The Immortal Zugzwang Game: Sämisch v Nimzowitsch, 1923

Aron Nimzowitsch won this game against Friedrich Sämisch in Copenhagen in March 1923. It ends with almost all the pieces still on the board and White unable to move anything without losing more. Zugzwang is the name for a position where having to move is a disadvantage; Nimzowitsch wrote in 1925 that the game's name came from "Danish chess circles".

White to move after 25...h6. Every move loses, and White resigned.

The whole game

Sämisch v Nimzowitsch, Copenhagen 1923. Step through all 25 moves; White resigned.

The opening is a Queen's Indian Defence. Black's plan is slow and deep: the knight that comes to c6 presses on c4 (Nimzowitsch's "ghost"), the queenside pawns advance to cramp White, and then Black turns to the kingside.

What Stockfish makes of it

We had Stockfish check every move.

The opening: 9.cxd5 and 14.h3

White to move after 8...c6. Stockfish prefers 9.e4 to 9.cxd5.

9.cxd5 is usually marked as dubious, and Stockfish agrees: 9.e4 would have kept White about three-quarters of a pawn better, while after 9.cxd5 the game is level. 14.h3, also usually criticised, was only a small inaccuracy by Stockfish's measure: 14.Ne4 was a little better.

16...f5 and 17.Qd1: one slip each

Nimzowitsch's 16...f5 is often given an exclamation mark. Stockfish prefers 16...b4, which would have left Black nearly a pawn better; after 16...f5 White is slightly better. Sämisch gave it straight back: 17.Qd1 let 17...b4 drive the knight to b1, and Stockfish rates Black about a pawn better after it. 17.Nb1, retreating at once, kept the balance.

20.e4 and 20...fxe4!: the sacrifice

20.e4 fxe4! 21.Qxh5 Rxf2: Black gives up the knight on h5.

20.e4 was White's real mistake: Stockfish prefers 20.a3, about three-quarters of a pawn worse for White, against more than two pawns after 20.e4. Nimzowitsch replied with 20...fxe4!, leaving the knight on h5 to be taken. It was the only strong move: the next best leaves the position about level. For the piece he got, as he put it, two pawns, a rook on the seventh rank and a queenside White could not untangle.

25...h6!!: the zugzwang

24...Bd3 25.Rce1 h6!!: Black makes a waiting move, and White has nothing useful to play.

After 25...h6 Black threatens 26...R5f3, winning White's queen. White has more than 20 legal moves, and Stockfish rates every one of them as losing by more than four pawns. If White's king or g-pawn moves, 26...R5f3 follows, and Black can then make waiting moves with the king. Stockfish rates 25...h6 as Black's strongest move, though it isn't the only winning one. Sämisch resigned.

Is it really zugzwang?

Not everyone accepts the name. Andrew Soltis argued that White could give back material and play on, and that a real zugzwang loses because of the victim's own move, not because of a threat; here Black threatens ...R5f3 anyway. Wolfgang Heidenfeld made the same point, and Raymond Keene preferred to call it "total paralysis". Nimzowitsch himself regarded it as zugzwang, and Wikipedia's article on the game notes that any White piece move loses more than the threat does.

How we checked

The moves are from Wikipedia's page on the game. We replayed every one with python-chess: all are legal.
The engine: Stockfish 16.1 checked every move at depth 22, and the key moments again at depth 24 to 26. "A pawn better" is Stockfish's measure of the position, not a count of material.
The history and the named opinions (Nimzowitsch, Soltis, Heidenfeld, Keene) are from the Wikipedia article.

Try it. Real positions, set at your level, every answer checked against the board. No account needed to start.

Start training