Talk:Instant-runoff voting

Tactical voting resistance

I've restored the sentence "Research suggests that instant-runoff voting is very resistant to tactical voting" for now. The referenced paper shows a significant gap in worst-case resistance between IRV and Condorcet-IRV methods, and most other methods.

If "very" is too strong, perhaps something like "uncommonly resistant" could work. I think the important part to show is that there's a notable difference between the two method classes. Wotwotwoot (talk) 20:57, 19 February 2025 (UTC)

Can you pull the word / phrase from the cited article itself? I don't know what part of the article you are referring to, but I see "substantially less frequent" and "relatively strong resistance" and "unusually resistant". If Wikipedia uses words from the article, it is easier to argue that they are not over- or under-hyped.

Also, for what it is worth, the cited article uses some form of "strategic" 41 times, but some form of "tactical" only once. If that's relevant then you could edit accordingly. —Quantling (talk | contribs) 22:39, 19 February 2025 (UTC)

I think the conclusion is the most relevant here, assuming you're talking about "Four Condorcet-Hare Hybrids" as the paper. Quoting:

Table 5 summarizes the results from sections 5–7. HRSV and HRSN are abbreviations for ‘highly resistant to strategic voting’, and ‘highly resistant to strategic nomination’. (Of course, reducing the simulation results to a binary score requires the imposition of a somewhat arbitrary cut-off, but in general, the methods deemed ‘highly resistant’ in each category perform substantially better than the others.)"

The table in question shows IRV and the Condorcet-IRV hybrids as HRSV; and everything but IRV, Plurality, and Borda as HRSN.

So "highly resistant to strategic voting" should do. It might be relevant to note IRV's failure of HRSN later as well, when discussing its strategic nomination incentive; and that, in turn, suggests that we should eventually describe or mention Condorcet-IRV hybrids somewhere since they resist both strategic voting and nomination. Wotwotwoot (talk) 11:54, 20 February 2025 (UTC)

Instant runoff voting not in "plurality-with-elimination family" of voting methods

article says "Instant runoff voting falls under the plurality-with-elimination family of voting methods" but actually IRV is a majoritatian system where intent, and often achievement, is electing member with majority of votes. FPTP = plurality-with-elimination family of voting methods, where winner may or may not have majority of votes. IRV is not like FPTP. all election systems are based on first preference votes to some degree or are 100 percent. list PR for example. that does not mean it should be lumped in with FPTP. STV for example elects the most-popular candidates, the most-popular at the point in time they are elected. that does not mean STV is a plurality system, as usually defined. 68.150.205.46 (talk) 06:10, 26 February 2025 (UTC)

> all election systems are based on first preference votes to some degree

this is not true

the "intent" of FPTP is also to elect the member with the majority of votes.

STV is most certainly plural! Affinepplan (talk) 12:39, 26 February 2025 (UTC)

plurality does not imply first preferences. You can use plurality rule with other preferences (like approval, etc), and you can have a non-plurality rule based on first preferences.

instant runoff is by definition a plurality-by-elimination method, since (first-preference) the plurality rule is used to establish an order and then elimination is based on that order. Rankedchoicevoter (talk) 15:38, 26 February 2025 (UTC)

I think that the stronger argument is that the term "plurality-rule family" itself might fail WP:Notability as while the idea is intuitive I haven't seen much mention of such a "family" from credible sources. so if https://en.wikipedia.org/wiki/Plurality-rule_family is put up for deletion I might support that, but so long as that article exists I think it's reasonable to categorize IRV as under it. Affinepplan (talk) 16:51, 26 February 2025 (UTC)

That plurality rule family article is definitely in need of more sources. It's entirely based off a single paper that was published only last summer. 180 Degree Open Angedre (talk) 02:15, 28 February 2025 (UTC)

I haven't done the research to have a strong opinion on the particular words to use. However, I think that there has to be prominent mention in some way of the fact that the way IRV chooses which candidate to eliminate is a key reason that IRV fails the Condorcet criterion. Whatever IRV is doing wrong there (in the sense of losing the Condorcet criterion), let's be sure to describe it well. —Quantling (talk | contribs) 17:00, 26 February 2025 (UTC)

Does quoting a source need to imply that the author agrees with what's being quoted?

The reference from Nanson's book of Condorcet rejecting IRV is being removed. From the edit history, 68.150.205.46 states that "Nanson brought up Condorcet only to say he was wrong". But that doesn't seem relevant, unless quoting a source implies that the author agrees with the claim being backed up (in this case, that Condorcet rejected IRV). I don't see why it should, but what do you think? Wotwotwoot (talk) 11:33, 2 March 2025 (UTC)

Elimination in each round is based on the redistributed vote, not the first preference

The statement in the first paragraph "in each following round, the candidate with the fewest first-preferences (among the remaining candidates) is eliminated" is not correct, at least in terms of how we do it in Australia. The elimination is based on the fewest votes in the latest round of counting after redistribution. The first preference vote only matters for the first round of counting.

It is not that uncommon that the candidate with the second highest first preference vote is eliminated before the third ranked candidate where there are three strong candidates contesting a seat. The outcome of a ballot in some electorates is often decided by who ends up running third and this is not known after the first preference count, but only after the fourth and lower candidates have been eliminated after several rounds of counting. Often the candidate running second on first preferences slips back into third place and is eliminated, but preferences from this candidate now decide the outcome. 121.200.5.244 (talk) 16:05, 23 April 2025 (UTC)

"first-preferences (among the remaining candidates)" is intended to mean exactly the same thing as "votes in the latest round of counting after redistribution": preferences that are first in the remaining ordering of candidates. There's no reason to think of it as redistributing anything: just look at the orderings and count how many voters have the candidate top in their remaining ordering. Redistribution is a crutch to explain it to people more used to systems where you only get to vote for one candidate. —David Eppstein (talk) 17:14, 23 April 2025 (UTC)