Pilot skill v. Aircraft Automation

[Zapphod Beblebrox]

So the discussion circles back to the underlying question: does a company really care about the "skill" of its pilots, beyond their ability to operate reasonably safe, on-time, and customer-courteously?

What interest is it of Acme Airlines if their pilots are all Al Hayneses or Cesley Sullenbergers, or if they are all John Doe who can achieve a mediocre-but-safe level of performance at the FAA-set average bar of achievement? What is the benefit to their business?

I have a distorted view, as I was trained and raised in an environment (military fighter community) where there was both objective and subjective value put on actual pilot "skill". Since the military cares about results and effects -- and generally not fiscal responsibility -- it puts training, testing, and performance bars much higher than what is required to perform the job 99% of the time. In fact, a huge amount of time, effort, and money is put toward training for those 1%-of-the-time Black Swan moments that a pilot or unit may never, ever see during an entire career. And it is one's performance there at the edges of the envelope that is used to measure a pilot's skill, because a pilot's ability to execute all of the other 99% is assumed to be perfect (or at least excellent).

Unfortunately, the commercial air world does care about money a lot more than the military does, and doesn't want to expend the time or resources required to focus on "1%" training when focusing on "99%" training is more than adequate the vast majority of the time (99% of it?).

When I put myself into the shoes of 121 airline management, and try to look at the big picture of keeping costs low to make profits big, I have a tough time seeing why I should really care about the skill of pilots on the line, so long as they meet the FAA standards and aren't making the company look too stupid in front of the paying passengers.

That is an excellent analysis. It's from a standpoint I would not previously have considered. Yes the company's want the lowest cost and don't really consider the degradation of piloting skills. How do you measure that and plot it on a spreadsheet? The aircraft manufacturers tout the benefits of the automation.

One new interesting fact is the fleet statistics are now much tighter. That is to say if the company looks at a particular flight between any given city pair the statistical plot now shows a much closer tolerance on fuel burn with a highly automated airplane. This makes those folks who read spread sheets (accounting types) for a living very happy. The dots line up better with fewer outliers.

This will continue until a major accident in the US, blamed on pilot proficiency at a major airline, or until someone can find a way to make stick and rudder skills pay.

 

[🦈💜]

The absence of an accident does not imply that an operation is safe. It is a necessary, but not sufficient condition.

"Absence of evidence is not evidence of absence," as the saying goes.

-Fox

 

[ppragman]

Hope actually is a plan.

Or better said, statistical outliers are not a valid training program or SOP.

In that case, why do we do much of our training on engine failures then? Because even those are statistical outliers. Nah, there's a little more to it to that, it's a risk management thing.

You have to evaluate what to train for by comparing probability with severity of the occurrence.

 

[Autothrust Blue]

In that case, why do we do much of our training on engine failures then? Because even those are statistical outliers. Nah, there's a little more to it to that, it's a risk management thing.

You have to evaluate what to train for by comparing probability with severity of the occurrence.

The engine failure at V1 is truly the worst-possible-scenario in terms of that particular malfunction combined with the need to precisely extract performance from the airplane (although I would say that it's actually slightly more difficult at 400' or on the go-around at initial thrust/power application from an approach setting). Mishandled, particularly in the presence of a complex departure procedure to clear terrain, it can easily lead to a less than satisfactory amount of clearance from said terrain. (For instance) Our CRJ program goes to RNO in the sim on a regular basis to fly those complex departures.

 

[ppragman]

The engine failure at V1 is truly the worst-possible-scenario in terms of that particular malfunction combined with the need to precisely extract performance from the airplane (although I would say that it's actually slightly more difficult at 400' or on the go-around at initial thrust/power application from an approach setting). Mishandled, particularly in the presence of a complex departure procedure to clear terrain, it can easily lead to a less than satisfactory amount of clearance from said terrain. (For instance) Our CRJ program goes to RNO in the sim on a regular basis to fly those complex departures.

