partner. After which each player rearranges the stones in their partners territory to make them easier to count, places the stones that they captured into their partner's territory and then the total score for each player is compared. The winner is decided by the greater territory and the degree of winning (mostly for personal interest) by the difference in score.

-- Redwall?

Question About Japanese Scoring last edited by MrTenuki on July 9, 2006 - 01:53
RecentChanges · StartingPoints · About
Edit page ·Search · Related · Page info · Latest diff
[Welcome to Sensei's Library!]
Search position
Page history
Latest page diff
Partner sites:
Go Teaching Ladder
Login / Prefs
Sensei's Library