With all the recent talk of primaries and “undemocratic caucuses” and the like, it’s pushed the idea of alternative voting systems to the front of my mind. I’ve written about them in the past, and generally posited that Instant Runoff Voting (IRV) was wildly superior to plurality voting in both choosing a winner and encouraging diversity in candidates. I made this conclusion based on limited test cases, and the fact that IRV is simple to understand and likely appealing to voters.

Well, it turns out I was wrong. Not completely, mind, but I was very wrong about IRV’s superiority in choosing an election winner. In fact, IRV fails to behave in a logical way in a surprising number of cases.

Someone recently pointed me to some voting simulations by Ka-Ping Yee. I won’t try to explain the work here – he does a fine job, tho it may take some time to comprehend – but it’s very good, and very enlightening. Based on the simulation results, it seems that Borda is superior to other methods tested, in the sense that it seems to do the best job of selecting the candidate that should best represent the overall consensus of the electorate in most cases. The job for the voter is essentially the same as for IRV; rank the candidates in order of preference. Unfortunately, no election system is foolproof and immune to cases where it fails to deliver the best result, but most of the alternatives Yee examines seem plainly preferable to plurality voting (the system we use now).

Unfortunately, this kind of election reform is likely to meet massive resistance, mostly because it opens the floor to additional players (thus “stealing” a portion of the power away from existing parties). However, increased diversity in the “marketplace of ideas” is a good thing, and while fringe candidates (such as Kucinich) still will not win, voters will be able to express approval for the positions of those candidates without worrying about a “spoiler” effect. That should have a positive result on the public discussion. And it will reduce the ability for fringe groups like the Religious Right to disproportionately effect the outcome.

I would love to see Ohio consider moving away from plurality voting for single-winner elections, and to institute an alternative such as the Borda or Condorcet methods.