Written by GM Mauricio Flores RiosGrandmaster and author of Chess Structures: A Grandmaster Guide

Rook and Pawn vs Rook

Rook and pawn versus rook is the most common concrete endgame in chess, and its result is decided by one question above all others: where is the defending king? If it stands in front of the pawn, the defender draws with known technique; if it is cut off from the pawn, the attacker wins with known technique; and in the territory between those poles, a small family of named defenses decides everything. This page is the map of that territory — every boundary on it has been verified against the endgame tablebase.

The Two Poles: Lucena and Philidor

Every rook and pawn versus rook position is drifting toward one of two destinations, and strong players evaluate these endgames quickly because they navigate by those poles rather than by calculation.

The winning pole is the Lucena position: pawn on the seventh, attacker's king in front of it on the promotion square, defender's king cut off. The attacker wins by building a bridge — rook to the fourth rank, king out of the pawn's way, rook interposes against the checks. If you can reach it, you win; the full method lives on the Lucena page.

The drawing pole is the Philidor position: the defending king stands on the promotion square and the defending rook patrols its third rank, refusing the attacking king entry. When the pawn finally advances to the sixth, the rook drops to the first rank and checks from behind forever. Here is the mechanism in motion:

Philidor defense: third rank, then checks from behind
Philidor defense: third rank, then checks from behind

White's only real try is 1.e6, giving up the shelter square in front of the pawn — and Black instantly switches plans with 1...Ra1!. After 2.Kf6 Rf1+ 3.Ke5 Re1+ 4.Kd6 Rd1+ the king has nowhere to hide: its own pawn occupies the one square that would block the checks. The tablebase confirms every move of this line is a dead draw. Notice the timing — the rook stays on the sixth rank until the pawn arrives there, and not one move longer.

The Deciding Question: How Wide Is the Cut-Off?

When the defending king cannot reach the pawn's path, the attacker's rook cuts it off along a file, and the win becomes a matter of geography. The margin is startlingly precise. Take a white e-pawn still back on e4, with White's king beside it. Cut the black king off by one file, and the position is a tablebase draw:

King cut off by one file: e-pawn on e4 is only a draw
King cut off by one file: e-pawn on e4 is only a draw

Black's king on g7 is barred from the f-file, yet the defense holds: the rook on a8 harasses from behind and from the side, and one file of cut-off does not buy White enough time to escort the pawn up the board. Now move the barrier one file over — White's rook cuts on the g-file, the black king sits on h7 — and the same pawn wins:

King cut off by two files: the same pawn wins
King cut off by two files: the same pawn wins

1.e5! and the king simply walks through the checks: 1...Ra4+ 2.Kd5 Ra5+ 3.Kd6 Ra6+ 4.Ke7 Ra7+ 5.Kf6, using the pawn as cover, and the checks run dry with the rook driven to passivity. The Lucena is coming and Black cannot prevent it. One file of cut-off, draw; two files, win — that is how thin the margins in this endgame are, and why fighting over the cut-off file in the middlegame-to-endgame transition is worth real material.

As the pawn advances, the requirement eases — by the time the pawn stands on the seventh with its king in front, the Lucena wins with the single-file barrier of its classic diagram. But the defender's resources in the middle ranks are stubborn: even with the pawn on the fifth, a one-file cut can still be held. The practical rule I give students is simply: when in doubt, cut wider — an extra file of fence never hurt anyone.

When the King Is Driven Away: The Short-Side Defense

Between the poles lives the most important modern defensive idea. Suppose the defender's king has been pushed off the promotion square but not fully cut off — say White has an f-pawn on the sixth. The defender's king must flee to the short side (for an f-pawn, the g- and h-files) so that the rook gets the long side and its full checking range. Played correctly, this holds a tablebase draw; played with the king on the long side, the identical structure loses because the rook is left checking from a file or two away, where the attacking king simply walks up and attacks it. The complete method — including the beautiful detail of why the promotion square stays covered — is on the short-side defense page, and the underlying geometry is quantified on the checking distance page: a rook needs roughly three empty squares between itself and the enemy king to check forever.

The Special Pawns: Knight and Rook Pawns

The map changes at the edge of the board, always in the defender's favor.

Against a knight pawn or rook pawn, the humble back-rank defense — king in front, rook parked passively on the back rank — holds the draw. The tablebase confirms it: with a white pawn on b6 or a6 and the defending king on the queening square, the passive rook draws. Against a bishop pawn or central pawn the very same setup loses, because the attacker has room to outflank on both sides. That boundary is the back-rank page's whole subject.

Against a rook pawn the defender has even more resources: the attacking king gets trapped in front of its own pawn in the corner, and a rook on the sixth rank produces the celebrated Vancura position, holding from the side. Rook pawns are so much worse for the attacker that they get their own page: rook pawn endings. No Lucena exists with an a- or h-pawn; that is the single biggest exception in this material balance.

How to Use the Map at the Board

As the attacker: keep the defending king cut off, as wide as possible, before touching the pawn. Escort the pawn with your king, not your rook — the Tarrasch rule puts your rook behind the pawn where every advance gains it scope. Do not rush the pawn to the seventh until the Lucena conditions are actually assembled. And if you can choose which pawn survives into the ending, choose a central or bishop pawn over a knight pawn, and anything over a rook pawn.

As the defender: the king sprints for the queening square first; reaching it means Philidor and a clean draw. If you cannot get in front, go short side with the king and long side with the rook, and never let your checking distance shrink. Passive defense on the back rank is a real tool — but only against knight and rook pawns; against the rest it is a slow loss, so stay active instead. Every tempo the attacker spends repositioning is a tempo you spend improving your king.

Half of all rook endgames with more pawns eventually simplify into this one, which is why it repays study better than any opening line. Master the two poles and the three defenses between them and you will never again sit at the board wondering whether to offer the rook trade.

Frequently Asked Questions

Is rook and pawn versus rook always a draw?

No. Roughly speaking it is a draw when the defending king reaches the pawn's path and a win when that king is cut off, with precise boundaries in between. Both results are common at every level, which is exactly why the theory matters.

What is the fastest way to know if the position is won?

Ask two questions: is the defending king in front of the pawn or on the short side with an active rook, and how many files separate it from the pawn if not. King in front means draw with correct play; a wide cut-off means the attacker wins by marching toward a Lucena.

Which pawn is best for the attacker?

Central and bishop pawns win most often, because the attacker can outflank on either side and every defense is at full stretch. Knight pawns allow the passive back-rank draw, and rook pawns are the worst of all — no Lucena win even exists with them.

Why is the cut-off measured in files?

Because the defender's king needs to cross those files to reach the pawn, and each file costs at least one tempo under rook harassment. The tablebase boundary for an e-pawn on the fourth rank sits at exactly two files: one file of cut-off draws, two files win.

Should the defender check from behind or from the side?

From wherever the checking distance is longest. Behind the pawn is ideal when the attacking king has no shelter; side checks from the long side are the tool when the pawn is far advanced. Checks from two squares away lose because the king simply attacks the rook.

When should I trade into rook and pawn versus rook?

Trade down when the map says the resulting position is drawn — for example, your king controls the queening square or the pawn is a rook pawn. Avoid the trade when your king would be cut off, even by a single file, since the margins are one tempo wide.