On Wed, Nov 30, 2005 at 03:17:23PM -0800, rob brown wrote: > I believe that condorcet elections intentionally ignore "strength of > opinion" information for the exact same practical reason. Since there is > no way to avoid collecting some strength of oopinion information (while > still collecting the information we *do* need), we have to consciously, > intentionally ignore that information in the tabulation. This is NOT a > bad thing.
I agree completely. If you want range voting to work, I think you have to provide voters with an incentive not to amplify: i.e., cast a ballot with high-strength preferences when the actual preference is weak. For example, it could simply be made inconvenient to do so (maybe you have to visit multiple voting booths). Or by charging $X or X% of your yearly earnings. Or by preventing the voter from voting in the next N elections. I don't think these make sense for US political elections, but I'm sure there are some plausible scenarios where this kind of incentive mechanism would be workable. -- Andrew ---- election-methods mailing list - see http://electorama.com/em for list info
