The positional exchange sacrifice · For club players · 15 min read
Positional Exchange Sacrifice: Petrosian's 25...Re6 and Stockfish
Petrosian gave a rook for a bishop against Reshevsky at Zurich 1953. Stockfish 19 shows the famous sacrifice was a rescue, and how to make the same trade work for you.
- threat
- plan
- key square
- outpost
- pawn chain or line of force
In short
- A positional exchange sacrifice gives up a rook for a bishop or knight, not to win material or hit the king, but to stop a pawn, win squares and leave the opponent with a piece that has no job.
- Petrosian played one against Reshevsky in round 2 of the 1953 Candidates tournament in Zurich: 25...Re6. The game was drawn after 41 moves.
- Stockfish 19 finds that the famous move was a rescue, not a plan chosen from a level position. Black was more than two pawns worse, 25...Re6 was the only good move, and White's 25.Rfe1 and 26.a4 gave most of the edge back.
- The trade works when it blocks a pawn, fixes the opponent's remaining bishop behind its own pawns and hands your pieces outposts nobody can challenge.
How we know what we say
- Verified From the record: Stockfish's numbers, the players' own notes or interviews, or a cited source.
- Inferred Read from the moves and the position: a reasoned reconstruction, not the player's words.
- Speculative A guess about mood, intent or practical motives: plausible, not proven.
Part 1Zurich 1953: Petrosian against Reshevsky, round 2
The 1953 Candidates tournament in Zurich was a double round-robin of fifteen players, held from 29 August to 24 October. Vasily Smyslov won it with 18 points. Samuel Reshevsky, Paul Keres and David Bronstein shared second place with 16, and Tigran Petrosian finished fifth with 15. [4] Verified
The game was played on 31 August 1953, in the second round. Reshevsky had White, Petrosian had Black, and the game ended in a draw after 41 moves. It is remembered for one quiet move on the 25th turn. [1][2] Verified
Reshevsky was an American grandmaster; Petrosian, then a young contender, would win the world title ten years later in 1963. Neither needed a brilliancy here. The game is a lesson in squares and pawn structure.
| White | Black | |
|---|---|---|
| Player | Samuel Reshevsky | Tigran Petrosian |
| Event | Candidates tournament, Zurich, round 2 | same |
| Opening | Nimzo-Indian Defence, Rubinstein System | same |
| Result | Draw in 41 moves | same |
| Tournament place | Joint 2nd, 16 points | 5th, 15 points |
Part 2What is a positional exchange sacrifice?
A positional exchange sacrifice is the surrender of a rook for a bishop or a knight ("the exchange") in return for lasting positional gains rather than material or a quick attack: a blocked pawn, a dominant outpost, a ruined enemy bishop. A rook is worth about two pawns more than a minor piece on the usual scale, so the side that gives it must collect that difference in squares. Inferred
It is not a gamble on the king. After the trade the position is usually closed, the pawn structure is fixed and nothing is hanging. The side with the extra rook finds that it has no file to use, no entry square, and a minor piece that bites on its own pawns. The side that gave the rook finds that its remaining pieces stand on the best squares on the board.
- plan
- outpost
- Black's outposts
- pawn chain or line of force
- threat
- target
- Open files
- pawn chain or line of force
Left: pawns fixed, outposts for the minor pieces, the rook's files closed. Right: an open board where a rook is simply worth more.
