5 ms·

Did you actually run the numbers? The 737 MAX statistics are abysmal [1]. It stands at 3.08 crashes per million flights, compared to 0.06 for the 737 NG. It's t

by Xixi 7y ago

Did you actually run the numbers? The 737 MAX statistics are abysmal [1]. It stands at 3.08 crashes per million flights, compared to 0.06 for the 737 NG. It's two orders of magnitude worse. Are cars really two orders of magnitude more dangerous than planes? [1] http://airsafe.com/events/models/rate_mod.htm http://airsafe.com/events/models/rate_mod.htm

umvi 7y ago
As another commenter stated, you can mince statistics any way you want to paint any picture you want. Only two 737 MAX have ever crashed (both outside of the US), so the 3.08 figure is an extrapolation, not necessarily reality. After the software update it might have gone another 500K flights (or more) without a crash, leaving the figure at 2.0 (or smaller). Let's say you need to travel from NYC to San Diego. Let's assume a car death rate of 1.25 deaths per 100 million vehicle miles (a figure that popped up in Google). But that includes motorcyles and pedestrians, so let's play it safe and call it 1 death per 100 million miles. By now we are comparing apples (chance of death by trip) to oranges (chance of death per mile traveled) but let's press on. Distance from NYC to San Diego is 2,800 miles. So your chance of dying en route to San Diego is about 1 * 2,800 / 100,000,000 or ~30 in a million chance if using a car. Whereas with the MAX it is ~3 in a million (if you take a direct flight), or an order of magnitude more safe.
- Xixi 7y ago
  Indeed I looked into it a little bit more and found that planes are about three orders or magnitude safer than cars [1] (fatality per miles traveled). So even assuming that the MAX is two orders of magnitude less safe than the 737 NG (based on limited data), it would still be an order of magnitude safer than traveling by car. [1] I found 11 fatalities per trillion miles for planes, and 12.5 fatalities per billion miles for cars (in the US).
  - ulfw 7y ago
    But how does travel per million miles make sense? Nobody decides to fly from Singapore to Sydney or drive. It's just not comparable. It's hard to find a good statistic that would make modes of transport comparable. But if I'd pick one I'd pick time spent (are 10 hours on a plane more or less safe than 10 hours in a car)
    - WWLink 7y ago
      https://www.esurance.com/info/car/where-car-accidents-happen-most https://www.esurance.com/info/car/where-car-accidents-happen...
      - PunksATawnyFill 7y ago
        I find the whole "most accidents happen within X miles of home" argument so tired. That's like saying "most electrons are found within the vicinity of their nuclei." I'd be more interested in knowing where FATAL accidents occur, on the suspicion that most people do not live on highways and local streets are traveled at lower speeds.
    - swfsql 7y ago
      If someone has crash statistics for planes/cars by time spent in a plane/car, please ping me to it! I'm super curious edit: nvm I found it on wikipedia, for United Kingdom 1990–2000: https://en.wikipedia.org/wiki/Aviation_safety#Transport_comparisons https://en.wikipedia.org/wiki/Aviation_safety#Transport_comp... air/car deaths/h = 4.2x so... on this metric, not like "1000x safer"
      - acqq 7y ago
        > air/car deaths/h = 4.2x That number confused me (I expected to be zero point something). For the reference, the actual numbers in Wikipedia are: deaths per billion hours traveled: air = 30.8 ; car = 130 30.8 / 130 ~ 0.24
        swfsql 7y ago
- Spooky23 7y ago
  Statistically, you’re more likely to choke and asphyxiate on your dinner than to die in a plane crash. That doesn’t make a defective flight system a greater or lesser problem. It’s irrelevant. Aggregate data is tough to interpret anyway for a cross country trip. Traffic deaths per 100M miles vary from 1.83 (South Carolina) to 0.54 (Massachusetts). Also, motor vehicle aggregate numbers include all trips — if you compare common carriers, busses and rail are dramatically safer than private cars.
  - umvi 7y ago
    > That doesn’t make a defective flight system a greater or lesser problem. It’s irrelevant. Oh, I totally agree. I was just trying to put into perspective the fact that even defective flight systems are incredibly safe and that our fear of flying is often irrational...
- a3n 7y ago
  > As another commenter stated, you can mince statistics any way you want to paint any picture you want. That sounds like something lawyers would study rather than engineers.
- Iwan-Zotow 7y ago
  who cares about miles? compute it per hour traveled and see
cjbprime 7y ago
This is odd: that table suggests the MAX has flown 600k flights, but https://randy.newairplane.com/2018/05/22/737-max-a-year-of-serving-the-globe/ https://randy.newairplane.com/2018/05/22/737-max-a-year-of-s... suggests the MAX had only flown 41k total flights six months before the first crash. Adding >500k flights before the grounding seems implausible.
- kayfox 7y ago
  That article was published exactly one year after the introduction, at the time about a third of all delivered aircraft to date were delivered, theres some lag from delivery to first revenue flight, so its plausible that less than 100 had been in revenue service. So, with all the lag in getting the aircraft into service, its plausible that in the next 10 months the type could rack up 500,000 flights, since thats just a little more than 4 flights a day per aircraft on average.
  - cjbprime 7y ago
    Thanks. It still seems very surprising to me, but I agree it's plausible. I suppose someone with full access to data from e.g. FlightAware could try to see how many total logged flights there were.

