Figure[White to move]

Finally, observe how similar principles can be used to win material even where the mating idea fails. Black’s own bishop on f8 and pawn on f7 prevent his king from retreating to the east, just as would be true if the king were at the edge of the board on h8. You see that your bishop cuts off the e7 and d8 squares as well. This might seem to call for a smothered mate, so naturally you imagine Nf6+ and Nc7+—and you see that neither of them work. Nc7 fails because Black guards the square with his queen.

The important question is how your train of thought runs from here. Don’t give up on observing that Black can extinguish the mate threat. Reflect instead on how urgent it is for Black to keep his queen trained on c7. This means the queen is vulnerable; it can’t afford to move much. So the next impulse should be to menace Black’s queen, and this you can do ostentatiously with 1. Qb5. Notice what a fix this creates for Black. If he plays QxQ, he gets mated by Nc7. Nor can he move his queen to safety on White’s side of the board. He has to keep the queen on a square from which it can protect c7, and there turns out to be no such square that is safe. His best play is 1. …Qd8; 2. BxQ, RxB, and White has won a queen for a bishop with another pawn capture or two still to come.

It all starts by seeing the idea of the smothered mate—and then not being deterred by the inability to make it work. Mating ideas that don’t work can be useful and powerful; the trick is to study the reasons they don’t work and ask whether you can exploit the vulnerabilities they create. Sometimes a piece that foils a mating threat is vulnerable in something like the way a pinned piece would be.