@ChristiaanJ - I'm not talking about "on the edge" training in sims for line flight crew, I'm talking about basic manual handling - turns, pitch management, stall warning recoveries and the requisite amount of sidestick manipulation for each.

If the PF had apparently lost control because of external forces beyond his ken, then you'd be right, and there wouldn't be much that more training could accomplish. But to all intents and purposes it looks like this guy controlled a perfectly flyable airliner (albeit with minor systems degradation) into a full stall -which he neither recognised nor was able to recover from - by significantly overcontrolling at altitude.

Someone needs to tell the "beancounters" just how expensive crashes can be, not to mention the amount of business they stand to lose when it turns out that their training regime produced a fully-qualified pilot who ended up completely out of his depth making basic course corrections at altitude!

Let's not forget, that but for the grace... The a/c could well have gotten there on its own (technical problems beyond his or its ken? ). Cuts both ways, seems to me.

The unknown is by its very nature, the unknown. To train for it is an oxymoron, and a waste of time. But let's be honest, and speculate to include the a/c as possibly the problem, No?

Thank you both for your patience

@DozyWannabe - I'm not querying the pilots' basic manual handling skills either (none of them were tyros), but I have my doubts about their skills at handling the 'brown matter hitting the ventilator' at FL350 entering the ITCZ.....

A long way away from the 'engine failure at V1' in the sim.... which seems to be the usual thing they get thrown at them during a sim check (I'm exaggerating of course, but I would be sure they'd never seen anything even remotely resembling the "AF447 scenario" in the sim, ever.)

So far, I agree with your remarks ..... and I admit to being equally baffled.....

Maybe this was the first time an A330 ended up in a fully developed stall, and the PF was not a test pilot?
Maybe, like in the olden days, an early A330 should have been taken into a full stall, and the data recorded, with a large parachute in the tail, and an escape chute and parachutes for the pilots? We had both on Concorde.....

Well I hope pilots will never seen (or learn) the AF447 scenario in the sim
What is the instructor who would teach pilots to get a airplane stalling ???
I hope that instructors will show in the sim has not go to a stall
Do not pull up on the stick like crazy is probably the first advise they will teach

BS. They are clearly stating that the airplane will fly safer than any pilot. And therefore implying pilts are not safe.
What he doesnt say of course is that their philosophy without all the protections would be hell, but whatever im not even gonna argue, ab has gone away wih so much stuff its ridiculous.

What is worrying is the amount of pilots defending people like the above from the quote.

I think you're reading what you want to read there. Taking 1 sentence out of something rarely captures the gist of it. I think what he was saying is that there is a time to practice hand flying, but it isn't while cruising with a hundred odd passengers in the back. I'd be inclined to agree with that. There is definitely a need to keep up manual skills but that doesn't mean one should start disconnecting the AP somewhere over the atlantic at FL370 to see what happens.

To use your own phrase, dlcmdrx, I think your interpretation is BS.

The meaning of the original comment is that sometimes a computer is better at handling an aeroplane than a human - which is spot-on. If your ego can't handle that, don't sit in a cockpit, because it's that attitude which leads to accidents.

Would you care to enlighten us as to what you think Airbus has "gone [sic] away with"?

I just flew from CDG to the US in the right seat of a 777. Weather good at both ends. Hand-flew to 18,000 on climbout. Hand flew from 3000' down on the visual approach. I consider it my duty to keep my hand-flying skills sharp. To do less would be lazy. The day I can't do this I will find another job.

Hand-flying only during recurrent? Go play the piano once every six months and see what happens to your performances.....

That said, I don't hand-fly at cruise.

EDITED later:
@ testpanel
I may have misread your post, or misunderstood its context. A Flightglobal article seemed to allude to something that your pithy observations may have been rooted in.
If that's what you were getting at in re the PILOT monitoring function, then your post makes more sense to me than the way it read the first couple of times.
@testpanel
You need the warning since the aircraft feedback system doesn't provide cues. Attempting to solve this by training perfect prevention is a good way to fill up a few more graves.
While I completely agree that stall avoidance is the general best practice, stall warning is particularly handy in conditions near stall (approach) where abrupt changes (gusts) or malfunctions can take your margins from comfortable to too darned close ... so yes, stall warning is a good design feature.
1) make sure it works
2) design as well as can be to eliminate spurious stall warning
You know that is NOT what CRM is all about. As to AF447, the PNF was indeed doing that, and more. He had to. (If you were arguing that he wasn't doing it well enough, that was a curious way to state the case).

The PILOT not flying or PILOT monitoring is a role which sometimes requires action ... back to CRM 101 ... for example, when the PILOT flying is cocking it up. (See the CVR extracts for this mishap). I don't recall ever teaching that one keeps monitoring failure. One works to correct error or failure.

PNF (LHS) noted errors and was at one point playing the helpful copilot role to support the PILOT flying (in terms of trying to get him to recognize and correct a pitch up error and a climb), a PILOT whose flying was breaking down. (Were his verbal inputs optimal? Separate discussion, and as I don't speak French, no further comment).

What more do you want this monitoring PILOT to do?
Monitor the screw up all the way down to the ground?
(Come on, you can't feel that way.)

I could make the complaint is that he didn't take the aircraft soon enough, but he also was struggling with his scan and puzzling out airspeed issues ... one can also argue that he didn't prompt the PILOT flying that they needed to get the UAS procedures underway ... but if the guy can't fly the bird straight and level, can you begin to work that QRH? There was apparently a fundamental flying problem underway.

The LHS guy is a PILOT, a role that sometimes requires action.
Amen, Deacon.

Also: I had to edit again, I realize my tone was way off base from appropriate.

