5 ms·

The level of certainty required for an action changes depending on the downside risk of action vs inaction. For example, if I am 20% certain that you have a pul

by dokein 8y ago

The level of certainty required for an action changes depending on the downside risk of action vs inaction. For example, if I am 20% certain that you have a pulmonary embolism, the correct action is to start treatment immediately; the risks of treatment are low relative the harm of waiting for further testing. The FAA and NTSB are rightly very conservative and cautious with respect to approving new designs, since the risk of premature approval is high and the harm of waiting a bit longer for more information is low. But this is the opposite scenario - the risk of grounding planes is operational inconvenience, versus the harm of another plane crash. The correct move here is to act even if they think the certainty is low.

js2 8y ago
> the risk of grounding planes is operational inconvenience Is it possible there is a safety risk to grounding the 737 Max that outweighs the risk of allowing it to fly?
- mikeash 8y ago
  Yes, if the alternative is even less safe. This is important when the alternative might be driving, as driving is orders of magnitude more dangerous. Safety measures that annoy people too much can end up with a net loss of life because of this. If there are other airliners that can be put into service to take up the slack, though, they’re very unlikely to be more dangerous.
  - Skunkleton 8y ago
    I doubt that enters into the FAA's calculus here. It is more likely that there is not enough information to show that the plane is unsafe (a sample size of 2 is not statistically significant). The FAA has already approved this design, and to say that there is a design flaw would cast a shadow on their own approval process, and by extension their authority. Of course, over the last 10 years or so the FAA has delegated much of the approval process to the corporations themselves. Perhaps the system really is rotten?
    - mikeash 8y ago
      No, it doesn’t enter into this decision because there are other planes that can step in, and grounding this type won’t encourage more driving. That was just an example of how safety measures can potentially backfire. The FAA absolutely does consider this in general. For example, this is a major reason they don’t require child safety seats: their analysis is that the extra expense would shift some travel to cars and ultimately kill more kids than it saves. https://www.faa.gov/news/press_releases/news_story.cfm?contentKey=1966 https://www.faa.gov/news/press_releases/news_story.cfm?conte... As for two crashes not being significant, I don’t buy it. Against such a low background rate (2017 recorded zero fatalities across all jet airliners of all types, for example), two crashes of the same type in a few months is huge. Six years ago, two non-fatal battery fires aboard 787s was enough for the FAA to ground the type until a fix could be made. In hindsight, that seems to have been the correct action. This 787 Max problem appears to be far more severe.
      - js2 8y ago
        If I had to hazard a guess at the logic behind not grounding the Max it’s something like this: We believe there may be an MCAS issue due to the preliminary Lion investigation but we haven’t seen it in our fleet and our pilots are confident they are prepared for it anyway. We don’t yet have enough data to draw any conclusion on the Ethiopian crash and so we’re currently treating it as an independent incident and a single incident doesn’t warrant grounding the plane. Would I personally feel safer on a new Southwest Max or on an older Southwest 737? I don’t know. Statistically, there have been no 737 Max crashes in the US.
  - mehrdadn 8y ago
    I think it's a fallacy that people compares chances of death/injury like a computer would. Things like being in control of your own fate matter too.
    - mikeash 8y ago
      People in general do not, but agencies like the FAA absolutely do, and that’s what I’m talking about.
  - jsharf 8y ago
    By the way, the difference between the safety factors of driving and flying are not quite so straightforward. This is one of my favorite charts, which shows that per-mile, flying is safer, but per-journey, driving is safer (which makes some sense, since flying journeys are much longer, but 80% of accidents happen in the 5 minutes during takeoff/landing, so really in some sense, takeoff and landing are more dangerous than driving to work). https://en.wikipedia.org/wiki/Aviation_safety#Transport_comparisons https://en.wikipedia.org/wiki/Aviation_safety#Transport_comp... Also, note the final entry in that table: Space Shuttle. Per-kilometer, it's hardly that much more dangerous than driving. But per Journey....
    - mikeash 8y ago
      Those numbers are for the UK in the 1990s. Commercial aviation is much safer today. The fatality rate per journey for US airlines over the last ten years is approximately 0.14 per billion journeys, several hundred times better than the number from your link. (Roughly 700 million journeys per year on average, one fatality during that period.) Cars have gotten safer too, but not nearly as much.
      - rootusrootus 8y ago
        Is there a point at which the statistics are too muddy to be useful? We have had 1 fatality in the last ten years, but in the last 10 years, 1 month, and 1 day we have had 51 fatalities.
        mikeash 8y ago
        Definitely! If you take the fatalities for the last ten years and convert that to a risk estimate, the result will have gigantic error bars. We can be confident that the real risk is greater than zero, and it's pretty clear that it's small, but the exact number is pretty much impossible to know. The farther back you look, the less noise you get in your data, but you also incorporate data that's less and less relevant to the current environment. Ten years seems like a decent cutoff, although it is a cutoff that happens to have had a wild swing recently. My original point still stands if you want to take the data back a bit further: even if you include Colgan, the resulting number is way below the 117 per billion listed on that Wikipedia page.
  - js2 8y ago
