3 ms·

I don't think that LW post says what you think it's saying? It's coining the term "fallacy of large numbers" to apply to the situation where someone is incorre

by jefftk 3y ago

I don't think that LW post says what you think it's saying? It's coining the term "fallacy of large numbers" to apply to the situation where someone is incorrectly asserting their situation has sufficiently large numbers to apply "law of large numbers" style reasoning. I agree that 5 minutes to each of 1000 people, especially 5min on your phone on a plane, are less valuable than 5k minutes to one person. But I don't think they're so massively less valuable that we can ignore them as trivial!

FrustratedMonky 3y ago
Maybe I'm mistaking that with "Law of truly large numbers" With a huge number, it is possible to make something look reasonable or un-reasonable. Wasting billions of hours of peoples time. Sounds bad, we should get rid of that. But that is also the time when people should be watching the safety video. Which can impact a few lives. Like remembering where the safety doors are. Maybe this is more along those lines. https://en.wikipedia.org/wiki/Law_of_truly_large_numbers https://en.wikipedia.org/wiki/Law_of_truly_large_numbers The law of truly large numbers (a statistical adage), attributed to Persi Diaconis and Frederick Mosteller, states that with a large enough number of independent samples, any highly implausible (i.e. unlikely in any single sample, but with constant probability strictly greater than 0 in any sample) result is likely to be observed.[1] Because we never find it notable when likely events occur, we highlight unlikely events and notice them more. The law is meant to make a statement about probabilities and statistical significance: in large enough masses of statistical data, even minuscule fluctuations attain statistical significance. Thus in truly large numbers of observations, it is paradoxically easy to find significant correlations, in large numbers, which still do not lead to causal theories (see: spurious correlation), and which by their collective number,[4] might lead to obfuscation as well
- jefftk 3y ago
  > that is also the time when people should be watching the safety video I thought we were talking about the most dangerous parts of a flight, takeoff and landing, which are not when the safety video (or presentation) happens? > With a huge number, it is possible to make something look reasonable or un-reasonable. Wasting billions of hours of peoples time. Sounds bad, we should get rid of that. I agree that we're not good at thinking about this kind of scale. I think a better way to make this at a scale that works well is to look at everything on a per-person basis. So if it saves one life per year (which I suspect is high) then we're balancing 10min of boredom against 0.0002 micromorts [1], in the same range as traveling 270ft by car. Someone who would rather ride 100mi in a car with their phone than 101mi in a car without their phone is already ok with this kind of risk. [2] [1] 1 in 4.5B chance of death [2] 230mi per micromort, so 100mi is 0.435 micromorts and 101mi is 0.439 for a difference of 0.004 micromorts. If the journey is 1.5hr then this is 9 rounds of trading off 10min of boredom for 0.0004 micromorts. Which is twice the risk of death we're positing in the airplane example.
  - FrustratedMonky 3y ago
    Yes, take off and landing. Typically the phones have to be off by the time the safety video is happening. I think they do this on purpose so you are paying attention. Seems like they occur very close together. I might be assuming too much overlap, like if there is longer delay waiting on tarmac. There probably is a lot of variance in this, so maybe it isn't really good point.

by jefftk 3y ago

FrustratedMonky 3y ago

Maybe I'm mistaking that with "Law of truly large numbers" With a huge number, it is possible to make something look reasonable or un-reasonable. Wasting billions of hours of peoples time. Sounds bad, we should get rid of that. But that is also the time when people should be watching the safety video. Which can impact a few lives. Like remembering where the safety doors are. Maybe this is more along those lines. https://en.wikipedia.org/wiki/Law_of_truly_large_numbers https://en.wikipedia.org/wiki/Law_of_truly_large_numbers The law of truly large numbers (a statistical adage), attributed to Persi Diaconis and Frederick Mosteller, states that with a large enough number of independent samples, any highly implausible (i.e. unlikely in any single sample, but with constant probability strictly greater than 0 in any sample) result is likely to be observed.[1] Because we never find it notable when likely events occur, we highlight unlikely events and notice them more. The law is meant to make a statement about probabilities and statistical significance: in large enough masses of statistical data, even minuscule fluctuations attain statistical significance. Thus in truly large numbers of observations, it is paradoxically easy to find significant correlations, in large numbers, which still do not lead to causal theories (see: spurious correlation), and which by their collective number,[4] might lead to obfuscation as well

jefftk 3y ago
> that is also the time when people should be watching the safety video I thought we were talking about the most dangerous parts of a flight, takeoff and landing, which are not when the safety video (or presentation) happens? > With a huge number, it is possible to make something look reasonable or un-reasonable. Wasting billions of hours of peoples time. Sounds bad, we should get rid of that. I agree that we're not good at thinking about this kind of scale. I think a better way to make this at a scale that works well is to look at everything on a per-person basis. So if it saves one life per year (which I suspect is high) then we're balancing 10min of boredom against 0.0002 micromorts [1], in the same range as traveling 270ft by car. Someone who would rather ride 100mi in a car with their phone than 101mi in a car without their phone is already ok with this kind of risk. [2] [1] 1 in 4.5B chance of death [2] 230mi per micromort, so 100mi is 0.435 micromorts and 101mi is 0.439 for a difference of 0.004 micromorts. If the journey is 1.5hr then this is 9 rounds of trading off 10min of boredom for 0.0004 micromorts. Which is twice the risk of death we're positing in the airplane example.
- FrustratedMonky 3y ago
  Yes, take off and landing. Typically the phones have to be off by the time the safety video is happening. I think they do this on purpose so you are paying attention. Seems like they occur very close together. I might be assuming too much overlap, like if there is longer delay waiting on tarmac. There probably is a lot of variance in this, so maybe it isn't really good point.