> > *STV has also been described as the most proportional system.[21]:83 *The > system tends to handicap extreme candidates because, to gain preferences > and so improve their chance of election, candidates need to canvass voters > beyond their own circle of supporters, and so need to moderate their > views.[25][26] Conversely, widely respected candidates can win election > with relatively few first preferences by benefitting from strong > subordinate preference support.[19] > This is on Wikipedia. This sentence alone has probably been read more times than every article on ElectoWiki combined. The bar is low, guys. (Lower than the bar on ElectoWiki, to be honest.) Does anyone want to do something about it? Take the theorems from Campbell, Kelly, and Maskin here: https://www.sciencedirect.com/science/article/abs/pii/S0165176511001972 https://scholar.harvard.edu/files/maskin/files/strategy-proofness_iia_and_majority_rule_manuscript_04.03.2020.pdf So. What you have here is a mathematical proof that everyone on this mailing list is completely right about everything. IRV and other non-Condorcet ranked methods are just plain worse than Condorcet. These results are well-known and well-cited, and show there's no reason anyone should ever use RCV as anything but a Condorcet tiebreaker. The proofs are intuitive and literally one sentence long. And yet there were zero words on this in Wikipedia before the start of this year, when I added the description to the article on Arrow's theorem! There's *still* no article on the Condorcet dominance theorems! (I haven't had time to write one, unfortunately.) I feel like I'm going crazy. There's so much important information that nobody's even *trying* to communicate to the average person! Special thanks to @Kristofer Munsterhjelm <km_elmet@t-online.de> for his past improvements to the spoiler effect article, and to @Richard, the VoteFair guy <electionmethods@votefair.org> who's provided a bit of help the past few months.

Hi CLC, For what it's worth, you're probably not going to find anyone that will help you dispute whether STV is "the most proportional". Intuitively, it seems to do a pretty good job with proportionality. I haven't read the Maskin paper nor the Merrill paper (and I'm not about to anytime soon), but can you explain why you believe that claim is so outrageous? Rob On Mon, Sep 16, 2024 at 11:38 AM Closed Limelike Curves < closed.limelike.curves@gmail.com> wrote: > *STV has also been described as the most proportional system.[21]:83 *The >> system tends to handicap extreme candidates because, to gain preferences >> and so improve their chance of election, candidates need to canvass voters >> beyond their own circle of supporters, and so need to moderate their >> views.[25][26] Conversely, widely respected candidates can win election >> with relatively few first preferences by benefitting from strong >> subordinate preference support.[19] >> > This is on Wikipedia. This sentence alone has probably been read more > times than every article on ElectoWiki combined. The bar is low, guys. > (Lower than the bar on ElectoWiki, to be honest.) > > Does anyone want to do something about it? Take the theorems from > Campbell, Kelly, and Maskin here: > https://www.sciencedirect.com/science/article/abs/pii/S0165176511001972 > > https://scholar.harvard.edu/files/maskin/files/strategy-proofness_iia_and_majority_rule_manuscript_04.03.2020.pdf > > So. What you have here is a mathematical proof that everyone on this > mailing list is completely right about everything. IRV and other > non-Condorcet ranked methods are just plain worse than Condorcet. These > results are well-known and well-cited, and show there's no reason anyone > should ever use RCV as anything but a Condorcet tiebreaker. The proofs are > intuitive and literally one sentence long. > > And yet there were zero words on this in Wikipedia before the start of > this year, when I added the description to the article on Arrow's theorem! > There's *still* no article on the Condorcet dominance theorems! (I > haven't had time to write one, unfortunately.) > > I feel like I'm going crazy. There's so much important information that > nobody's even *trying* to communicate to the average person! > > Special thanks to @Kristofer Munsterhjelm <km_elmet@t-online.de> for his > past improvements to the spoiler effect article, and to @Richard, the > VoteFair guy <electionmethods@votefair.org> who's provided a bit of help > the past few months. > ---- > Election-Methods mailing list - see https://electorama.com/em for list > info >

My point is just that describing one particular system as "the most proportional", without first defining what is meant by "proportional" or even offering an argument for this position is ridiculous. On the Condorcet dominance theorems: these are just straightforward results about how replacing a ranked voting rule X with Condorcet//X makes that voting rule "strictly better" in some way, e.g. always being more resistant to spoilers or strategic voting. On Thu, Oct 3, 2024 at 5:43 PM Rob Lanphier <roblan@gmail.com> wrote: > Hi CLC, > > For what it's worth, you're probably not going to find anyone that will > help you dispute whether STV is "the most proportional". Intuitively, it > seems to do a pretty good job with proportionality. I haven't read the > Maskin paper nor the Merrill paper (and I'm not about to anytime soon), but > can you explain why you believe that claim is so outrageous? > > Rob > > > On Mon, Sep 16, 2024 at 11:38 AM Closed Limelike Curves < > closed.limelike.curves@gmail.com> wrote: > >> *STV has also been described as the most proportional system.[21]:83 *The >>> system tends to handicap extreme candidates because, to gain preferences >>> and so improve their chance of election, candidates need to canvass voters >>> beyond their own circle of supporters, and so need to moderate their >>> views.[25][26] Conversely, widely respected candidates can win election >>> with relatively few first preferences by benefitting from strong >>> subordinate preference support.[19] >>> >> This is on Wikipedia. This sentence alone has probably been read more >> times than every article on ElectoWiki combined. The bar is low, guys. >> (Lower than the bar on ElectoWiki, to be honest.) >> >> Does anyone want to do something about it? Take the theorems from >> Campbell, Kelly, and Maskin here: >> https://www.sciencedirect.com/science/article/abs/pii/S0165176511001972 >> >> https://scholar.harvard.edu/files/maskin/files/strategy-proofness_iia_and_majority_rule_manuscript_04.03.2020.pdf >> >> So. What you have here is a mathematical proof that everyone on this >> mailing list is completely right about everything. IRV and other >> non-Condorcet ranked methods are just plain worse than Condorcet. These >> results are well-known and well-cited, and show there's no reason anyone >> should ever use RCV as anything but a Condorcet tiebreaker. The proofs are >> intuitive and literally one sentence long. >> >> And yet there were zero words on this in Wikipedia before the start of >> this year, when I added the description to the article on Arrow's theorem! >> There's *still* no article on the Condorcet dominance theorems! (I >> haven't had time to write one, unfortunately.) >> >> I feel like I'm going crazy. There's so much important information that >> nobody's even *trying* to communicate to the average person! >> >> Special thanks to @Kristofer Munsterhjelm <km_elmet@t-online.de> for his >> past improvements to the spoiler effect article, and to @Richard, the >> VoteFair guy <electionmethods@votefair.org> who's provided a bit of help >> the past few months. >> > ---- >> Election-Methods mailing list - see https://electorama.com/em for list >> info >> >