Oops...I made a typo In the email I just sent on monotonicity...a couple of places I typed "otherwise loser" where I meant "otherwise winner." I hope it makes sense now.
Terry ----- Original Message ----- From: "Terry Bouricius" <[EMAIL PROTECTED]> To: "EM" <[email protected]> Sent: Thursday, August 07, 2008 11:17 AM Subject: Re: [Election-Methods] Can someone point me at an example of thenonmonotonicity of IRV? Kathy, On the monotonicity criterion and IRV. IRV does indeed fail the most common definition of the monotonicity criterion, as do two-round runoffs and all other methods (I believe) that satisfy the later-no-harm criterion. It is a trade off ... Which is more important in the real world, monotonicity or later-no-harm. Different experts have different opinions on that. The monotonicity criterion is often miss-understood by people who hear it for the first time. A candidate can never be hurt in IRV simply by getting more first preferences (although that is the impression people take away upon learning about the criterion). For example if a few more voters participate and only rank that one candidate as their first choice, that can never cause that candidate to lose. It is counter-intuitively possible, however, for a winner to become a loser if some voters reverse their ranking and elevate the otherwise winning candidate to first instead of lower on their ballots. The cause of the loss is the fact that some OTHER candidate may now face that otherwise loser in the runoff. The winner could just the same lose if those voters did not rank the otherwise loser first. In other words it is never the fact that the otherwise winner receives more first preferences that CAUSES her to become a loser. It is the change in ranking of OTHER candidates, changing who makes it into the runoff, that creates the non-monotonic effect. Thus, it is incorrect to say that under IRV, a candidate receiving more first preferences can cause the candidate to lose. A short essay I wrote on this, which includes some simple examples you requested, is at http://fairvote.org/monotonicity/ Terry Bouricius ----- Original Message ----- From: "Kathy Dopp" <[EMAIL PROTECTED]> To: "EM" <[email protected]> Sent: Wednesday, August 06, 2008 6:38 PM Subject: [Election-Methods] Can someone point me at an example of thenonmonotonicity of IRV? Someone called me today and pointed out in a way that it finally sunk in how awful it is that voters can actually cause their own first choice candidate to LOSE by casting votes for their first choice candidate in IRV elections. Could someone please point me to a fairly simple (if possible) example of the nonmonotonicity feature of IRV elections where adding say two more votes for candidate A causes candidate A to lose whereas candidate A would have won the election without the addition of the two more first choice votes? I find it absolutely astounding now that I think about it, that anyone could support a method where the voters going to the polls cannot know if their first choice votes for a candidate would hurt or help their candidate! Thanks. Pretty soon, if I find one more flaw of IRV, I'm going to have to revise the title of the paper to "20 flaws & 3 benefits of IRV..." -- Kathy Dopp The material expressed herein is the informed product of the author Kathy Dopp's fact-finding and investigative efforts. Dopp is a Mathematician, Expert in election audit mathematics and procedures; in exit poll discrepancy analysis; and can be reached at P.O. Box 680192 Park City, UT 84068 phone 435-658-4657 http://utahcountvotes.org http://electionmathematics.org http://electionarchive.org How to Audit Election Outcome Accuracy http://electionarchive.org/ucvAnalysis/US/paper-audits/legislative/VoteCountAuditBillRequest.pdf History of Confidence Election Auditing Development & Overview of Election Auditing Fundamentals http://electionarchive.org/ucvAnalysis/US/paper-audits/History-of-Election-Auditing-Development.pdf Voters Have Reason to Worry http://utahcountvotes.org/UT/UtahCountVotes-ThadHall-Response.pdf ---- Election-Methods mailing list - see http://electorama.com/em for list info ---- Election-Methods mailing list - see http://electorama.com/em for list info
