When I said that, with 3 candidates, Middle's preferrers have no reason to approve anyone else, Lomax posted his example. Though I told of its fallacy, what if we suppose that the voters _do_ believe as Lomax said--that the Democrat and Republican have equal probability of winning, and that that probability is about .5?
Then, if so, as Lomax said, and I eventually agreed, some Green Democrats will have reason to approve the Green. The advantage of no Middle-preferrers (in a 3-candidate election) approving anyone else would be that that would make it easier to judge the size of the Green and Republican factions--They'd be equal to those candidates' approval count totals. But suppose people believe that p(D)=p(R)=.5, and so some Green Democrats approve the Green: As I was saying before, when Approval's new voting freedom begins to show the viability of non-Republocrat parties and candidates, then statisticians and honest poll-takers of all political persuasions will be very interested in estimating the sizes of the various factions and parties. They'll get good estimates by studying the approval count totals in the previous election, or poll results, or both. When I said that, with only 3 candidates, Middle preferrers have no reason to approve anyone else, I emphasize that I was speaking only of situations in which there are only 3 candidates, or at least in which only 3 candidates are perceived as viable and relevant. My discussion based on that assumption has no application otherwise. My comments about results based on better-than-expectation strategy don't assume or require only 3 candidates. For instance, my demonstration that Approval quickly goes to the voter median and stays there makes no assumptions about the number of candidates. If I've posted a lot in reply to Lomax's example, it's because I always reply thoroughly to anything calling for a reply. Mike Ossipoff Mike Ossipoff ---- Election-Methods mailing list - see http://electorama.com/em for list info