I do think that if you're going to call it out as an issue, you should do say 100 test-unlocks, or even say 20, something to increase the sample size (of possibly "it annoyed me a few times last week"...
), to something that makes this whole thing a bit more empirical. Or maybe it's just from having written too many unit-tests, I dunno'.
NOT saying this isn't happening, to anyone, just wonder how widespread it is, because I sure haven't seen it reviewed this way, in a published sense.
Fingers are different, and h/w tolerances are just that, and s/w algos that "interpret" the newer under-screen readers aren't perfect, no question, all of this feeds into things.
Maybe it's a large, un-amalgamated issue, of some sort, it wouldn't be the first time.
I just got an 18/20 success rate, as a starter, testing 4 fingers, 5 times each. It was a single thumb that failed twice, so that's at least interesting.