How could an accurate simulation be created of unknown situations unless the AB test pilots intentionally put the aircraft into those situations, which, in my opinion, might lead to vacancies in the test pilot ranks?
If you're talking aerodynamically, then of course, behaviour outside the normal flight enveleope can be simulated. All the data e.g. mass/inertial distributions, lift, drag pitching moment of indiviudal components and the whole integrated a/c as well as all the pertinent stability derivatives, thrust and even aerolastic effects - would be available from initial CFD work, w/t refinement/cross checks and corrected within the flight tested envelope, a fair bit wider than certified limits.
The predicted beahviour could be simulated well past any flight test/certification flight limitations
There wouldn't be too many what ifs behaviour-wise, the
whats are what the
actual character and level of disturbance was, what were the auto and/or crew responses/inputs.