    So the argument is to ground the Max because there is a known MCAS issue and have people fly on proven older equipment. But Southwest hasn’t seen the MCAS issue in its fleet and its own pilots union feels its prepared to handle an MCAS issue should it occur. Could the reasoning be the Max is believed to be safer with US pilots at the helm than flying on the older equipment the Max is replacing? If the argument is that this is just about dollars and inconvenience to the airline, what would a 737 Max crash, now, cost Southwest in law suits and reputation? Maybe SWA is weighing this against the likelihood of another flight 1380 type incident.
    - mikeash 8y ago
      The rationale could be that it’s still safer with a pilot who knows what to do. Given how safe other aircraft are, it seems unlikely. As for financial damages, it’s human nature to underestimate risks that have a large cost to mitigate. Even if grounding the Max is more cost effective, I wouldn’t be confident in them reaching that conclusion.
    - LeifCarrotson 8y ago
      There's an additional confounding factor: Now that this has happened not just once but twice, causing hundreds of deaths, every pilot on the planet is acutely aware of this issue. Instead of Boeing publishing information about the systems and requiring pilots to receive training in the new models, the pilots themselves have essentially come up with their own pilot training course, and used the tragedies plastered across global news as the change notification system.
- djsumdog 8y ago
  People are most likely going to be shuffled around to other planes. 737s use to be for a lot of short haul flights, but they're being used increasingly for mid and even long haul flights. They're often far enough it's impractical to drive, and at least in the European market, trains are always a more expensive, but available alternative. If the FAA took action, in the US people would probably just postpone trips unless they had the personal days to drive instead (or a lot of personal days to AmTrak instead). Most people will be shuffled onto larger flights to hubs with connections via smaller aircraft; so they'll be switching to airplanes with proven safety records. It increases demand for the seats and may raise cost, but safety would actually go up if it turns out there is a problem with the 737-max8.
  - kenneth 8y ago
    Flying American or Southwest is going to be an absolute shitshow in the coming weeks, if they are truly losing such a large portion of their fleets it's going to be hard to accommodate so many passengers at once.
xbmcuser 8y ago
Who will be responsible for a crash FAA Boeing or the airline. Even if FAA doesn't ground them airlines might start doing it themselves although these are the same airlines that manhandle overbooked passengers out the plane so cant be sure they would.
irq11 8y ago
You skipped a step: demonstrate that the risk estimate is real, and not merely a guess. I can’t help but draw the parallel to anti-vax logic: ”if it’s at all possible that my child might be seriously injured from a vaccine, then I must avoid vaccines!” Setting a standard of “any risk is unacceptable because the downside risk is absolute” is also incorrect.
Scoundreller 8y ago
> the risk of grounding planes is operational inconvenience I disagree. The EU grounded planes returned to their starting point, and will have to fly yet another leg: a double dose of flight risk. How will airlines fill the gap from their MAX8 fleet? By dusting off their parked planes. Why are they parked? They could be lemons or reduced safety (e.g. Minimim Equipment Lists). Or by deferring maintenance, which is done to keep planes safe. Health care is trying to reduce its unnecessary testing habit: if they cost money and time without value, then resources are taken away from other life-saving interventions.
- linuxftw 8y ago
  > How will airlines fill the gap from their MAX8 fleet? Europe will probably be fine. Unfortunately here in the states, the Airlines colluded with "capacity discipline" [1]. If you've been on a US flight recently, you'll notice that a whole plane canceled would require another entire plane to be found. And since airlines don't seem to invest too much into their workforce, well they just don't have the extra captains and crew to run extra-hour flights late into the evening either. 1: https://www.latimes.com/business/la-fi-antitrust-airlines-lawsuit-20180618-story.html https://www.latimes.com/business/la-fi-antitrust-airlines-la...
  - ams6110 8y ago
    American passengers, assisted by SV technology, have been encouraged to be and demonstrated that they are totally indifferent to anything but price and will book with whichever carrier is least expensive. So don't complain that in order to stay in business, airlines are cutting costs to the bone and not carrying extra capacity anywhere that it's not essential.
    - linuxftw 8y ago
      You're not cutting costs by colluding with other airlines, you're keeping prices artificially high by ensuring you don't undercut each other.
      - 8y ago
cc439 8y ago
>The correct move here is to act even if they think the certainty is low. What is the correct action though? The root cause of the first crash is understood to be a complete lack of training on a single aspect of the flight control system in a particular set of circumstances (one that happens to occur in the most dangerous part of the flight envelope). From what I understand about this issue, the correct action is to institute an immediate crash course (pun intended?) to train pilots on how to recognize and manage the actions of this subsystem unless further, currently unspeculated issues with the sensors and software are discovered. >(T)he risk of grounding planes is operational inconvenience I think you're overlooking the risks involved in adding significant pressure to flight schedulers (who have to get planes in place to make up for missed routes), maintenance crews (who have to provide more flight hours for the remaining fleet to make up for the missing planes), and pilots (who will have to fly something different than what they've recently been flying). Think of how many opportunities for a mistake to happen would arise from changing something like the route, time, and (sometimes) vehicle used for the daily commute by every worker in a 2,000 person company. Tighter timetables, increased flight hours on the remaining fleet, and the need for pilot's to shake the rust off their skillset flying an airframe they haven't been in for a while are all significant risk factors that have played a part in the vast majority of flight accidents in the modern era of air travel.

