On Mon, 13 May 2002, Donald Davison wrote: > It is not proper to use any of the considered methods as a standard, not > Plurality nor Irving nor any of variants of the junk methods, Approval and > Condorcet. It would be as if in the American Beauty Constest Miss > California (or Miss Condorcet) was regarded to be the standard by which all > the contestants are to be judged including Miss California - and the winner > is Miss California - of course, who else - more deceit of the Condorcet > people.
Saying that "Condorcet Winner" is a code word for a particular method is like saying that "Majority Winner" is a code word for some particular method. There are many methods including IRV that satisfy the Majority Criterion. Some of these methods (but not IRV) also satisfy the Condorcet Criterion. The Condorcet Criterion is a criterion, in fact a generalization of the Majority Criterion, not one particular method. Like the Majority Criterion the Condorcet Criterion stands on its own two feet independent of any particular method. A method satisfies the Condorcet Criterion iff it always chooses the candidate preferred by a majority over any other candidate when there is such a candidate. Methods like Approval and IRV that fail the Condorcet Criterion may or may not have good excuses for failing it. The only way a method can fail the Condorcet Criterion is to choose some candidate X as winner when there is some other candidate Y that beats X and every other candidate by majorities of votes. IRVists claim that IRV chooses "the" majority winner. But the IRV winner has to win only one majority preference to claim that title. To earn the title "Condorcet Winner" a candidate has to beat all comers head-to-head including the IRV winner. Whenever there is a Condorcet Winner and IRV doesn't choose it, then the CW beats the IRV winner by a majority. The Loring Ensemble Rules recognize the superiority of the CW over the IRV winner when there is one. If there is no CW, then those rules say, "Fine, give the win to the simple runoff winner." Forest ---- For more information about this list (subscribe, unsubscribe, FAQ, etc), please see http://www.eskimo.com/~robla/em
