Apply the current logic to this position. White has two pieces trained on the general territory of Black’s king: his queen on b2 and his rook on d6. Indeed, they both are aimed at f6; White almost is able to play Qxf6+ and then perhaps mate soon afterwards. The obstacles to this are two. Black’s bishop is in the way on e6, and Black’s knight on d5 defends the wanted square, f6. The payoff of seeing these obstacles is not so much in efforts to actually execute the mating possibility just sketched; it is in the vulnerability of the knight and bishop that their defensive roles suggest—especially because the bishop protects the knight. Maybe the knight can be taken with BxN, since Black dares not recapture BxB; with the knight then off the board and the bishop off the sixth rank, White would be able to play the dreaded Qxf6+. But first account for all the knight’s defenders. There appears to be another in the pawn on c6, yet on inspection it, too, turns out to be pinned—to Black’s queen. So while Black’s knight appears to be defended twice, as a practical matter it is defended not at all. Both of its defenders are pinned by the rook on d6. White is free to play BxN.

If Black replies to BxN with c6xB, White wins a queen for a rook with 2. RxQ, RxR. But it must be said that if Black were to instead reply BxB and White then played Qxf6+, careful play by Black can avoid mate, though he nevertheless suffers a material calamity. This is important to emphasize because before going through with a sacrifice on the assumption that an enemy piece is pinned to a “mating square,” you need to make sure it really is a mating square or that you otherwise will have enough gains to offset your losses when the smoke clears. Obviously not every position where you can land a protected queen next to the king will yield a mate. Here Black’s king would be able to retreat to h6. White then has QxR, which is profit enough to justify the sequence. Or he can play Qf8+ and continue to try hunting down Black’s king with checks and captures.