... when someone doesn't have a solid knowledge of aerodynamics then it is very easy to attempt a visualization based on what you "think" air does. Finding an optimal shape to enclose our package under these circumstances can lead you down a path with poor results or worse, a car that is unstable. An incorrectly shaped tail section can be as bad as a KAMM effect tail. This discussion along with observation, trial and error and asking a lot of questions are the only options for many of us ...
I agree, aerodynamics (like electronics) is not intuitive to anyone outside of the field. I can give you some basic guidelines though:
First, test, test, TEST!!! You are absolutely correct that optimized shapes take trial and error. What is not commonly known is that we can tuft test at low speed (literally 40 mph) and see much of the separation we will see at 400 mph. I know Speed Week comes but once a year, and most LSR vehicles are not street legal, but yarn, a video camera, a pickup and a deserted 1/4 mile of pavement is all we need for low speed. Put yarn under the car too and prepare to be shocked that a flat belly with a skirt is full of reversed flow and drag. Duct tape the camera under the tail of the car to get this shot.
Second, longer in the back is better than short. It is not guaranteed that a long tail will not separate, it
is always harder to keep flow attached on a short tail. Just as important, stability in yaw (fishtailing or spinning) is best resisted by a longer, not taller, tail. Look at Bub-7 and try to avoid those tall verticals too close to the rear axle. Closing the tail out means covering the chute tubes (or better, pointing them up), and makes a push bar difficult. IM<HO, it's worth it.