On page 62 of David Deutsch's excellent The Beginning of Infinity he writes:

a typical child born in the United States today is more likely to die as a result of an astronomical event than a car accident

According to the National Safety Council the lifetime probability of dying in a motor vehicle accident is huge: over 1%, and by far the largest of the non-intentional non health related causes. Is the lifetime chance of dying from an astronomical event greater than that?

  • Or his he referring to the UK, where the odds are about a third of those in the US. (medicine.ox.ac.uk/bandolier/booth/Risk/trasnsportpop.html)
    – raxacoricofallapatorius
    Nov 5, 2011 at 13:23
  • 1
    Or could it be that "astronomical event" includes things like coronal mass ejections (books.nap.edu/catalog.php?record_id=12507#description) and even skin cancers?
    – raxacoricofallapatorius
    Nov 5, 2011 at 20:40
  • Wikipedia says that about 30,000 people die each year in the USA from road accidents. en.wikipedia.org/wiki/….
    – Jase
    Nov 30, 2012 at 13:39

3 Answers 3


Here's what the book's website says:

Correction: That should read "an aeroplane accident". A true comparison with airplane accidents appears on p207, using slightly different statistics.

Note: The probability of dying in a car accident is at least 12 times and perhaps as much as 100 times as great as that of dying from an astronomical event. As road safety improves, this will presumably continue to fall rapidly during the lifetimes of children born today -- but then, so will the probability of dying from an astronomical event.

So it is, in fact, an error.

  • 2
    and even that seems extremely fishy as aircraft accidents are exceedingly rare (far more so than are fatal car crashes), especially ones involving Americans (most of them happen in areas Americans don't tend to travel in large numbers, like 3rd world countries).
    – jwenting
    Nov 7, 2011 at 13:22
  • 1
    If you spend as much time flying as you do driving, then both activities are almost equally dangerous.
    – ESultanik
    Nov 7, 2011 at 14:47
  • 1
    @jwenting: doesn't that rareness make it less fishy that another improbable event (astronomical event) is more likely?
    – Oddthinking
    Nov 7, 2011 at 15:17
  • 1
    I think the comparison is if you are in an airplane accident you are likely to die in X% of them. Where as you are likely to die in Y% of automobile accidents. And X% > Y%. So apples are not as juicy as oranges.
    – Chad
    Nov 7, 2011 at 15:43
  • 1
    @Oddthinking that sounds more plausible.
    – jwenting
    Nov 9, 2011 at 8:31

According to Bad Astronomy, astronomer Alan Harris estimates that the probability of dying due to an asteroid is about 1 in 700,000.

As for other forms of astronomical events, one could hypothesize that non-Earth based radiation could cause deaths in the form of cancer, but that would likely be quite difficult to deduce, unless that person was an astronaut, which is pretty low.

Bottom line is, I think it's pretty unlikely to occur, much lower than the probability of dying in a car accident.

  • 1
    So (according to that estimate) a car is about 8,000 times more likely to kill you.
    – raxacoricofallapatorius
    Nov 5, 2011 at 15:03
  • 1
    All cars as a whole. Not some particular homicidal car. :)
    – Andrew
    Nov 5, 2011 at 17:28
  • Could "astronomical event" be more broadly defined than just asteroids (see comment above)?
    – raxacoricofallapatorius
    Nov 5, 2011 at 20:03
  • 4
    @Jonathan. Tell that to the dinosaurs.
    – LarsTech
    Nov 6, 2011 at 17:45
  • 1
    @larsTech thankfully that was an anomaly :)
    – Jonathan.
    Nov 6, 2011 at 23:12

To give an answer to this question we must first ask ourselves what we are comparing to what. If we are comparing events occurred in the last 50 years then you might be right to say that the chance is quite low for us to dye from an astronomical event than say a car accident.

All depends on the statistical sample you are considering. The statement that you are trying to understand is based on much bigger statistical sample going back to the history of all earth. If you estimate the amount of fatal car accidents or calculate the fatal accidents involving a passenger jet and compare this with the number of catastrophic astronomical events occurred during the existence of earth you will reach to the same conclusion that you are questioning.

For proper explanation I suggest you visit http://star.arm.ac.uk/impact-hazard/ address and follow the links to the specific presentations and documents.

  • 2
    You are arguing that the intention of the author was "Over all of human history, more people have died from astronomical events than car accidents." However, the author clearly states that he is referring to "a typical child born in the United States today". I suggest your interpretation is not tenable given this wording.
    – Oddthinking
    Nov 7, 2011 at 0:40
  • OK, this is quickly getting into mode of forum. To be precise I'm not arguing about one or the other point. I'm simply stating a fact that if you are talking about "a chance" of this or that you have to bare in mind also the statistical sample over which you are calculating the chance. The point is although the astronomical events are rare but they are more devastating than the car accident. I really urge you to follow the links and read the background of this all story. And why was this question migrated to skeptics? Nov 7, 2011 at 14:48
  • I am sorry if I have misinterpreted your point. I'm not sure why the Astronomer mods felt it didn't fit their format, but it does fit ours. Part of that is we generally like people to extract and summarise the key parts of their references, to protect against link rot and to make their arguments clearer. In this case, I am still at loss as to why the original claim isn't sufficient to specify the predicted range of risk. I am inviting you to clarify that for everyone in your answer.
    – Oddthinking
    Nov 7, 2011 at 15:05

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .