Solved

no-setup is faster. at least is not slower at all.

The driver is capable of making errors, and qualifying seems to be a tad too unpredictable still, so perhaps this is behind the mixed results?

Try making a terrible setup (as far from ideal as possible) and running a test on it, then run a test on an ideal setup and see the difference. There is quite a difference in lap time in my experience, over a second sometimes, but usually several tenths at least.

Just tested, this very moment: Laps 1,5 = ideal setup. Laps 2-4 = bad setup. Difference = 0.8s per lap on a short track.

You definitely weren't crazy to bring it up during beta testing, because this was actually a bug then. It was addressed after launch recently in this thread, though. So, this was a known issue until less than 2 weeks ago. I wonder how many people are remembering results from around that time?

Even in the present, all comparisons of finishing positions can be false positives. It doesn't actually prove anything about whether your performance was better with an ideal setup or without one. Finishing position isn't necessarily tied to how good your setup is *at all*. It could be totally unrelated. If nobody else prepares for example, then everyone prepares the next race, you'll have totally erratic finishing positions. It's easy to be mislead in to thinking this was tied to what you did, when it was really tied to what everyone else did.

We've been clear that in the new iGP certain characteristics suit different circuits better too. Like in F1, one week Red Bull are all over Mercedes because the circuit suits their car, then the next they aren't close because they lack performance in another area. This too could explain erratic finishing positions and a team that doesn't prepare performing very well at a given track. Depending on where teams assign their design points between each race there might be fluctuations in performance and finishing positions.

That said, I still think some of the posts here are unusual enough that I'm going to look in to this and apply a patch if I find anything before the races tonight. Looking at my post above, #8, it's clear there isn't an issue in practice, and it would seem to be related to races only.

I do think I found one potential issue with the default setup in races. It would not have been as good as an ideal setup, but it would have been good nonetheless. I've applied a patch now, so let's see if it improves in the races tonight.