I’m not a rocket scientist, but the Navier-Stokes equations are nasty; the best we can do is model what happens, assuming that we can exactly predict what will happen, which we can’t. Temperature, pressure and humidity of the air the rockets will fall through will not be exactly known, and all will affect how the rocket moves.
And even the best of modeling will not guarantee that reality will behave as predicted. Suspension bridges, for example, often need some ‘tuning’ to remove resonance after construction.
Bridge engineers, knowing that, often won’t even try to spend time and money to exactly predict bridge behavior. This may be similar: they may just have thought it would be faster or easier, maybe even cheaper to do the experiment than to work on exactly modeling their engine.
(The center engine is different from the others, but it also could just be statistics, with the amount they used giving a probability slightly below 1 of igniting one of these engines, and this being their unlucky day (given that they already have quite a bit of data from earlier launches, that is not too likely, but it is something I would let an engineer look at)
And even the best of modeling will not guarantee that reality will behave as predicted. Suspension bridges, for example, often need some ‘tuning’ to remove resonance after construction.
Bridge engineers, knowing that, often won’t even try to spend time and money to exactly predict bridge behavior. This may be similar: they may just have thought it would be faster or easier, maybe even cheaper to do the experiment than to work on exactly modeling their engine.
(The center engine is different from the others, but it also could just be statistics, with the amount they used giving a probability slightly below 1 of igniting one of these engines, and this being their unlucky day (given that they already have quite a bit of data from earlier launches, that is not too likely, but it is something I would let an engineer look at)