Figure 4.1.8.2[White to move]

Your train of thought this time might begin with the observation that Black has pinnable pieces on two axes: his knight on e7 and his bishop on f5. At first pinning the knight seems the more plausible idea because White can do it right away with Bd6, but then there is no decisive follow-up: the knight itself is guarded, so taking it won't be profitable. It might seem that White could take advantage of the knight’s paralysis by playing NxB with impunity (or by starting with BxN, inviting the recapture KxB—and then playing NxB). But remember that after White imposes the pin Black has a move to play, and he can spend it taking his bishop out of danger.

So now think about the other possible target for a pin: Black’s bishop on f5. At first no pin might seem possible because White’s rook is out of range and the bishop is guarded, but be more precise about the obstacles: the bishop has a guard in the e7 knight; and the White rook’s path to the pinning square (f3) is blocked by the knight on e3. When these problems are seen together their solutions suggest themselves. Again we vacate the obstructing piece violently with NxB. Black replies NxN, and now both problems are gone. The protected target has been replaced with a loose one (the knight), and the rook’s path to f3 is clear. Rf3 will win the knight a move later. Another way to see this idea is to start by experimenting with any exchanges you can force and their consequences. Here White can play NxB, inviting the reply NxN—a standard swap of minor pieces. What makes it interesting is the aftermath, as now the way is clear for a pin.