Part 3How the Nimzo-Indian produced the structure
Petrosian chose the Nimzo-Indian Defence. With 8.a3 Bxc3 9.bxc3 Black traded the bishop to double White's c-pawns; White got the two bishops and the central break e3-e4 in return. After 10.cxd5 exd5 the c-pawns were reduced to one pawn on c3, and Black's task was to keep the centre fixed. Inferred
1.d4 Nf6 2.c4 e6 3.Nc3 Bb4 4.e3 O-O 5.Bd3 d5 6.Nf3 c5 7.O-O Nc6 8.a3 Bxc3 9.bxc3 b6 10.cxd5 exd5 11.Bb2 c4 12.Bc2 −0.10about equal
The opening as played. Stockfish rates the final position as level. Verified
The key move is 11...c4, which attacks the bishop on d3 and fixes White's centre. The bishop retreats to c2, a passive square where e3-e4 is the only way to bring it back to life. Black's pawn on c4, supported by d5, takes away b3 and d3. The d3 square becomes a hole: neither the pawn on c3 nor the pawn on e3 can ever go back to cover it. Inferred
- idea not played
- key square
- Squares c4 controls
- pawn chain or line of force
Part 4White's centre against Black's holes
After 12.Bc2 Bg4 13.Qe1 Ne4 14.Nd2 Nxd2 15.Qxd2 Bh5 16.f3 Bg6 17.e4 Qd7 18.Rae1 dxe4 19.fxe4, White has the centre and the space. The plan is e4-e5, the queen to f4 and the bishop to g4. Black's plan is to resist in the centre and use the hole on d3. Inferred
- plan
- White's centre
- plan
- key square
- Black's holes
The same position after 19.fxe4, two plans. White has the central pawns and the space; Black has the squares. Stockfish sees White slightly better.
The race is about timing. If White gets e5 and the queen to f4 unopposed, Black's pieces are driven back and the hole on d3 is out of reach. The next section shows where Black lost that race, and where White gave it back.
Part 5Where the game was decided
Move 19: the knight that Black did not play
Critical moment 1 of 4 · Move 19, Black to play
How does Black challenge the centre?
- key square
- White's centre
White has just rebuilt the centre with fxe4: pawns on d4 and e4 with the queen behind them on d2. Black's pieces have little room, and the hole on d3 is there. What is Black's most active reply?
Think it through on the boardWeigh the candidates
| Move | Stockfish | Verdict |
|---|---|---|
| 19...Ne5 20.Bc1 Rad8 21.Qe2 Nd3 | +0.33about equal | Stockfish's choice. White cannot take with dxe5 because the queen on d2 would fall to ...Qxd2, so the knight reaches d3 and White keeps only about a third of a pawn. Verified |
| 19...b5 20.h4 a5 21.Bc1 | +0.80White slightly better | Gains queenside space, but White keeps about four-fifths of a pawn. Verified |
| 19...Rad8 20.h4 b5 21.Bd1 | +1.48White clearly better | The other rook move: no better than the move played, with White about a pawn and a half ahead. Verified |
What Petrosian played
19...Rfe8+1.60White clearly better
Petrosian put a rook on the open e-file with 19...Rfe8. It is the natural move: the rook joins the file where White's pawn will advance.
Stockfish preferred 19...Ne5 (+0.21); the game move gave up about 1.4 pawns. Depth 23.
What it revealed
The natural rook move overlooked a tactical point. The queen on d2 stands undefended behind the pawn on d4, so a knight on e5 cannot be captured. The rook move did nothing about e4-e5, and White's space grew.
What Stockfish adds Stockfish puts the played move about a pawn and a third below 19...Ne5: White's edge grows from about a fifth of a pawn to about a pawn and a half. Black was worse after the move, not beaten.
- Verified After 19...Ne5 the capture 20.dxe5 loses the queen to 20...Qxd2, because no white piece guards d2. Stockfish rates 19...Ne5 as leaving White about a third of a pawn better.[1]
- Inferred The rook was not badly placed by itself. The error was the move not played: nothing in Black's reply stopped e4-e5.
Move 23: a pawn move that loosens the queenside
Critical moment 2 of 4 · Move 23, Black to play
White has played e5. Where does Black spend the next tempo?
- key square
- White's grip
White's pawn has just reached e5 and the queen stands on f4. Black is under pressure on the kingside. Where should Black spend the next tempo?
Think it through on the boardWeigh the candidates
| Move | Stockfish | Verdict |
|---|---|---|
| 23...Rd8 24.Re3 h6 25.Bc1 Bd3 | +1.92White clearly better | Stockfish's choice: the rook comes to the d-file first and the bishop heads for d3. White keeps close to two pawns. Verified |
| 23...Bd3 24.Rf3 Rd8 25.h4 | +2.04White winning | Takes the outpost at once; a little less accurate, with White about two pawns ahead. Verified |
What Petrosian played
23...a5+2.36White winning
Petrosian replied 23...a5, supporting the b5 pawn and preparing ...b4 or ...a4. A queenside pawn move is the usual reaction to a closed centre.