That's basically exactly what I was saying, the probability of an engine failure at V1 (though like you said, I think it's more difficult at 400' or so too during an ODP) is very low but measurable and the severity is high enough is that it poses an unacceptable level of risk to not train for it - in contrast, training to ditch in the Hudson is basically a waste of money. In the 135 world, I don't think we train against CFIT, LOSA and loss of control enough. The prevailing attitude seems to be, "well, I mean, if you stall the airplane or crash into a mountain your kind of a dumbass and have no business being in an airplane" and yet we still seem to be stalling airplanes and pranging into mountains under part 135 at statistically significant rates...

 

[Autothrust Blue]

That's basically exactly what I was saying, the probability of an engine failure at V1 (though like you said, I think it's more difficult at 400' or so too during an ODP) is very low but measurable and the severity is high enough is that it poses an unacceptable level of risk to not train for it - in contrast, training to ditch in the Hudson is basically a waste of money. In the 135 world, I don't think we train against CFIT, LOSA and loss of control enough. The prevailing attitude seems to be, "well, I mean, if you stall the airplane or crash into a mountain your kind of a dumbass and have no business being in an airplane" and yet we still seem to be stalling airplanes and pranging into mountains under part 135 at statistically significant rates...

I think we have the same problems in 121, honestly, as far as what we do train for and not. Part of the problem is that we have to check this this and that box because The Regulator says so and oh by the way do it in (small number) of sessions so that we can crush The Competition through low training costs.

(never mind guys getting extra sim anyway just to get through all the box-checking)

So, yeah.

 

[Richman]

"Absence of evidence is not evidence of absence," as the saying goes.

-Fox

Correlation is not causation, either.

The big picture problem is that you need to pay bucks to get Buck Rogers.

But if all you want is Twikki, then the bill isn't as much, and some places are OK with that.

Flying is one of those things that takes equal doses of art and science. Take a look at any GOOD competition glider pilot and you'll see what I mean. The best way, usually, to get there is to start with the basics, learn the basics, love the basics, and carefully nurture a career full of learning the sights, smells, sounds and touches of flying. In the end this grants the ability to avoid a lot of issues from the get-go ("no, Mr. Mechanic, I don't think that defering the fetzer value is appropriate under the MEL under these conditions....and, please, use Quaker State), but also the ability to see THROUGH problems to determine the ultimate issues. It's also these skills that make visual approaches a relief rather than a burden.

But all of the above takes time and money. It's much easier, shorter and cheaper to go for the 80% rote solution.

I hear lots of complaining from newbs about the 1,500 hour rule. "What's 1,000 hours instructing the pattern going to teach me? 300 hours in a Airbus sim would be much better".

The secret is the that all the gobbley-gook at the end is the EASY part. FMSes, ACARS, glass and all that nonsense is just that. Someone gilding the lily to make themselves and their master's thesis project sound important. I can take any competent IFR current pilot and spin them up on a FMS in a day. Glass even less. I don't even know why those are requirements at some places. If you are a good, solid, experienced pilot, those things are only power tools, and luxury one's at that.

But back to the point. Those "1,000 hours in the pattern" really do mean a lot...you learn to read the airplane, conditions, patterns, your student, your co-workers, the weather, the airport(s), and yes, even your boss. If you're not learning anything on almost every flight, then you don't belong in the profession. You learn to smell BS when it's out there. There is NO easy button if you're doing it right.

That's the difference between two "3,000 hour" pilots. One can pull solutions seeming out of their bum, the other one gets to the end of the ECAM, and that's that.

Richman

 

[BEEF SUPREME]

So how much $$$ did AF447 cost AF?

What is the cost cutoff for good training? No low time cruise pilots etc.

Was AF447 above or below the cost threshold of properly staffing and training for the A340 program.

*I think AF447 was a 330 or 340, but I'm not feeling nerdy enough to look it up.


