I just read this answer in astronomy.stackexchange, where a sobering point was made that the overall to-date chance of losing ones life in a spacecraft is about 4%.

While I don't want to dwell on that, is it roughly correct? What are the numerator and denominator in that ratio, and is it improving over the years? It seems to me that most, if not all, incidents have been within the atmosphere, and so far none actually in space?

  • 2
    Incidental comment (not worth a separate answer), you are right in that all fatalities have occurred during launch or re-entry/landing, or possibly during spacecraft separation before actual re-entry (Soyuz 11). No actual fatalities during the more passive regions of spaceflight (though major incidents such as fires or tank ruptures remind us we shouldn't be complacent.) – Andy Apr 25 '16 at 8:24
  • 2
    Keep in mind that the mortality rate for US Presidents in office is ~20%. – David Grinberg Apr 25 '16 at 14:37
  • 1
    @DavidGrinberg ...so you'd think someone would tell these guys to wear their seat belts, but no... – uhoh Apr 25 '16 at 15:11
  • Update No, it's definitely less! Please see all answers and comments below, but it looks like the number is roughly about 3.3% for a person who's been in a spacecraft bound space at least once, and more like 1.5% (18/1228) per individual spaceflight attempt. Thanks to @Eoin's comment here - also see here. – uhoh Apr 27 '16 at 11:31
up vote 27 down vote accepted

It depends on how you count.

  1. Flights or individual astronauts:

302 manned spaceflights, 4 of which ended with a total of 18 fatalities: 1.3% of flights.

544 people have been in Earth orbit, 18 fatalities: 3.3% of astronauts have died during spaceflight. But many astronauts have made more than one flight. Thanks to @Eoin: the number of person-flights is 1228, for a fatality rate of 1.4%.

  1. When do you take a sample: Just after the Columbia disaster, the number of astronaut fatalities could well have been 4%.

  2. Who do you include? In my view, a spaceflight includes everything from the countdown to the landing. Everything before that is a training accident. Apollo 1 is an unusual case: a training accident that took place in the spacecraft itself.

  • 2
    When you include that, do you also want to include other accidents during training? And what do you count as 'space' - does it include X-15 flights? What about SpaceShipTwo? See en.wikipedia.org/wiki/… – Hobbes Apr 25 '16 at 7:29
  • 4
    For the 544, this counts people who have been to orbit more than once as just one. I think the more relevant number is the number of person-journeys. – Gremlin Apr 25 '16 at 8:49
  • 1
    @uhoh According to en.wikipedia.org/wiki/… , it's 1228 – Gremlin Apr 25 '16 at 8:58
  • 3
    @uhoh Apollo 1 was a rehearsal four weeks before launch - even if it was in the actual spacecraft, I think that's pushing "bound for space" somewhat. – Andrew Apr 25 '16 at 10:14
  • 1
    "Currently, it's more dangerous to be ground personnel in a space program than it is to be an astronaut." I would dispute that. While it is true that by far more people on the ground have died in spacecraft accidents, there is vastly more ground personnel than astronauts involved in a mission. Also, if you can't really say "currently", because the last accident where ground personel died preparing a rocket was 12 years ago in Brazil. "historically" may be the better word here. – Rikki-Tikki-Tavi Apr 25 '16 at 14:00

(this post assumes that by going to space you mean going to orbit, if you also include high alititude suborbital missions things get more vauge).

What are the numerator and denominator in that ratio,

The numerator is pretty small, especially if you count by missions rather than by people (which makes sense because generally when a spacecraft fails it kills everyone onboard). At another answer says if you go from "countdown to landing" you get four accidents. If you also include testing on the pad prior to countdown you get five.

and is it improving over the years?

It's hard to say because the numbers are so small. 4-5 fatal accidents total on orbital missions is about one per decade on average. The last fatal accident was in 2003 but whether that means safety is improving or is just good luck is difficult to say.

Footnote on suborbital missions.

I'm aware of two suborbital craft that were considered spacecraft by their promotors and broke up killing thier pilots. I'm not aware of any comprehensive list of such craft though. I'm also not positive if the virgin galactic craft was actually on a suborbital mission at the time (my memory of the news articles at the time is it was but I may be misremebering and the wikipedia article on the crash doesn't say)



  • Peter, this is a decent attempt to provide a supplemental answer. Unfortunately, though you say "if you also include high alititude suborbital missions there were a few more accidents", you don't really bring in any new numbers. This answer would be better if you did. – called2voyage Apr 25 '16 at 14:31
  • 1
    Wikipedia mentions the X-15 crash, there was also the virgin galactic crash though i'm not positive if that test flight was intended to be suborbital or merely an in-atnosphere test. There may be others I don't know about. – Peter Green Apr 25 '16 at 14:38
  • 1
    Right, that is the type of data that would be good to bring into your answer with links to your references. – called2voyage Apr 25 '16 at 14:39
  • That's a good point! It's good to keep "sub-orbital" and "space" as separate concepts. The difference is the first one won't be there for long, while the second one has a chance. I'm trying to figure out if the SpaceShip2 (Virgin Galactic) was "bound for space", regardless of whether it would be sub-orbital or not. Was its planned mission to pass the von Kármán line? Wait - should it be called the "Kármán line" or ""von Kármán line? – uhoh Apr 25 '16 at 15:02
  • 1
    @uhoh it is surprisingly hard to determine from the accident report whether the SS2 flight was meant to carry out the full flight profile or not - ntsb.gov/investigations/AccidentReports/Reports/AAR1502.pdf – Andrew Apr 25 '16 at 17:40

Your Answer

By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.

Not the answer you're looking for? Browse other questions tagged or ask your own question.