Stockfish preferred 23...Rd8 (+1.81); the game move gave up about 0.6 pawns. Depth 21.
What it revealed
The pawn move was slightly too slow. It does nothing against White's kingside pressure, and Black's knight and bishop still need their squares. It is a small slip in a position that was already difficult.
- Verified Stockfish rates 23...a5 about half a pawn below 23...Rd8. The position was already about two pawns better for White either way.
- Inferred The pawn move is a reflex: a closed centre sends the player to the wing. Here the e-file and the d3 square were the places to look.
Part 6Move 25: White lets the advantage go
Critical moment 3 of 4 · Move 25, White to play
White to move with a big edge: which piece goes where?
- key square
- Black's cramped pieces
After 24...Rd8 White has rooks ready on the e- and f-files, a bishop on g4 and a pawn on e5, and Black's pieces are cramped. White has a big edge. Which move keeps it?
Think it through on the boardWeigh the candidates
| Move | Stockfish | Verdict |
|---|---|---|
| 25.Bf3 25...f6 26.Qh4 Re6 27.Ree1 | +2.43White winning | Stockfish's choice. White keeps about two and a half pawns, and the same rook move to e6 would now lose to d5. Verified |
| 25.h4 25...h5 26.Bf3 f6 | +1.88White clearly better | Also strong: nearly two pawns for White, and clearly better than the move played. Verified |
What Reshevsky played
25.Rfe1+1.06White clearly better
Reshevsky doubled rooks with 25.Rfe1, a natural move that adds a second rook to the e-file.
Stockfish preferred 25.Bf3 (+2.36); the game move gave up about 1.3 pawns. Depth 23.
What it revealed
The move let Black answer with the rook to e6. Here the push 26.d5 no longer wins anything: after 26...Rxd5 27.Bxe6 Qxe6 Stockfish sees the position as level, so White's grip on the position was gone.
What Stockfish adds This is the largest swing of the game, bigger than Black's mistake on move 19. White gave away well over a pawn of advantage with a move that looks natural.
- Verified Stockfish rates the position before 25.Rfe1 at more than two pawns for White and the position after it at about one pawn.
- Inferred Doubling rooks on a file that White's own pawn on e5 has closed adds little. The bishop and queen needed to find squares first.
After 25.Bf3 the same exchange offer would fail. In the line below 25...Re6 is met by 26.d5, which attacks the rook and the knight at once, and Stockfish sees White winning. Verified
25.Bf3 Re6 26.d5 +6.22White completely winning
With the bishop on f3 instead of g4, 25...Re6? runs into 26.d5. The reason 25.Bf3 is better is that it leaves the exchange offer nothing to work with. Verified
27.Bxe6 fxe6 28.Qf1 Nd5 29.Rf3 Bd3 30.Rxd3 cxd3 31.Qxd3 b4 32.cxb4 axb4 +0.08about equal
From 26...Ne7 the game goes: Bxe6 fxe6 Qf1 Nd5 Rf3 Bd3 Rxd3 cxd3 Qxd3 b4 cxb4 axb4. Black wins d5 and d3, trades the pressure away and reaches a level ending. Verified
Part 7How Black defended: the quiet rook
Critical moment 4 of 4 · Move 25, Black to play
White threatens e6. What does Black do?
- threat
- key square
- weakness
White's pawn on e5 wants to advance, and the bishop on g4 eyes the squares e6 and d7. Black has rooks on e7 and d8 and a queen on e8, in a cramped position. What is the most resilient reply?
