Iterative development like that isn’t uncommon in engineering as a whole. Simulation can get you a long way but there’s a hard limit to that. You don’t think spacex designed a starship to use without running extensive simulations to try and figure it out before hand right?
Sometimes you need to test in the field just to find out what bits you missed. Structural engineers will simulate and calculate extensively but they’ll still build scale models and test pieces because it’s the most reliable and effective way to ensure you’re covering as many bases as possible.
Its not an either/or situation here. They’re doing the testing and simulation and applying it IRL to find out where things break.
Leave a bowl out with a sign that says “if the bowl is empty, please knock.” You don’t even have to fill the bowl with anything.