If you don’t mind, I’m curious what sort of procedure and training changes occurred after the 2009 incident. The streak of no major domestic crashes (and basically no crashes at all) is really impressive and I think it should be talked about more. If the procedure and training changes are responsible for the streak we should be holding a parade or something for those responsible. Speaking of parades, I wonder what changed to make them seem lame these days. Are they just boring relative to film?
The biggest one was requiring both pilots to have ATP certificates, effectively moving the bar to co-pilot from 250 hours to 1500 hours of flying. Lots of other changes that taken on their own are small, but overall it was a pretty massive set of changes.
It's not clear whether the 1500 hour rule directly improved safety. Both the Colgan captain and the FO had more than 1500 hours at the time of the accident. What the rule did unambiguously is severely restrict the supply of new pilots to regional airlines. This significantly improved quality of life for regional pilots, who prior to Colgan were effectively making minimum wage or worse. (That may have indirectly improved safety, by making pilots more well-rested and fed...)
Another point of comparison is Europe, which has a not-dissimilar safety record, but still allows FOs to have 250 hours. That said, from what I understand, they have a very different training culture where students start in a multi-crew environment very early.
I remember reading about the typical lifestyle of a new regional pilot. Low wages, long hours, and often flying many hours as a passenger to get to work (living in Pitts, grabbing the first flight to NYC to work, etc). It sounded miserable.
Fatigue rules were changed to be more strict. Requirements for regional airlines pilots were changed to require at least a minimum of 1500 hours for both pilots, previously only one pilot was required to have 1500 hours. Standards for grading stall recovery were changed to avoid incentivising the behavior observed in the pilots in the 2009 crash.
> One engineer made a preliminary estimate that the chance of another Max crash was more than 13 times greater than FAA risk guidelines allow.
> However, this document was not completed and did not go through managerial review due to lack of detailed flight data
A plane crash is too significant for the FAA not to even disposition the analysis, meaning formally receiving, reviewing, and deciding what to do. Incomplete data would just be part of the decision process, it's not a legitimate excuse for not looking at it. Killing the report like this where no one is on record opposing it is a pretty common tactic to at least buy some time when you know what the result will be.
They were hoping to find a reason not to ground the plane. They didn't find one, so they eventually ground it.
The "safe" recommendation is to keep all planes in the sky regardless of suspected safety issues because alternative transportation is not safer than a dangerous airplane.
If the alternative to flying is not taking the trip at all, then that's certainly safest. And I imagine that's the case for many trips, especially those where the flight is more than an hour and a half or so.
(Consider that I can fly from SF to LA in an hour and a half, but it's a 6-hour drive. If my flight was canceled and I couldn't find an alternative flight, I just wouldn't go at all. Certainly there are some extreme situations where I might bite the bullet and drive, but that'd be rare.)
Put another way: when we know something is unsafe, we shouldn't keep it going just because of the theoretical risk of alternatives. And yes, I know that the risk of driving isn't "theoretical", but we can't fully predict what humans will do absent their first choice of transportation.
If I have to fly for a one-day or two-day business meeting, and I can’t find a flight, I don’t go. Driving ensures I miss the meeting. Staying put, I can at least join virtually.
Air transportation may be overall safe, but the 737 Max was a dangerous aircraft that resulted in hundreds of deaths. Youer position is extraordinarily peurile.
How about, wildly dangerous relative to other commercial passenger jets? Second in fatal accidents per million miles flown only to the Concorde, which didn't fly very much. Why would you want to continue flying a plane that was statistically 5-10x more fatal than peers? Surely there's no good reason to subject passengers to that, even if it were still relatively safer than automobile fatalities or whatever else you might compare to, not when safer aircraft are generally available?
(Note this data is only up to 2019 for 737 models, but that's exactly the time period at which decisions had to be made, of whether to continue flying it or not).
The whole point of the 737 max was that the pilots wouldn't be trained on the "max" part of the 737 -- they were qualified for the 737 and there's some hidden emulator making the "max" fly like a 737.
As soon as the emulator got bad feedback from the one sensor that was used to map "max" airframe behavor to "737" controls, all bets were off. And because the pilots weren't allowed to know they were in a Max not a 737 (which would require further training that would void the advantages of the Max) there was no SOP for how to deal with "your Max is blind and can't pretend to be a 737"
The "poorly trained" pilots weren't test pilots -- pilots aren't supposed to be able to improvise and react to things off book.
“Undertrained” because Boeing told them extra training g wasn’t necessary.[1]
Both pilots had multiple thousands of flight hours (well beyond the minimum 1500 required to co-pilot in the US). Only the co-pilot of the Ethiopian Air flight would have been deemed too new to fly for a US airline (and even then, only post-Colgan Air 3407).
