How to react to it is an individual choice, depending on the person and on the situation. Sometimes (but hopefully not) the defection could be so blatant and offensive, and the principle of not being taken advantage of could be so important, that you'd refuse to top-rate an acceptable in a u//a election. In other situations (hopefully usually) you're more interested in electing an acceptable instead of an unacceptable, and if Compromise's voters are rivalry-minded and unlikely to support Favorite, though much preferring Favorite to Worse, and they're likely to defect, you might top-rate Compromise anyway, because electing an acceptable is what is important.
For example, in recent presidential polls, I much prefer Roseanne Barr to Jill Stein. They both support the GPUS (Greens) policy platform, but Barr is more outspoken and forthright. (Some people, who only know Barr from a tv show, don't realize that she's a serious candidate. They haven't read what she's said--for instance in reply to a Green candidacy-questionnaire.) But in every Approval or Score poll, I've top-rated both candidates. Will the Stein-preferrers top-rate Barr? I hope so, but even if they aren't going to, I still top-rate Stein, because the important thing is to elect an acceptable. Of course I'd do the same if the official public election were by Approval or Score. In minguo's Plurality poll, I voted sincerely for Barr. In Plurality, of course, the u/a strategy is to vote for the acceptable who is most able to win with your help--and it's esssential that people who share a common acceptable-set combine their votes on the same _one_ acceptable. But, since we have Approval, Score and Emocracy polls, I regard the minguo Plurality poll as only a question about favoriteness, where election strategy doesn't apply. We can always add up the progressive votes to find out if, combined, they'd have elected a progressive. But, even though I prefer Barr, I advise people to vote for Stein in November, because, having won the GPUS nomination, and being on the ballot in a large majority of the states, Stein is the acceptable on whom progressives should combine their votes. (Barr's nomination by California Peace & Freedom might only be a vice-presidential nomination--I'm not sure. But the Peace & Freedom ticket and the GPUS ticket shouldn't be running against eachother, splitting the progressive vote, even if PFP is only in California. This inclination of progressives to split their vote is indicative of a, probably mistaken, belief that they can't win.) By the way (I mentioned this in a recent reply), when two factions are mostly not defection-inclined, so that nearly all of them top-rate both of their candidates, of course even just _a few_ defectors will decide which one of them wins. If they're winnable, then the one with the most defectors will be the one who wins. For one thing, as I said, If they're both acceptable, then one would hope that you wouldn't care about that, and would top-rate both anyway. Let the defectors decide which one wins--it doesn't matter. Still, there's an easy solution, whereby both factions can essentially fully support eachother's candidate, but the larger faction will still elect its own candidate: Each faction gives, to the other faction's candidate, a rating just barely, insignificantly, less than max. If the method is Approval, then approve Compromise with a probability of .99 If the method is 0-100 Score, give Compromise 99. If the method is 0-10 Score, then give Compromise 10 instead of 9, with a probability of 9/10.l Of course if you don't give Compromise 10, you give hir 9. Mike Ossipoff ---- Election-Methods mailing list - see http://electorama.com/em for list info
