The problem (or really, nuance) comes when you try to define what "run exactly at 100%" means.

Yeah, I love the expression the "Flaw of the Averages".

Depends on what you're counting, among other things. (Not to mention when you took the samples being averaged).