The most disappointing aspect of this entire thread is that the idea Hambini proposed intrigues me. What if the basic premise is correct? How do we learn more? And from which…credible…source?
I think we’ll learn more once we get a handle on the best practices on using all the new “aero sticks” to test equipment. If there is a significant difference between “real world airflow” and “non-louvered wind tunnel flow” it should eventually emerge.
I’m of half a mind to just do my own Chung-style testing of a Flo6 vs a Bontrager Aeolus 6 using an outdoor course and doing enough repetition to account for wind gusts, etc… If there is indeed a 25W difference @50km/h per the Hambini graph, it should pop right out in the data. (But if I do this I will present my data as a hobbyist, not an expert).
If so, the measurements of CdA we find in the wind tunnel would not be usable for predicting power or speed in the real world, nor the measurements of CdA found by widely varying speed in field tests. In addition, applications like Best Bike Split would be way off.