Scoring at the end game

How is the scoring for this game at the very end. what was the right way to do it?

by doing last move #92 , opponent lost the game, would have one otherwise?

What should the scoring have been?


Japanese rules - eyes in seki should not count.
Black wins anyway, it seems.


If White did not capture the black stone at the end, then white would score 0 points in that area under Japanese rules because it’s a seki and the black stone is NOT removed during counting as a dead stone. By capturing the stone White gets 1 point for the capture and no territory points as it’s a seki. So capturing was a good move which gained 1 point. If White didn’t do that black would have won by 1.5 instead. The need to capture dead stones in seki to get their points is something even experienced players get wrong sometimes.

Edit: above was based on me being lazy and not manually counting and assuming OGS counted correctly. But it did not! It incorrectly gave black a point for the eye in seki group, and White 2 points for the seki group, but they should all be zero points. The proper score for this board position should be back +1.5 not 0.5. and if White didn’t capture would have been black +2.5.

Manual count to confirm:
White on board: 7
White captured 7 black stones
Komi 5.5
White total 19.5

Black on board: 9
Black captured 12 White
Total 21


its not OGS, its what players chose. Kata on OGS says ~1.5


Wow! An interesting edge case of Japanese rules. Thanks for sharing @andreas.v :+1:

I dare say most if not all of us at our Go club would have gotten that wrong if it had come up IRL. In fact, we usually are not explicit about which ruleset we are using :roll_eyes: This got me wondering… How would this game be scored under other rulesets? :thinking:

Based on this page: Comparison of Some Go Rules | British Go Association it appears that all other rulesets are essentially equivalent in terms of scoring this situation so I will just consider Chinese. From that page and memory, the important points are:

  • Area scoring
  • Eyes in seki do count
  • Prisoners do not count (hence dead stones are just removed)
  • Dame are worth half a point for each player (so doesn’t make any practical difference)

Thus white has:

Right 22 + left 15.5 + Komi 5.5 = 43

And black has:

Bottom 13.5 + top 30 = 43.5

So result under Chinese (and AGA, BGA, Ing, and NZ) is B+0.5 and the score is the same regardless of move 92 :nerd_face:

1 Like

KataGo was correct.

But did the players do anything wrong? Move 92 was correct under Japanese rules.

Did the players determine the scoring of A3, B1 and F2, or is it just the OGS scoring algorithm that wrongly decided to score those?


you are referring to black stone captured at B1? So even if the final shape the same, the score can be different depending on how it is played. Interesting.

1 Like

Surely @uberdude was talking about the capture of A3 on move 92. That move is neccessary under Japanese rules to score any points in that corner, because seki scores no points. Even uncaptured stones left inside a seki group are not scored as prisoners (they are left on the board during counting). So white needed to capture A3 before passing (when playing under proper Japanese rules, instead of some deviation as implemented on IGS and possibly OGS?).


No, black stone at a3 with move 92.

1 Like

Got it, in seki under japaness ruleset, take the dead stone out of board gives white a point. Easy to remember. Thanks!


Even though move 92 was correct under official Japanese rules, I suppose that white could have won by 0.5 if they had passed on move 92, thus exploiting a bug in OGS’ implementation of Japanese rules?