* First, you'll need an incompetent crew (the one flying blindly straight into a giant storm);
* Second, you'll need a bunch of incompetent designers (for all those unreliable flight envelope protection systems);
* Third, you'll need a good deal of very bad luck convenientely applied with the right timming (updraft here, downdraft there, and bits of ice scattered over the whole dish).
As a matter of fact, pitots icing at FL350 would not instantaneously switch the airspeed from Mach 0.82 to Mach 0.87, triggering a pull up to 60,000 feet (due to rogue C-3PO in command), then declare all the pitots frozen, then quit, handling over the whole situation with zero speed left (but displaying mach 1 and 90 degrees AoA) to the poor doomed crew which will try to figure out what to do next into the storm until contact with the ocean surface, flying backward, tail first, as per Bearfoil's analysis.
Takata, take a deep breath and turn your internal critical judgement filters back on.
Bearfoil's concept is a wild hare and should be disregarded. We do have actual wreckage this time and it is located very close to LKP pretty much down track, and indicating a descent to the water that almost guarantees it was in a deep stall. The aircraft is not supposed to fall out of the sky while in full cruise, and the flight control system is designed to guard against departures from controlled flight, so something undeniably happened that the flight control design folks did not forsee.
Tubby's concept may be slightly over-elaborated but has significant potential to explain what happened. Your expectations of engine stall at high AOA may be unreasonable.
About the best we can do with the data we have is try to match endpoints and any event data that ACARS provided until we see actual aircraft data.
It all fits in a context. AF447 did not do the impossible. We just have to be clever enough to figure out how it did what it did.