I think you've over simplified things a bit. You presume a given RR and a given drag, but those things aren't given.
I'm using the values gives on the Roues Artisanales site
Also, it is not conclusive that clinchers in fact have a better RR in all given conditions. So far, the data points that way; however, the data has not necessarily been collected in a way that accounts for varying operating conditions.
So you think that the truth most likely lies in the opposite direction to all evidence accumulated so far?
So, it's a mistake to make that assumption. It is known, however, that aero factors dominate over about 15mph (24kph).
No it's not, that's a rule of thumb... it will depend on what sort of aero and what sort of rolling resistance we're talking about. The measurements suggest the cross-over occurs much higher for wheels and tyres
Also, you've forgotten that aero drag is quadratic with velocity whereas rolling resistance is only linear.
No, I haven't. The measurements on Adrien's site claim the aero drag was reduced by almost exactly 1.95 times between 50 & 30km/h regardless of wheel type. That makes it a mix of linear & quadratic.
At 50km/h, the measurements (for what they're worth... if you've got better, go ahead...) suggest 23W for Boras, 33.2 for Ksyrium ES, so 10.2W different at 50km/h, or 5.2 at 30km/h (or, by fitting a general quadratic through those two points and zero, 7.7W at 40km/h).
Now from the RR tables, at 30km/h, Michelin Pro Race 2: 29.2W, Veloflex (best of the tubulars) 34W, so 4.8W.
Hence siding solo at 30km/h you are pretty damn close to break-even: 0.4W in favour of veloflex + Bora vs Ksyrium + Michelin.
Rolling at 50km/h and assuming RR stays constant so that dissipated power increases linearly, the RR losses would differ by 8W, just slightly more than the differences in aero losses at 40km/h.
So if running in a pack reduces aero drag by about 10km/h worth (which is probably conservative, given time vs distance results for a pack vs ITT), then using an $80 pair of Michelins will give more benefit than a $2200 pair of Boras.
You're assuming that RR overtakes aero at some point in your model, and that cannot be assumed. You need to account for initial conditions, i.e. actual drag numbers, RR numbers, etc.
Actual numbers above...
IF your assumptions are correct, there is still a lot that is not accounted for. It's doubtful that wheel flex accounts for any significant energy loss at all, given the very tiny amount of frontal deflection any wheel demonstrates. And in a pack, it's unlikely that there's any significant lateral flex. There's really nothing empirical to suggest that a wheel's hysteresis in lateral flex is significant in the pack. The greatest lateral loads are in sprinting, entering a corner, exiting a corner under power, and also in the case where a wheel might hit a large obstruction.
I think this is probably true: my point was to bring up the "a lot that is not accounted for"