Mr. Cary inadvertently sent a query to me instead of to the list, but with the [EM] subject tag, and my reply then automatically went back to him. This is a copy of my response, and will be followed by another.
At 08:20 PM 3/6/2007, David Cary wrote: >--- Abd ul-Rahman Lomax <[EMAIL PROTECTED]> wrote: > > > However, if we assume sincere voters, what is the ideal election > > method, or the best among the options we know? > > . . . > > However, is Range ideal with sincere voters? If not, why not? > >What exactly is a sincere Range voter? >How does a sincere Range voter vote and based on what? >Ditto both questions for Approval voting? A sincere Range vote is one that is not distorted by considerations of what will "win." We have come to consider two kinds of voting sincerity with Range. Full sincerity would have, with Range of sufficient resolution, the rating of A > B whenever the voter prefers A to B. What might be called ordinary sincerity or truncated sincerity allows the voter to push ratings to the extremes. Thus ratings of Abe Lincoln = Mahatma Gandhi = max rating > George Bush = Adolf Hitler = min rating could be considered ordinarily sincere even if the voter actually prefers Bush to Hitler or, say, Gandhi to Lincoln. There is no particular precision in the definition of sincere in the question I was asking; rather, let it be defined as it is understood by those who criticize Range on the allegation that it only works with sincere voters and fails if voters exaggerate their preferences or otherwise vote strategically. From my point of view, Range doesn't break down because of Approval style voting. It merely collects somewhat less information and thereby makes a somewhat less sophisticated choice. But if voters actually distorted preferences in the sense of reversing them, all bets are off. I'd claim that these voters are harming themselves more than they harm others, but it's a different matter than the "exaggeration" insincerity. One might consider a vote sincere if *significant* preferences are reflected in ratings, again assuming Range of sufficient resolution (Range 100 ought to be quite adequate). In other words, a fully sincere voter who really would prefer, say, Nader over Gore, would show this in the ratings and not distort the ratings out of fear that therefore Bush might win. Pushing the Nader and Gore ratings to max would be strategic in this sense, and not fully sincere. ---- election-methods mailing list - see http://electorama.com/em for list info
