The Ko Rule in Go: Why You Cannot Take Back Right Away

When you have just learned go, it can happen that one of your stones is captured, you try to take one back at once, and the move is refused. You can play anywhere else, just not there. This is ko.

In Go Rules for Beginners I only wrote that you cannot immediately recapture in the same spot. In this post I show what the shape looks like and what to do when you want to take back.

The word comes from the Japanese コウ (劫, kō). According to the English Wikipedia article on the ko fight, it comes from a translation of the Sanskrit word kalpa, an aeon. If the taking back and forth were allowed, it would never end.

The ko shape

The diagrams show part of the board. ● is Black, ○ is White, and ・ is an empty point.

・●○・
●○A○
・●○・

The single white stone in the middle is surrounded by black stones above, to the left and below. Its only liberty is A. If Black plays at A, the white stone is captured.

・●○・
●B●○
・●○・

This is the position after the capture. Call the point where the white stone stood B. Now the black stone played at A is surrounded by white stones above, to the right and below, and its only liberty is B. It looks as if White could play at B and capture it back.

But if White recaptures at B, the board returns to exactly the first diagram. Then Black can take at A again, and White at B again, forever. This shape, where each side takes one stone in turn, is called ko.

The ko rule

As Wikipedia puts it, the ko rule is a special rule that prevents repetition of position. The player whose stone was just taken in a ko may not take back immediately. They play somewhere else first, and once the opponent has answered, the recapture becomes legal.

So White cannot recapture at B straight away. White first plays elsewhere on the board. If Black answers, the board is no longer the same as before, and White may now take at B. Then it is Black who cannot take back at A immediately.

The purpose of the rule is to stop the whole board from repeating. If you remember it as “you may not bring back the same board”, you are less likely to get confused in other shapes.

Ko threats

The move you play elsewhere so that you can take back is called a ko threat. A ko threat is a move that demands a reply. If the opponent answers it, you can recapture the ko.

A ko threat only works if the opponent cannot ignore it, for example a move that would capture the opponent’s stones if left alone. If the opponent answers, you take back the ko. If the opponent ignores it and ends the ko instead (for instance by filling the point where the captured stone was), you gain something where you played the threat. Weighing those two is what a ko fight is about.

Where beginners get stuck

There are three common places.

The first is not knowing the rule and trying to recapture at once. Once you know it, this stops.

The second is playing a ko threat that the opponent simply ignores. Playing anywhere on the board is not enough. The opponent will just end the ko. A ko threat has to be a move that costs the opponent something if left unanswered.

The third is answering every ko threat by reflex. If you answer the opponent’s threat, the opponent retakes the ko. Compare how much you lose by ignoring the threat with how much you gain by winning the ko. If the ko is bigger, ignoring the threat and ending the ko is a real choice. This is a hard comparison at first, but just knowing that you can choose already helps.

Japanese and Chinese rules

An ordinary ko works the same way under every ruleset. The difference shows up only in much rarer shapes, such as three kos on the board at once.

According to the English Wikipedia article on the rules of go, Japanese and Korean rules have no superko rule, and a triple ko that never ends leads to an undecided game. Rulesets derived from the Chinese rules, including those of the American Go Association and the New Zealand Go Society, have a superko rule that forbids repeating any previous board position. In normal play you will hardly ever need to think about this difference.

Ko on a 9x9 board

On a 9x9 board, stones run into each other quickly. Capturing fights appear early in the game, so the chance to learn ko also comes early.

Also, because the board is small, there are fewer places to find ko threats. On a large board, weak groups here and there serve as threats, but on 9x9 there are fewer of them. When a ko starts, try counting how many threats you and your opponent have. It makes the outcome of the ko easier to see.

Playing in MindSports

StudyX Mindsports is an app with 12 strategy games, including go. Go is played on a 9x9 board, against a computer opponent from Lv.1 to Lv.50.

In MindSports go, you cannot play a move that returns the board to any position that has appeared before. If you try to recapture a ko immediately, the stone is not placed and the app shows “Superko: repeating a previous board position is not allowed”. For an ordinary ko, this behaves exactly like the rule above. Play once somewhere else, and after the opponent answers you can take back.

The beginner go lesson has a section on ko. You can check on the board, by tapping answers, that you cannot take back at once and how a ko threat lets you take back afterwards. The games and beginner lessons are free, with ads.

How to use the app is in Getting Started with StudyX Mindsports.