User talk:Psephomancy: Difference between revisions
no edit summary
No edit summary |
|||
Line 92:
: Quite a lot could be done with such an idea. Consider for example that if a voting method passes the majority criterion and ordinal cloneproofness, it must also pass mutual majority, since a mutual majority's preferred set of candidates are essentially a clone set where if you eliminate all but any one of them, the remaining candidate is the majority's 1st choice, and must win. So an automatic criterion checker could suggest this for voting methods that are proven to pass the first two criteria, for example. [[User:BetterVotingAdvocacy|BetterVotingAdvocacy]] ([[User talk:BetterVotingAdvocacy|talk]]) 19:58, 21 February 2020 (UTC)
:: A mutual majority set isn't a clone set. Suppose voters constituting a bare majority rank some permutation of {A,B,C,D} above everybody else, and the remaining 49% voters rank every candidate in a random order. Then {A, B, C, D} is a mutual majority set but not a clone set. So a method of the type "determine what candidates are ranked consecutively in some order by absolutely everybody, replace them with a single candidate, and run FPTP on the result" would be cloneproof and pass majority but would fail mutual majority.
:: I get what you mean, though. The absolute deluxe option would be to have something like a theorem prover and then being able to specify implications (e.g. "Smith criterion" and "plurality criterion" => not "Mono-add-top criterion", or max two of "Mutual majority criterion", "Later-no-harm", "Monotonicity"), and have the system automatically fill criteria it can infer from already specified pass/fail data. Then Electowiki could become a theorem prover library for criterion compliances. But I have no idea how you'd integrate such a thing with the Mediawiki software, and it may well be overkill. [[User:Kristomun|Kristomun]] ([[User talk:Kristomun|talk]]) 20:52, 21 February 2020 (UTC)
[[User:Kristomun]],
Line 112 ⟶ 115:
: That's fine with me. You mean evidence that is not a citation, though? What form would that take? — [[User:Psephomancy|Psephomancy]] ([[User talk:Psephomancy|talk]]) 05:44, 12 February 2020 (UTC)
:: Things like the proofs I've given on the [[weighted positional method]] page, where the proof is given directly on the page. It might come back to bite us if people start submitting very involved or very subtly wrong proofs, but we could try to deal with that if/when it happens. [[User:Kristomun|Kristomun]] ([[User talk:Kristomun|talk]]) 20:52, 21 February 2020 (UTC)
| |||