Essentially the same concept, but a move earlier. As yet there is no pin in place. White’s thinking might be stimulated by the presence of his bishop and rook on the same file leading toward e8—almost a mating pattern, but Black’s queen is in the way. Another approach to the position is to just consider every check and its consequences. White has Qa3+. To this suppose Black replies Qe7, interposing his queen and walking it into a pin. Now Black threatens QxQ, so what is White to do? Again, take the offensive by adding a vertical pin: here as in the previous case the rook is brought into the fray by simply moving the bishop on e4 out of the way, and here as before the best placement for the bishop is determined by the demands of the cross-pin White wants to achieve. The point is to imagine Black playing QxQ and to place the bishop so that it has the maximum possible usefulness afterwards. That means Bc6 is best, because then after Black if QxQ White will have RxB#, just as he had a similar mate in the previous position.

We have been assuming Black would interpose his queen in reply to White’s Qa3. Now we can see that would be a blunder; his better option would be to remove his king to g8. But then you should see another pattern: a classic discovered attack via Bxh7+, followed a move later by RxQ—the start of a winning sequence for White, though not so devastating as if Black had played Qe7. Hence the importance of seeing the kernel of the discovery on the e-file from the very beginning: however it gets used, the potential for it dominates all the variations that might follow from White’s check.