I did
A-B testing only on the undertray (because it was difficult to re-install, as you said).
It showed about a 2% improvement.
The numbers I have from the undertray + the boat tail suggest the combination of the two is greater than the sum of their individual effects taken separately.
Meaning: if I were to leave the boat tail on, and test just the undertray, we'd see greater than a 2% improvement from it (because the best maximum performance of the boat tail depends on the smooth flow provided by the undertray).