This seems like a bit of a Pascal’smugging, unless you are planning on buying a LOT of tickets (and in that case, it would probably be -EV as others would likely be doing the same, so it will be more likely you’ll be sharing the prize).
Interested in arguments as to where I’m going wrong here (re the -ve agreement karma). Why isn’t this basically just throwing away money? What other +EV things should we be doing that have ~million:1 chances of success?
I didn’t downvote you. I think you’re using Pascal’s Mugging idiosyncratically.
Pascal’s Mugging is normally for infinitesimal odds and astronomical payouts, with both odds and payouts often being really uncertain.
Here odds and payout are well-defined. The odds while extreme aren’t infinitesimal.
I think we should be doing lots of things with one in a million chances. Start-ups that could change the world, promising AI research paths, running for president or prime minister. :)
I say “seems like a bit of a”; I get that it’s not literally infinitesimal odds (or an astronomical payout[1]), but it’s small enough that it’s similarly not worth going for imo.
Start-ups that could change the world, promising AI research paths, running for president or prime minister
I don’t think any of these are one-in-a-million chances for most people, let alone most EAs, and if anyone goes into them thinking they are, they should be doing something else! Hundred-to-one or even thousand-to-one shots are reasonable for EAs to be making I think, but not million-to-one (or worse). This is especially true given the current size of the community (it would make more sense going for million-to-one opportunities if there were of the order of a million other EAs going for them).
And the odds are certain, and the payout more certain than with most Pascal’s Muggings, but still not certain, as it depends on the number of other potential winners
I think personally preventing extinction for most individuals working to do so, or having an impact on how many animals are farmed through your own diet alone have similar or lower probabilities of success.
Maybe “small” donations to some of the large animal welfare charities, too.
Maybe it’s millions-to-one if you take the end goal (e.g. x-safety or abolition of factory farming), but most people are aiming at accomplishing things in smaller increments, that overall build up to the ultimate success. This action of buying lottery tickets is not cumulative in the same way. It’s 0 or 1 (and overwhelmingly likely to be 0). Donating what would be spent on tickets to EA charities (or otherwise enabling direct work) seems better.
I think buying many lottery tickets is similar to many people working on x-risks, but if you’re passing your donations/grants through a lottery first, you might significantly decrease the probability of making any difference with that money. In some cases, you might not, if you can almost guarantee an impact with the extra money. I could imagine it not changing the probability that the money helps farmed animals too much if the alternative was donating directly to a big animal charity.
OTOH, people should worry about decreasing marginal returns with the money.
This seems like a bit of a Pascal’s mugging, unless you are planning on buying a LOT of tickets (and in that case, it would probably be -EV as others would likely be doing the same, so it will be more likely you’ll be sharing the prize).
Interested in arguments as to where I’m going wrong here (re the -ve agreement karma). Why isn’t this basically just throwing away money? What other +EV things should we be doing that have ~million:1 chances of success?
I didn’t downvote you. I think you’re using Pascal’s Mugging idiosyncratically.
Pascal’s Mugging is normally for infinitesimal odds and astronomical payouts, with both odds and payouts often being really uncertain.
Here odds and payout are well-defined. The odds while extreme aren’t infinitesimal.
I think we should be doing lots of things with one in a million chances. Start-ups that could change the world, promising AI research paths, running for president or prime minister. :)
I say “seems like a bit of a”; I get that it’s not literally infinitesimal odds (or an astronomical payout[1]), but it’s small enough that it’s similarly not worth going for imo.
I don’t think any of these are one-in-a-million chances for most people, let alone most EAs, and if anyone goes into them thinking they are, they should be doing something else! Hundred-to-one or even thousand-to-one shots are reasonable for EAs to be making I think, but not million-to-one (or worse). This is especially true given the current size of the community (it would make more sense going for million-to-one opportunities if there were of the order of a million other EAs going for them).
And the odds are certain, and the payout more certain than with most Pascal’s Muggings, but still not certain, as it depends on the number of other potential winners
I think personally preventing extinction for most individuals working to do so, or having an impact on how many animals are farmed through your own diet alone have similar or lower probabilities of success.
Maybe “small” donations to some of the large animal welfare charities, too.
Maybe it’s millions-to-one if you take the end goal (e.g. x-safety or abolition of factory farming), but most people are aiming at accomplishing things in smaller increments, that overall build up to the ultimate success. This action of buying lottery tickets is not cumulative in the same way. It’s 0 or 1 (and overwhelmingly likely to be 0). Donating what would be spent on tickets to EA charities (or otherwise enabling direct work) seems better.
I think buying many lottery tickets is similar to many people working on x-risks, but if you’re passing your donations/grants through a lottery first, you might significantly decrease the probability of making any difference with that money. In some cases, you might not, if you can almost guarantee an impact with the extra money. I could imagine it not changing the probability that the money helps farmed animals too much if the alternative was donating directly to a big animal charity.
OTOH, people should worry about decreasing marginal returns with the money.
The choice is:
a) Donate $2000 to an EA charity of your choice.
b) Spend $2000 on lottery tickets with a 99.9% chance of losing[1], but an expected value of, at best, ~$2200.
Who is choosing b?
This is only factoring in the ≥$50k prizes, but I don’t think it changes the argument that much factoring in the lower prizes.