EM list: Yesterday I made a typo when defining the Independence from Tied Voters Criterion (ITVC). When I said "tie", I meant "tie between all of the candidates". So it should have read: (For a method to meet ITVC): If some subset of the ballots, if counted without the others, would result in a tie between all of the candidates, then the removal of those ballots from the count shouldn't change the count's winner. (Only one balloting is conducted). *** But someone could also say: If some subset of the ballots, if counted without the others, would result in a tie between 2 or more candidates, then the removal of those ballots shouldn't cause one of those candidates to win when another of those candidates would have won without the removal. I'm talking about single winner outcomes. But if one wanted to cover tied outcomes before & after the removal, one could say that if, without the removal, a candidate in that candidate subset would win, or some candidates in that subset would be the only ones in a tie for winner, then, after the removal, there shouldn't be a different candidate from that subset who is the single winner, or a different subsubset of candidates from that subset who are the only candidates in a tie for winner. *** Myself, I don't believe either of those "should" statements. I'm just saying 2 criteria that someone could define. The 2nd one is a generalization of the 1st one. I feel that if Saari's example with the symmetrical circular tie had any appeal, it would be in terms of those 2 criteria. *** Mike Ossipoff ______________________________________________________ Get Your Private, Free Email at http://www.hotmail.com