Thanks to those who liked my post (my first post here -- long time lurker, thanks for being kind!)
Again, i'm not a pilot, but am a cognitive psychologist, so I'm sure I come at this from a different viewpoint.
To me, the question is this:
Given an identical situation, what % of professional pilots (or perhaps '3 man groups of pilots') would flub it and crash the plane?
There seems to be a contingent of forum members saying "0% : the AF guys were idiots".
There are others saying "100% : the AB design is at fault".
In my opinion, this is actually a very nice scientific question, amenable to research. Let's calculate this %age empirically.
Imagine this:
Get 100 x 3-man crews, and put them in a multi-day full-experience simulator. In this simulator, they fly full flights, sleep odd hours in weird hotel rooms, etc. Each crew does this for 30 days. 99% of the flights are uneventful.
At some point, each crew will get on 1% of their flights an AF447-type scenario. No warning, it just happens.
From this study, we calculate the ultimate data point: what % of crews survive. And perhaps more interestingly: what % of crews survive
for the right reasons.
Expensive as hell? Sure.
Would it cost more than one AF447 tragedy? (Not snark: that's a serious question - how long would it take to do this study? I have no idea.)
The outcome would be quite interesting.
Reading the BEA report, the most interesting findings to me was the comparison to events "similar" to the AF447 event, which the BEA summarizes as follows (p 106, english edition)
- 'Calling on the "unreliable airspeed" procedure was rare'
- 'The triggering of the STALL warning was noticed. It was suprisging and many crews tended to consider it as inconsistent"'
So I propose (a modest proposal, being on a forum where we do nothing but argue) that we don't need to argue theory : The BEA has collected the pilot (ahem) study data for us...
Let's take the next step, do the real experiment and see what the data shows.
If, for example, 99% of fully-trained expert pilots flub this, then the "bad pilots" chorus would probably have to rethink their position.
If, for example less than 1% of fully-trained pilots flub this, perhaps that's more of an example of bad apples needing to be better-trained or weeded out?s
Or maybe in either case, we need to consider the human-machine interface as the thing that must change?
We could of course include variations and get lots of data : set the simulator up with A vs. B style controls. 2 vs 3 vs 4 man crews. Time since last slept. Time since woken up. Aurual vs. visual warnings. Fullt-time AOA sensors.
We could even include some stooges (experimental confidants) to mess with the results: what if half the simulated AF447 scenarios, the PF was in fact a trained experimenter who just holds nose up, and we see if the other 2 PNFs figure it out and overrule him?
Sounds like a fun experiment to me!
(NB - I'm sure a lot of this research has been already done, right? I'm not a researcher in this field, so I'd be surprised if these ideas are novel).