It’s opposite of current NASA that ended up with way over budget, way behind schedule SLS. When it finally did fly, it’d been component testing simulated to billions of dollars and it flew as expected
Conversely Space X built a bunch of original Falcons, didn’t absolutely know they would work and indeed some RUD’ed
They eventually moved on to the next generation of Falcons and tried to land them Again not knowing if they could land them and the initial results were uh, interesting.
Space X is doing the same with Starship…well it ought to work better than last time, we made some changes, let’s give it a go. Simulations can’t simulate everything so nothing like real world data