Réti v Alekhine, Baden-Baden 1925: the gem of gems

Alexander Alekhine played this game against Richard Réti at the Baden-Baden tournament on 25 April 1925; he won the tournament without losing a game. Alekhine named it, with his win over Bogoljubow at Hastings 1922, as his most brilliant tournament game, and regretted that neither won a brilliancy prize. C. J. S. Purdy called it "the gem of gems", and Max Euwe described it as the deepest strategic combination with rooks and minor pieces.

Black to move after 26.axb5. Alekhine's next move starts a deep combination.

The whole game

Réti v Alekhine, Baden-Baden 1925. Step through all 40 moves; White resigned.

A disputed repetition

Around move 18 the bishops moved back and forth, and Alekhine claimed a draw by repetition. Réti protested, and the tournament director ruled that the repetition hadn't yet happened. Purdy argued the claim was wrong, since the position had occurred only twice, and suggested Alekhine may have made it on purpose, to make Réti overconfident. Alekhine wrote that accepting a draw would have been a moral defeat for White, so Réti played on, and that retreating the bishop to h1 gave Black a tempo.

The published versions of the game differ only in how many times the bishops go back and forth, so after move 17 their move numbers differ from ours by two either way. We follow the version in Alekhine's own book of his best games, the most common one.

What Stockfish makes of it

We had Stockfish check every move.

Up to move 26: White slightly better

For most of the game Stockfish rates White slightly better, by up to half a pawn. Réti himself later criticised some of his moves, suggesting e4 instead of 25.b5, and Stockfish agrees that 25.e4 was better.

26...Re3: the rook goes in

26...Re3 27.Nf3? cxb5 28.Qxb5 Nc3 29.Qxb7 Qxb7 30.Nxb7 Nxe2+ 31.Kh2 Ne4!!: the combination.

Alekhine's rook goes to e3, where the f-pawn can take it, and it stays there. Réti didn't take it. He played 27.Nf3, the move John Nunn marks as the critical mistake. Réti later suggested 27.Bf3, and Stockfish agrees: about level after 27.Bf3, against Black more than a pawn better after 27.Nf3. 28.Qxb5 made it worse; Stockfish prefers 28.Qd4.

31...Ne4!!: the knights take over

After the queens come off, Black's knights and rook attack White's king. 30...Nxe2+ and 31...Ne4 were each the only strong move: the next best leaves the position about level. Garry Kasparov gave 31...Ne4 two exclamation marks and counted the game among the greatest attacks without queens.

The finish

34...Ng4+ 35.Kh3 Ne5+ 36.Kh2 Rxf3 37.Rxe2 Ng4+ 38.Kh3 Ne3+ 39.Kh2 Nxc2 40.Bxf3 Nd4: White resigned.

34...Ng4+ and 40...Nd4 were also each the only winning move. 40...Nd4 forks White's rook and bishop, and Réti resigned.

How we checked

The moves are from Edward Winter's article on the game, which sets out the three published versions, and A. J. Goldsby's analysis page, which agrees with the one we use. We replayed every move of all three versions with python-chess: all are legal.
The engine: Stockfish 16.1 checked every move at depth 22, and the key moments again at depth 26. "A pawn better" is Stockfish's measure of the position, not a count of material.
The history and the named opinions (Alekhine, Réti, Purdy, Euwe, Nunn, Kasparov) are from Winter's article and Goldsby's page.

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

Start training