[EM] DH3 pathology, margins, and winning votes

[Rob LeGrand]

Warren wrote:
> Sorry, my last email was in error:  BTR-IRV can entirely eliminate the
> Smith set and elect some nonmember.

I don't think it can.  BTR-IRV never eliminates a candidate that pairwise
beats all the other remaining candidates because such a candidate
wouldn't lose the pre-elimination "runoff".  So even if all but one of
the Smith set had been eliminated, the last member will last until the
end and win.  So BTR-IRV satisfies Smith, but it does not satisfy
Schwartz (at least in my current implementation, where a tied pairwise
runoff eliminates each of the two candidates with equal probability).

Chris Benham wrote:
> As I understand it, BTR-IRV was invented by Rob Le Grand purely as a
> gimmick to try to sell Condorcet to IRV supporters. I don't think he
> ever seriously suggested  it was good, and I don't know how anyone
> else got that idea.

I did originally suggest it as an IRV-like Condorcet method; I don't
claim it to be clearly the best Condorcet method by any means, but I like
it as a Condorcet method that seems to encourage strategy less often than
some others.  Also, given sincere votes, it scores higher on social
utility in my simulations than the two other IRV-like Condorcet methods I
include: "eliminate the candidate with the fewest first-place votes until
a Condorcet candidate among the remaining candidate emerges" and "choose
the Condorcet winner if one exists and the IRV winner otherwise".  That
said, I would not choose to have it (or any other purely ranked-ballot
method) used in public elections.

--
Rob LeGrand, psephologist
rob at approvalvoting.org
Citizens for Approval Voting
http://www.approvalvoting.org/

__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com