Rook and Pawn versus Rook
The most common endgame on the board — appearing in roughly one in eight master games — and the one whose theory is too rich to be a single position. A map of the cases that arise, the ones that win, and the ones that draw.
Roughly one master game in eight ends with a rook and a pawn against a rook, and roughly half of those games are drawn despite the material imbalance. There is no other endgame where the gap between “I am up a pawn” and “I will win this” is so wide. Rook endings are the technical centrepiece of every endgame book ever written, and the reason is not that rooks are interesting — it is that the rook-and-pawn-versus-rook position arises so often and resolves so unpredictably that the player who has not studied it will lose half-points every tournament.
Why R+P vs R is everywhere
Master games reach rook endings by attrition: queens come off through trades that favour the side ahead in development, minor pieces follow as the position simplifies, pawns get traded in fights for files and squares. By move 50, in many games, the remaining material is two rooks and a handful of pawns on each side. By move 70, after further pawn exchanges, the typical structure is one extra pawn for one side and nothing else.
The result is R+P vs R, and the result of that is where the half-points are won and lost.
When R+P vs R wins
The winning cases reduce, in the end, to one general principle: the attacker wins if her king can reach the queening square’s vicinity, and her rook can prevent the defender’s king from doing the same. The mechanical instantiation of that principle is the Lucena position, where the technique called “building the bridge” converts an extra pawn into a queen.
The conditions for a Lucena-style win:
The attacker’s pawn is on the seventh rank, one move from promotion.
The attacker’s king is on the eighth rank, blocking the pawn’s forward path.
The defender’s king is cut off — usually two files or more from the pawn — by the attacker’s rook.
The pawn is not a rook pawn (not on the a- or h-file).
If all four conditions hold, the attacker wins by the canonical bridge-building technique. If even one fails, the position needs separate analysis.
A second winning case worth knowing: the active king and rook against a passive defender. If the defender’s rook is tied to defending a weak pawn or blockading a passed pawn, the defender effectively plays without his rook. The attacker can convert by king activity even when the material situation looks drawn.
When R+P vs R draws
The drawing cases are the more numerous, and the Philidor position is the canonical example of how the defender holds.
Conditions for a Philidor-style draw:
The defender’s king sits on the queening square (or the file of the pawn) — not cut off.
The defender’s rook holds the third rank in front of the attacker’s pawn until the pawn advances.
When the pawn advances past its supporting king, the defender switches to back-rank checks.
The technique works for any pawn except (in some variants) the rook pawn, and is reliable enough that strong players reach it on autopilot.
The second major drawing case is the Vancura position against a rook pawn. When the attacker has an a- or h-pawn, the Philidor’s back-rank-check phase fails because the defender’s king has no room. The Vancura method — defender’s rook attacking the pawn from the side, defender’s king in the opposite corner — holds these positions despite the apparently winning material.
A third drawing case: perpetual checks against an exposed king. If the attacker tries to win without first getting his king to the eighth rank — that is, if he advances the pawn before the king — the defender’s rook gives endless checks from behind, and the attacker’s king has nowhere to hide.
What every player must know
The minimum knowledge required of any tournament player below master strength:
The Lucena position as the winning attacker technique.
The Philidor position as the canonical defensive technique.
The Vancura position for the rook-pawn exception.
The geometric idea of “cutting the king off” — the attacker’s rook on a file the defender’s king cannot cross, forcing the defender to play without his king’s support.
The activity heuristic: rooks belong behind passed pawns, both attacking and defending. The Tarrasch rule “rooks belong behind passed pawns” remains the most useful single piece of rook-endgame advice ever written.
These five concepts — three positions and two principles — give the player a chance in roughly 90% of practical R+P vs R endings. The remaining 10% need book-deep analysis and engine confirmation, but those are rare enough that the player who masters the basics will outperform her rating in rook endings for years before she runs out of cases.
The honest summary of rook endings as a study subject: the work pays back at the rate of about a hundred rating points per twenty hours invested, more reliably than any other endgame topic and probably more reliably than any opening study above the club level. Rook endings are where games are won and lost in the long, late, technical phase of competitive chess. Every hour spent on them is a half-point you will not lose three years from now.