by dokein 8y ago

js2 8y ago

> the risk of grounding planes is operational inconvenience Is it possible there is a safety risk to grounding the 737 Max that outweighs the risk of allowing it to fly?

mikeash 8y ago
Yes, if the alternative is even less safe. This is important when the alternative might be driving, as driving is orders of magnitude more dangerous. Safety measures that annoy people too much can end up with a net loss of life because of this. If there are other airliners that can be put into service to take up the slack, though, they’re very unlikely to be more dangerous.
- Skunkleton 8y ago
  I doubt that enters into the FAA's calculus here. It is more likely that there is not enough information to show that the plane is unsafe (a sample size of 2 is not statistically significant). The FAA has already approved this design, and to say that there is a design flaw would cast a shadow on their own approval process, and by extension their authority. Of course, over the last 10 years or so the FAA has delegated much of the approval process to the corporations themselves. Perhaps the system really is rotten?
  - mikeash 8y ago
    No, it doesn’t enter into this decision because there are other planes that can step in, and grounding this type won’t encourage more driving. That was just an example of how safety measures can potentially backfire. The FAA absolutely does consider this in general. For example, this is a major reason they don’t require child safety seats: their analysis is that the extra expense would shift some travel to cars and ultimately kill more kids than it saves. https://www.faa.gov/news/press_releases/news_story.cfm?contentKey=1966 https://www.faa.gov/news/press_releases/news_story.cfm?conte... As for two crashes not being significant, I don’t buy it. Against such a low background rate (2017 recorded zero fatalities across all jet airliners of all types, for example), two crashes of the same type in a few months is huge. Six years ago, two non-fatal battery fires aboard 787s was enough for the FAA to ground the type until a fix could be made. In hindsight, that seems to have been the correct action. This 787 Max problem appears to be far more severe.
    - js2 8y ago
      If I had to hazard a guess at the logic behind not grounding the Max it’s something like this: We believe there may be an MCAS issue due to the preliminary Lion investigation but we haven’t seen it in our fleet and our pilots are confident they are prepared for it anyway. We don’t yet have enough data to draw any conclusion on the Ethiopian crash and so we’re currently treating it as an independent incident and a single incident doesn’t warrant grounding the plane. Would I personally feel safer on a new Southwest Max or on an older Southwest 737? I don’t know. Statistically, there have been no 737 Max crashes in the US.
- mehrdadn 8y ago
  I think it's a fallacy that people compares chances of death/injury like a computer would. Things like being in control of your own fate matter too.
  - mikeash 8y ago
    People in general do not, but agencies like the FAA absolutely do, and that’s what I’m talking about.
- jsharf 8y ago
  By the way, the difference between the safety factors of driving and flying are not quite so straightforward. This is one of my favorite charts, which shows that per-mile, flying is safer, but per-journey, driving is safer (which makes some sense, since flying journeys are much longer, but 80% of accidents happen in the 5 minutes during takeoff/landing, so really in some sense, takeoff and landing are more dangerous than driving to work). https://en.wikipedia.org/wiki/Aviation_safety#Transport_comparisons https://en.wikipedia.org/wiki/Aviation_safety#Transport_comp... Also, note the final entry in that table: Space Shuttle. Per-kilometer, it's hardly that much more dangerous than driving. But per Journey....
  - mikeash 8y ago
