Hacker News Reader

Comment by astrocat | Hacker News Reader

I found this extremely insightful, specifically Arrow's impossibility theorem [1]

It really seems like this should get more attention: Every voting system is flawed. You cannot design a voting system that will, in every possible case, be "fair."

[1] https://en.wikipedia.org/wiki/Arrow%27s_impossibility_theore...

View on HN

But that doesn't at all mean that every possible voting system is equally not fair.

You have to look very closely at what the definition of "flawed" is. The dictator rule in Arrow's thm is not what it is commonly portrayed to be. All it says is that there should be no voter whose preferences are the same as those produced by the choice function. That doesn't really sound all that objectionable. Often people word the dictator rule as "one person decides the outcome" but that is not what it is at all.

And like the other response said, the fact that no scheme is "perfect" doesn't mean all schemes are equally flawed.

I'm pretty sure your characterization of the non-dictatorship rule is wrong. According to wikipedia, the rule states that there should be no voter whose preferences are always the same as those produced by the choice function:

> There is no individual, i whose strict preferences always prevail. That is, there is no i ∈ {1, …, N} such that ∀ (R1, …, RN) ∈ L(A)N, a ranked strictly higher than b by Ri implies a ranked strictly higher than b by F(R1, R2, …, RN), for all a and b.

It does not mean always as in "the same voter every time." What it means is that if more than two voters have a preference ranking which is digested by a social choice function, the outcome should not be exactly the same as any individual voter. I argue that doesn't make any sense, because there are plenty of situations where the best choice of ranking might coincide with the preferences of a particular voter, completely by coincidence.

Reply on news.ycombinator.com