Hacker News Reader

Comment by Alex3917 | Hacker News Reader

> What would life look like if we tried to be 100% prepared for all contingencies?

No one is suggesting that it would make sense to be 100% for all contingencies. But most people wear their seat belts, even though your chances of dying in a Carrington-type event are greater than your chances of dying in a car crash. So it would seem rational to invest at least as much into preventing Carrington-type events as we invest into car safety.

View on HN

> even though your chances of dying in a Carrington-type event are greater than your chances of dying in a car crash

That can't be right. Or am I misunderstanding something? Your chances of dying in a car crash are higher than dying in a global extinction level disaster. How likely are car crashes with fatalities? How likely are extinction level disasters?

In the US, 0.01% of the population dies in car crashes each year (I didn’t misplace a decimal, it’s actually a fairly rare event).

If there’s a 1% annual chance of a single event where 1% of the population dies all at once, that makes the risk dying of dying in such an event the same as dying in a car crash. If if either of those numbers are higher than 1%, it’s more likely than dying in a car crash. I’m not sure if the comment you responded to is correct about the numbers for this particular risk, just showing how rare extreme events can be important risks.

I think people tend of overrate rare events that only hurt a few people at a time (like shark attacks and lightning strikes) and underrate rare events that hurt a lot of people at once (like pandemics, extreme weather events, or other society-scale disasters).

It hinges massively on how correct the worst-case predictions about impact are. I don't agree with the worst-case predictions but let's entertain them for a moment as a thought experiment: let's say grid power is almost entirely knocked out for a full year in and that leads to mass starvation and a general breakdown in society. 30% of the population dies of either starvation, heat/cold exposure or from th troubles.

If, and it's a big if, that came to pass just once in five hundred years then sure, the chances of dying in such an event turn out to be higher than dying in a car crash. Much higher, at that, so the real figures don't need to be anything like this extreme. Low-frequency high-impact risks can lead to that situation and it's one reason why 'the odds' is often too simplistic to be useful in assessing risk. Humans have a general cognitive blind spot when it comes to assessing this kind of thing.

We're generally terrible at predicting how high-impact, low frequency events will impact us and play out and part of that is that our collective reaction is the most unpredictable element, with potential to multiply or mitigate the impacts to a huge degree in either direction. The current pandemic illustrates that quite well, as does Texas. A CME if/when it hits will open us up to a similar opportunity (or risk) of reacting in a way that either minimizes or intensifies the damage.

> If, and it's a big if, that came to pass just once in five hundred years then sure, the chances of dying in such an event turn out to be higher than dying in a car crash.

Wait, do you mean "if you were involved in a car crash you'd be less likely to die than if you were involved in a CME"?

Is that what you and the OP were saying? If so, I could agree with that, rephrased as "the lethality of a CME is higher than that of a car crash".

But what I was thinking is that CMEs are far less likely, so their lethality doesn't matter as much when predicting what you'll die of. I'm far more likely to be involved in a car crash than in a CME, and since fatal car crashes actually occur, I'm far more likely to die in one than in a CME. Simply because the CME is very, very unlikely to happen.

To rephrase it in terms of things that actually happen in our lifetimes more than once: the lethality of plane crashes is very high: you're very unlikely to survive one. Their lethality is higher than that of car crashes. However, if I were to ask you: "is this person, Jane Random, more likely to die of a car crash or a plane crash?", you'd have to bet on the car crash -- simply because plane crashes occur far less often.

(This is purposefully excluding other, likelier causes of death, of course).

> Wait, do you mean "if you were involved in a car crash you'd be less likely to die than if you were involved in a CME"?

Car crashes kill around 35k Americans per year. Whereas let's say a CME would kill 300M Americans. Even if a CME only happens once every 500 years, your chances of dying in a CME would still be vastly greater than your chances of dying in a car crash, because 35,000 * 500 < 300,000,000.

Now granted a CME might not kill 85% of Americans, but they also happen more like every 100 years, rather than every 500. So yeah, the math should actually be pretty terrifying no matter what assumptions you plug in.

Thanks, you said it far better than I was able to!

It’s pretty weird reading that Atlantic article in a world with covid.

Reply on news.ycombinator.com