When straight opposition stops working, every square on the board has one right answer. Learn to read the map that decides blocked king-and-pawn endings.
✓ Interactive boards ✓ Step-by-step ✓ Free forever
Two squares correspond when a king arriving on one of them forces the enemy king to occupy the other — or lose the position. That is the whole idea, and it is bigger than it sounds: corresponding squares are the general theory behind the opposition. In a blocked king-and-pawn endgame every useful square for the attacking king has exactly one right answer for the defender, and the board becomes a map of paired squares. Read the map correctly and you know the result — won or drawn — before either king takes a step. The method matters precisely where the mechanical rules stop working: when the pawns are frozen, the entry squares are scattered across two wings, and 'take the opposition' points the defending king to a square that loses. The attacker reads the map to find the one square the defender cannot answer in time; the defender reads it to keep stepping on the right square, move after move, forever. Opposition, key squares, triangulation, zugzwang — every one of them is a chapter of this single map, and this page builds it from the simplest pair of squares up to a full network.
Correspondence is never abstract — it is always about concrete target squares. In a blocked pawn ending the attacking king wants to reach the handful of squares from which it wins a pawn or escorts its own pawn home: the key squares of a passed pawn, or the squares beside a frozen enemy pawn. With pawns locked on f5 and f6, for example, the white king's entry squares are e6 and g6 — the black king is barred from both by the white pawn, so they are doors that only open one way. Before you can pair a single square, you must list these entry points. Everything else in the method is calculated backwards from them.
Two squares correspond through obligation: when the attacking king stands on its square, the defender must be standing on the partner square, or the position collapses. The purest example is the opposition — king on e5 with a pawn behind it, defending king on e7. Whoever must move gives way: the pair e5 and e7 answer each other, and having the move is the only thing that separates a draw from a loss. A correspondence chart is nothing more than this obligation repeated: for every square the attacker can probe, the defender has one post that meets every threat from it.
It is not enough that each attacking square has an answer — the defender must be able to travel between his answers in step with the attacker. When the attacking king slides from one probing square to a neighbouring one, the two answering squares must also be neighbours. This is where defences break: two duties can meet on a single hub square that has to cover both entry points at once, or the defender's answering squares fail to touch each other while the attacker's squares form a connected triangle. When the routes do not match, the attacker manoeuvres until the defender runs out of correct squares — and zugzwang does the rest.
King and pawn against king: the squares e5 and e7 answer each other. With White to move it is a draw — the advance squares d6 and f6 are covered, and if White steps sideways, Black mirrors. With Black to move, Black must leave e7, the white king walks past the pawn, and the pawn queens. One pair of squares, and the whole result turns on who must move.
The key squares of the e4-pawn are d6, e6 and f6: if the white king reaches any of them, the pawn promotes no matter whose turn it is. Here the king already stands on e6, so White wins with either side to move — no opposition, no tricks, nothing to calculate. Correspondence theory starts from squares like these: the defender's entire job is to keep the enemy king off them, and every corresponding square is deduced backwards from that fight.
Blocked pawns on f5 and f6, material dead level. White's king may enter at e6 or g6 — Black is barred from both by the white pawn — and Black's king patrols the doors from the hub square f7. With best play this is a draw whoever moves, but the defence hangs by a thread: when the white king reaches h5, only the g7-square answers it, and the plausible retreat toward the centre loses the f6-pawn to an entry on the wing. No opposition rule names these squares; the correspondence map does.
White to move wins — but not by force of entry. The black king covers the entry squares, and the direct pawn push is even a dead draw after the king simply captures on c7. The winning method is the tempo dance: White's king walks the triangle of d5, d4 and c4 and arrives back on d5 with Black to move. Black's answering squares around c8 and d8 do not connect the way White's triangle does, so the same position returns with the defender on turn — zugzwang — and the white king finally enters on b6 or d6 to decide the game.
Test yourself with these engine-checked positions
King and pawn against king: White king on e5, pawn on e4, Black king on e7. It is White to move. Can White make progress?
Blocked pawns: White pawn d5, Black pawn d6. The white king has arrived on b5, and Black has just met it by putting the king on d7, directly covering the entry squares c6 and e6. Was that the right square?
White pawns on a5 and c6, Black pawn on a6, kings on c5 and c7 — and this time it is Black to move. White's triangulation has just returned the position with the defender on turn. Assess.
Solve these positions to test your understanding
White to move and win. The pawn is itching to run — but only one plan works.
Black to move and hold the draw. Material is level, but White's king on the wing eyes both entry squares.
White to move and win. Head-on there is no way in: the black king covers both entry squares. Find the manoeuvre.
These openings' structures funnel into blocked pawn endings
The French locks the central pawns early — chains with pawns fixed on e5 against e6 and d4 against d5 survive mass exchanges remarkably often. When a French middlegame simplifies all the way down, you inherit precisely the terrain of this page: frozen central pawns, kings manoeuvring for one or two entry squares, and a result decided by who knows the corresponding squares. If you play the French from either side, blocked king-and-pawn endings are not a rarity but a structural destiny worth preparing for.
View opening pageThe Caro-Kann's calling card is a sound, symmetrical pawn structure that survives into deep endgames. Games in these lines frequently trade to level king-and-pawn endings with one blocked pair in the centre — positions that look trivially drawn and are anything but. A single tempo, or a defending king that steps to the plausible square instead of the corresponding one, flips the result. The correspondence method is the difference between holding these endings on understanding and holding them on luck.