    Those numbers are for the UK in the 1990s. Commercial aviation is much safer today. The fatality rate per journey for US airlines over the last ten years is approximately 0.14 per billion journeys, several hundred times better than the number from your link. (Roughly 700 million journeys per year on average, one fatality during that period.) Cars have gotten safer too, but not nearly as much.
    - rootusrootus 8y ago
      Is there a point at which the statistics are too muddy to be useful? We have had 1 fatality in the last ten years, but in the last 10 years, 1 month, and 1 day we have had 51 fatalities.
      - mikeash 8y ago
        Definitely! If you take the fatalities for the last ten years and convert that to a risk estimate, the result will have gigantic error bars. We can be confident that the real risk is greater than zero, and it's pretty clear that it's small, but the exact number is pretty much impossible to know. The farther back you look, the less noise you get in your data, but you also incorporate data that's less and less relevant to the current environment. Ten years seems like a decent cutoff, although it is a cutoff that happens to have had a wild swing recently. My original point still stands if you want to take the data back a bit further: even if you include Colgan, the resulting number is way below the 117 per billion listed on that Wikipedia page.
- js2 8y ago
  So the argument is to ground the Max because there is a known MCAS issue and have people fly on proven older equipment. But Southwest hasn’t seen the MCAS issue in its fleet and its own pilots union feels its prepared to handle an MCAS issue should it occur. Could the reasoning be the Max is believed to be safer with US pilots at the helm than flying on the older equipment the Max is replacing? If the argument is that this is just about dollars and inconvenience to the airline, what would a 737 Max crash, now, cost Southwest in law suits and reputation? Maybe SWA is weighing this against the likelihood of another flight 1380 type incident.
  - mikeash 8y ago
    The rationale could be that it’s still safer with a pilot who knows what to do. Given how safe other aircraft are, it seems unlikely. As for financial damages, it’s human nature to underestimate risks that have a large cost to mitigate. Even if grounding the Max is more cost effective, I wouldn’t be confident in them reaching that conclusion.
  - LeifCarrotson 8y ago
    There's an additional confounding factor: Now that this has happened not just once but twice, causing hundreds of deaths, every pilot on the planet is acutely aware of this issue. Instead of Boeing publishing information about the systems and requiring pilots to receive training in the new models, the pilots themselves have essentially come up with their own pilot training course, and used the tragedies plastered across global news as the change notification system.
djsumdog 8y ago
People are most likely going to be shuffled around to other planes. 737s use to be for a lot of short haul flights, but they're being used increasingly for mid and even long haul flights. They're often far enough it's impractical to drive, and at least in the European market, trains are always a more expensive, but available alternative. If the FAA took action, in the US people would probably just postpone trips unless they had the personal days to drive instead (or a lot of personal days to AmTrak instead). Most people will be shuffled onto larger flights to hubs with connections via smaller aircraft; so they'll be switching to airplanes with proven safety records. It increases demand for the seats and may raise cost, but safety would actually go up if it turns out there is a problem with the 737-max8.
- kenneth 8y ago
  Flying American or Southwest is going to be an absolute shitshow in the coming weeks, if they are truly losing such a large portion of their fleets it's going to be hard to accommodate so many passengers at once.

