Quote:
Originally Posted by someone123
If you know the distribution you calculate expected value for cost-benefit (assign what cost you want to the tail risk). If you don't you can't do this.
|
You never know the real distribution except in toy examples or physics (and even then, you often don't really...). You always have to marginalize over uncertainties. If they are large, then yes, your tail risk blows up and that sucks....
Quote:
Maybe you will implement countermeasures anyway if they are cheap. If they're not cheap and you keep doing that for every low probability threat you will run out of resources and cause more harms eventually. The "odds don't matter, do everything you can" scenario doesn't work well in the real world with many different goals and scarce resources. Obviously there's a lot more to covid and the omicron scenarios but that's one basic intuition to have.
|
And if you ignore every low probability threat eventually (for some value of eventually...) we all die. I agree you should consider catastrophic economic impacts in decision making if they are potentially ruinous (and this calculus differs between societies), but we don't always have good choices. In many interesting ways, this mirrors climate change mitigations, though there we are further complicated that the majority of the risk is in the indeterminate future...
Anyway, at least around here, nobody is proposing any particularly expensive countermeasures for omicron (except perhaps for those who refuse to vaccinate). I don't see the Christmas gathering mandate as particularly economically onerous either (though it is inconvenient, no doubt... this will be my third Christmas away from family in a row, the first because I was at the South Pole, the second because of 2020, and now because of omicron + having a baby due soon).