The species absolutely determines the answer lol

Also markets without resolution criteria are kinda inneffective right?

Maybe polls asking for people's credences would be better or at least cheaper?

My layman guesses:

P(injury | black bear) = ~50%

P(death | black bear) = ~25%

P(injury | grizzly bear) = ~75%

P(death | grizzly bear) = ~50%

P(injury | polar bear) = ~90%

P(death | polar bear) = ~85%

The reason for the non-100%'s for the later ones isn't that you beat it, it's that it happens to not be with its cubs, not hungry, not startled, or assumes you are dead after maiming you for a while like in The Revenant

@TheAllMemeingEye What sort of poll are you thinking of? On here or on something like Mechanical Turk? And how would you break it down by species? Tyyy!

(For the sake of the market, it probably makes sense to assume that I'll encounter any of the species at the rate at which they populate Alaska – to pick a place that has all three.)

@DawnDrescher I'm thinking, for each species + injury/death combo, a poll with options for either every 5% or 10% increment, with a title like "What is the probability of [injury/death] if I encounter [species] alone in the wild? (poll because will never actually happen)".

If you want you can also make an unlinked multiple-choice market with an option for each combo, for people to predict the mean values of the polls, with a title like "What will Manifold polls predict is the probability of etc..."

@TheAllMemeingEye Okay, I can try that! Might start with the black bear. I heard they are least dangerous.

Thanks for the broken-down odds! Those are much higher than I thought, especially for the black bear! :-o

@TheAllMemeingEye Posted the first poll: https://manifold.markets/DawnDrescher/what-is-my-risk-of-serious-injury-o-6a3be78b9c56

@DawnDrescher tbf I've never met a bear or looked at official statistics so I could be vastly overestimating haha

@TheAllMemeingEye I think you are vastly overestimating, at least for black bears. Black bear encounters are very common but injuries/deaths are not. Black bears are just not that interested in picking a fight humans.

My impression is that grizzlies and polar bears are somewhat more aggressive/dangerous. I suspect that P(altercation | encounter) is still lower than your estimates (idk maybe depending on how hungry the bear is?).

If you somehow do end up in a fight with a bear, even a black bear, unless you are armed I think it's ~100% for major injury, and death likely depending on how far help is.

e.g. just googling for bears in grocery stores
https://www.huffpost.com/entry/bear-supermarket-california_n_6111f279e4b0da973b3c7393
https://www.nbcnews.com/video/watch-bear-browses-produce-aisle-in-lake-tahoe-supermarket-90942533564

@jcb haha fair enough, yeah I was basically assuming P(it attacks) = P(injury) lol