This experiment look pointless, for starter AI physics and tires are different from player's , and second :which reading would you do to the results ? The only thing we know is that Studio 397 released content officially partnered with the only people on the planet that know those cars. About the performance issue, for me the main disconnection from real times and simulated times are linked to the road surface. There is an enormous gap from Sebring and Silverstone, and I don't think any track before Sebring is detailed enough, even considering that Sebring is an old track with a weathered surface. Track surface impact almost everything, from ride height, suspension stiffness, best lines, grip and so on.
I don't even think that the reasoning that F1 cars are always faster than their older versions is sound. Often F1 rules were implemented specifically to slow down cars and reduce performances.