<DIV><BR>> From: "C.Benham" <CBENHAMAU@YAHOO.COM.AU><BR>> To: election-methods-electorama.com@electorama.com<BR>> Subject: [EM] Enhanced DMC<BR></DIV>
<DIV>> Forest,<BR>> The "D" in DMC used to stand for *Definite*.</DIV>
<DIV> </DIV>
<DIV>Yeah, that's what we finally settled on.</DIV>
<DIV><BR>> <BR>> I like (and I think I'm happy to endorse) this Condorcet method <BR>> idea, <BR>> and consider it to be clearly better than regular DMC<BR>> <BR>> Could this method give a different winner from the ("Approval <BR>> Chain <BR>> Building" ?) method you mentioned in the "C//A" thread (on 11 <BR>> June 2011)?</DIV>
<DIV> </DIV>
<DIV>Yes, I'll give an example when I get more time. But for all practical purposes they both pick the highest approval Smith candidate.</DIV>
<DIV><BR>> <BR>> >Initialize a variable X to be the candidate with the most approval.<BR>> ><BR>> >While X is covered, let the new value of X be the highest <BR>> approval candidate that covers the old X.<BR>> ><BR>> >Elect the final value of X.<BR>> ><BR>> >For all practical purposes this is just a seamless version of <BR>> C//A, i.e. it avoids the apparent <BR>> >abandonment of Condorcet in favor of Approval after testing for <BR>> a CW.<BR>> ><BR>> ><BR>> >Assuming cardinal ballots, candidate A covers candidate B, iff <BR>> whenever B is rated above C on more <BR>> >ballots than not, the same is true for A, and (additionally) A <BR>> beats (in this same pairwise sense) some <BR>> >candidate that B does not.<BR>> > <BR>> ><BR>> <BR>> Your newer suggestion ("enhanced DMC") seems to have an <BR>> easier-to-explain and justify motivation.<BR>> <BR>> Chris Benham<BR>> <BR>> <BR>> Forest Simmons wrote (12 July 2011):<BR>> <BR>> >One of the main approaches to Democratic Majority Choice was <BR>> through the idea that if X beats Y and <BR>> >also has greater approval than Y, then Y should not win.<BR>> ><BR>> >If we disqualify all that are beaten pairwise by someone with <BR>> greater approval, then the remaining set P <BR>> >is totally ordered by approval in one direction, and by <BR>> pairwise defeats in the other direction. DMC <BR>> >solves this quandry by giving pairwise defeat precedence over <BR>> approval score; the member of P that <BR>> >beats all of the others pairwise is the DMC winner. <BR>> ><BR>> >The trouble with this solution is that the DMC winner is always <BR>> the member off P with the least approval <BR>> >score. Is there some reasonable way of choosing from P that <BR>> could potentially elect any of its members?<BR>> ><BR>> >My idea is based on the following observation: <BR>> ><BR>> >There is always at least one member of P, namely the DMC <BR>> winner, i.e. the lowest approval member of <BR>> >P, that is not covered by any member of P.<BR>> ><BR>> >So why not elect the highest approval member of P that is not <BR>> covered by any member of P?<BR>> ><BR>> >By this rule, if the approval winner is uncovered it will win. <BR>> If there are five members of P and the upper <BR>> >two are covered by members of the lower three, but the third <BR>> one is covered only by candidates outside <BR>> >of P (if any), then this middle member of P is elected.<BR>> ><BR>> >What if this middle member X is covered by some candidate Y <BR>> outside of P? How would X respond to <BR>> >the complaint of Y, when Y says, "I beat you pairwise, as well <BR>> as everybody that you beat pairwise, so <BR>> >how come you win instead of me?"<BR>> ><BR>> >Candidate X can answer, "That's all well and good, but I had <BR>> greater approval than you, and one of my <BR>> >buddies Z from P beat you both pairwise and in approval. If Z <BR>> beat me in approval, then I beat Z pairwise, <BR>> >and somebody in P covers Z. If you were elected, both Z and <BR>> the member of P that covers Z would have <BR>> >a much greater case against you than you have against me."<BR></DIV>