I don't know the physics of why those results came out the way they did. So any guess about your question would be just that, a guess. One here at TFP questioned the results based on the setup of the test. Others here swear it makes no worthwhile difference. And certainly not for one or two turns, right? So I wouldn't worry about it.
I think the advantage to a sweep is cumulative, so if we were talking about a dozen or so bends, like there were in the solar system I built, pushing water up to a roof and through a thousand little tubes, then, yah, I'd want to use sweeps, which is what I did. And I used them whenever I had to rebuild a run on my pad, because I had the leftovers from the solar project. But I wouldn't take something apart just to replace it with a sweep. And I'm not going to worry about all the regulars left in my circuit, both the ones I can see and those that are buried. In your case, if reg's were sitting right in front of me, and I had to get sweeps from another country, for one or two bends, I'd...
Glue away!
We're only really talking about the energy difference between the two, and so a few pump RPMs lost and at such a small amount the money saved probably wouldn't pay off the shipping cost for decades!
Yes. good point. Head loss per fitting x 1000 fittings is something to bat an eye at. 1 or 2 fittings, not so much.