Suggestion by Tideman: “Bottom Two Elimination Runoff”

cfrank

Tideman suggests a Condorcet-compliant version of IRV, where the bottom two least (top)-preferred candidates undergo a majoritarian preference runoff, with the loser being eliminated and the process repeating. This complies with the Smith criterion as well, I’m not sure what other criteria it satisfies (ex: I doubt it is monotonic, although it does seem to have more resistance to producing non-monotonic results than IRV--could it be monotonic?) but I thought it was an interesting concept. I think with exact ties, quasi/weak-Condorcet winners can still be eliminated which is potentially problematic, but I like the idea of this method because it can be used to shrink large candidate pools to any desired size in a simple and consistent way.

Here is an example with three candidates and a comparison with the results of other methods:

7: A>B>C
6: B>C>A
2: C>A>B

In this case, there is no Condorcet winner. This new method by Tideman, his older method Ranked Pairs, IRV, Dodgson and Young voting all agree with the ranking A>B>C.

A and B are both Bucklin winners, so depending on how a single-winner is chosen it could give A>B>C or B>A>C. Black voting (AKA Condorcet/Borda) on the other hand ranks B>A>C.

An argument against the ranking B>A>C is that if C were removed, A would beat B, meaning that using Black voting has made C a spoiler for A. In any case it isn't actually clear which ranking is preferable, the ambiguity is related to the independence of irrelevant alternatives criterion.

Andy Dienes

@cfrank Will just note that Nobel laureate Eric Maskin recently also endorsed BTR at his recent talk this summer at a conference at Columbia in How to improve ranked-choice voting and democracy

Generally speaking, he endorses any kind of Condorcet method; for a while he and another world-class economist had together written some essays proposing what is essentially Copeland//Borda (now rebranded by EVC as Ranked Robin)

rob

@cfrank Not long ago I posted this, I assume we are talking about the same thing? I usually call it BTR. Tldr: I like it.

https://www.votingtheory.org/forum/topic/230/bottom-two-runoff-condorcet-irv-hybrid