Sent from my iPhone using Tapatalk

 

[Beep]

Correlation is not causation, either.

The big picture problem is that you need to pay bucks to get Buck Rogers.

But if all you want is Twikki, then the bill isn't as much, and some places are OK with that.

Flying is one of those things that takes equal doses of art and science. Take a look at any GOOD competition glider pilot and you'll see what I mean. The best way, usually, to get there is to start with the basics, learn the basics, love the basics, and carefully nurture a career full of learning the sights, smells, sounds and touches of flying. In the end this grants the ability to avoid a lot of issues from the get-go ("no, Mr. Mechanic, I don't think that defering the fetzer value is appropriate under the MEL under these conditions....and, please, use Quaker State), but also the ability to see THROUGH problems to determine the ultimate issues. It's also these skills that make visual approaches a relief rather than a burden.

But all of the above takes time and money. It's much easier, shorter and cheaper to go for the 80% rote solution.

I hear lots of complaining from newbs about the 1,500 hour rule. "What's 1,000 hours instructing the pattern going to teach me? 300 hours in a Airbus sim would be much better".

The secret is the that all the gobbley-gook at the end is the EASY part. FMSes, ACARS, glass and all that nonsense is just that. Someone gilding the lily to make themselves and their master's thesis project sound important. I can take any competent IFR current pilot and spin them up on a FMS in a day. Glass even less. I don't even know why those are requirements at some places. If you are a good, solid, experienced pilot, those things are only power tools, and luxury one's at that.

But back to the point. Those "1,000 hours in the pattern" really do mean a lot...you learn to read the airplane, conditions, patterns, your student, your co-workers, the weather, the airport(s), and yes, even your boss. If you're not learning anything on almost every flight, then you don't belong in the profession. You learn to smell BS when it's out there. There is NO easy button if you're doing it right.

That's the difference between two "3,000 hour" pilots. One can pull solutions seeming out of their bum, the other one gets to the end of the ECAM, and that's that.

Richman

Exactamundo!

 

[Richman]

Exactamundo!

I'm here to say that the 1,500 hour rule was the most (inadvertently?) brilliant idea the FAA ever put into action in recent memory.

If someone actually sat down and said "Ya know, it's not just 1,500 hours, but how you get there, maybe have a little fun, and with this rule we'll force people to learn something before it's done", then I would be doubly impressed.

Richman

 

[BobDDuck]

So how much $$$ did AF447 cost AF?

What is the cost cutoff for good training? No low time cruise pilots etc.

Was AF447 above or below the cost threshold of properly staffing and training for the A340 program.

*I think AF447 was a 330 or 340, but I'm not feeling nerdy enough to look it up.


Sent from my iPhone using Tapatalk

I'd guess it cost them way, way less than the alternative of having a pipeline and training system that would put pilots in the seat that could have avoided that. Now, if they do it over and over again, I'd think the economics start to swing the other way.

 

[killbilly]

Correlation is not causation, either.

The big picture problem is that you need to pay bucks to get Buck Rogers.

But if all you want is Twikki, then the bill isn't as much, and some places are OK with that.

Flying is one of those things that takes equal doses of art and science. Take a look at any GOOD competition glider pilot and you'll see what I mean. The best way, usually, to get there is to start with the basics, learn the basics, love the basics, and carefully nurture a career full of learning the sights, smells, sounds and touches of flying. In the end this grants the ability to avoid a lot of issues from the get-go ("no, Mr. Mechanic, I don't think that defering the fetzer value is appropriate under the MEL under these conditions....and, please, use Quaker State), but also the ability to see THROUGH problems to determine the ultimate issues. It's also these skills that make visual approaches a relief rather than a burden.

But all of the above takes time and money. It's much easier, shorter and cheaper to go for the 80% rote solution.

I hear lots of complaining from newbs about the 1,500 hour rule. "What's 1,000 hours instructing the pattern going to teach me? 300 hours in a Airbus sim would be much better".