Think it through on the boardWeigh the candidates
| Move | Stockfish | Verdict |
|---|---|---|
| 25...Ra7 26.e6 f6 27.a4 | +2.29White winning | Gives up the blockade: about a pawn worse than the move played. Verified |
| 25...Na7 26.h4 Nc8 27.h5 | +3.70White winning | Sends the knight away from the centre; White is winning, more than three pawns up in Stockfish's numbers. Verified |
| 25...a4 26.h4 h5 27.Bf3 | +4.04White completely winning | A pawn push with no point; White wins easily. Verified |
What Petrosian played
25...Re6+1.19White clearly better
Petrosian played 25...Re6, putting the rook where the pawn wanted to go and inviting the bishop to take it.
Stockfish's first choice too. Depth 30.
What it revealed
The move stops the e-pawn. If White takes with Bxe6 fxe6, the knight reaches d5 and the bishop d3, while White's bishop on b2 bites on its own pawns. The exchange has bought squares.
What Stockfish adds Stockfish agrees that 25...Re6 is Black's only good move. The next best, 25...Ra7, is already about a pawn worse. Even after the best defence White stays about a pawn up at this point.
- Verified Mikhail Tal said the move made an indelible impression on Tal because it was a purely positional sacrifice, a quiet move without checks or obvious threats.[3]
- Inferred The threat being met is e5-e6, which hits f7 and d7 with the bishop on g4 behind it. The rook on e6 blocks the pawn and the file in one move.
This purely positional sacrifice (a quiet move, without any checks or obvious threats!) made an indelible impression on me.
- plan
- outpost
- Black's outposts
- pawn chain or line of force
White could have kept more. Instead of 26.a4, which let Black complete the regrouping with 26...Ne7, Stockfish prefers 26.Rg3 or an immediate 26.Bxe6: the edge was about a pawn then, and about a third of a pawn after the move played. The defender's rule is when you cannot stop the pawn with a piece, block it with the rook and take what the structure gives. The rook on e6 was a blocker first, and a sacrifice only if White chose to take it. Verified
Part 8A fresh angle: a rescue, not a masterpiece
The usual story presents 25...Re6 as a strategist's idea conceived in a balanced position. Stockfish tells a different story. Before the move Black was more than two pawns behind, and the engine puts 25...Re6 as the only good move. The sacrifice did not turn a level game into a good one; it saved a bad position, and White's play over the next two moves turned the rescue into a draw. Verified
| Move | What happened | Stockfish's view |
|---|---|---|
| 19...Rfe8 | Rook to the e-file instead of the knight to e5 | White's edge grows from about a fifth of a pawn to about a pawn and a half |
| 23...a5 | A slow queenside pawn move | The edge rises to about two and a third pawns |
| 25.Rfe1 | Rooks doubled instead of Bf3 | The edge falls to about one pawn |
| 26.a4 | A pawn move instead of Rg3 or Bxe6 | The edge falls to about a third of a pawn |
| 28.Qf1 | The queen retreats | Level |
The lesson for a club player is the other half of the story. Reshevsky had a clear advantage from move 19 to move 25 and gave it away with two natural-looking moves. The best sacrifice of the game was made by the side that was already worse, and it worked because the other side did not look for the strongest continuation. Compare the bad-bishop theme in Fischer's bishop against knight win, where the same logic of fixed pawns and a restricted piece decides a game.
Part 9How to use the exchange sacrifice yourself
A four-step recipe
- Find the pawnLook for an enemy pawn that is about to advance and gain space. If a rook can block its path and be defended once it is taken, the trade is available.
- Count the squaresAfter the trade, list the squares your pieces will own: an outpost for a knight, a diagonal for a bishop. If you cannot name two, do not play it.
- Look at their bishopCheck the opponent's remaining minor piece. If it stands behind its own pawns on the same colour, it is the piece that has lost the exchange.
- Make sure the position stays closedCheck that the trade does not open a file for the extra rook. Closed pawns are the reason the trade works.
- Does the sacrifice stop a pawn or block a file?
- Does the enemy bishop become a bad piece after the trade?
- Do my pieces gain a square nobody can challenge?
- Can the opponent open a file for the extra rook?
