Capture go problems ⚫

Capture go is arguably the simplest possible go variant. The only rules are:

  1. Black and white take turns placing stones. Passing is not allowed.
  2. The first player to capture one (or more) of the opponent’s stones wins.

Capture go is a great way to introduce new players to go, since you don’t have to explain scoring or ko. But it is also an interesting variant for experienced players. Many of the tactics from normal go no longer works, most obviously snapbacks and throw-ins. This means that it is generally easier to live: a single two-point eye is enough to guarantee life.

In this thread I’m going to share some of my favourite capture go problems. I’ll start with a super easy one:

Link to demo board

Black to play. There is only one winning move.


You are probably correct in assuming that it is a great way to introduce new players to go.
I think that it also has a disadvantage: learning new players that go is only about capturing stones. This might handicap them later in their attempt to improve their playing.

Every advantage has its disadvantage.


How is this a super easy problem? I am completely stumped. I’ve tried every single move and nothing works.

(Even though I am also in the pro-capture go teacher group.)

Edit : Ooh, passing is not allowed. Ok, I’ve got it. Nice problem.


The concept of territory also arises naturally by disallowing passing. If nothing is captured, your are eventually forced to fill in your own eyes, so the player who has secured more territory wins. So the game is actually closer to normal go than it might seem at first glance.

It is definitely a valid concern that you are learning a different game: the tactics involved are quite different. But there is also a lot of overlap.

In any case, I’m not here to advocate for it as an introduction right now, I just want to share some cool problems :slightly_smiling_face:


So white doesn’t just always win by mirroring blacks moves, even though not being able to pass gets black killed a lot.

White can’t mirror blacks A2, but I can still see it getting to a symmetrical position where black loses by not passing.

Feels very tricky

1 Like

You’re on the right track, finding a variation where white cannot mirror is what’s needed :slightly_smiling_face:

1 Like

Yeah I think I found the variation. I could see the position I needed to get to win, then just needed the right order of moves!


Offered some variations on the demo board

Solution diagrams for problem 1




Problem 2:
Black to play as usual.


Blacks first move is B3. This protects A4 from capture and threatens to surround C4. If White pushes through C3, black shoud play Atari at A2, followed by White B1 and Black C2, leaving White at a shortage of liberties and with no moves.




The interesting part of this problem is that black must refrain from starting with the atari (A2 B1 B3 B2 A3 C3 loses). Just like in normal go, forcing moves aren’t always good, even if they are absolute sente!


Another nice, witty problem. We want more!

1 Like

Very well, here’s another one :smiley:

Problem 3:

Demo board


These problems feel much more tactical than the usual tsumego, thinking about them give me the feeling of chess compositions.

My solution

Too many variations to show but I believe B2 is the solution.


Indeed! Atari is similar to a check (much more so than in normal go), and you can talk about a “capture in N” problem just as we have “mate in N” problems in chess. (also the whole game is built upon “zugzwang”)

Response to proposed solution

B2 A2, what next? :slightly_smiling_face:

1 Like
I have no idea

I thought C2, but now I see that I have problems.
B2 A2 C2 D2, C1, A4, B5.
Now I thought A1 failed to A5 A3 but I missed a liberty here.
So I guess the answer is A2 instead? But I don’t see a way after A2 B2. White wins by one move.

B2 A2 B1 could be tesuji, but next C2 C1 A4 B5 D2 and again White wins by one.
So many different tight variations. I’ll think about it some more.

1 Like
My (other) solution

Looking back at it today, B1 and C1 both seem to work. I read a few variations where black seems to win.
For example B1 B2 C1 A4 B5. Next White and Black share the dame and Black wins by one.
It’s a bit frustrating as I might just be missing a key variation and I don’t have any intuition why this is significantly different from B2.

1 Like

In the variation B1 B2 C1 A4 B5 black actually loses by one move in the end.

I might have ramped up the difficulty too quickly, I’ll make the next one after this a bit easier :sweat_smile:

1 Like

Oh yeah, That’s why I couldn’t see the difference between the variations.
In France we have a funny saying, “To be a dan player you have to be able to count up to three”
Yeah, I’m not a dan player. Give me a last try.


I think I've got it

A2, finally. A2 A3 B2 is clearly lost for white, but A2 B2 A3 B1 A4 A1. I thought that this was lost for black, but C1! and white doesn’t actually have 4 liberties.


I am a dan player, but I still can’t count to three sometimes :wink:

You got it!

I saw that you almost had it yesterday, but didn’t want to spoil C1 for you :slightly_smiling_face: Recognizing that a 2x2-shape actually is worth 0 moves is quite a common pattern in these problems. Might be worth noting that black can actually just as well B5 instead of C1, and no matter where white plays next inside the 2x2-shape, black will get another move there also!

1 Like