In this case, radar shows a climb to 45,000. At some point the airplane will stall due to not enough airflow over the wings. A wing dips and nose pitches forward
A "good" pilot can recover at that altitude within 2 to 3 thousand feet. In the simulator almost all pilots can recover before losing 10,000 feet. That would be considered a failure of this recovery.
According to the radar, it shows a loss of 25,000 feet. Soooooo alot of other crap is going on. But they apparently recover. Or Did they?
I have been involved in only two accident investigations. In both cases, we saw anomalies in the Air Traffic control radar data. Meaning it can have a burst of inaccurate data. But the whole body of information usually smoothes out the anomalies.
Did the altitude changes occur before the transponder was no longer working. Hmmmmm according to reports all is normal until the point that the transponder is no longer working.
If the altitude changes occurred AFTER the transponder was no longer working, they are all full of crap, because they would not know with any certainty altitude heights.
This answers some questions about when the transponder and ACARS were turned off or stopped transmitting. It came out an hour or so ago.
http://mobile.reuters.com/article/idUSBREA2G14020140317?irpc=932