You've mentioned Arrow's theorem in the past so I won't bother.
My own personal preference is neither IRV or Condorcet. (I think of IRV as being like the Hugos without "No Award" and Condorcet as being like international figure skating circa 1999-2003.) My preference is actually the system that figure skating used internationally until around 1998 and in the US until last year. (Not part involving the scores, but the part involving the ordinals.)
It too has its quirks, but it's very good at generating a consensus winner. The results only get truly wacky when there is no consensus. It's also been studied and written up favorably academically.
If someone has a majority of the 1st place votes, he obviously wins. Otherwise, we sum the 1st and 2nd place votes and so on to determine the final ordering. (There are something like 4 or 5 levels of tie breakers.) You can see here how the results can be a bit unintuitive. The person with the most first place votes may not win if he did not get a majority of them and lots of people gave someone else their 2nd place vote since you keep summing lower placed votes until someone ends up with a majority.
It's actually a fairly simple system. I should point out that one of the reasons they changed it was because it was deemed "too complicated to explain." This doesn't explain why they replaced it with something even more complicated.
Like I said, we all know Arrow's Theorem so for any of these voting systems, someone can cook up a case where it just falls apart.
My main criticism of these systems is that to work really well people are going to have to rank all of the candidates. (Imagine what the election that put the Governator into office would have been like!) I think I remember hearing on NPR about some communities which have IRV style voting except that you only get to rank your top two. I don't think I like that very much either though.
no subject
My own personal preference is neither IRV or Condorcet. (I think of IRV as being like the Hugos without "No Award" and Condorcet as being like international figure skating circa 1999-2003.) My preference is actually the system that figure skating used internationally until around 1998 and in the US until last year. (Not part involving the scores, but the part involving the ordinals.)
It too has its quirks, but it's very good at generating a consensus winner. The results only get truly wacky when there is no consensus. It's also been studied and written up favorably academically.
If someone has a majority of the 1st place votes, he obviously wins. Otherwise, we sum the 1st and 2nd place votes and so on to determine the final ordering. (There are something like 4 or 5 levels of tie breakers.) You can see here how the results can be a bit unintuitive. The person with the most first place votes may not win if he did not get a majority of them and lots of people gave someone else their 2nd place vote since you keep summing lower placed votes until someone ends up with a majority.
It's actually a fairly simple system. I should point out that one of the reasons they changed it was because it was deemed "too complicated to explain." This doesn't explain why they replaced it with something even more complicated.
Like I said, we all know Arrow's Theorem so for any of these voting systems, someone can cook up a case where it just falls apart.
My main criticism of these systems is that to work really well people are going to have to rank all of the candidates. (Imagine what the election that put the Governator into office would have been like!) I think I remember hearing on NPR about some communities which have IRV style voting except that you only get to rank your top two. I don't think I like that very much either though.