Stewart,
That's an interesting theory; certainly worthy of some discussion, I think!
I'd like to comment on that, if I may, and remember; this is only my opinion... I am no expert at anything, much less dynamometer engine testing, but it seems to me that...
If you're not specifically dyno-testing for just "peak" horsepower, but are testing for a horsepower CURVE, then the quicker you set the acceleration rate, the less horsepower will be left for the dyno to "see," because using a quicker acceleration parameter, the weight of the reciprocating AND rotating assembly will be creating more inertia-based parasitic drag, and will result in a smaller horsepower output number, as the engine accelerates its rpms.
BUT, I don't think that will be the case in a "peak horsepower-only" dyno test, using a static (peak) RPM, because there's no changes in the inertia at a steady rpm.
NHRA bases their factors on peak horsepower, probably because that's the way engines have tradirtionally been rated by the factories... the maximum output IS the rated horsepower. There's no easy way around that that I can see, insofar as assigning a factored horsepower rating is concerned.
So, the different accelerations in a dyno test, and the different output numbers they produce, probably wouldn't affect the peak rating, where the factors come from.
That's just my 2-cents... Your mileage (and, horsepower) may vary....
Bill