Bishop versus Knight
The minor-piece debate that has divided endgame writers for two centuries — when the bishop's range outperforms the knight's hops, when the knight's leap defeats the bishop's lines, and the practical heuristics that decide each case.
Whether the bishop or the knight is the stronger piece is the oldest question in chess endgame literature. Philidor argued for the knight; Lasker for the bishop. Capablanca said it depended on the position and gave a list of conditions. Modern engine analysis confirms Capablanca’s list and adds a few. The honest answer, the one every working player needs at the board, is that the two pieces are different rather than unequal — each dominates the other in a specific kind of position, and the player who recognises the position quickly is the player who wins the minor-piece ending.
The minor-piece question
The mechanical difference between a bishop and a knight is well known. The bishop moves diagonally, any distance, on a single colour. The knight moves in an L-shape, two squares one way and one the other, but reaches any square eventually. The bishop sees longer, but only half the board; the knight sees less but every square.
The positional consequences fall out of those mechanics:
The bishop gains in open positions where its long lines have nothing in the way. It loses in closed positions where its diagonals are blocked.
The knight gains in closed positions where it can hop over the blockading pieces. It loses in open positions where the bishop’s range outpaces it.
The bishop is faster across the board from one wing to the other. It gains when there are pawns on both flanks.
The knight is faster at short range and excels at outposts — squares from which it cannot be evicted. It gains when there are pawns on one flank and the position is condensed.
These four rules cover most positions. The remaining detail is in the conditions for each, which we turn to next.
When the bishop is better
The classical statement of the bishop’s case comes from Steinitz: the bishop is to be preferred in endings where there are pawns on both wings. The reason is the bishop’s ability to attack pawns on either side without travelling — a single bishop move from b2 to h8 controls squares on both wings simultaneously; the knight needs four moves to make the same transit.
In endings where the attacker has a passed pawn on one flank and the defender’s king is on the other, the bishop dominates: it stops the defender’s counter-passed pawn while the attacker’s king walks across the board to support the bishop’s pawn. The knight in the same position is overworked — it cannot watch both flanks at once.
Open central pawn structures favour the bishop. So do bishop-pair situations against a single knight, where the bishops cover both colours and the knight cannot find a safe outpost. Capablanca’s win against Janowski in 1916 is the canonical example: a clean two-bishop versus bishop-and-knight ending where the bishop pair’s coverage of both colours simply ran the knight out of squares.
When the knight is better
The knight’s case is the reverse. Closed positions, where the centre pawns interlock and the bishop’s diagonals are blocked, leave the knight in command. The knight hops over the obstructions and finds outposts in the holes left by pawn breaks.
The single most decisive structural feature is the outpost — a square deep in the opponent’s position, defended by a pawn, where the knight cannot be challenged by a pawn or driven away by a bishop of the wrong colour. A knight on the fifth or sixth rank on a strong outpost is worth a small piece more than its nominal material value; the position is, for practical purposes, a piece up for the side with the outpost.
The Nimzo-Indian and the Stonewall structures classically end in knight victories. The same is true of any opening where Black accepts doubled or backward pawns in exchange for outpost squares — the knight finds the squares, the bishop has nothing to do. In games where one side plays for a closed centre against a fianchetto bishop, the knight usually outscores its opponent by a wide margin in the resulting endings.
Practical rules of thumb
The working player carries three rules from study to game.
Capablanca’s exchange rule. If you are worse, trade your bishop for the opponent’s knight. If you are better, trade your knight for the opponent’s bishop. Capablanca’s premise was that the bishop is generally the stronger piece in the abstract; he refined this to “the stronger side wants to keep his bishop, the weaker side wants to remove it.” Modern engines disagree on the abstract claim but agree on the practical consequence in most positions.
Pawn-flank count. Count the flanks. Both flanks have pawns? Bishop. One flank? Knight. This single heuristic gets most endings right at first glance and is a useful tiebreaker between candidate moves earlier in the game.
The “good” and “bad” bishop distinction. A bishop is “good” if the player’s own pawns are on the opposite colour to the bishop — leaving the bishop room to operate. It is “bad” if his own pawns are on the same colour as the bishop — blocking the bishop in. A “bad” bishop is often worth less than a knight in the resulting ending; a “good” bishop is usually worth more. This explains many of the apparent exceptions to the flank-count rule.
The minor-piece ending is the one place in chess where the side ahead in material can lose because the side behind has the right piece. A pawn or even two extra pawns will not save a player whose minor piece is wrong for the position. The fastest way to improve endgame results is to learn which piece you would rather have in each kind of position, and to steer trades, even at the cost of a tempo or a pawn, toward the structure where your piece outclasses your opponent’s.