3 ms·
A better metric IMHO is "crashes involving at least one fatality" per flight. For large aircraft the fatalities are often everyone on board or nearly so. So peo
by phkahler 3y ago
A better metric IMHO is "crashes involving at least one fatality" per flight. For large aircraft the fatalities are often everyone on board or nearly so. So people care about "what is the probability that the flight I'm getting on has a fatal accident?" This is largely independent of number of passengers or miles flown.
- gruez 3y agoI don't get it, is your point that the "deaths per 100M passenger miles" is underestimating the likelihood of death? Doesn't the metric already normalize for number of passengers on a plane?
- phkahler 3y agoIt's obfuscating it behind enormous numbers. Most accidents happen on takeoff or landing, so per-flight statistics are ideal for them. Some accidents happen in flight, so I'd maybe include hours in flight somehow. But miles and people counts don't serve any purpose.
- gruez 3y agoDoes this matter for the average flyer? Sure, if you're in Hawaii and flying daily for island hopping purposes you should probably estimate your risk as being higher, but the average flyer is still going to have average risk per miles flown (by definition).