Now we extend the progress of our current logic a bit farther. Sometimes a piece can become overworked by defending no pieces—but two mating squares.
Examine the Black king’s position here and you see that White’s pawn on h6 seals off g7 as a flight square; Black’s king is stuck on the back rank. White can put a piece there with RxR, which would be mate were it not for the protection supplied to e8 by Black's queen and knight. Look for ways to disrupt the work of those pieces and you find QxN+. The move threatens Qg7#, which is what will happen if Black replies by moving his king to g8. Nor does Black have any interpositions to offer. The best he can do is capture White’s queen with his own; yet then e8 has been left loose, so White plays RxR+ and mates anyway a move later (Black only has a useless interposition with his bishop).
In a sense Black's queen was overworked in a familiar way: it guarded two Black pieces (his rook and knight). But the real significance of its defensive role involved not those pieces but the squares on which they sat. The position works the same way with the Black pieces gone from e8 and f6.
Now one way to blunder in this position is to start with RxR+, inviting QxR; you imagine then playing QxN+—not because it’s a queen fork that picks up Black’s bishop, but because it allows you to mate with Qg7 next move. But this fails because the initial move RxR+ is instead met by Black with NxR, where the knight still guards f6 from its new square, e8. The morals are several. First, the idea of capturing first with your less valuable piece obviously is less important when you're going for mate. Second, move order matters—but you knew that. Last, you want to carefully look at all the ways your opponent can recapture, not just the first that comes to mind.