At 01:03 AM 10/9/2007, Juho wrote: >"Insincere" is good, at least for scientific purposes. Term >"strategic" that you picked up is good as well. For me the difference >is maybe that that sincerity assumes that the voters have been asked >to vote in some way (maybe in order to guarantee that the method will >provide the intended result) but they will not. Term "strategic" >assumes only consideration of different voting alternatives. > >It is also possible that they are asked to consider their strategic >options and then vote strategically. In this case the two terms can >differ a bit. One could say that the voters are now sincere but >strategic.
Here are the tricky parts: Rank the candidates in preference order. Equal ranking not allowed. This causes voters to exert, in a condorcet method, *equal* voting power in every pairwise contest, even if the voter is actually indifferent or nearly so. Rank the candidates in preference order. Equal ranking allowed. This increases the freedom of the voter. Indeed, this instruction reduces to approval if every one uses only the first rank, and to a pure ranked method if nobody uses the ranking privilege. My contention is, by the way, that it is offensive to take information provided by the voter, information which has a clearly discernable meaning, on the grounds that some formal rule -- but not any democratic principle -- was violated. When I name Approval "Count All the Votes," I do it to emphasize this principle, which I think will be a winning argument politically, if it be effectively made. Now, there are situations where a "count all the votes" principle would, in fact, be unjust. In STV, with some methods of distributing the votes, extra votes could in fact give the voter more representational power. But this is not the case with single-winner Approval. And, in fact, it would not be the case with a ranked method PR method if proper vote redistribution methods are used. This is an argument, really, for fixing the method so that equal ranking can be allowed. It brings the vote closer to a real representation of voter preference. Borda with equal ranking allowed, I believe, is equivalent to Range with the same number of ranks as candidates.... And thus it should perform better. IRV with equal ranking allowed is, to my mind, a vastly improved method. And all we have to do is count the votes! This, then indicates a clear and simple path to voting reform; it may fork, but I think the forks would converge. It starts with beginning to use all vote data collected. One pathway: (1) Count All the Votes (2) Make it IRV with equal ranking, which continues to count all the votes, until it discards exhausted ballots. (3) Stop discarding ballots, consider exhausted ballots to be No votes against all remaining candidates, true majority required to win (i.e, what I'm claiming Robert's Rules recommends). (4) Now that we have STV-single-winner in place, introduce proportional representation. Either prohibit overvotes (bad idea to my mind) or provide rules that maintain one-person, one-vote. That is, in the end, every voter has contributed at most one and only one vote toward winning seats, all other votes were moot, cast for losers. And these are few in a good PR system. (5) Allow fractional votes. Probably not necessary, since multiwinner STV with many winners is so nearly equivalent, I think, that it's overkill. (I could describe how to do this, it seems pretty simple to me, but I think others have already done it.) The other path: (1) Count all the votes. (2) Make it range by allowing fractional votes. (3) Use a form similar to Reweighted Range Voting for PR. (4) Add pairwise analysis and require a runoff if majority consent is not apparent from the vote. And then it all gets trumped, for PR, by Asset Voting, which is terminally simple and can provide for what might be called direct representative democracy. *No* wasted votes, no *dependence* on any party system, though parties may continue to function with full freedom. ---- Election-Methods mailing list - see http://electorama.com/em for list info
