Behind your own passer to push it, behind the enemy's to freeze it: the one guideline that decides most rook endgames.
✓ Interactive boards ✓ Step-by-step ✓ Free forever
Rooks belong behind passed pawns — behind your own to support their advance, and behind the enemy's to restrain them. That two-part maxim is the Tarrasch rule, the single most useful guideline in rook endgames, and the formulation appears in Siegbert Tarrasch's book on the St. Petersburg 1914 tournament. The logic is beautifully symmetric. A rook behind its own passed pawn gains scope with every step the pawn takes: the file behind the pawn keeps getting longer, and the rook never has to move out of the way for the pawn to promote. A rook in front of its own pawn suffers the exact opposite fate — its corridor shrinks square by square, it physically blocks the promotion, and it ends up entombed on the promotion square doing guard duty. The same geometry works in defence: a rook behind an enemy passed pawn freezes it, because the pawn cannot safely advance unless the opponent escorts it every single step. Like every rule of thumb, it has known exceptions — but you should know the rule cold before you learn when to break it. (Not to be confused with the Tarrasch Defense, an opening in the Queen's Gambit.)
When your rook stands behind your own passed pawn, everything improves as the pawn advances. The rook defends the pawn on every square of its journey, so the pawn can creep forward without ever hanging. The file behind the pawn grows longer with each push, so the rook's scope increases rather than shrinks. And crucially, the rook never obstructs the promotion square — when the pawn is ready to queen, the rook does not have to lose a tempo stepping aside. The pawn and rook form a self-sufficient unit, freeing your king and other forces for duty elsewhere.
The defensive half of the rule is just as important. Park your rook behind an enemy passed pawn and the pawn is paralysed: every square it advances to is attacked from behind, so it can only move when the opponent's rook or king babysits it the whole way. Meanwhile your rook keeps its full freedom — it attacks the pawn along the file but can leave for one move to give a check or grab material, then return. The opponent's pieces, not yours, are chained to the pawn. A frozen passer often turns from an asset into a long-term liability.
Why is the front of the pawn the wrong side? Tempo and activity. A rook in front of its own passer must eventually abandon the file to let the pawn promote, losing time at the critical moment. Worse, as the pawn advances, the rook is squeezed into fewer and fewer squares until it sits on the promotion square itself, where its only job is to defend the pawn — a five-point piece reduced to a bodyguard. A defending rook that merely blockades in front is a little better, but it, too, is pinned to one square while the stronger side improves king and pawns at leisure. In rook endgames, the passive side usually loses the manoeuvring battle.
The ideal setup. White's rook on a1 defends the a-pawn on every square of its march and never has to move for the pawn to promote — its file grows longer with each push. Black's rook, blockading in front on a8, is condemned to guard duty: every advance of the pawn shrinks its scope. White's king is free to go wherever it pleases.
The same material, the roles reversed — and White is struggling to win at all. White's rook stands in front of its own passer on a8, where it blocks the promotion square and has almost no moves along the file. To let the pawn queen it must eventually step aside, losing a decisive tempo, and Black's active rook is ready to slip behind the pawn and freeze the whole operation.
The defensive half of the rule. Black's rook sits behind White's passed pawn and attacks it from a file that only gets longer: the pawn cannot take a single step without being lost, so White's rook is chained to defending it from the side. Black's rook keeps full freedom — it can leave for one move to give a check and return — while White's pieces do the babysitting.
The rule works before a passed pawn even exists. White's queenside majority will produce a passed b-pawn, and the rook on b1 is already behind the file where that passer will be born. When the pawns roll forward, the new passed pawn is supported from its first step — while Black's rook watches from the wrong side of the board. Place the rook behind the majority, then advance.
Test yourself with these positions
A pure rook endgame. White has a passed a-pawn on the fifth rank, the black rook already blockades on the a8-square, and White's rook on d1 must pick a post. It is White to move. Where does the rook belong?
White is a pawn up with a passed b-pawn on the fifth rank — but Black's rook has already taken up residence behind it on the b1-square, and White's rook defends from the front on the b8-square. It is White to move. How should this ending be judged?
White's rook sits in front of a passed pawn on the a6-square, and Black's rook is NOT behind the pawn — it attacks it from the side along the sixth rank. It is White to move. Is Black's unorthodox setup a mistake?
Solve these positions to test your understanding
White to move. The passed a-pawn is under attack from the black rook in front of it. Find the one square from which your rook supports the pawn forever.
Black to move, defending against a dangerous passed pawn already on the sixth rank. One move neutralises it — find the square that freezes the pawn for good.
These openings frequently lead to rook endgames where the rule decides
The Grünfeld's central trade — Black exchanges the d-pawn and strikes at White's big centre — routinely leaves Black with a queenside pawn majority. Deep into the endgame that majority becomes an outside passed a- or b-pawn, and whole Grünfeld endings are decided by which rook reaches the file behind it first. If you play this opening from either side, the Tarrasch rule is not optional theory; it is the standard operating procedure of your typical rook endgame.
