The journal project sounds better as Adrian proposed it. Jameson: I'm not sure about the wisdom of just dealing with what we agree on--criterion-compliances and equivalences, compiance-implication relationships among criteria. Hasn't that mostly been done anyway? You spoke of all the many methods that haven't been evaluated by all of the many criteria. But surely only the methods actually being advocated need be looked at in terms of their criterion compliances. And not by all criteria either. Only by the ones that someone is advocating as important.
What we _don't_ agree on is the importance of the various different properties that can be asked of voting systems. That's the important thing that hasn't been done, on which zero progress has been made. Want to do something new and valuable? Evaluating the criteria, inviting people to justify the criteria that they advocate--that would be something new and valuable. For that matter, there's no need to wait for a new website (though I'm in favor of Democracy Chronicles hosting Jameson's journal project). I mean, there's no need to wait till that's set up, before the criteria's importance is discussed. If you want to discuss something important, than right now start a discussion about how we justify the properties that we claim are desirable, important, or necessary for voting systems. That's what would turn a cacophony into a discussion. I've told why I consider FBC to be important. It's available in Approval, Score, and Symmetrical ICT. Later-No-Help simplifies bottom-end strategy, and is available in Approval, Score, and Symmetrical ICT. Though the chicken dilemma is only a nuisance, one that can be dealt with, in Approval and Score, it can be easily gotten rid of, in a rank method. Because the chicken dilemma is the nearest thing to a "problem" for Approval, I suggest that any method claimed to improve on Approval must be one that gets rid of the chicken dilemma. So I claim that these properties listed below are important: ================================================= A) For any method: FBC, Later-No-Help ----------------------------------------------------------------------------- B) For a rank method (to justify using a rank method): No chicken dilemma And to fully justify rank-balloting: Condorcet's Criterion (preferably legitimately-defined) ------------------------------------------------------------------------------- C) Available in Approval & Score, and very much justified as "how it should be", on the grounds that it sounds a bit nonsensical to violate these: Mono-Add-Top Participation Consistency IIAC Bristow's "Pairwise-Social-Choice" criterion .......(If more people vote X>Y than Y>X, then don't elect Y) =================================================== Approval and Score have all of the above-listed properties except for group B.. I've read that Majority Judgment fails Participation and Consistency. Majority Judgment shares the chicken dilemma with Approval and Score. Condorcet methods don't meet the group C criteria. Symmetrical ICT has all of the above-listed properties, except for group C. =================================================== Which of the above-listed properties are had by Beatpath, Ranked-Pairs, or VoteFair? None? =================================================== I don't consider the group C criteria to be necessary, but it makes more sense to meet them than to fail them. Mike Ossipoff ---- Election-Methods mailing list - see http://electorama.com/em for list info