That doesn’t really avoid the issues in Arrow’s Theorem, merely blunts them, assuring us that we shouldn’t actually care about IIA. However, the fact that this karma scale is one-dimensional combined with the assumption that people have a singly-peaked preference function does show that this is one of those cases where Arrow’s Theorem doesn’t apply. Median is a good choice because it’s not terribly gamable.
Actually, the point of the linked article was that irrelevant alternatives aren’t. Rather, they reveal information about relative strengths of preferences IF, as Arrow’s Theorem’s assumes, you are restricted to voting methods involving ordinal ranking of the options.
Therefore, you can avoid the claimed problems by being able to express the magnitude of your preference, not just its ranking against others, which is the idea proposed here.
Arrow’s Theorem seems relevant...
Or not.
That doesn’t really avoid the issues in Arrow’s Theorem, merely blunts them, assuring us that we shouldn’t actually care about IIA. However, the fact that this karma scale is one-dimensional combined with the assumption that people have a singly-peaked preference function does show that this is one of those cases where Arrow’s Theorem doesn’t apply. Median is a good choice because it’s not terribly gamable.
Actually, the point of the linked article was that irrelevant alternatives aren’t. Rather, they reveal information about relative strengths of preferences IF, as Arrow’s Theorem’s assumes, you are restricted to voting methods involving ordinal ranking of the options.
Therefore, you can avoid the claimed problems by being able to express the magnitude of your preference, not just its ranking against others, which is the idea proposed here.
Sweet, thanks.
“One-dimensional” preferences are a special case, and I think solvable.