I've had to be away from the computer for a while. I haven't yet had a chance to look at the many interesting postings about finding the best candidate, and right now I have just enough time to post this quick note:
My objection to Blake's notion of the best candidate has been that he didn't define it in terms of anything else, but says that it's possible to say that it's the absolute best. Best candidate needs an operational definition, a definition that , at least in principle, is usable, at least in a thought-experiment. I'm going to post some suggestions about such a definition within a few days, after I've read the "best candidate" postings that have been posted since I was last on the computer. Maybe my suggestions have already been posted by someone else. But my point today is that, even if we have an operational definition of "best", then, if that definition is any good, Blake's Ranked-Pairs(margins) won't find the best in public political elections. RP(m) causes insincere voting. It forces defensive strategic voting, and it tempts offensive stategic voting. The notion of it finding the best candidate depends on sincere voting. And if we assume sincere voting, then as has been pointed out, CR is what finds the best candidate. Borda is best if rankings must be used, if everyone votes sincerely. Another thing, Blake is mistaken to assume that the people's ballots point to the best candidate, even if there is one: Have you ever known the American people, during your lifetime, to elect the best person. Have you ever known them to do other than to elect one of the 2 or 3 worst candidates? I mean, in terms of your own notion of "best". And don't say "George Washington", or "Abe Lincoln". I don't know who else was running then, and we're more interested in present & future elections. Mike Ossipoff _________________________________________________________________ MSN Photos is the easiest way to share and print your photos: http://photos.msn.com/support/worldwide.aspx