xbmcuser 8y ago

Who will be responsible for a crash FAA Boeing or the airline. Even if FAA doesn't ground them airlines might start doing it themselves although these are the same airlines that manhandle overbooked passengers out the plane so cant be sure they would.

irq11 8y ago

You skipped a step: demonstrate that the risk estimate is real, and not merely a guess. I can’t help but draw the parallel to anti-vax logic: ”if it’s at all possible that my child might be seriously injured from a vaccine, then I must avoid vaccines!” Setting a standard of “any risk is unacceptable because the downside risk is absolute” is also incorrect.

Scoundreller 8y ago

> the risk of grounding planes is operational inconvenience I disagree. The EU grounded planes returned to their starting point, and will have to fly yet another leg: a double dose of flight risk. How will airlines fill the gap from their MAX8 fleet? By dusting off their parked planes. Why are they parked? They could be lemons or reduced safety (e.g. Minimim Equipment Lists). Or by deferring maintenance, which is done to keep planes safe. Health care is trying to reduce its unnecessary testing habit: if they cost money and time without value, then resources are taken away from other life-saving interventions.

linuxftw 8y ago
> How will airlines fill the gap from their MAX8 fleet? Europe will probably be fine. Unfortunately here in the states, the Airlines colluded with "capacity discipline" [1]. If you've been on a US flight recently, you'll notice that a whole plane canceled would require another entire plane to be found. And since airlines don't seem to invest too much into their workforce, well they just don't have the extra captains and crew to run extra-hour flights late into the evening either. 1: https://www.latimes.com/business/la-fi-antitrust-airlines-lawsuit-20180618-story.html https://www.latimes.com/business/la-fi-antitrust-airlines-la...
- ams6110 8y ago
  American passengers, assisted by SV technology, have been encouraged to be and demonstrated that they are totally indifferent to anything but price and will book with whichever carrier is least expensive. So don't complain that in order to stay in business, airlines are cutting costs to the bone and not carrying extra capacity anywhere that it's not essential.
  - linuxftw 8y ago
    You're not cutting costs by colluding with other airlines, you're keeping prices artificially high by ensuring you don't undercut each other.
    - 8y ago

cc439 8y ago

>The correct move here is to act even if they think the certainty is low. What is the correct action though? The root cause of the first crash is understood to be a complete lack of training on a single aspect of the flight control system in a particular set of circumstances (one that happens to occur in the most dangerous part of the flight envelope). From what I understand about this issue, the correct action is to institute an immediate crash course (pun intended?) to train pilots on how to recognize and manage the actions of this subsystem unless further, currently unspeculated issues with the sensors and software are discovered. >(T)he risk of grounding planes is operational inconvenience I think you're overlooking the risks involved in adding significant pressure to flight schedulers (who have to get planes in place to make up for missed routes), maintenance crews (who have to provide more flight hours for the remaining fleet to make up for the missing planes), and pilots (who will have to fly something different than what they've recently been flying). Think of how many opportunities for a mistake to happen would arise from changing something like the route, time, and (sometimes) vehicle used for the daily commute by every worker in a 2,000 person company. Tighter timetables, increased flight hours on the remaining fleet, and the need for pilot's to shake the rust off their skillset flying an airframe they haven't been in for a while are all significant risk factors that have played a part in the vast majority of flight accidents in the modern era of air travel.