The original version (and many of these, potentially) being so framing-dependent in one’s answers is an interesting case of responding to framing being rational: you know most people would be irrationally more unwilling to step into a physical blender than press a button that does the same thing, so in that framing there’s a very likely strong red majority hence reason stronger reason to choose red. In this sense “Here’s the problem, btw everybody the Schelling point is [red/blue]” would be the most “honest” framing effect.
This may or may not point to broader principles of how framing effects work; I’ll have to think on it!
With respect to the Decision Theory Befuddler:
I think the correct game-theoretic answer would be to flip a coin.
This might be analogous to cases of the division of labor, though that’s a case where people are able to explicitly coordinate.
I regret naming that graph The Decision Theory Befuddler, because in retrospect I think it’s likely that a population of perfectly rational FDT agents would all independently settle on the coin toss strategy.
In a population of irrational agents, I’d assume that you’re just trying to vote against whichever side you expect to be the majority.
The original version (and many of these, potentially) being so framing-dependent in one’s answers is an interesting case of responding to framing being rational: you know most people would be irrationally more unwilling to step into a physical blender than press a button that does the same thing, so in that framing there’s a very likely strong red majority hence reason stronger reason to choose red. In this sense “Here’s the problem, btw everybody the Schelling point is [red/blue]” would be the most “honest” framing effect.
This may or may not point to broader principles of how framing effects work; I’ll have to think on it!
With respect to the Decision Theory Befuddler:
I think the correct game-theoretic answer would be to flip a coin.
This might be analogous to cases of the division of labor, though that’s a case where people are able to explicitly coordinate.
I regret naming that graph The Decision Theory Befuddler, because in retrospect I think it’s likely that a population of perfectly rational FDT agents would all independently settle on the coin toss strategy.
In a population of irrational agents, I’d assume that you’re just trying to vote against whichever side you expect to be the majority.