If you are on the other side of the board, do not take a sacrificed exchange on reflex. Ask what your opponent gets and whether a waiting move such as a rook lift keeps the tension. Petrosian's habit of asking what the opponent wants first is the subject of the Karpov and Unzicker prophylaxis study.
Your turnPractise it
Set each position up, decide your plan before you look, then play it out against Stockfish to see whether it holds.
Black to play after 19.fxe4: how do you challenge White's centre and use the d-file?
Hint
Look at the queen on d2 and the pawn on d4 in front of it.
A good answer
19...Ne5, Stockfish's first choice. White cannot take it because dxe5 would lose the queen to ...Qxd2, and the knight can go to d3.
White to play after 24...Rd8: how do you keep the advantage?
Hint
Where should the bishop on g4 stand so that a rook cannot come to e6?
A good answer
25.Bf3, Stockfish's first choice. The move played, 25.Rfe1, allows 25...Re6, while after 25.Bf3 the rook move would lose to 26.d5.
Black to play after 29.Rf3: how do you use the blockade?
Hint
Which square can no white pawn ever attack?
A good answer
29...Bd3, Stockfish's first choice. The bishop takes the outpost on d3 and the game stays level.
Part 10Takeaways
- A positional exchange sacrifice buys squares, not material. Count the squares before you give the rook.
- Block the pawn. The best moment to give the exchange is when a rook on a blocked file is the only way to stop an advancing pawn.
- Aim at the bishop. The trade pays when the opponent's remaining bishop is shut in by its own pawns.
- Defend with it. Petrosian's famous move was a rescue from a worse position, not an attacker's plan.
- Do not take on reflex. The side that accepts the exchange should ask what the sacrifice gains.
Questions readers ask
What is a positional exchange sacrifice in chess?
A positional exchange sacrifice is giving up a rook for a bishop or knight to gain lasting positional advantages instead of material: a blocked pawn, an outpost for a piece or a bad enemy bishop. It usually happens in closed positions.
When should you sacrifice the exchange?
Sacrifice the exchange when it stops a dangerous pawn, fixes the opponent's remaining bishop behind its own pawns and gives your pieces outposts that cannot be challenged. In open positions with files for the rook, it rarely works.
Was Petrosian's 25...Re6 against Reshevsky really a sacrifice?
It was an offer. White did not have to take the rook, but after 27.Bxe6 fxe6 Black had the better pieces for the exchange and the game was drawn. Stockfish rates 25...Re6 as Black's only good move, and White's 26.a4 gave more of the edge away than 27.Bxe6.
Is a rook worth more than a bishop or a knight?
On the usual scale a rook is worth five points and a minor piece three, so the exchange is worth about two pawns. In closed positions where the rook has no open file, the minor pieces can be worth more than the rook.
How do you play against an exchange sacrifice?
Do not take it on reflex. Ask what the sacrifice gains, check whether a waiting move such as a rook lift keeps the tension, and think about a plan that opens a file for the extra rook. In this game 26.Rg3 kept more of the advantage than the 26.a4 White played.
Sources and method
The game score was checked against two independent sources. Every position, line and candidate in this article was replayed for legality, and every engine number was produced by Stockfish 19 WASM Multithreaded for this article (depth 16 to 32, deepest at the critical moments), not copied from elsewhere. What the players thought is reconstructed from the moves, their published notes and later analysis; it is labelled as verified, inferred or speculative, and nothing is presented as a quotation unless a source is cited for it.
Read next
Kasparov's knight on d3 in game 16 of the 1985 match, what Stockfish says about it, and how to build an outpost no pawn can attack.
Bishop Pair Strategy: Kramnik v Tiviakov 2001, Checked by StockfishKramnik, the new world champion, turned the two bishops into a win at Wijk aan Zee 2001. Stockfish 19 shows what the pair was worth and where both players let the advantage slip.
King in the Centre: Morphy's Opera Game, Checked by StockfishMorphy's famous 1858 win against two amateurs, and what Stockfish 19 says about when the uncastled king was really lost.
Turn reading into skill
Strategy sticks when you use it. Practise real positions with Stockfish's verdict at once, or play the plans out from any diagram above.