IIRC the FO (copilot) of the Ethiopian Air flight correctly identified MCAS as the source of the problem, but was struggling to trim the airplane manually because it was so far out of trim that the load on the horizontal stabilizer jackscrew made it difficult or impossible for him to adjust.
This possibility was mentioned in the AFM, but the recommended response - pitch nose-down to decrease the load - was not feasible in the situation.
I thought the story was that the normal training should have sufficed, and that it did in some cases where the same issue occurred and the pilots were able to handle it (themselves not having any special training either).
The argument was that the problem could be handled as any other pitch-trim runaway situation, for which pilots were already trained.
There were, however, reasons to be skeptical. Firstly, unlike in most other cases, it was not designed to halt when opposed by the pilot’s use of the control yoke (to do so would defeat the purpose of MCAS.) Secondly, it was found in flight testing that MCAS needed to be made much more aggressive, yet the original decision was allowed to stand. Thirdly, it could stop on its own accord, only to start up again a few seconds later.
Consequently, while MCAS failure is nominally handled as a form of trim runaway, it did not present itself to pilots like the trim runaway scenarios they had trained for.
So it seems like successfully handling the situation hinges on how much the pilots have abstracted the training vs applying it by rote. This abstract vs rote distinction is a common theme in Westerners looking down on other cultures, so I guess it lines up with the “third world” comments on the situation.
Except in the second crash, as I recall it, the problem was correctly IDed, but manually adjusting the trim was impossible without adding nose down (to unload the control surface). This led the co-pilot to re-enable auto-trim, which re-enabled MCAS, which then drove the plane into the ground.
There may have been a path to saving the aircraft, but it definitely didn't seem at all intuitive. And figuring it out while the plane is actively trying to crash itself is a pretty big ask.
If a successful resolution of the problem depended on pilots making the right abstractions from their training for prior versions, then that would, in itself, establish that it was a serious error to withhold information about how MCAS operated.
I am also very familiar with both catastrophic failures of 737 MAX and I can be extremely certain that undertrained pilot is only one of the factor. Boeing's extremely fail-unsafe and irredundant design of MCAS causes the failure in the first place, and they do not provide enough training material related to this system (in the first incidence they providie not training at all. MCAS is not even mentioned in the pilot's manual, nor does most maintainence crew know the existence of such system). Blaming people, especially pilot, is the easiest way to end an air crash investigation, but the industry will never be safer if nobody take the root cause seriously.
Except no, at the time of the second crash, there had been over 20 reports of un-commanded nose down inputs by pilots. The two flights just didn't manage to diagnose the issue or turn off MCAS before it crashed the aircraft.
There was IIRC 8~16 hours of training that was intended to be mandated, but Boeing told airlines that training was not necessary. So many (not all) pilots never learned the deeper mechanisms of MCAS. This would reduce costs, and increase desirability of the aircraft.
But it gets even worse! The FAA approved Boeing's request to remove a description of MCAS from the aircraft manual! So no, most pilots wouldn't even be able to learn about it even if they wanted to!
There was 0 redundancy. The MCAS was directed by ONE SINGLE AoA (Angle of Attack) SENSOR! No backups!
Even worse! Boeing was aware of the problems as of 2016! They even knew MCAS violated Boeing's own design documents and rules for the max8!
And yet here you are blaming the pilots instead of the money grubbers at the top.
I doubt this is a feasible calculation. Firstly, grounding one type will not shut down air transport (though it will degrade service, complicate operations, and possibly have unanticipated consequences.) Secondly, the travel being disrupted will not all be completed by other means. Thirdly, it is not possible to estimate the risk of not grounding as the cause of the problem is not well-understood.
> The "safe" recommendation is to keep all planes in the sky regardless of suspected safety issues because alternative transportation is not safer than a dangerous airplane.
You're presumably taking the average deaths per miles and applying it in the sense of "those people would drive and that's more dangerous."
...that's....not how that works.
You assume:
* they would take alternate transportation, instead of canceling the trip or booking another flight not on a MAX. At the time there were just 387 MAX aircraft in the world in service.
* that said transportation would be passenger cars. Trains and busses are 10x (or more) safer than passenger cars per passenger mile. They're so safe that the danger to a passenger is basically insignificantly higher compared to commercial jet flights (and by the way, the industry very conveniently separates out "commuter" flights, which are FAR more dangerous.)
* that the overall safety of the commercial jet fleet is applicable to flights on a specific model of jet, which were clearly more dangerous, with not just two crashes in close proximity, but several close calls and almost certainly numerous other incidents that went unnoticed or unreported
The “safe” recommendation is to ground all the planes of that model any time there is a suspected issue.
That said, the FAA has a long history of not implementing safety recommendations. NASA had this issue with the space shuttle.