How Does RCV Compare?
Ranked Choice Voting compares very favorably against other single-winner and multi-winner voting methods. Although there is no perfect voting method, we believe it to be the best available for political elections.
The following chart compares RCV when used to elect single-member offices to other single-winner voting methods, using criteria we believe to be most important for political elections. The criteria and their rationale are more fully explained below the chart.
How well does the method reflect the will of cohesive political majorities? RCV is perfect in this regard, as it satisfies the Mutual Majority Criterion, so politically cohesive majorities will always elect one of the options they support. Not all Condorcet methods satisfy Mutual Majority, but the most popular ones do, including the Schulze method, and the rest only violate it in the rare case of a majority rule cycle. Two-round runoff does not satisfy Mutual Majority in general, but it will satisfy mutual majorities whose support is divided between at most 2 candidates, which is notable. Plurality only respects political cohesive majorities that are unanimous in favor of a single candidate, a weaker property known as the Majority Criterion. Approval and Range do not satisfy any criteria related to majority cohesion.
How well does the method prevent a minor candidate from causing a similar major-candidate to lose due to vote-splitting? This property is formally captured by the Independence of Clones and Independence of Spoilers criteria defined in our full Spoiler Resistance Analysis. RCV satisfies both criteria, making it free of the spoiler effect. There are Condorcet methods that satisfy Independence of Clones but at the time of this writing, and while none known to satisfy the stronger Independence of Spoilers property. Two-Round Runoff is resistant to many but not all spoilers in practice and does not formally satisfy either criteria. Plurality fails both these properties outright.
Approval and Range are more resistant to spoilers than plurality, because under some assumptions, voters can score the front-runner they like best the top score on the ballot to prevent that candidate from being "spoiled." However, the expectation that voters will behave in this fashion depends on three assumptions, which are sometimes true but often not. First, voters need to know who the front-runners are, so they require access to accurate polling data in advance. Second, there must be no more than 2 clear front-runners, otherwise the question of how best to vote to avoid spoilers is further complicated. Third, voters must be comfortable insincerely giving a front-runner the same score as their honest favorite. Whenever any of these 3 assumptions are not true, the spoiler effect remains.
How resistant is the method to efforts to manipulate the result? Every method is vulnerable to some form of strategic manipulation, but they do differ in terms of how often the method makes strategic voting advantageous and how likely voters are to use the strategy. Our Strategic Resistance Analysis finds Ranked Choice Voting to be the method that is most resistant to strategic manipulation. This concurs with practical experience with Ranked Choice Voting where strategic voting is not a concern among the jurisdictions and voters that use it.
In contrast, strategic voting under plurality is quite common, as supporters of minor candidates often strategically "compromise" to vote for a front-runner. Two-round runoff eliminates much of the incentive to compromise, but not entirely, especially in crowded fields. Condorcet voting methods are vulnerable to a number of strategies, the burying strategy in particular. Approval and Range voting are highly vulnerable to bullet-voting, compromising, and burying strategies. See the full Strategic Resistance Analysis for more details.
How often does the method elect "beats-all winners," a candidate that would win head-to-head against every other candidate in the race, when such a candidate exists? A method that always elects the beats-all winner when one exists (assuming sincere voting), is said to meet the Condorcet Criterion, a property that be definition is only satisfied by Condorcet methods. While RCV doesn't formally satisfy the Condorcet criterion, data from real RCV elections suggests it elects Condorcet candidates in nearly every single election. As of this writing, there have been about 200 Ranked Choice Voting elections held in the United States since 2004, for which full ranked ballot data is available, and the "beats-all" winner only lost one, for a Condorcet efficiency rate for RCV of over 99% for political elections in practice.
Two-round runoff likely performs relatively well in this regard, too. If that same ballot data from RCV elections is used to simulate a traditional runoff between the top-two candidates, it produces the same winner as RCV nearly every single time. However, that analysis of two-round runoff assumes that the population of voters is the same between both rounds, and we know in practice that separate runoff elections often experience severe drops in voter turnout.
Plurality, Approval, and Range also fail the Condorcet Criterion, but they also fail a far weaker property known as Condorcet Loser Criterion. Where as the Condorcet Criterion requires the "beats-all winner" to be elected, Condorcet Loser requires the candidate that would lose to every other candidate in the race head-to-head, to not be elected. RCV, Condorcet methods, and Two-round runoff satisfy Condorcet Loser.
How simple is the vote tabulation to conduct? Plurality, Two-Round Runoff, and Approval are the best in this regard, as they only require incrementing each candidate's tally by 1 for each vote. Range voting is more complicated to count, as it requires incrementing each candidate's tally from a range of scores, but it still ultimately just performs a simple sum of them. Ranked Choice Voting and Condorcet Methods use counts that are more complex than a simple arithmetic sum, and are therefore harder to explain and implement.
Does the method have an accepted version or analog method for multi-winner elections that ensures fair representation? Methods that have a natural generalization to the election of multiple candidates with proportional representation allow for single-member and multi-member offices to coexist on the same ballot in an intuitive and coherent way for the voter. Multi-winner RCV, also known as the "Proportional Representation by the Single Transferable Vote," is an accepted and well-tested method for ensuring proportional representation for multi-member districts. While there have been theoretical proposals for proportional analogs to Condorcet, Approval, and Range, they have seen scant or non-existent use and little study or advocacy. Plurality has a semi-proportional analog in the form of the Single Non-Transferable Vote (SNTV), but not a fully proportional one. It's possible to imagine a kind of semi-proportional analog of Two-Round Runoff in which SNTV is used in both rounds, but not a true proportional method.
Has the method been tested in real, competitive, political elections? This matters both to it's political viability for adoption and the degree to which the voting system is a "known quantity." RCV, Two-Round Runoff, and Plurality have all been used extensively for competitive elections around the world. Condorcet, Approval, and Range are not used for governmental elections anywhere in the world, so any claims about how they will or will not behave in practice are largely unproven. Reforms that are not well-tested face an additional political hurdle, because jurisdictions are reluctant to be "guinea pigs" for a method without a substantial track record.