The limits of the HHT for me are based on the hair itself. You need to be using the same hair every time, and I ask my barber for some when a suitable client gets their hair cut. It needs to be dark, not colored, and straight.
Even so, there's enough variability in hair to make a point or more difference in HHT from hair to hair, so I always use 2-3 if there's any difficulty performing the test. A hair can also vary a little along the shaft it seems.
But overall it works for me. I also use it as a secondary bevel set test. My primary test off the bevel setter is the cherry tomato test, but I generally use a fairly coarse suita after the bevel setter, probably 4-6k synth equivalent. Off this suita, I palm strop just enough to clean the bevel and the edge should cut hair cleanly without pulling or extraordinary effort. If it does not, the bevel is not set properly.
Using the HHT this way also allows you to test very small parts of the edge, for example where a chip used to be (hope you marked the location) and the ends of the edge right out to the last mm.
The OP is correct about the weaknesses, but with experience and with consistent hair it can also be useful. I've never had an edge fail the test and shave well, but I have had edges that passed but weren't smooth. I actually am looking at the behavior of the hair as it contacts the edge, does it pop immediately, require some force or lateral movement, split, hang, etc.
Hope this helps.
Cheers, Steve