The "Should" Manifesto

I used to talk about aerodynamic modifications on my cars—and remember, I had no idea if they worked or not because I had never tested anything—using the word “should” a lot. “Removing the mirrors should be reducing drag,” for example. “Adding this tail should make it a lot more streamlined.” “Putting these lowering springs on should lower drag.” “Taping vortex generators here should keep flow attached over this panel.”

 
Banish this word, “should,” from your vocabulary whenever you talk about aerodynamic modifications. The air doesn’t care about what you think it should do, and you’re most likely wrong anyway.

This guy got it right.

Mindset
 
That word, “should,” betrays a mindset of hoping that your modifications work without testing them to see if they do or not. If fear—of finding out they don’t work, of learning that you failed—is holding you back, see if you can approach aerodynamic modification with a different attitude. The Last Jedi had its problems, but it did have one good line: “Failure, the greatest teacher is.” Failure is not a result, it’s a starting point. I can’t say I look forward to failing, but I do welcome the opportunity to learn something from it. Which I do, every time I fail.
 
For example, I fitted air curtain ducts ahead of the front and rear wheels on my Prius before I learned how to test; “These should speed the air up over the wheel openings and reduce drag,” I reasoned. Fast forward several months: I measured the pressure change along the side of the body, with and without the ducts in place. The front ducts, it turned out, weren’t doing anything measurable—exactly the same pressure on the front and rear doors with the ducts installed as without them. But the rears were; there was a measurable increase in pressure on the side of the rear bumper cover.

Version 1 on the left; version 2 on the right.

So, back to the drawing board I went to make another version. I fitted the newly-designed, larger front ducts to the car and went out for a test drive; I immediately noticed that the front of the car was dramatically more stable. I subsequently tested the car in crosswinds and passing heavy trucks, with and without ducts, and measured the pressure difference side-to-side in crosswinds; sure enough, there was a measurable difference. These ducts were doing something cool, although it was not what I had intended (and further testing showed no change in drag with or without them—so in that sense, I had failed again!). I never would have discovered this without failing.
 
Another example: last year I was reading an old Autospeed article in which Julian Edgar reported taping vortex generators under the front of a car to see if the air could be sped up there. He wrote that the car felt more stable with the VGs in place than without. So I fabricated some VGs from scrap metal sheet, taped them under my bumper, and went out for a drive. No change. Absolutely no discernible difference in the car’s handling. Failure! But ever since then, I’ve been thinking about other things I could experiment with—for instance, a converging duct to accelerate the air under the car—that I never would have had cause to come up with if I hadn’t tried something and failed.
 
Could
 
Rather than using should, replace this word with could. “Could” allows for the possibility that a change works or doesn’t, subject to testing and measurement; it removes any theory-based expectation that may or may not turn out to be real. “Could a tapered tail at the back of my car reduce its drag?” “Could adding vortex generators at the top of the rear window keep the flow attached on the trunk lid?” “Could blocking the grill reduce drag enough to measure?”

A grill block undergoing testing.

All scientific progress has been made this way: by finding out what happens when something is done or changed and then figuring out why it happened. Observations of physical phenomena are the foundation of the scientific method; observation precedes hypothesis and theory. So don’t get that turned around! Especially with something as complex as aerodynamics, where air behaves non-intuitively and cannot be predicted with any certainty, you cannot proceed from theory to application without testing. Go the other way—application and observation first, then theory. Tape tufts to your car and observe where the flow separates; then make a change and observe the effect on the tufts. Measure surface pressure somewhere on your car; then make a change and measure in the same spot again to see what effect it has. Record a throttle-limited top speed; then make a change and see if it goes up or down.
 
Along with that, don’t go online and BS about your modifications working when you haven’t ever tested them and are relying solely on a “should” mentality, guessing and hoping. I did this for years, and it was only later that I realized it is incredibly dishonest. If something you built works, you should have test results to back that up. We have a variety of methods available to measure things on the road; use them!

Because testing showed it works--reducing drag around 4%--I built a permanent version. No "should," no guesswork.