Knight Endgames — How to Win the Tricky Ones

The Short Version

> Quick answer: Knight endgames behave a lot like pure pawn endgames — the side with an extra healthy pawn usually wins — but with one huge twist: a knight can't lose a tempo or triangulate, so zugzwang works differently. The trickiest pattern to know is that a knight plus a rook pawn (a- or h-pawn) on the 7th rank defended by the knight is a fortress draw against a lone king. Master king activity and outside passed pawns, then practice the conversions against the bot on CheckmateX.

Knight endgames have a reputation for being dull, and I get why — there's no flashy mating net, just slow, patient maneuvering. But they're some of the most practically useful endings to understand, because they come up all the time after a pair of minor pieces gets traded and the position simplifies. I've won and lost more rating points in knight endings than I'd like to admit, almost always because of one small technical detail I didn't know.

The good news is that knight endgames are governed by clear principles. Once you understand how the knight's quirks change the usual pawn-endgame logic, most positions become readable. I'll cover the big rule (they're basically pawn endings), the knight's one giant weakness (it can't switch flanks fast), and the two patterns that decide most close games: the rook-pawn fortress and converting an extra pawn.

Knight Endings Are Pawn Endings — Almost

The single most useful mental model: knight versus knight endgames play very much like king-and-pawn endgames. The side with an extra sound pawn usually wins, the same way it does without the knights on the board. King activity matters enormously. Passed pawns are gold. Pawn majorities convert into passers.

So your first job in any knight ending is to think like it's a pawn ending. Get your king active and centralized. Create or push a passed pawn. Trade into a winning pawn structure when you can. The knights add complications, but they don't overturn the fundamentals.

Where it gets interesting is the one place the analogy breaks. In a pure pawn ending, you can win key positions by triangulating — wasting a move with your king to put your opponent in zugzwang. With knights, that tool partly disappears, and that single fact decides a shocking number of endings.

If you're still shaky on the underlying pawn-ending logic, it's worth shoring that up first. I broke down opposition, the rule of the square, and zugzwang in chess endgame basics, and every one of those ideas carries straight into knight endings. You can't understand the knight version until the pawn version is automatic.

The Knight Can't Lose a Tempo — Why That Matters

Photo by Unsplash on Unsplash

Here's a quirk that took me embarrassingly long to internalize: a knight cannot lose a tempo. Every other piece can waste a move and return to the same square having burned a turn — a bishop shuffles on a diagonal, a rook slides back and forth, a king triangulates. A knight can't. From any square, a knight needs an even or odd number of moves to return, and it always changes the color of the square it stands on with each move. So it can never just "pass."

This means a knight can't triangulate to hand the opponent the zugzwang. In king-and-pawn endings, that's often how you win — you lose a tempo with your king and force your opponent into a losing move. Take the knights off the logic and you have to find a different way to make progress.

The practical upshot is that in a tight knight ending, you often want to maneuver your knight to a square where it both does its job AND happens to put the opponent in zugzwang on the move that arrives — you can't engineer the tempo separately, so you have to bake it into a useful move. It's fiddlier than triangulation and it's why these endings reward calculation over intuition.

The knight's other big weakness is range. A knight is short-legged — it takes several moves to cross the board. So a knight hates defending on two flanks at once. If you can create a passed pawn on one wing while the enemy knight is tied to the other, the knight often can't get back in time. Creating a second front, an outside passed pawn far from the action, is one of the most reliable winning plans in these endings.

The Rook-Pawn Fortress and Other Must-Know Draws

This is the pattern that's saved me half points and the one I most often see people not know. A knight plus a rook pawn — an a-pawn or h-pawn — that has reached the 7th rank and is defended by the knight is a fortress draw against a lone king. The defending king simply can't make progress; the pawn is stuck one square from promotion and the knight holds it. If you're a pawn down in a knight ending, steering toward this exact setup is often your get-out-of-jail card.

The flip side, when you're trying to win with knight versus a passed pawn alone: the corner pawn is the knight's worst enemy. A knight struggles to stop a rook pawn because it can't easily get in front of it. If the a-pawn is racing to promote, your knight needs to reach c8 or b5 in time; against an h-pawn, the key squares are f8 and g5. Miss those squares and the pawn queens.

The general rule with knight versus a passed pawn that hasn't reached the 7th rank: the knight can usually stop it alone if it gets to a controlling square in time, but rook pawns are the dangerous exception. Central pawns are far easier to blockade because the knight has squares on both sides.

When I'm grinding a knight ending I genuinely don't fully calculate, I lean on these reference points: extra pawn usually wins, activate the king, make an outside passer, and remember the rook-pawn-on-7th fortress as both a goal and a danger. The endgame theory here is well documented — the Wikipedia chess endgame overview is a solid free reference if you want to go deeper into the exact tablebase results. But honestly, the four ideas above cover the vast majority of what you'll face over the board.

Knight vs Bishop — Which Do You Want?

A question that comes up the moment you're choosing what to trade into: in the endgame, is the knight or the bishop better? The honest answer is "it depends on the pawns," and knowing when each shines has won me a lot of games before the endgame even started.

The bishop is the long-range sniper. In open positions with pawns on both wings, the bishop usually wins the argument because it can influence both sides of the board in a single move — exactly what the short-legged knight can't do. So if you have the bishop and pawns are spread across the board, trade pieces and steer toward an open ending.

The knight shines in closed positions and where the action is concentrated on one side. With pawns locked and the fight happening in one zone, the knight's ability to hop over blockades and land on outpost squares outweighs its lack of range. The knight also loves fixed pawn chains where it can sit on a square the enemy bishop can never attack. If your bishop is bad — hemmed in by its own pawns on its color — you'd often happily trade it for an active knight.

There's a structural rule worth burning into memory: in same-colored-bishop endings, the side with an extra pawn wins roughly nine times out of ten, while opposite-colored-bishop endings are drawish — nearly half of bishop-and-two-pawns-versus-bishop positions end in a draw when the bishops are on opposite colors. That's a giant practical signpost. If you're up a pawn, steer away from opposite-colored bishops; if you're down a pawn, steer toward them. Knight endings don't have a direct equivalent of the opposite-colored-bishop draw, which is part of why an extra pawn converts more reliably with knights than with opposite-colored bishops.

I sort these decisions out in my head one or two moves before the trade, not after. By the time the position simplifies, the structure is already locked, and you can't fix a bad minor piece once the pawns are fixed. A knight that lands on a protected outpost the enemy bishop can never challenge is worth more than its nominal three points, and a bishop hemmed in behind its own fixed pawns is worth less — so judge the pieces by the actual board, not by the textbook value. If you want to feel the difference, set up the same pawn structure with a knight versus a bishop and play both out on the CheckmateX puzzle trainer and against a sparring opponent — a few games teaches you more about piece value than any rule of thumb. And if you want the bishop side of this story, I broke down the long-range piece's strengths in bishop endgames — how to win with your bishop.