Apologies for that. :(

You didn't get what I meant... Pilot errors were most frequent (primary) causes of accidents even 40 years ago. And they are now as well, of course. The important thing to notice is that todays total number of accidents (per number of flights or flying hours -- whatever you choose) is significantly (i.e. few times!) less than just the accidents caused by pilot errors those 30-40 years ago. That means than (so) many pilot errors went away. It's either because pilots are so much better today (which is rather hard to belive) or their errors either get corrected or are unable to happen. IOW: those protections do work. Of course CRM progress has helped as well, but all those protections, GPWSes, TCASes, FDs, etc. are primary factors.

It's not BS. BS is your overinterpretation :=. And those protections are the primary reasons that from about 2500 heavy FBW planes out there only 4 (or 5 if flight testing is included) were lost while on duty (compared to ~40 out of 1500 of most popular earlier design).

Maybe they're grown up pilots, who know their limitations and that their duty is to without undue risk get those few hundred souls behind them to their destination not their destiny. There are both old pilots and there are bold pilots, but there are no bold old pilots, my son.

I hope you're not proposing that pilots should train stall-recovery (or even approach to stall) while there few hudred pieces of Self Loading Cargo behind them :E .

There is no other option than sims to train dangerous things.

Well, AF447 would not happen if PF handled properly UAS. If that guy simply kept that bird flying (mostly) straight there would be no issue of approach to stall, and no need for stall recovery.
Then I think that sims are capable enough to even train propmpt stall recovery, when the plane is just stalled (with AoA ~7deg at FL350) not when stall was allowed to progress to >40deg AoA.

BESIDES: it looks like that PF was also a glider pilot... So you're apparently right that glider training is no substitute for type training.

Stall warning systems have saved a lot of lifes over the years. People make mistakes (even pilots), and if pilot error did cause an inadvertent approach to the stall would you prefer they discovered about it by means of a stall warning before it happened, giving them adequate time to avoid the stall, or would you prefer they discovered it only after the aircraft actually stalled?

I find your two factor analysis wanting.

The entire system has improved (in some ways, degraded in others), which includes over fifty years of working at crew skills and crew habits, and the improvement thereof. Likewise, there are a pile of things, inputs that pilots must consider, that inform crew decisions. What 40 to 50 years of continual change has also done is make it more common to fly at the edges of performance (and weather and airspace capacity) margins.

THAT change helps the bottom line, as fewer flights are cancelled or diverted, with associated cost, and along with this benefit the system as a whole gets more comfortable with operating at a higher risk profile. (Take down ATC radar and see what that does to the tightly timed and coreographed spacing around LaGuardia, for example.)

You might want to consider that in your accident and causation analysis.

Also, as another poster raised this point in the other thread ... what behaviors and decisions are rewarded by the company, and by the industry (by pay or strokes or reinforcement) and what are discouraged?

That will influence decision making, and thus your mishap rate.

Amusing, yet to be determined if those fabulous systems of yours have killed 200 people and you still will affirm those things. Whatever. I guess perpignan or quito wasnt enough

Go and check your history, dlcmdrx, especially regarding Perpignan. And Quito, for that matter. When you're better educated on both accidents, come back.

That's a guess, and nothing more.

It's a fact that flight safety has improved. Why it has improved in terms of specific interventions is a different question. One of the problems with comments like yours is that it ignores the role of luck in the equation and attributes all the statistical improvements to mankind.

Some people in the airline industry often think that they are the only ones dealing with issues surrounding extremely unlikely events. They are not. Lotteries are another field where this discussion rears its heard all the time. There is a nice non-academic article that talks about this issue here. I especially like the article because it breaks down the thinking of the two different camps nicely. One camp is willing to put it down to luck. The other camp “When something this unlikely happens in a casino, you arrest ‘em first and ask questions later.”

The question "Did she cheat?" is no different than the question "Who is to blame for the crash of AF447?" My response is that she has the money; you don't. AF447 crashed. If you are going to take that money away from her the burden of proof is on you to prove that that she cheated. If you are going to argue that CRM improved safety, prove it. If you are going to argue that bad CRM was to blame for AF448, prove it. You can't. All you can say is something along the lines "it's possible that she won fairly but it's more probable that she cheated." That's why the findings in an accident report are based upon the probable cause of the accident, not the definitive fact. I await the day, however, when the NTSB or CAA or BEA finds the probable cause of an accident to be "bad luck".

tespanel:
Would you please enlighten us all, as to what "indications" do you think the AF447 pilots had (apart from the stall warning) when AP failed?
thanks

Buffet, as indicated in the BEA report?

Pitch-up, as probably indicated on the attitude display?

Sudden increase in altitude, despite the flying pilot mentioning to both the captain and the second co-pilot that the REC MAX was only FL375?

There may be others.

Thanks JH:
Buffet could be one of them, although could also be confused with turbulence;
Pitch-up would have been another indication, if not for the "probably indicated" interjection you have used in your answer, because (like me) you are not sure. I haven't seen a good explanation for this message, yet 2:13:14FLR/FR0906010211 34123406IR2 1,EFCS1X,IR1,IR3,,,,ADIRU2 (1FP2),HARD;
and although I recognize they have completely obliterated their previous concerns regarding climbing restrictions, FL375 would have given them a 1,3G protection (in steady flight circumstances), so FL380 would have given them, what? 1,1G, 1,2G?.
But there are more, I'm sure. And altogether would have meant to a clear minded pilot, that he would be approaching the edge of stall...

Further to what I have said on the previous post, my question still is:
Would the PF, if flying a yoke equipped non-FBW aircraft, have stated at time: 02:11,58
(my bold and underlining)

An important "indication" of a stalled aircraft is given to the pilot via yoke or stick feed-back. And "that one indication" he didn't have...