I like the idea that expectation hurts a jumpscare, but I don’t think expecting a scare = automatically ineffective.
Anticipation can increase stress. That’s different from a cheap surprise that comes out of nowhere and feels unfair. Take FNAF 1 for example, you expect a jumpscare when the power runs out, but the uncertainty/randomness of the timing when the scare actually is, creates dread which makes the jumpscare work well, something I've done as well.
And sometimes keeping the player in constant tension is better, where they know something bad is coming but don’t know exactly when, it can be more effective than letting them fully relax first. In the projector example, the goal isn’t to surprise after safety, but to stretch the stress and to make the players hair gray, and maybe if they're lucky the shift ends and they win before the jumpscare appears, making it feel much more rewarding.