Hello all,

just encountered the following situation against the computer. Which makes me wonder what part of the rules
I have misunderstood

1. The computer ruled my group, marked as dead - and I do not understand why - the white groups marked and could have been dead as well?
As far as I can tell neither black or white can kill each other?

2. A minor point that has happened often in other games: if the white and black groups would indeed stay on the board, since nobody can kill each other, will A17 (marked X) give a point to white?
My understanding is yes (since it is surrounded by white stones, hence white's territory). But GoDroid always rules such points as dame.

Click Here To Show Diagram Code

[go]$$Bc
$$ +---------------------------------------+
$$ | . Z Z Z O . O . O X X O . O . X . . . |
$$ | @ @ Z O O O . O O O X O O O O X X . . |
$$ | x @ Z Z O O O O . O X X X O X X . X X |
$$ | W Z Z Z Z O X O O X X X X X X X X X O |
$$ | W W W . Z O X X O X X O X X O X X O O |
$$ | W X W W Z O X O X X X O O O O O O O . |
$$ | X X X X O O X O O X O O . . . . . . . |
$$ | . . X X O X X O O O X . . . . . . . . |
$$ | X . X O O X O O . X . . X O . . . . . |
$$ | . X X X X X X O O X . . O . . X . . . |
$$ | X X O X . . X X O . O . O X . . . . . |
$$ | O O O O X X X O O O . O O O O O O O . |
$$ | . O O X X . X X O . O X O X X X X O O |
$$ | . . O O X X O O O O X X X . . X . X O |
$$ | . . . O X . X X X X . . . . . . X . X |
$$ | . . O X X X X . . X . . . . . X X . . |
$$ | . . O X O O X X . . . . . X . . . . . |
$$ | . O . O . O X . X . . . . . . . . . . |
$$ | . . O . O O O X . . . . . . . . . . . |
$$ +---------------------------------------+[/go]

Thank you for any answer
Disagreements about alive/dead/inbetween - how are those resolved over the board?

//T

1. Neither are dead. It's seki.

2. Depends on ruleset. Either eyes count in seki, or no points at all.

_________________
be immersed

Ok, thanks for clarifying.

Guess I'll drop the author of the go engine an email

Not exactly.

It could be a (highly?) non-trivial problem.

Currently, Tygem, IGS, and KGS -- all require human agreement at the very end to mark live/dead groups.

Does anyone know the state-of-the-art (AlphaGo, Zen, etc.) situation in this area ?

It is a non-trivial problem because among other things, "end of the game" isn't so well defined. A game is TENTATIVELY ended, but a dispute could require resumption of play. But a seki isn't seki until the game is ended. For example, suppose in this dispute a ko is involved and the value of winning that ko greater than the value of the area in presumed seki. A play there (changing seki into death) could be the final ko threat.

The status of the group before playing the ko threat was still seki then. I don't get your point of attributing the status of the group only once the game ended.
Just imagine talking to a (newer) player about a game.
"Yea, that group on the lower right is seki once the game ends. Right now it doesn't really have a status, because either player could still play a ko threat there or just make a plain stupid move."

Edit:

Oh yea before I forget, nice one discovering a seki in a game of yours before knowing the rules and then wondering about it, Tapani. I usually explain the rule to players after they've played a few dozen games because it is good to know and it happens only so rarely, but I really like if people discover it on their own and then think about it first, before getting the explanation.

On WBaduk, scoring was also done manually in the past. However since a certain upgrade the server automatically scores the game immediately after both players passed. In my games, I haven't yet encountered a problem with that, however I'm not quite sure how it works when the server counting function has messed up...

If white says the black stones are dead after both players pass, white has to play additional moves try to kill them. If white succeeds, they are dead, otherwise they are alive. Either way you reset the board to how it looked after the passes, and score it from there.

Japanese rule / Korean rules : no. the territory is defined as the eyespaces completely surrounded by chains that are not in seki. Here, the chains are living in seki, so they don't surround territory.
Chinese, AGA, New Zealand, Ing, French and UK rules : yes, all empty intersections completely surrounded by stones of the same colour count.



That's right, at least in japanese / korean rules. And thanks to the people in this forum, I've learnt that White has to kill Black when it's White to play, but also when it's Black to play. Even if it would mean that Black had played two moves in a row. And ko bans are lifted, even if it would mean that a player recaptures immediately in a ko. This isn't written anywhere in the rule, at least in its english translation.
Also, in official japanese rules, it is forbidden to recapture in a ko while trying to prove that you can kill, unless you announce that you pass your turn for this ko (which obviously, let your opponent solve it). Rules are different during the game and during the demonstration of the death / alive status.
This rule is not supposed to be applied at the World Amateur Championship, that still uses the old japanese rule, which features a list of 24 figures and how to count them, but it is applied by moderators in kgs.

In chinese, Ing New Zealand, AGA, french, and UK rules, if both players can't agree, the game resumes until all stones are captured. Anything that remains on the board is then assumed to be alive.

Chinese, New Zealand and Ing rules use special ways of counting that allows you to kill any dead group without loosing points : they count one point per stone still on the board, plus one point per surrounded empty intersection. Prisoners don't count.

AGA, french and UK rules use pass stones to keep the japanese way of counting the score (empty intersections minus prisoners) while allowing to capture dead groups without loosing points : each time you pass, you must give a prisoner to your opponent, and White must make the last move. This way, once the prisoners have been used to fill territories, for games without handicap, there is exactly the same number of black and white stones on the board, and the two ways of counting (stones + empty intersections, or empty intersections minus prisoners) give exactly the same result.

There is a difference for handicap games between the two methods, that is partially compensated by extra points given to white in chinese, Ing and NZ rules, equal to the number of handicap stones (at least in chinese rules). It would exactly compensate the difference if it would be the number of handicap stones minus one.

Anyway, the komi is different for both family of rules : 6.5 in Japan (not sure about Korea), 7.5 in China, America, France and UK.
I don't know for Ing rules and New Zealand.

The main thing to remember is that in chinese, AGA, french, Ing, UK and NZ rules, you have to fill all the dame at the end of the game if you don't want to loose points (since each extra stone one the board is worth one point, or passing costs you one point). You might even want to fill the last dame before connecting the last ko.