At 07:43 AM 4/29/2007, Juho wrote: > > Range *also* allows the majority to rule. > >Ok, if the voters go use the Approval strategies.
I just want to underscore this. The majority has a choice. It can insist upon its first preference and prevail by bullet-voting for it. Or it can allow a *better* overall outcome by more accurately estimating its satisfaction with other candidates. I assume that, in doing this, the majority will only allow significant "leakage" of power to non-preferred candidates who are acceptable to the majority. In voting in this way, a self-aware majority can decide which is more important: getting its first preference, or maximizing the unity of the society around the election outcome. That unity has a power and value all its own. Now, what is the best vote if you don't know if you are in the majority? There is a risk involved in underrating a candidate who is not your first preference, just as there is a risk in overrating such a candidate. My sense is that the most powerful strategy, in the zero-knowledge case, is to simply vote sincerely. I.e., max rate your favorite, min-rate your least-liked, and let the rest fall where they may, depending on how strong the preferences are. I can see a common voting pattern as being *mostly* Approval style. But it is the exceptions that are interesting in Range! I'm coming to think of "partial votes." In Approval, you can segregate candidates into two classes; you vote for one class and not for the other, the effect being that you abstain from all pairwise elections between members of one class and participate fully in all pairwise elections between members of differing classes. Range is really quite similar, except that, in addition to these two options, there are intermediate options, where one *partially* abstains, or partially participates, same difference. You can fully participate in any election where your preference is very strong and fully abstain where your preference is very weak. But where your preference is between these extremes, you can partially participate. From the point of view of decision-making theory -- entirely aside from politics -- this is so obviously superior that it's a wonder that we haven't been trying to apply this to politics long ago. The idea, we so commonly see, that Range allegedly rewards strong preference ... is absolutely true, and this is exactly as it should be. The corollary is that Range ignores weak preference. Isn't that precisely correct? More accurately, Range *respects* strong preference. It does not necessarily reward it, it can actually punish it. If you expressed strong preference for A over all others, leaving B, a reasonable second choice, out in the cold, and, as a result, C, whom you hate, wins, Range has punished your strong preference. Because it was inappropriate. You lied, and Range took you at your word. It's ironic, don't you think, that even Range critics often claim to accept that Range is a great method for sincere voters. (But, of course, blah, blah, blah.) One of the ways of encouraging sincerity is to take people at their word. Give them what they ask for, if you can afford it. And if you are a member of the majority who allows some rating strength to your second preference, you can afford it, assuming that the second preference is good enough for you. If someone else bullet-votes for that second preference, have you been a "sucker," as Mike Ossipoff so liberally claims? I don't think so. You simply allowed someone else's expressed strong preference to prevail. That's actually reasonably rational behavior under most circumstances. When you can afford it. ---- election-methods mailing list - see http://electorama.com/em for list info
