Hacker News Reader

Comment by gruez | Hacker News Reader

Statistically speaking it's not worth worrying about. In 2021 there were less than 0.001 deaths per 100M passenger miles. Assuming these aircraft are 10x more dangerous than the alternative, that works out to 0.01 deaths per 100M passenger miles, or 0.00000026 deaths NY to SF flight. For an average american aged 38.9 years[2] with 40 years of life left[3], that works out to a loss of life of 5.4 minutes. I don't think most americans would worry over this, especially compared to other far more deadly threats like obesity.

[1] https://injuryfacts.nsc.org/home-and-community/safety-topics...

[2] https://www.census.gov/newsroom/press-releases/2023/populati...

[3] https://www.ssa.gov/oact/STATS/table4c6.html

View on HN

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.

I 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?

It'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.

Does 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).

>Statistically speaking it's not worth worrying about.

Those statistics don't really work out at all. They take number of miles and deaths, and nothing else. Let's say flown miles and driven miles are exactly the same for arguments sake. If the number of miles flown on planes were traveled by 4 million people, whereas the number of miles traveled by car were from 4 billion people, your likelihood of dying in a plane crash is FAR higher than by car if you're one of the 4 million who fly on a plane.

Miles and deaths alone are pretty poor metrics. There's also the part where I may only drive once or twice a week vs. the average which takes into account people who commute daily. So my odds of dying in a car crash are still lower. The frequency of travel seems just as pertinent as the miles traveled.

> They take number of miles and deaths, and nothing else. Let's say flown miles and driven miles are exactly the same for arguments sake. If the number of miles flown on planes were traveled by 4 million people, whereas the number of miles traveled by car were from 4 billion people, your likelihood of dying in a plane crash is FAR higher than by car if you're one of the 4 million who fly on a plane.

It's passenger miles, so the number of passengers is already factored in.

Excellent analysis.

Reply on news.ycombinator.com