LaviChess

The Tarrasch Trap: castling too early, and the trap Marco fell into in 1892

Black castles, just as everyone learns to do. But in this very position it costs a pawn, and trying to win it back loses more. Siegbert Tarrasch had written the whole trap down in advance.

abcdefgh87654321
The position before 7…O-O. Black to move, and the natural move loses.

The trap

The trap arises in the Steinitz Defense of the Ruy Lopez: 1.e4 e5 2.Nf3 Nc6 3.Bb5 d6. Black supports e5 with a pawn, and White keeps pressing it with 4.d4 Bd7 5.Nc3 Nf6 6.O-O Be7 7.Re1.

Now Black castles, 7…O-O, and it costs a pawn. After 8.Bxc6 Bxc6 9.dxe5 dxe5 10.Qxd8 Raxd8 11.Nxe5 White has won the pawn on e5.

1. e4 e5 2. Nf3 Nc6 3. Bb5 d6 4. d4 Bd7 5. Nc3 Nf6 6. O-O Be7 7. Re1 O-O 8. Bxc6 Bxc6 9. dxe5 dxe5 10. Qxd8 Raxd8 11. Nxe5 Bxe4 12. Nxe4 Nxe4 13. Nd3 f5 14. f3 Bc5+ 15. Nxc5 Nxc5 16. Bg5 Rd5 17. Be7

abcdefgh87654321

7. Re1

White defends e4 and waits. Now the pressure on e5 is too much.

  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.…
Open on the analysis board

Why it's easy to fall for

Castling early is one of the first things you learn. Here it also looks safe, because Black's pieces guard e5 with both a knight and a pawn.

But White can trade off the defenders in the right order. First the bishop for the knight on c6, then the pawns on e5, then the queens. Then the pawn on e5 hangs.

The second mistake comes after that. With 11…Bxe4 Black wants to win the pawn back, because 12.Nxe4 Nxe4 13.Rxe4?? fails to 13…Rd1+ and mate. But White plays 13.Nd3! instead, and Black loses material anyway.

What Black should have played

Play 7…exd4. Black gives up the center pawn but doesn't lose material. White is slightly better, no more.

If the pawn is already lost, Black shouldn't try to win it back with 11…Bxe4. Black is badly off a pawn down, but after 11…Bxe4 it gets worse still.

abcdefgh87654321

7… exd4

Right. Black gives up the center pawn but doesn't lose material. White is only slightly better.

  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
Open on the analysis board

How it ends

White wins first a pawn and, after 11…Bxe4, the exchange. After 17.Be7 the bishop forks the rook on f8 and the knight on c5. The rook on d5 protects the knight, but after 17…Re8, for example, 18.c4 wins the exchange.

After castling, Siegbert Tarrasch wrote that Black was lost. It isn't that bad: after 7…O-O White is clearly better with an extra pawn, but the loss only came with 11…Bxe4.

How often does it happen?

Our numbers reach as far as 5…Nf6, which 43,536 of the 319,640 games between club players in that position played (14%). White won 58% of those games. After that the position is too rare to be among the positions we have counted, so we can't say how often the mistake itself is made.

Our numbers reach as far as 5…Nf6, which 9,414 of the 43,704 games between strong players in that position played (22%). White won 56% of those games. After that the position is too rare to be among the positions we have counted, so we can't say how often the mistake itself is made.

Move by move: club players

MoveGames in the positionPlayed the moveThen White wonDrawBlack won
4…Bd74,049,0871,158,255 (29%)55%5%39%
5.Nc31,164,662172,236 (15%)63%5%32%
5…Nf6319,64043,536 (14%)58%6%36%
Move by move: strong players
MoveGames in the positionPlayed the moveThen White wonDrawBlack won
4…Bd7559,027105,718 (19%)58%6%36%
5.Nc3106,98327,971 (26%)63%6%31%
5…Nf643,7049,414 (22%)56%8%36%

"Played the move": of the games in the position, how many continued with the move of the trap line; then how those games ended. A position missing from the table is too rare for our numbers. Source: games on Lichess in blitz, rapid and classical, fetched 6 and 7 November 2026. Club players means games where the players average 1600 or more, strong players games where they average 2000 or more.

Explore the position on the analysis board: the moves club players choose here and how the games tend to end.

Try it

abcdefgh87654321

Black has just taken 11…Bxe4 to win the pawn back. White to move: show that it doesn't work!

Your opponent moves…

Six moves to find

The history

Tarrasch published the analysis in the magazine Deutsche Schachzeitung in February 1891, in his notes to a game between Isidor Gunsberg and Wilhelm Steinitz. In his textbook Das Schachspiel (1931) he was milder himself: after castling, Black loses “certainly a pawn and perhaps the Exchange”, as the English edition has it.

A year and a half later, at the German Chess Federation's congress in Dresden in 1892, Georg Marco fell into the trap against Tarrasch himself. Tarrasch later wrote in his book Dreihundert Schachpartien that the game was an exact copy of his analysis. Marco resigned when he saw that the exchange would be lost. The sources disagree on exactly when: according to Tarrasch's book straight after 17.Be7, according to others after 17…Re8 18.c4.

Tarrasch also gave his name to a trap in the Open Variation of the Ruy Lopez. Two players who challenged Steinitz for the world title, Johannes Zukertort and Isidor Gunsberg, fell into that one.

Frequently asked questions

What is the Tarrasch Trap?

A trap in the Steinitz Defense of the Ruy Lopez. If Black castles with 7…O-O, White wins a pawn, and if Black tries to win it back with 11…Bxe4, Black loses even more after 13.Nd3!, the exchange in the game against Marco. Siegbert Tarrasch analyzed it in 1891, and Georg Marco fell into it against him in 1892.

What should Black play instead of 7…O-O?

7…exd4. Black gives up the center but keeps the material.

Every move on this page has been checked against the rules of chess, and every judgement of who stands better has been checked with a chess computer.

Updated November 14, 2026