You know, You're right. And respect for actually doing some research before just spouting out an answer. Now, I definately have to rework my numbers. The question is how do I filter it down to not include simple turbulance-related injuries like these.

Most of these inferential calculations are assuming plane crashes occur randomly. They don't. Frequency analysis is still useful, but not necessarily predictive. You're on the right track by trying to find correlations and causes first. You might find that on some airlines your chance is close to zero (for example) notwithstanding acts of god, because their maintenance and training and security is that good. There are actually dozens of airline companies who have never had a crash.

http://www.airsafe.com/events/nofatals.htm

Southwest Airlines has something like 17 million flights and never a crash, ever. They had one plane go off the end of a runway once, injuring two people. I'm not saying they are the safest, as they have had some maintenance violations, maybe they've just been lucky.

The OP uses the expression "Crash."

This could mean almost anything. There are fatal accidents, non fatal accidents, incidents and ... whatever.

Another poster said "Longer flights are more likely to crash than shorter ones."

I am a retired airline pilot ... 6 years USAF and 32 years with a major airline.

The longer flights are usually operated by the more senior/more experienced crews and are very safe. This does not mean the shorter flights are unsafe, but experience does count.
The shorter flights are operated by the commuters, Colgan Air at Buffalo comes to mind.

Anyway, a "crash", "accident", "incident" are about as common as having your quad Kings cracked by a straight flush.

My tip is to take a bomb on board with you as the chances of there being two bombs on board are effectively zero, infact there has never been a case of a flight with more than one bomb on board (excluding military flights).

It is more possible to die in a car crash

i've heard this before, but it still made me laugh.

what about the guy who gets on his first plane and it crashes and he dies? a plane can go down without it being your 2000th flight or whatever.

sick bump.

you mean sick first post

Also for consideration is the fact that the data you have used is from the past. Every year that goes by both air safety standards and planes are getting better. The chance of a crash this year should be sightly higher than the chance of a crash next year. (I would suspect)

There are also many other factors that would come into play like the economic crisis and rising fuel costs and everything else mentioned in other posts like the different airlines. These (and many more factors) all change the stats

I think that it would be impossible to correctly calculate the chances of dying in a plane crash although we are doing a lot better than the whole 1 in 11M BS that the simple minded have made up.

You need to keep in mind that when a plane is forced to land they consider that a "crash", even if it didn't explode.

You said 36.05(national transportation safety board) crashes, well I'd guess only 6-7 of those "crashes" actually caused fatalities, and even then not everyone on those planes died.

There is no propaganda at work, planes are safe.

Let's say the 270,000-1 applies across the board and Southwest has been lucky. What % chance did Southwest Airlines have of going 17,000,000-0 ?

[1 - (1/270,000) ]^17,000,000 = 4.52358 x 10^-28 or

0.000000000000000000000000000452358

Essentially 0: 1 in 2.2 * 10²⁷

But if we change it by a factor of 10 to 2,700,000-1, then it jumps to just 1 in 543.

The OP didn't distinguish between the big name airlines and smaller charter planes which have been much riskier, by like a factor of 10 IIRC. I don't know if this is still the case, but a few years ago they didn't have to conform to the same safety requirements. This was true even for airlines which by their name and colors seemed to just be extensions of a large airline, but in fact were run by completely different companies.

The above posts are the reason I say that airline safety isn't random and subject to the laws of probability, it's a matter of running a safe airline.

Account for adverse selection. If you remove flights in Alaska, you get completely different results. Alaskan pilots have a reputation for flying people in anything to anywhere.

http://www.youtube.com/watch?v=KeYf-rhMQIQ

Yes, the probability for that has been concluded to be non-zero, thank you for confirming it.

What would actually be interesting is looking at how the safety of air travel compares to other forms of travel (I.e. Cars) on an hourly basis. I suspect people travel by car many more hours than airplanes, which probably means the probability of dying in a given hour would be interesting. If I remember correctly Taleb briefly mentions this in Fooled by Randomness

I think it's fairly well known that driving is much more dangerous than flying on a deaths-per-passenger-mile basis. My guess is that should still be true for deaths-per-passenger-hour since the difference factor is only about 10x in distance/hour.

I've heard that you are far more likely to die on the drive to the airport than on the plane. And you are more likely to die driving to buy a lottery ticket than winning the lottery. At least that's what Charlie Epps said on NUMB3RS. I wouldn't be surprised if those were true.

Yes, but are you more likely to die driving home after buying a winning lottery ticket?

I've been looking for the real data on this. In freakonomics the authors claim that you are about equally likely to die in a plane or a car per the amount of time you spend in each vehicle. I would imagine it would have tons of factors like length of flight, weather condintions in the area you live etc.

I do think that people misrepresent the statistic frequently though by saying "You are more likely to die in a car than in a plane." Well no kidding. I'm also more likely to die in a car crash than in a fight with a lion, however I am far more likely to die every hour I spend fighting a lion than every hour I spend in a car.

Die or be killed?

I was always under the impression that driving is very dangerous, and flying in a large commercial jetliner is very safe. I'll have to re-evaluate.

I think they were talking about being killed in a crash... but that is a good point because people are probably way more likely to have heart attacks/strokes etc. on airplanes. It kind of bothered me that they just mentioned it without citing data because I thought it was really interesting. I would like to believe them because they did do a lot of good analysis and brought up interesting points in the book, however I also know that a lot of their data was discredited (FWIW I think they admitted it and were not trying to skew anything intentionally)