Manbearcat
Legend
If the DM violates what's already been established--either in game-facts, or in play-style--they're risking their credibility. If you're leaning hard enough into Skilled Play (as a play style) that the PCs are able to establish an ability to take a Long Rest, you're violating established stuff to prevent them from doing so; if you've established that the PCs know Strahd's capabilities, you're violating established stuff to change them.
But this assumes that you're leaning hard enough into Skilled Play. Yes, I agree, if you're leaning hard into Skilled Play, this would be a GM risking their credibility. But 5e GM's don't have to do that. There are two other Styles of Play cited on page 34. One doesn't lean into Skilled Play. The other is halvsies.
I made my post to address all 5e GMs...not just the ones that are leaning hard enough into Skilled Play. I was curious about how the voting would turn out given how much leaning into the other two play styles cited on page 34 I see in 5e threads.
My point is that they're very rarely at odds in the way the OP suggests.
I don't suggest anything about frequency in the lead post. There is nothing there about frequency at all.
In fact, I don't even believe this would be a frequent thing (I've GMed 5e for probably 100 Long Rests? At all levels 1-20...I've seen/felt the tension of this maybe a handful of times?).
My question was "when this does happen (even if its only 1 in 50 long rests)...what do you do?"
@loverdrive had the most interesting answer (none of the above). "My bad guys, please don't take the Long Rest or it will screw play up" (again, a D&D version of "hold on lightly").