[11101010001100]

Sure, you can also say we start with the ZFC axioms. QED.

Calculations provide predictions, experiments provide verification.

[amelius]

I mean, computational fluid dynamics is pretty mature now.

And if for some reason it doesn't work here, then that would be nice to know.

Besides, the experiments themselves are simulations.

[11101010001100]

I just spoke with an YC company. The CFD needed for their applications doesn't scale. There are plenty of reasons for experiments.

[mhh__]

CFD is actually really hard to get right.

F1 teams for example constantly struggle getting their simulations to line up with their measurements.