by Xixi 7y ago

umvi 7y ago

As another commenter stated, you can mince statistics any way you want to paint any picture you want. Only two 737 MAX have ever crashed (both outside of the US), so the 3.08 figure is an extrapolation, not necessarily reality. After the software update it might have gone another 500K flights (or more) without a crash, leaving the figure at 2.0 (or smaller). Let's say you need to travel from NYC to San Diego. Let's assume a car death rate of 1.25 deaths per 100 million vehicle miles (a figure that popped up in Google). But that includes motorcyles and pedestrians, so let's play it safe and call it 1 death per 100 million miles. By now we are comparing apples (chance of death by trip) to oranges (chance of death per mile traveled) but let's press on. Distance from NYC to San Diego is 2,800 miles. So your chance of dying en route to San Diego is about 1 * 2,800 / 100,000,000 or ~30 in a million chance if using a car. Whereas with the MAX it is ~3 in a million (if you take a direct flight), or an order of magnitude more safe.

Xixi 7y ago
Indeed I looked into it a little bit more and found that planes are about three orders or magnitude safer than cars [1] (fatality per miles traveled). So even assuming that the MAX is two orders of magnitude less safe than the 737 NG (based on limited data), it would still be an order of magnitude safer than traveling by car. [1] I found 11 fatalities per trillion miles for planes, and 12.5 fatalities per billion miles for cars (in the US).
- ulfw 7y ago
  But how does travel per million miles make sense? Nobody decides to fly from Singapore to Sydney or drive. It's just not comparable. It's hard to find a good statistic that would make modes of transport comparable. But if I'd pick one I'd pick time spent (are 10 hours on a plane more or less safe than 10 hours in a car)
  - WWLink 7y ago
    https://www.esurance.com/info/car/where-car-accidents-happen-most https://www.esurance.com/info/car/where-car-accidents-happen...
    - PunksATawnyFill 7y ago
      I find the whole "most accidents happen within X miles of home" argument so tired. That's like saying "most electrons are found within the vicinity of their nuclei." I'd be more interested in knowing where FATAL accidents occur, on the suspicion that most people do not live on highways and local streets are traveled at lower speeds.
  - swfsql 7y ago
    If someone has crash statistics for planes/cars by time spent in a plane/car, please ping me to it! I'm super curious edit: nvm I found it on wikipedia, for United Kingdom 1990–2000: https://en.wikipedia.org/wiki/Aviation_safety#Transport_comparisons https://en.wikipedia.org/wiki/Aviation_safety#Transport_comp... air/car deaths/h = 4.2x so... on this metric, not like "1000x safer"
    - acqq 7y ago
      > air/car deaths/h = 4.2x That number confused me (I expected to be zero point something). For the reference, the actual numbers in Wikipedia are: deaths per billion hours traveled: air = 30.8 ; car = 130 30.8 / 130 ~ 0.24
      - swfsql 7y ago
Spooky23 7y ago
Statistically, you’re more likely to choke and asphyxiate on your dinner than to die in a plane crash. That doesn’t make a defective flight system a greater or lesser problem. It’s irrelevant. Aggregate data is tough to interpret anyway for a cross country trip. Traffic deaths per 100M miles vary from 1.83 (South Carolina) to 0.54 (Massachusetts). Also, motor vehicle aggregate numbers include all trips — if you compare common carriers, busses and rail are dramatically safer than private cars.
- umvi 7y ago
  > That doesn’t make a defective flight system a greater or lesser problem. It’s irrelevant. Oh, I totally agree. I was just trying to put into perspective the fact that even defective flight systems are incredibly safe and that our fear of flying is often irrational...
a3n 7y ago
> As another commenter stated, you can mince statistics any way you want to paint any picture you want. That sounds like something lawyers would study rather than engineers.
Iwan-Zotow 7y ago
who cares about miles? compute it per hour traveled and see

cjbprime 7y ago

This is odd: that table suggests the MAX has flown 600k flights, but https://randy.newairplane.com/2018/05/22/737-max-a-year-of-serving-the-globe/ https://randy.newairplane.com/2018/05/22/737-max-a-year-of-s... suggests the MAX had only flown 41k total flights six months before the first crash. Adding >500k flights before the grounding seems implausible.

kayfox 7y ago
That article was published exactly one year after the introduction, at the time about a third of all delivered aircraft to date were delivered, theres some lag from delivery to first revenue flight, so its plausible that less than 100 had been in revenue service. So, with all the lag in getting the aircraft into service, its plausible that in the next 10 months the type could rack up 500,000 flights, since thats just a little more than 4 flights a day per aircraft on average.
- cjbprime 7y ago
  Thanks. It still seems very surprising to me, but I agree it's plausible. I suppose someone with full access to data from e.g. FlightAware could try to see how many total logged flights there were.