The secret is the that all the gobbley-gook at the end is the EASY part. FMSes, ACARS, glass and all that nonsense is just that. Someone gilding the lily to make themselves and their master's thesis project sound important. I can take any competent IFR current pilot and spin them up on a FMS in a day. Glass even less. I don't even know why those are requirements at some places. If you are a good, solid, experienced pilot, those things are only power tools, and luxury one's at that.

But back to the point. Those "1,000 hours in the pattern" really do mean a lot...you learn to read the airplane, conditions, patterns, your student, your co-workers, the weather, the airport(s), and yes, even your boss. If you're not learning anything on almost every flight, then you don't belong in the profession. You learn to smell BS when it's out there. There is NO easy button if you're doing it right.

That's the difference between two "3,000 hour" pilots. One can pull solutions seeming out of their bum, the other one gets to the end of the ECAM, and that's that.

Richman

You just succinctly and precisely described the gist of an incredibly difficult process I went through in deciding not to go to a pilot mill in 2006. It took a couple of years of probing JC and some other websites to figure out what I wanted. It's what you wrote up there, sorta.

Do you still teach?

 

[Richman]

Not currently. I've done a LOT of 61, 141 and 121 in the past. Taught college. CFI still current.

250-1200 hours is incredibly important in the gestation of a pilot. How you get from one end to the other determines a lot.

Richman

 

[Zapphod Beblebrox]

Yesterday was January 15. On that day in 2009, I watched US Airways flight 1449 successfully ditch in the Hudson River. Chellsely Sullenberger was the right guy at the right time to be in command of that flight. I had flown with him before and considered him to be a little stiff; but an outstanding pilot.

He used to sit on a crew room break and review the QRH. He had memory items for stuff that didn't require a memory item. There were no dual flame out memory items for the 319/310/32, at that time, unless you were at heigh altitude. He knew that without asking.

US 14449 was the ultimate "Black Swan" but he was a former ALPA airline (original PSA) safety chairman, an Airforce Academy graduate and an Airforce F-4 guy who has a checklist and memory item for everything. For the rests of us mere mortals, how do we balance automation and hand flying proficiency?

Sully once stated that he had been putting deposits on his safety account for many years so that when needed he could make a quick withdrawal from his experience account.

Is this a realistic approach. It works; but knowing Sully it is not the norm.

 

[Hacker15e]

I have it on good authority that "hope" is the proper way to address ensuring your airmanship bucket is both deep and full at the moment you really need to reach in it and pull something out to save the day (and your bacon).

 

[killbilly]

Ferchrissakes, can any of you spell his name right, given how high we hold him in esteem?

It's Chesley Sullenberger. Chesley.

Jesus. Philistines, most of you....

 

[derg]

Chelsey Handler?

 

[killbilly]

Chelsey Handler?

Heretic! Kill him! Kill him with fire!


(ok, not really. BTW - my CFI and I had a good laugh about the "cloud sight glass" today as we got our arses kicked while he was trying to teach me 8s on pylons today)

 

[Zapphod Beblebrox]

I kin fly arplanze, I jess caint spell good. That's why I become a pilut.

 

[juxtapilot]

I'm not so much concerned about our ability to hand fly an RNAV departure so much as a high and fast visual from the downwind. The knowledge of when to turn off the AP and then manipulate the aircraft by hand in those situations is more useful, and likely tells you more about the pilot's hand flying skills than almost any other situation.

Use of automation and managing the other pilot is the challenge in these cases. Often I'll leave the autopilot on and manipulate the FD panel myself.. Once the autopilot is off the PF is essentially useless-leaving the other guy to manage the panel.. The newbies don't have the muscle memory of switch position so they're slow in button pushing and knob turning. If I am planning on disconnecting the AP early I'll tell the other guy my plan and have things set up as much as possible so they have very little to do. One thing I hate seeing is the FD's "plan" differ from the PF's actions.