5 ms·

A similar thing thing happened in 1962, during the Cuban missile crisis.[0] The math is sobering. Even if we assume 'only' a 1% chance per year, that implies a

A similar thing thing happened in 1962, during the Cuban missile crisis.[0] The math is sobering. Even if we assume 'only' a 1% chance per year, that implies a 63% probability per century. [0] https://en.wikipedia.org/wiki/Vasily_Arkhipov https://en.wikipedia.org/wiki/Vasily_Arkhipov

mach1ne 4y ago
1% per year is too high though. I’d estimate it’s closer to 1% per risk-year, e.g. Cuban Missile Crisis.
- wizofaus 4y ago
  It's not so hard to believe we're living in a 50%- probability world though (assuming roughly 99% chance of no nuclear strikes for each year since 1945). Just the missile crisis and the Petrov incident alone had a decent chance of swinging the other way. The question is if it could really be demonstrated we had barely 50% chance of making it through the next 70 odd years without some sort of self-inflicted civilization-ending event based on current human behaviour patterns, would it actually be enough to force us to change those patterns?
  - mach1ne 4y ago
    Yeah and survivorship bias extends here too. I think we could actually do the math to derive the maximum probability for a nuclear apocalypse within a given timeframe from the fact that we’re not in the middle of it.
anhner 4y ago
> a 1% chance per year, that implies a 63% probability per century. I'm not sure I'm following the math here?
- Synaesthesia 4y ago
  Maybe cause nuclear weapons only existed for 70 years ...?
  - schiffern 4y ago
    Who said otherwise? Putting it in terms of "probability per century" is intended for forward-looking projection, not backwards.
- tralarpa 4y ago
  1 - (1 - 0.01)^100
  - anhner 4y ago
    Thanks, now I understand how they reached the number but not why it's so convoluted. I only took one stats course in college. So it's P(bomb exploding in a century) = 1 - P(bomb not exploding in a century) = 1 - (P(bomb not exploding in a year)) ^ 100 = 1 - (1 - P(bomb exploding in a year)) ^ 100 Why is it not something simpler, like P(bomb exploding in a year) ^ 100 ?
    - tralarpa 4y ago
      > Why is it not something simpler, like P(bomb exploding in a year) ^ 100 ? That would be "bomb exploding EVERY year (= this year AND next year AND ...)". What we actually want is "bomb exploding this year OR bomb exploding next year OR bomb exploding this year and next year OR bomb exploding in two years OR ...". But that's too complicated, so we use the trick with 1 - P(bomb not exploding in a century).
      - anhner 4y ago
        That makes sense. Thanks for taking the time to clear up a stranger's confused mind! :)

mach1ne 4y ago

1% per year is too high though. I’d estimate it’s closer to 1% per risk-year, e.g. Cuban Missile Crisis.

wizofaus 4y ago
It's not so hard to believe we're living in a 50%- probability world though (assuming roughly 99% chance of no nuclear strikes for each year since 1945). Just the missile crisis and the Petrov incident alone had a decent chance of swinging the other way. The question is if it could really be demonstrated we had barely 50% chance of making it through the next 70 odd years without some sort of self-inflicted civilization-ending event based on current human behaviour patterns, would it actually be enough to force us to change those patterns?
- mach1ne 4y ago
  Yeah and survivorship bias extends here too. I think we could actually do the math to derive the maximum probability for a nuclear apocalypse within a given timeframe from the fact that we’re not in the middle of it.

anhner 4y ago

> a 1% chance per year, that implies a 63% probability per century. I'm not sure I'm following the math here?

Synaesthesia 4y ago
Maybe cause nuclear weapons only existed for 70 years ...?
- schiffern 4y ago
  Who said otherwise? Putting it in terms of "probability per century" is intended for forward-looking projection, not backwards.
tralarpa 4y ago
1 - (1 - 0.01)^100
- anhner 4y ago
  Thanks, now I understand how they reached the number but not why it's so convoluted. I only took one stats course in college. So it's P(bomb exploding in a century) = 1 - P(bomb not exploding in a century) = 1 - (P(bomb not exploding in a year)) ^ 100 = 1 - (1 - P(bomb exploding in a year)) ^ 100 Why is it not something simpler, like P(bomb exploding in a year) ^ 100 ?
  - tralarpa 4y ago
    > Why is it not something simpler, like P(bomb exploding in a year) ^ 100 ? That would be "bomb exploding EVERY year (= this year AND next year AND ...)". What we actually want is "bomb exploding this year OR bomb exploding next year OR bomb exploding this year and next year OR bomb exploding in two years OR ...". But that's too complicated, so we use the trick with 1 - P(bomb not exploding in a century).
    - anhner 4y ago
      That makes sense. Thanks for taking the time to clear up a stranger's confused mind! :)