14
$\begingroup$

The chances of a commercial flight crash are about 1 in 1 200 000 - pretty low. If each plane makes about 40,000 flights in it's lifetime, about 1 in 30 airplanes will be involved in a crash before it reaches retirement.

Is this an accurate estimate? 1 in 30 seems like quite a large amount of airplanes to crash before they are retired.

I'm also assuming here that after a crash an airplane is written off. This may not be the case - I don't know how much a plane can be repaired.

$\endgroup$
  • 3
    $\begingroup$ 1 in 1,200,000 where did you get that number? $\endgroup$ – Simon Nov 1 '15 at 18:08
  • $\begingroup$ @Simon curiosity.com/playlists/… click "show more" - or use Ctrl + F for 1.2 Their citation is now broken. $\endgroup$ – Tim Nov 1 '15 at 18:15
  • $\begingroup$ @mins yes that answers one part - how many crash. That doesn't answer this question - what percentage of airplanes crash before they finish service. $\endgroup$ – Tim Nov 1 '15 at 18:17
  • 1
    $\begingroup$ Great question! So, statistically, flying is very safe if you are a passenger, but things look different when you are a plane (or a pilot!) $\endgroup$ – szulat Nov 2 '15 at 11:02
10
$\begingroup$

I just ran some numbers based on Wikipedia statistics. I don't know how they define "incident or accident" but hull loss is pretty straightforward. That means it was written off.

chart

All of these were introduced in the late 1960's. As you can see, the DC-10, L-1011, and 727, which are mostly retired, have pretty bad numbers. But the aircraft that are still being made and flown have much better stats. I think this is probably due to the fact that air travel gets safer all the time. The planes that had their heyday in the 70's and 80's fared poorly but the longer they continue to be produced, the less likely they are to crash.

$\endgroup$
  • 1
    $\begingroup$ Probably the best aircraft for measuring lifetime hull-loss rate is the 737 Classic: out of production long enough for most to have retired, but recent enough for much of modern air-safety practice. It's suffered 50 losses out of 1,988 made, for a 2.5% loss rate. $\endgroup$ – Mark Nov 1 '15 at 19:30
  • $\begingroup$ The list is missing Airbus a/c. I'm reading a document from Boeing (which link was provided by @Peter Kämpf in this question). Page 14, for the period 1959-2014: 1890 accidents, 616 fatal, 958 hull losses. For 2005-2014: 404 accidents, 72 fatal, 165 hull losses. This is for "jet aircraft". $\endgroup$ – mins Nov 1 '15 at 19:33
  • $\begingroup$ @mins Yeah, I didn't think of the A300. It's been around for a long time too. Your link to the document actually links to the question. Are you talking about the document linked by Mark in that question $\endgroup$ – TomMcW Nov 1 '15 at 19:41
  • $\begingroup$ @TomMcW The Boeing document link. Thanks for the feedback. In my previous comment I forgot to include the conclusion... hull loss seems not synonymous for fatality (which I expect to be more or less equivalent to crash). $\endgroup$ – mins Nov 1 '15 at 19:43
  • $\begingroup$ @mins That chart is great. Problem is it's broken down by departures and the OP asked about per plane. I noticed that the hull loss numbers are way different than the Wikipedia numbers I used. Wiki says 118 hull losses for 727; Boeing claims only 94. Wiki says 32 for DC-10; Boeing shows 27. 11 vs 4 for L-1011. That's a pretty big discrepancy. I can see where "accident" can have different definitions, but hull loss is hull loss. $\endgroup$ – TomMcW Nov 1 '15 at 19:57
9
$\begingroup$

In case nobody can find official figures, let's try to make a rough estimate:

  • From the data in the IATA study cited in the question What are the statistical probabilities of commercial aircraft accidents?, we know for year 2014 there was:

    • 100.000 flights per day,
    • A total of 12 accidents with casualties,
  • According to this study, there were about 23.000 aircraft in service in 2014,

  • Let's assume a mean 30 years aircraft lifespan for the aircraft in service in 2014,

  • Mean flights per day per aircraft = 4.3,

  • 360 fatal accidents during a period equivalent to an aircraft lifespan.

That's a rate of 1.56 % aircraft with a fatal accident, or 1 / 64.

During this time:

  • The aircraft involved in the accident will have flown 47,000 times,
  • The 64 aircraft will have totaled 3 millions flights.

Said otherwise: an significant airline with 320 (5 × 64) aircraft:

  • Will face a disaster every 6 years,
  • After 3 millions flights (based on 2014 statistics).

(updated with a worldwide fleet of 23.000 aircrafts instead of 44.000 initially estimated.)

$\endgroup$
  • $\begingroup$ Again, seems quite a high chance so I guess it is significantly overestimated. I'd be interested in an official figure if it exists, but +1. $\endgroup$ – Tim Nov 1 '15 at 18:28
  • $\begingroup$ That's not so much. A company with 122 aircraft will face a crash each 30 years. Air France with their 350 a/c will face 3 crashes every 30 years. This seems to be the correct order of magnitude. $\endgroup$ – mins Nov 1 '15 at 18:36
  • $\begingroup$ Yeah I suppose. Better than 1/30. $\endgroup$ – Tim Nov 1 '15 at 18:37
  • $\begingroup$ @Tim: I've changed the number of aircraft, and now the result is more 1/64, closer to your figure. But that's on a base of 4.3 flights per day, which seems quite high. $\endgroup$ – mins Nov 1 '15 at 21:40
  • $\begingroup$ @mins I wouldn't say 4.3 is high. Doesn't work for international widebodies but for short haul and regionals the aircraft do more legs per day. I've done plenty of 4-leg days with another crew taking the airplane after us to do more flying. $\endgroup$ – casey Nov 1 '15 at 23:35
2
$\begingroup$

The thing about statistical chances is that they reset every time the event did not happen. If a passenger has a chance of one in a billion flight hours to die in a crash, it does not mean that his chance of dying increases every time he flies.

In the casino, it is useless to track how many times red came out at the roulette - eight out of ten last balls came out red so now black has a higher chance, is not valid. Every time there is a 50/50 chance, history does not matter.

So if the plane has a chance of 1 in 1,200,000 to crash, it has that every time it takes off. It's not counting down the take-offs until it reaches 1,200,000 and then crash. It's number of crashes divided by a denominator like number of flights, number of flight hours, number of male flight attendants on board...be very very careful with denominators, and with projecting their meaningfulness.

$\endgroup$
  • $\begingroup$ On the other side, if your plane crash on the first time you fly, to you it's 100%. $\endgroup$ – Antzi Jun 15 '17 at 1:41
  • $\begingroup$ Yes, game over. $\endgroup$ – Koyovis Jun 15 '17 at 1:42
  • $\begingroup$ "There are lies, damn lies, and statistics" -Mark Twain. Your point is very valid. $\endgroup$ – FreeMan Jun 15 '17 at 12:00

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.