From takata
Everything in the fault sequence analysis makes me think that the PF display was the first affected. The PROBE fault recorded imply that the 3 probes readings were different as well as out of range with previous median values. Hence, no single ADR could be rejected but the three altogether. We are also told of the recorded sequence implying PNF speed decrease followed by ISIS: the first value then to drop should have been the one on the PF display.
There is also an indirect proof of an altitude drop and range: the following reported TCAS fault by ACARS. AFS/FMGES (autoflight system) ADR altitude monitoring fault treshold is set at 400 ft instead of 3,000 ft at EFCS level (flight control monitoring). Hence, TCAS should have faulted because of that. So, could it be that uncorrected static pressure was dropping to the point of displaying an over 400 ft of instantaneous altitude change, Mach going down from 0.81 to about 0.18? Could it be that static pressure was also affected by icing?
If this was the first information taken by the PF in addition to the roll at AP disconnection, he could have effectively feared that some kind of spiral dive could follow. Likely, his pitch rate wasn't his first concern, then without speed, after ignoring the first stall warnings as spurious, he might have lacked the correct info necessary to understand how much energy was lost during the climb with an altitude under reading, then he also would be spatially disoriented. There was also no mention of thrust change during this climb and this would rather fit with a PF trying to slow down than one fearing of stalling.
FWIW after following this and other thread for many months and wanting to believe the PF reacted to indications (not terrified rabbit pulling up), takata's theory is what I believe happened. That in my view is saying a lot after all the posts and credits takata with a conclusion that has taken so many months of analysing, questioning and thinking.
My worry now is the PF (RHS) instrument datas were not recorded and the BEA may have difficulty concluding about why the PF inputs were as recorded. But I fundamentally believe the PF was reacting to what he saw on instruments (outside was "black", no horizon). I do however think he failed to integrate his feelings from the seat, i.e. G forces over time (which probably began before AP disconnect with subtle descent). I can excuse him from that in the urgency of the situation, but am still trying to understand ~30 sec of 7000fpm climb.
My hope is that the A330 will become even safer (it probably already is due to modified UAS procedures promulgated). Surely we must find a more resiliant way to measure airspeed than those pitot's. But if takata is right, we must find a better way to reliably measure altitude at all times as well.