PJ2
sorry, but the (non aviation) engineer in me sais differently:
yes, the accident would have easily prevented by basic airmanship.
It would never have happened if the A/T and A/P systems were, by design, aware of their mutual modes, which they seem not to be. I think "mode confusion" is the term for this.
We all know that two things are safe: those you never do and those you always do. This was something that you seem to be doing extremely rarely. Not an inop radalt, but the fact that A/T disconnects as a result and A/P bleeds off speed, trying to maintain GS. As I said earlier, the systemic failure of this accident is terribly similar to the Birgenair 301 accident in 1996. In both cases, a faulty sensor was the single input to an automation system that subsequently flew the airplane into a stall. In both cases, the crew failed to notice the loss of energy and, when they finally did, were unable to recover.
Everyone will seek the cause in the realm where they can personally improve, and that's good. Pilots who are aware of their own fallability are better pilots (I do, however, sometimes wonder if some of the "idiots forgot to fly the plane" claims here don't have an underlying "couldn't have happened to me").
Engineers must, at the same time think about something else, I believe: Pilots are on the flight deck when things happen, they can adapt to the situation and react accordingly. Engineers aren't. Their decisions are, however, at the heart of the actions taken by the automation systems. How they respond to peripheral failures is decided upfront by what the design engineers could imagine. Failures they couldn't think of will often be handled poorly or not at all (that's a general observation, not specifical for aviation at all)
So what I'm saying is: yes, both accident should have turned out as non-events had the trend being spotted and identified early enough.
But: both events shouldn't have developed to anything serious in the first place, because at least theoretically, all relevant information was available to the automation systems to make a decision to either continue in a coordinated, safe mode (A/T and A/P feeding off a known good data source) or to completely hand it off to those humans who, in ambiguous situations, should know better what to do than the engineer behind his desk.
The thing I'm wondering about is whether it wasn't a full handoff to manual by the A/T system communicated to the crew by the thrust levers moving all the way back. Was that an ambiguous signal, since it was what they expected to happen when they were at an energy state that was too high to begin with?
Quite a few slices of cheese, quite a few holes lining up.