View opening pageQueen's Gambit structures — the Carlsbad above all — revolve around fixed central pawns and pawn majorities on opposite wings. After heavy trades the endgames turn into king geometry: whose king reaches the blocked centre's entry squares first, and whether the defender can answer probes on both wings at once. Because the pawn skeleton is often locked by move fifteen, these are the openings where a middlegame decision quietly commits you to a corresponding-squares ending forty moves later.
View opening pagePitfalls to avoid
When the attacking king approaches, the instinct is to plant your king directly in front of the pawns, covering the entry squares you can see. In a real correspondence network that square is often wrong: the attacker is not planning to walk through the front door but around the wing, and the chart usually names a mirror square — sometimes a full file away from the pawns — as the only answer. Defenders lose these endings not by miscalculating but by never asking which square corresponds; the plausible post and the correct post are different squares, and only one of them draws.
A pawn move feels like a free way to change whose turn it is — and in a corresponding-squares position it is almost always self-destruction. Every pawn advance redraws the entire map: new entry squares appear, old pairs dissolve, and a reserve tempo that might have won the ending later is gone forever. The classic version is the attacker pushing a protected passed pawn one square too early, letting the defending king capture or blockade it and turn a mapped win into an instant draw. In these endings tempo battles belong to the kings; the pawns are the terrain, not the army.
Opposition is only the first and simplest pair of corresponding squares, and it is reliable exactly when the position is symmetrical and the entry squares sit in a straight line. Near a board edge, or with entry squares spread across two wings, the true corresponding square is frequently not the opposition square at all — mirroring the attacker's king can walk you straight into an outflanking. Players rated well above club level lose drawn endings this way: they play the mechanically correct opposition move and discover the attacker entering on the wing the mechanical rule ignored. When pawns are blocked, calculate the pairs; recite the rule only afterwards.
Learn the opposition first, then relearn it as a correspondence of one pair: the king on e5 forces the defender to e7. Every harder position on this page is that idea repeated.
Memorise the key squares of a passed pawn — two ranks ahead of it, three squares wide — and a king on the sixth rank ahead of its pawn wins automatically, rook pawns excepted. Corresponding squares are always calculated backwards from targets like these.
As the defender, hunt for the hub: the single square that watches both entry squares at once. If such a square exists and you can keep returning to it in time, the ending is usually a draw.
As the attacker, probe the wing where the defender has less room. Correspondence breaks near board edges, because the defender runs out of answering squares before you run out of probing squares.
Never move a pawn while the kings are still dancing. Pawn moves redraw the whole map and cannot be taken back — save them as reserve tempi for the moment the king battle is already decided.
Do not trust a quick engine glance in blocked pawn endings: shallow evaluations wobble between level and winning in these positions. Mate scores and tablebase-style certainty exist here — check the position properly or work out the squares yourself.
Everything you need to know about corresponding squares
Corresponding squares are pairs of squares in king-and-pawn endgames linked by obligation: when the attacking king steps onto one square of a pair, the defending king must occupy the partner square, or the position is lost. In blocked pawn endings every relevant square has such a partner, and mapping the pairs tells you the result in advance. The opposition is the simplest case — one pair of squares directly facing each other — and the corresponding-squares method is what you use when the position is too lopsided for the opposition rule to give the right answer.
The opposition is one special correspondence: kings facing each other with an odd number of squares between them, where whoever must move gives way. It works when the battlefield is symmetrical. Corresponding squares generalise the idea to any blocked position: the defender's correct square might be a diagonal mirror, a knight's-move away, or on the far side of the pawns — wherever the geometry of the entry squares dictates. Every opposition is a correspondence, but most correspondences are not oppositions, and the positions that decide games are usually the second kind.
Key squares are the squares that decide a pawn's fate: if the attacking king reaches any of them, the pawn promotes by force regardless of whose turn it is. For a pawn that has not crossed the middle of the board, the key squares lie two ranks ahead of it, one file to each side and straight ahead; once the pawn passes the middle, the three squares directly in front of it join in, and a king on the sixth rank ahead of its pawn always wins — the rook pawn is the one exception, because the defender can hold the corner. Key squares matter here because they are the targets the whole correspondence network is built around — the defender's squares are defined by which key squares they deny.
Triangulation is the attacker's way of cheating the map. If your king has three connected squares that all keep the position intact, while the defender's answering squares do not form a matching triangle, you can walk your king around the three squares and arrive back where you started — same position, other side to move. The defender, unable to copy the tempo loss, must now step off the one correct square, and zugzwang forces the entry. Triangulation and zugzwang are both covered in depth in their own Kingsights concept guides; corresponding squares are the theory that explains exactly when the trick works and when the defender can mirror it.
Yes. Kingsights reviews your games and flags the king-and-pawn endings where the result turned on a single square — drawn positions given away by a king stepping off the corresponding square, and wins missed because the pawn moved before the king claimed its key square. If simplifying into pawn endings you then misplay is a recurring habit in your games, Kingsights will surface it as a pattern, not a one-off. Enter your Chess.com username above to find out.
Kingsights scans your real games for king-and-pawn endings where one square — held or missed — decided the result.
✓ Interactive boards ✓ Step-by-step ✓ Free forever