# Talk:Participation criterion

Jump to navigation
Jump to search

## Mono-add-plump[edit | edit source]

User:BetterVotingAdvocacy, the restricted version of Participation that you claim IRV and Condorcet fail, "A can't be harmed by adding a vote that bullet-votes for A", is called mono-add-plump. Here's a link that shows that at least one Condorcet method passes it: [1]. I think the Minmax-y ones also do (Schulze, RP, etc). IRV passes mono-add-top, which is a stronger (more general) version of mono-add-plump, and thus also passes mono-add-plump. I don't know about STAR, but you should probably find a source given that the other two claims were wrong. I've removed the section for now. Kristomun (talk) 11:12, 13 May 2020 (UTC)

- Thanks for catching this. Edit: Actually, I didn't say "A can't be hurt by adding bullet votes for A", I said the voter can't be hurt by bullet voting. So consider that in IRV and STAR, adding a bullet vote could help a voter's favorite eliminate their lesser evil, and then their favorite loses later rounds. For Condorcet, while I am not admittedly not being rigorous there, I'm sure that adding bullet votes for Favorite could similarly start a Condorcet cycle that helps elect a Greater evil. BetterVotingAdvocacy (talk) 18:37, 13 May 2020 (UTC)
- Then that criterion is not a variant of Participation. Participation says "if you show up and cast a ballot that ranks X over Y, then that shouldn't change the winner from X to Y". A bullet vote for X ranks X above everybody else, so there's only a Participation violation if the outcome changes from X to someone who isn't X. Kristomun (talk) 13:38, 14 May 2020 (UTC)
- Thinking about it a bit more, it seems your criterion is closer to "If a voter shows up and equal-ranks a set S of candidates at the bottom, then that shouldn't change the outcome from someone in S to someone else in S". It's a sort of "reverse no favorite betrayal". Kristomun (talk) 13:46, 14 May 2020 (UTC)
- I came up with this variation because I wanted to point out that not only do many voting reforms fail Favorite Betrayal, just like FPTP, but they fail Participation even in the case where you limit your expressiveness to an FPTP-style vote, unlike FPTP. So with such methods, a plausible argument could be made that they're actually worse than FPTP (though I'm not saying that is actually true). Also, the thinking you just did is basically Irrelevance of Independent Ballots, which has been discussed at forum.electionscience.org as IIB. BetterVotingAdvocacy (talk) 18:15, 14 May 2020 (UTC)
- That would probably be Independence of Irrelevant Ballots, not the other way around :-) And that just says that an all-equal ballot shouldn't have an effect, not that a partially all-equal ballot should have no effect on the all-equal part. I was thinking that because your criterion is about some voters plumping for A when their honest fully specified ballot prefers X to Y, the plump-for-A ballot must not change the winner from X to Y. Since it's impossible to know who X and Y are from the ballot alone, the logical conclusion is that for any pair of candidates X and Y, the A-first ballot can't make X win instead of Y.
- In any event, my point is that it's inaccurate to say that a method "fails Participation even in the case" when "the case" isn't a special case of the Participation criterion itself. I'd either call it some kind of truncation resistance or a generalization of the IIB criterion you mentioned. I'll remove the example and clear up the description by User:Dr. Edmonds to make it more clear that the participation criterion is about cast (honest) ballots. Kristomun (talk) 18:49, 14 May 2020 (UTC)
- Alternatively, you could say it's a special case of a more strict Participation criterion failure, where the strict Participation criterion goes "adding a ballot must not change the winner from X to Y unless the ballot explicitly ranks Y ahead of X" (i.e. the ordering that must not be violated is the ranks-higher-or-equal instead of the rank-higher one). But this is nonstandard. Kristomun (talk) 18:59, 14 May 2020 (UTC)

- I came up with this variation because I wanted to point out that not only do many voting reforms fail Favorite Betrayal, just like FPTP, but they fail Participation even in the case where you limit your expressiveness to an FPTP-style vote, unlike FPTP. So with such methods, a plausible argument could be made that they're actually worse than FPTP (though I'm not saying that is actually true). Also, the thinking you just did is basically Irrelevance of Independent Ballots, which has been discussed at forum.electionscience.org as IIB. BetterVotingAdvocacy (talk) 18:15, 14 May 2020 (UTC)