View opening pageCatalan middlegames are famous for long-term queenside pressure and early queen or rook trades, so simplified positions with a distant passed pawn on the a- or b-file appear constantly. The side that remembers to put a rook behind the passer — before it even becomes a passer — usually converts or holds; the side that defends it from the front slowly runs out of moves. Placing a rook behind your queenside majority is one of the most thematic Catalan endgame plans.
View opening pageQueen's Gambit structures head for rook endgames more often than almost any other opening: heavy pieces come off along the open c-file, and the pawn majorities on opposite wings produce a passed pawn for one side. Whether you are nursing an extra queenside pawn or defending against one, the same question decides the game — whose rook is behind the passer? Training the Tarrasch rule pays off directly in the endings this opening produces.
View opening pagePitfalls to avoid
It feels safe to lead the pawn with the rook — the rook clears the path and shields the pawn from blockaders. But the geometry is against you: as the pawn advances, the rook's corridor shrinks, until the rook sits on the promotion square doing nothing but guard duty. Then comes the painful moment when the rook must step aside to let the pawn queen, losing a decisive tempo, often into a check or a capture. Meanwhile the defending rook slips behind the pawn and freezes the whole operation. If you notice your rook standing in front of your own passer, spend a move rerouting it — the earlier the cheaper.
The rule fails most often for the defender when the attacking king can shelter beside its own pawn. Sitting behind the passer, your rook has a limited number of checks from behind — and when they run out, the pawn walks in. In those positions the saving setup is lateral: attack the pawn from the side, keep at least three files of checking distance, and harass the king every time it approaches. Players who apply the Tarrasch rule mechanically choose the behind-the-pawn square, run out of checks, and lose an ending that a side attack would have held. Before parking the rook behind an advanced passer, count the checks: if the enemy king has shelter, restrain the pawn from the flank instead.
Memorise both halves of the maxim: behind your OWN passed pawn to push it, behind the ENEMY's passed pawn to freeze it. Most players remember only the attacking half.
Think of it as scope: a rook behind a pawn gains squares as the pawn advances, a rook in front of it loses squares. When in doubt, choose the placement whose future is growing, not shrinking.
The move to get behind the pawn is usually quiet and easy to postpone — a rook sliding along your first rank. Play it early; one tempo often decides who claims the file.
As the defender, freezing the enemy passer from behind keeps your rook free: it can leave for one move to give a check and then come straight back. A rook blockading in front has no such freedom.
In the middlegame, apply the rule before the passed pawn exists: put a rook behind your pawn majority, on the file where the passer will appear, and the future pawn is supported from birth.
Learn the exceptions after the rule: a king blockade favours side defence, a pawn that has not crossed the middle of the board tolerates a rook in front, and a defender may need lateral checking distance instead of the square behind the pawn.
Everything you need to know about the Tarrasch rule
The Tarrasch rule states that rooks belong behind passed pawns: behind your own passed pawn to support its advance, and behind an enemy passed pawn to hold it back. The formulation appears in Siegbert Tarrasch's book on the St. Petersburg 1914 tournament, and it remains the most quoted guideline in rook endgame play. The point is activity — a rook behind a passer gains scope as the pawn advances and never blocks the promotion, while a rook in front of the pawn becomes progressively more passive.
Because the two placements age in opposite directions. Behind the pawn, the rook defends it on every square of its journey, the file behind the pawn keeps getting longer, and the rook never has to move for the pawn to queen. In front of the pawn, the rook's corridor shrinks with every advance, it physically occupies the pawn's path, and at the decisive moment it must lose a tempo stepping off the promotion square. The same logic serves the defender: from behind, an enemy passer is frozen — it can only advance escorted — while your rook keeps its full freedom along the file.
Three classic cases. First, when the pawn is securely blockaded by the opposing king, protecting your pawn from the side is often stronger than from behind, because the rook also needs lateral flexibility. Second, when the pawn is still on its own half of the board, a rook in front of it is no tragedy — promotion is far away and the rook keeps other duties. Third, as a defender against an advanced passer escorted by the enemy king, lateral checking distance can outrank the file behind the pawn: attack the pawn from the side and check the king whenever it approaches. A saying often attributed to Tarrasch acknowledges the point with a wink — always put the rook behind the pawn, except when it is incorrect to do so.
Yes — and this is the part most club players miss. The rule guides rook placement before a passed pawn even exists. If you have a queenside pawn majority, the file where your future passer will appear is known in advance, so a rook placed behind that majority supports the passed pawn from the moment of its birth. The same holds defensively: a rook posted behind the file of the opponent's majority is already on duty when their passer arrives. Applying the rule one phase early is a cheap way to win the endgame during the middlegame.
Yes. Kingsights reviews your games and flags the recurring rook-endgame habits the Tarrasch rule is about — rooks left in front of your own passers, enemy passers you never froze from behind, and majority endgames where the rook arrived on the key file a tempo too late. If passive rook placement is costing you endings, the pattern will show up across your games. Enter your Chess.com username above to find out.
These openings frequently feature this concept
Kingsights scans your real games to find the rook endgames where the Tarrasch rule won — or would have won — the game.
✓ Interactive boards ✓ Step-by-step ✓ Free forever