So if the template can be scaled in the way you describe, isn't it basically just an angle? The angle from the front the windscreen header rail to the rear of the car?
And then when we look at cars with angles that differ like the 2020 Tesla Roadster and 2019 McLaren Speedtail, isn't it just
a range of that angle?
You say "It has had very good effect nearly all the time, on a variety of vehicles, in a variety of tests but usually testing has been limited to somewhat longterm fuel economy tracking."
I'd expect that to be the case if the flow remains attached - and that's great. But what I cannot see is why this is somehow regarded as
the best shape.
It's like the person who says a particular airfilter gave good engine power - but hasn't compared it to any others.
I just posted this in another thread. It's from Hucho 2nd ed p.201. It shows how a variety of shapes can be used to give the same low drag.
I think the idea that there is a 'best shape' is a bit simplistic - and especially in the real world, the thickness of the boundary layer (that in turn will depend on what is happening on the forward section of the car) will help determine if the flow remains attached.
As to swimming upstream - that's something I am very used to doing, especially in the often group-think environment of an online discussion group!