[Piwoslaw]

A-B-A-B testing is a good way to measure the Cd change of aeromods that can be taken off/put on on the road without too much fuss, but how does one go about quantifying improvements of irreversible mods, like bellypans, lowered suspension, roof chopping, etc.?

My guess would be to do a bunch of "A" tests before any mods. Each test would be 3-5 pairs of runs (1 pair=both directions), with weather details recorded. The next tests would be on differnet days, but on the same stretch of road with similar weather. After a certain number of tests on different days, the average from all runs in all tests is less sensitive to slight weather differences. This is the "old" Cd. After finishing the mods, new tests over many days will reveal the "new" or "B" Cd. The multiple tests over many days compensate for not being able to do a second "A" run after the "B".

Question: how many tests are enough to get a good Cd value? 5, 10, 15? Enough so that each next test effects the grand average by less than 5%? 1%?

Anything important to pay attention to that I didn't mention?