Ko
There’s a shape in Go where you capture one of your opponent’s stones, and they could immediately capture yours right back, returning the board to exactly where it was a move ago — and then you could recapture, and they could recapture, forever. So Go has a rule against it: you may not make a move that recreates the previous board position. After your opponent captures in such a spot, you can’t take it straight back. You have to go play somewhere else first. The shape, and the rule, are both called ko — borrowed from the Buddhist word for an aeon, an unimaginably long span of time, named for the endless loop it would otherwise create.
The rule turns a stalemate into something far more interesting. Because you can’t recapture immediately, you instead play a ko threat — a move somewhere else big enough that your opponent feels they must answer it. If they do, the board has changed, the prohibition lifts, and now you can take the ko back.
This is why a ko fight is one of the most whole-board things in Go. A fight that’s nominally about a single stone in one corner suddenly drags in moves from every other corner, because each side is hunting for threats large enough to force a response, and weighing the value of the contested ko against everything they’re spending to win it — a cascade of second order effects where the local fight is decided by resources stored all over the board. A local skirmish becomes a referendum on the entire board.
And the loops can get genuinely strange. With three kos interacting, you can reach a cycle where every resolution lets the opponent restart it — the position repeats with no legal way forward, and under Japanese rules the game is simply declared no result and replayed. There’s a famous piece of Go lore that a triple ko appeared the night before the warlord Oda Nobunaga was assassinated in 1582, in a game involving the founder of the Hon’inbō house — which is why tradition treats the triple ko as an omen of disaster. (The omen tradition is real; the 1582 anecdote is legend — the earliest reliably recorded triple-ko games only go back to the 1700s.)
The transferable core is the rule against immediate repetition, and how productively it changes a deadlock. Plenty of conflicts are ko-shaped: you undo my move, I undo yours, nobody can hold the contested ground, and we’d grind there forever if allowed — a loop with no exit and a war of attrition no one can win. The ko rule is a designed constraint: you can’t simply reverse the last exchange — you have to go generate leverage elsewhere first, and bring it back. Which is exactly how real standoffs get resolved, and the heart of any negotiation: not by re-fighting the same square, but by opening a second issue large enough that the other side trades the contested one to settle it.
It’s a designed escape from infinite loops, and a quiet lesson about them. When you and an opponent are stuck reversing each other’s moves, the way out is never another reversal. It’s to change the subject to something they can’t ignore, and use what you win there to finally settle what you were stuck on.