Seki not recognised?

The group in the corner is in seki or am I wrong? But it scored it as white being dead.

Opponent accepted the result and so did I but it’s not right is it? How come the auto scoring cannot see that?

It’s dead because once the outside liberties are filled completely, the White group will be in atari.

White will have to connect at the 1-3 point or be captured, which reduces the eye space to a 3-space eyeshape (with the group still in atari).

It’s actually a common non-seki dead shape one sees often in tsumego. :slight_smile:


You can always check AI analysis. If the scores match then it’s probably right.



Ah yes - missed that, thanks.


No worries :slight_smile:


Also here’s a fun recently posted game on recognising eyeshapes as alive (probably seki) unsettled, dead


Since noone has brought this up: You should be aware that marking dead and alive stones at counting falls on the players. The system may find dead stones and automatically mark them as a suggestion, but this is just a nice convenience. It can not solve tsumego for you.

So if you find a non-trivial situation where the auto-suggestion gets it wrong, that’s not a bug report.


Given the false expectations the auto-marking sets, and the many posts from OGS players who don’t know how to score, I think it might actually be better to not have any auto-marking feature at all.

1 Like

Spell check for Greek is usually subpar, and I need to correct things now and then, but it’s better than not having it. Especially for the majority of people who are useless with spelling.

People who think it’s the end of their effort are the problem, not the tool itself.


I like that players mark themself at first. That’s not such a big story.
If the situation is complicated (which is rare) then some automatic help, or/and moderators may interfere on request by a player.

On IGS, at the scoring phase, some stones are marked with “!” and a message says something like “stones marked with ‘!’ are probably dead, click on them to confirm”. In other words, all stones on the board are considered as living until they are marked dead by one of the players.

1 Like

I haven’t played on KGS for some years, but IIRC players always need to mark dead stones on KGS and there is no exclamation marks or anything to suggest which stones might be dead. So it’s quite similar to finishing an IRL game.

At the other extreme is FoD, where for some years now the server determines which stones are dead and players can’t do anything to change the server’s judgement. But that server does a good job, as it should when the players are at its mercy. I can’t remember it ever being wrong in any of my games.
This is convenient, because it avoids score cheating, but I wonder what it would do when the game isn’t quite finished yet (players passing prematurely). AFAIK this hasn’t happened yet in one of my games on FoD.

1 Like

Sorry, what is FoD?


What happens is you can get in a game if ping-pong - who blinks first - where one things the shape is dead and the other seki. Doesn’t happen much, and even if the AIv correctly recognises it, it may still happen, but the more the AI gets it right, it’s a nudge in the right direction.

I like that - “unclear status” as such?