Living with three groups in 9x9

I looked around but didn’t see any thread about living with a lot of groups in 9x9.

In 19x19 we have Five Groups Might Live But The Sixth Will Die. The equivalent for 9x9 seems to be 2 vs. 3 groups. Here’s one page about it (that I saw thanks to @mark5000’s link elsewhere):

When I notice my opponent (usually White) trying to make 3 groups in 9x9, I switch my focus to keeping them disconnected and connecting my own groups, and I almost always find that one of them dies naturally, swinging the balance dramatically in my favor.

However, in my last two correspondence games my opponents - one of each color - managed to live with three groups!

(The second was a handicap 2 game so I guess White really had to do something crazy like that to have a chance.)

I still won both but thought that was interesting and would be happy to hear others’ ideas or see more examples.


I think that’s rare, so nice to show!


Interesting stuff, thanks for sharing those games!

In both cases you missed multiple chances to kill. Keep working those tsumego problems :slight_smile:


Doesn’t it depend on the skill level of the players? Beginning players might be able to keep 4 or 6 groups alive, if their opponent can’t kill them. Advanced players, with better play, might have a natural 2 or 3 group limit. In my case, in between, I’ve sometimes kept 4 groups alive, and sometimes none, so my skill level proves nothing at all. Of course, with perfect play on both sides, there must be a fixed limit. But the OP doesn’t say that play is ideal or perfect.

1 Like

4 on 9x9?! Quite a performance. I can’t do that myself in a regular game.

1 Like

Indeed. But with reliable very good play, I can trust the AI to guide me in evaluating my moves, after a game has been played.

5 groups

After a bit of doodling I got 5 living black groups.
Five seems to be the maximum number of groups on 9x9.
But maybe smarter minds than me will prove me wrong.

White wins by 6.5 points.


I suppose it depends on how you define a “group”, but how about this?


I’m starting to get the feeling that a “formally prove upper and lower bounds for the number of live groups on an nxn board” thread is in the making, here, lol


I think this board has radically different scores in area and territory counting, because of eyes in seki.

Japanese rules I count 4 points black, 6.5 white (basically komi, no points on the board).

Chinese rules 45 black, 36.5 white.

But, as a constructed position, there are 37 black stones on the board and 30 white stones. So 7 difference in prisoners makes black ahead by 4.5 in japanese rules and ahead by 8.5 in chinese rules.

Difference of 4 and there are 4 black eyes in seki. Seems to add up…

Edited for math and score difference of 1 disappeared!