@Machinbird: Your comments on flat spin are interesting, and it is possible to put some figures on the g load that a pilot would encounter using some basic physics. The acceleration during circular motion is 4 Pi squared times the radius for the motion times the square of the number of rotations per second. If the pilots are about 15 metres from the centre of rotation during a flat spin, and the aircraft was rotating about a vertical axis taking 3 seconds per rotation which is not unrealistic in this scenario then this gives a radial acceleration which after normalising to 10 m/s/s for 1G, leads to a g load in a direction facing the front of the aircraft (eyeballs out as you said), of over 6G. Any faster rotation than 3 seconds per turn would give even higher g loads which would likely render the pilot unable to react to the controls as you have detailed in your post. This calculation confirms your analysis if the aircraft did end up in a flat spin, despite this being a significant departure from the normal flight envelope. Nevertheless the FDR data should ultimately provide evidence as to whether the aircraft did end up in a spin following the event sequence at the time of loss of contact.