Colin, > Asking what's wrong with bullet voting is equivalent to asking what's > wrong with FPTP. No it isn't. That's like comparing people choosing to remain silent with people not being allowed to speak. > The answer is not that it subverts the system, but that > it withholds information the system would use to good effect. Unlike FPP, voluntary bullet voting provides the "information" that those voters were given the opportunity to rank some candidates above equal-bottom and chose not to. The lower preferences of the supporters of candidates with sufficient first-preference support to get on the final ballot are (as far as I can see) completely irrelevant, so I can't imagine what "use to good effect" in this case that you have in mind. > But if this summary is pessimistic, voters might indeed fill in > ranked preference ballots to a reasonable depth. There is no "reasonable depth" other than exactly what the voter chooses. > In this case, it seems > to me that they're being put to unconscionable lengths for what is only > a primary, and they have no way of knowing where to stop. If they are really so clueless as to how the system works they can take advice from their favourite candidates/parties.  Most will suggest they just bullet vote. I would think very few would advise them to rank more than three candidates. > Such criticisms are futile unless it's possible to do better; but I > had hoped that my own method was better, in that it achieved roughly the > benefits of voting to depth four at roughly the cost of voting to depth > one. Your suggested method transfers some power from voters to candidates, which I find unconscionable.  In that circumstance (along with it being just a primary) the supposed "benefits of voting to depth four" are  imaginary. Chris Benham > *Colin Champion* colin.champion at routemaster.app > <mailto:election-methods%40lists.electorama.com?Subject=Re%3A%20%5BEM%5D%20Condorcet%20meeting&In-Reply-To=%3Cb8010a11-93a6-b05a-97ff-69e3a052cd49%40routemaster.app%3E> > /Tue Aug 29 11:31:42 PDT 2023/ > > > ------------------------------------------------------------------------ > Asking what's wrong with bullet voting is equivalent to asking what's > wrong with FPTP. The answer is not that it subverts the system, but that > it withholds information the system would use to good effect. The whole > of ranked voting theory is based on exploiting the information which > bullet-voters withhold. >    I think Chris's summary of how his system might work is fair. > Supporters of minor parties give their first preferences accordingly, > and compromise with a mainstream candidate for their second preferences. > Supporters of mainstream parties (the majority) bullet vote. They don't > consider the merits of little-known alternatives because it's too much > effort, and because minor parties get squeezed out by the election > method in any case. This is very much like PR based on plurality (with a > little compromising thrown in), and unlike PR by STV except insofar as > FPTP is its limiting case. >    But if this summary is pessimistic, voters might indeed fill in > ranked preference ballots to a reasonable depth. In this case, it seems > to me that they're being put to unconscionable lengths for what is only > a primary, and they have no way of knowing where to stop. >    Such criticisms are futile unless it's possible to do better; but I > had hoped that my own method was better, in that it achieved roughly the > benefits of voting to depth four at roughly the cost of voting to depth > one. >    CJC > One of the agenda items is "rules for reducing a large field of > candidates to a field of 2 to 5". This seems to me an important topic, > since voters cannot be expected to vote in the way ranked preference > methods assume if the number of candidates is large. Presumably > proposals have been made for addressing it; unfortunately I haven't seen > them. > > One which occurs to me is this. There are two rounds, the first of which > may have up to (say) 20 candidates and reduces the number to (say) 6, > and the second of which uses ranked preferences to find a single winner > from the 6. >    In the first round, each candidate nominates up to 3 "alternates" > from the rest of the field; these are listed in order against his name > on the ballot. The ballots use FPTP format - choose *one* option - but > are processed as multiwinner STV ballots in which a vote for A is > treated as a vote A>B>C>D, where B...D are his alternates. Surplus votes > are not transferred. Choose the 6 winners as the survivors to the second > round. >    I dare say there exist better ideas, but I thought I'd mention it. >       CJC