This method looks like one pretty natural way of measuring who should be elected.
The privacy concerns are a problem in some environments but not all. This method could thus well suit for some "real-world use" (if privacy in not a problem or if voting machines or vote counters can be trusted). Note that already e.g. basic rankings of typical Condorcet methods may violate privacy. If we want to be sure, we need large enough atomic voting areas and bullet votes, or not much more than that (e.g. short/truncated ballots and few discrete rating values only). The most efficient counting process could be one where you guess some cutoff level and then try to adjust it. One possible strategy could be that all parties (or wings) move towards exaggeration so that thy will give candidates of other parties 0 points. That way we could end up solving ties where the cutoff drops down to 0. Juho On 3.8.2011, at 19.05, Jameson Quinn wrote: > > > 2011/8/3 Peter Zbornik <[email protected]> > Hi Jameson, > > I like the slate-nominating feature it requires the nominators of the slates > to think about the "best" composition of the council and not about "their" > candidates. > This encourages deliberation and discussion across partisan "borders", I > imagine, in order to find the perfect mix. > > Slate nomination is used in Sweden a lot, where a nomination committee gets > the assignment to find "the ideal" slate. > By allowing everyone to nominate slates, this nomination committee might not > be needed, or would get some competition, I imagine. > > I like letting the voters do some deliberation and cross-partisan > communication in order to ease the pain of the computer in evaluating > zillions of slates. > > Peter > > Thanks for your positive comments. However, I have to admit that I anticipate > that in most cases, the supposedly NP-complete problem would be an "easy > case" which is resolvable using modern computation. So the winning slate > would be often be proposed not by cross-partisan deliberation, but by someone > who had a computer to evaluate zillions of slates. > > Note that another practical problem with this method is that it requires > publishing full ballot data. With even a fair number of candidates and rating > levels, that would be enough to make many individual ballots, opening up the > possibility of vote-buying and such. > > So while I think this method is quite beautiful in theory, I don't propose it > for real-world use. > > JQ > > ---- > Election-Methods mailing list - see http://electorama.com/em for list info
---- Election-Methods mailing list - see http://electorama.com/em for list info
