Ladders and Nets in Go: Two Ways to Capture a Stone

A ladder and a net are the two standard ways to capture a stone that is trying to escape. A ladder (shicho in Japanese) chases the stone in a zigzag across the board, keeping it at one liberty the whole way until it runs out of room. A net (geta) does the same job with one loose move that cuts off both escape routes at once. The net is usually the safer choice, because a ladder can be ruined by a single stone sitting far away on the other side of the board.

If liberties and capture are still new, the rules of Go covers them first, and the interactive tutorial walks through capturing a stone on a real board.

How a Go ladder works

Here is a white stone with two liberties left. Black’s stone in the top right of the group is the one that makes what follows possible.

A 9x9 board with a single white stone hemmed in by three black stones, leaving it two liberties and a black stone waiting diagonally on the escape route.
White's stone has two liberties. Black is about to take one of them.

Black takes the liberty below, putting the stone in atari (one liberty left). White has to run, and there is only one point to run to. Black takes the next liberty, White runs again, and a staircase starts to form.

The first four moves of a ladder: black puts the white stone in atari, white extends, black ataris again, and the white group steps diagonally.1234
Every black move leaves White with one liberty. Every white move gets it back to two.

That pattern is the whole technique. Black never has a choice about whether it works, only about which of two points to take, and White never has a choice at all. The staircase marches diagonally toward the corner, and eventually the board runs out. In the corner White has nowhere left to extend to, and the whole group comes off.

The end of a ladder in the corner of a 9x9 board: the white group has one liberty at the marked point and no way to extend further.
One liberty, at the marked point, and the edge of the board behind it. Seven white stones are lost instead of one.

That is why a ladder is worth so much when it works. White did not lose one stone by running; White lost seven, and every one of those moves was wasted.

Read the ladder before you start it

A ladder only works if nothing is sitting in its path. Before you play the first atari, follow the staircase with your eyes all the way to the corner and check whether either player already has a stone there. If your opponent does, the ladder fails, and it fails catastrophically, because by then you have handed them a free run of forcing moves.

The proverb is blunt about it: if you don’t know ladders, don’t play Go. It is one of the few things in the game where the answer is completely knowable in advance if you take twenty seconds to look.

A stone that spoils a ladder is called a ladder breaker. It does not have to be near the fight. On a full-sized board it can be fifteen lines away.

A 13x13 board showing a ladder about to start in the upper left and a white stone already sitting on the ladder's path near the lower right corner.
The marked white stone is nowhere near the fight, and it decides the fight.

Play the ladder anyway and this is what happens. When the staircase arrives, White joins up with the waiting stone, the group is suddenly out in the open, and Black’s chasing stones are riddled with gaps.

The same 13x13 ladder after it has run into the white breaker stone, with the white group connected and free and black's chasing stones full of cutting points.
White connects and is out. Black spent eleven stones building a wall full of holes.

On a 9x9 board this rarely comes up, because a ladder reaches the edge in a handful of moves and there is not enough room left for a breaker to matter. On 19x19 it matters constantly.

The net (geta)

A net captures the same stone without any of that. Instead of chasing, you play one move a small distance away that covers both escape routes at once. Here Black has one more stone in place than in the ladder diagrams, and that is enough.

A white stone with two liberties surrounded loosely by four black stones, with the marked point showing where a net would be played.
Black plays the marked point. It touches nothing, and White is already caught.

White has two ways to run and both fail in one move. Run to the right, and Black takes the liberty on the outside.

White tries to escape a net to the right, black answers on the outside, and the white group is left with a single liberty and no useful extension.123
1: the net. 2: White runs right. 3: Black blocks, and White's group has one liberty and nowhere to go.

Run downward instead and exactly the same thing happens on the other side.

White tries to escape the same net downward, black blocks below, and the white group again has a single liberty with no way out.123
2: White runs down instead. 3: Black blocks below. Extending again would not add a liberty, so White is captured.

In both cases White can keep placing stones, but every extension is surrounded the moment it lands, so the group never gets back to two liberties. The stone is dead where it stands. Nothing anywhere else on the board changes that, which is the point.

Ladder or net?

Use the net when it works. A ladder depends on a long diagonal staying clear, and that can change later in the game: your opponent can play a stone in the path as a forcing move, then come back and rescue the stone you thought you had. A net is settled the moment you play it.

The ladder is not obsolete, though. It works in positions where no net is available, and it is worth more when it does work, because you often get to pick up forcing moves along the way. The practical rule for a beginner: look for a net first, and only fall back on the ladder if there isn’t one, and then only after you have read it to the corner.

Both of these are captures, and captures are not the point of the game by themselves. A net that takes one stone while your opponent takes an empty corner is not obviously a good trade, which is the judgement covered in Go strategy for beginners. The related question of when a group is too weak to run at all is on life and death in Go.

Practising ladders

Ladders are the easiest thing in Go to practise on your own, because the answer is always knowable. Set up a game on a 9x9 board, cut one of the computer’s stones off, and chase it. On a small board you will hit the edge within a few moves, which makes the whole staircase easy to see in one glance. Once the pattern is automatic, try the same thing on a larger board, where the chase crosses half the game and you have to look ahead before you start.

If you want to see the shapes that make stones hard to capture in the first place, good shape in Go is the companion page, and the Go opening explains why a stone worth capturing was sitting there in the first place. You can also just start a game and try to get cut, which teaches this faster than reading does.

Frequently asked questions

What is a ladder in Go?

A ladder is a capturing technique where you put a stone in atari, the opponent extends, you put it in atari again, and the chase zigzags diagonally across the board until the escaping group runs into the edge and is captured. The Japanese name is shicho. It only works if no stone of the escaping player’s colour is sitting on the diagonal path.

What is a ladder breaker?

A ladder breaker is a stone in the ladder’s path that lets the escaping group connect and get free. It can be anywhere on the diagonal, including the far side of the board, and it does not have to have been played for that purpose. If a ladder breaker is in place, starting the ladder is much worse than not capturing at all, because every chasing stone becomes a weakness.

What is a net (geta) in Go?

A net is a loose capture: one move played a short distance from the stone you want, positioned so that both escape routes lead to immediate capture. It usually takes one move where a ladder takes many, and unlike a ladder it does not depend on anything far away, so it cannot be broken later.

Try it on the board →