I am pretty sure that they simulated a lot this descent. This failure means they do not have (yet ) the correct modell. Or the descent sequence has to many var

I am pretty sure that they simulated a lot this descent. This failure means they do not have (yet ) the correct modell. Or the descent sequence has to many variables and any slight change might propagate down the computation to a larger deviation ( like the exact time to stop the thrusters - seems they stopped earlier than needed ). Or we need to load more advanced algorithms/AI software that decides and computes a local solution. At this moment, we don't really know.