To all of you who have endorsed runoff voting (IRV): IRV only sounds good when compared to our current "first past the post" plurality system, the strawman of choice for recent critics who've ignored the nice things I've said about other voting alternatives.
Even after discounting less credible claims, IRV advocates have probably convinced most of us that we should replace the plurality system, but they haven't said word one on why IRV should be the replacement. And they can't, because the other proposed systems out there are vastly superior to IRV in many ways, and some are superior in every way.
My favorite example is Approval Voting (AV). There are other, even more sophisticated systems, but AV performs nearly as well while actually saving money and even being easy to explain: Instead of casting one vote per race as we do now, each voter casts one up or down vote per candidate. The winner is simply the candidate with the most yea votes. Nay votes are treated the same as unmarked candidates, the nays serving only to expose ballot tampering.
AV can use current balloting and tabulation equipment, avoiding the expense and ballot complication caused by ranked systems like IRV. Furthermore, AV can can select one winner from any field without spoiler effects, which can eliminate primaries, which even saves money over the current system. As a bonus, eliminating primaries could shorten the campaign season to something less nauseating. Moreover, AV determines a winner by simply counting votes, making it more transparent than IRV or any other ranked system.
Even better, AV raises the bar for victory: IRV, like our current wretched system, grants automatic victory at 50% + 1 votes; all other voters become "expendable", imperiling their rights and property. By contrast, under AV, a candidate approved by 55% of voters can still be defeated by another candidate approved by 70%... or even more. AV's theoretical (but virtually impossible) victory bar is raised to 100%, so no voters, not even rivals' campaign staff, are ever expendable.
Thus, AV motivates candidates to pursue "common good" policies that can win approval from voters of any and all stripes. For libertarians like me, this dynamic highlights AV like a neon sign among all possible voting systems because it inhibits authoritarian policies that rob (or imprison) Peter to buy Paul's vote.
Indeed, when nobody's vote is expendable, a candidate could lose merely by acting and speaking impolitely about rivals and their constituencies, so AV should grant yet another bonus: elevating campaigning to a more statesmanlike and collegial level than any polar system would give us.
Contrast that to IRV, with its heavy emphasis on first place votes. With IRV's bias being second only to plurality's all-or-nothing, candidates improve their chances of winning by splitting voters out of the middle rankings. In other words, IRV rewards a candidate for attracting a new first place vote from the middle even if it means losing some other mid-ranked vote to last place.
This is especially true when a candidate faces voters who rank a leading rival first. In IRV, one will never see their other rankings, not even their 2nd, rendering them all "expendable". Thus, IRV leads to (or fails to cure) unnecessary and even harmful polarization, authoritarianism and mud-slinging. Approval voting avoids or inhibits all three.
So, in every way, AV is superior to IRV. There's nothing to be said for IRV, except that it might have some misguided momentum, and even that could be readily shifted over to AV instead. Therefore, all champions of democracy should abandon all promotion of and material support for run-off voting, replacing it with Approval Voting in every argument, plank and reform initiative.