How Koga decides
- Before our move, an enemy piece must already be pinned — frozen on a line between its king (absolute pin) or a more valuable piece (relative pin) and one of our long-range pieces.
- Because it cannot legally step off that line, it cannot run away when we come after it.
- Our move must bring a NEW attacker to bear on the pinned piece: the piece we move ends up attacking the pinned square, and it was not already attacking that square from where it started.
- Adding a fresh attacker to a piece that cannot be defended by moving wins material — the pinned piece is stuck and must eventually fall.
- Not a capture — simply taking the pinned piece is just a capture, not 'attacking' it; this motif is about adding pressure, not grabbing it outright.
- Not an old attacker — if the moving piece already hit the pinned square before it moved, nothing new has been added.
- The pin must survive — the pinner, the pinned piece, and the piece behind it must all still be on the board after the move, and the pin must still hold; if the move breaks the pin there is nothing to exploit.
- The pressure must pay — after the move the pinned piece is winnable (capturing it gains material), or the engine line goes on to win it. Piling onto a pinned piece the defenders still hold wins nothing: …Qf6 hits the pinned c3 pawn, but Bd2 and Qc2 hold it and the line takes f2 instead.
Test positions (the examples that guard against regressions)
Pawn attacks a relatively pinned knight
The rook on e1 pins the black knight on e5 to the queen on e7 (a relative pin). The push f2–f4 brings a new attacker onto the frozen knight, which cannot step away without losing the queen behind it.
Pawn attacks an absolutely pinned knight
A constructed position. The rook on d1 pins the knight on d5 to the king on d8 (an absolute pin). After e3–e4 the pawn attacks the d5 knight, which is glued to the d-file and cannot escape.
Capturing the pinned piece is NOT this motif
A constructed position shown JUST BEFORE the move. The rook on e1 pins the knight on e5 to the queen on e7. The bishop on f4 simply takes the knight with Bxe5 — that is an ordinary capture, not piling a new attacker onto a frozen piece, so the motif does not fire.
The attacker was already hitting the square
A constructed position. The rook on e1 pins the knight on e5 to the queen on e7, and the bishop on f4 already attacks e5. After Bf4–g3 the bishop still attacks e5, but no NEW attacker has been added, so the motif does not fire.
A move that ignores the pinned piece
The same position as the motivating example: the knight on e5 is pinned by the rook on e1 against the queen on e7. But the move a3–a4 does nothing to the e5 knight, so even though the pin exists, the motif does not fire.
A move that abandons the pin
A constructed position shown JUST BEFORE the move. The rook on f1 pins the knight on f4 to the queen on f7 along the f-file. The move Rf1–a1 leaves the f-file and breaks the pin, so there is no longer a pinned piece to attack.
Pressure the defenders hold
chess.com 184201652418 (Avo's report). After 18.bxc3, the rook on c8 pins the c3 pawn to the queen on c2, and ...Qf6 adds a new attacker. But Bd2 and Qc2 still hold c3, so every capture there loses material, and the engine line takes f2 instead. Avo: "Not a tactic at all."
The line wins the pinned piece
Lichess uhw9B. The queen on d1 pins the g1 knight to the king, and ...Ne2 attacks it. Nothing is won yet, because Qg3 still guards g1. But Ne2 also hits the queen: after Qg4 the knight on g1 falls (...Nxg1).