<html>
  <head>
    <meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
  </head>
  <body>
    <font face="Helvetica, Arial, sans-serif">Thanks to Kevin and
      Michael for pointing out a feature of minimax I was unaware of. I
      had however seen Richard Darlington's paper [1] in which he
      referred to 'several studies' comparing margins with winning
      votes. He reports that margins 'was the big winner in all of
      them'. I suppose I'll have to look deeper.<br>
         Colin<br>
      [1]. <a class="moz-txt-link-freetext" href="https://arxiv.org/abs/1807.01366">https://arxiv.org/abs/1807.01366</a><br>
    </font><br>
    <div class="moz-cite-prefix">On 18/09/2023 07:57, Michael Ossipoff
      wrote:<br>
    </div>
    <blockquote type="cite"
cite="mid:CAOKDY5DNcvimxv+r6wOgfcViHsWyCG8kt0zUZgVddmd=Be34Hg@mail.gmail.com">
      <meta http-equiv="content-type" content="text/html; charset=UTF-8">
      <div><br>
      </div>
      <div><br>
        <div class="gmail_quote" dir="auto">
          <div dir="ltr" class="gmail_attr">---------- Forwarded message
            ---------<br>
            From: <strong class="gmail_sendername" dir="auto">Michael
              Ossipoff</strong> <span dir="auto"><<a
                href="mailto:email9648742@gmail.com"
                moz-do-not-send="true">email9648742@gmail.com</a>></span><br>
            Date: Sun, Sep 17, 2023 at 22:54<br>
            Subject: Re: [EM] Ranked Pairs<br>
            To: Forest Simmons <<a
              href="mailto:forest.simmons21@gmail.com"
              moz-do-not-send="true">forest.simmons21@gmail.com</a>><br>
          </div>
          <div class="gmail_quote" dir="auto"><br>
          </div>
          This was meant to be sent by “Reply All”, in order to post it.
          So now I’m forwarding it to EM.<br>
          <br>
          <div dir="auto">Forest—</div>
          <div dir="auto"><br>
          </div>
          <div dir="auto">But wv prevents truncation (strategic or
            otherwise) from taking the win from a CW.</div>
          <div dir="auto"><br>
          </div>
          <div dir="auto">…&, with, wv, refusing to rank anymore you
            don’t approve will cause offensive order-reversal by their
            preferrers to backfire.</div>
          <div dir="auto"><br>
          </div>
          <div dir="auto">I’d always take that precaution, & would
            advise others to.</div>
          <div dir="auto"><br>
          </div>
          <div dir="auto">When we discussed these guarantees years ago
            they seemed absolute, & we still have the
            guarantee-criteria based on them…met by wv versions of
            MinMax, RP, CSSD, & Smith//MinMax.</div>
          <div dir="auto"><br>
          </div>
          <div dir="auto">…&, with MinMax, whose winner can come
            from anywhere, not just from the top-cycle, & so,
            offensive order-reversal, when there are a fair number of
            candidates, is unpredictable & risky for its
            perpetrators, even if the precaution of deterrent-truncation
            isn’t taken.</div>
          <div><br>
            <div class="gmail_quote">
              <div dir="ltr" class="gmail_attr">On Sun, Sep 17, 2023 at
                21:17 Forest Simmons <<a
                  href="mailto:forest.simmons21@gmail.com"
                  target="_blank" moz-do-not-send="true">forest.simmons21@gmail.com</a>>
                wrote:<br>
              </div>
              <blockquote class="gmail_quote" style="margin:0px 0px 0px
0.8ex;border-left-width:1px;border-left-style:solid;padding-left:1ex;border-left-color:rgb(204,204,204)">
                <div dir="auto">
                  <div><br>
                    <br>
                    <div class="gmail_quote">
                      <div dir="ltr" class="gmail_attr">On Sat, Sep 16,
                        2023, 9:42 PM Michael Ossipoff <<a
                          href="mailto:email9648742@gmail.com"
                          target="_blank" moz-do-not-send="true">email9648742@gmail.com</a>>
                        wrote:<br>
                      </div>
                      <blockquote class="gmail_quote" style="margin:0px
                        0px 0px
0.8ex;border-left-width:1px;border-left-style:solid;padding-left:1ex;border-left-color:rgb(204,204,204)">
                        <div dir="auto">Is that RP(wv), or RP(margins) ?</div>
                        <div dir="auto"><br>
                        </div>
                        <div dir="auto">RP(wv) would thwart & deter
                          offensive strategy, an important property in
                          public elections.</div>
                        <div dir="auto"><br>
                        </div>
                        <div dir="auto">…&, actually, it seems to me
                          that MinMax(wv) would do that better.</div>
                        <div dir="auto"><br>
                        </div>
                        <div dir="auto">That’s because, choosing only
                          from the Smith Set RP, limits it’s choice to
                          the strategic top-cycle that created by the
                          offensive strategists.</div>
                        <div dir="auto"><br>
                        </div>
                        <div dir="auto">Suppose that the CW’s preferrers
                          don’t do defensive truncation (never rank
                          anyone you wouldn’t approve in Approval, or
                          whose preferrers you regard as likely to
                          offensively order-reverse) ?</div>
                        <div dir="auto"><br>
                        </div>
                        <div dir="auto">Knowing that RP will limit its
                          choice to their small  strategic top-cycle, it
                          would be easier for the strategists to be
                          fairly sur that their candidate would win in
                          that top-cycle.</div>
                        <div dir="auto"><br>
                        </div>
                        <div dir="auto">But, with MinMax, the winner is
                          chosen more broadly, & could be anywhere
                          in the candidate-set.  …making it more
                          difficult & risky to confidently do
                          offensive order/reversal.</div>
                        <div dir="auto"><br>
                        </div>
                        <div dir="auto">RP(margins) might the best
                          choice for a completely honest electorate, but
                          MinMax(wv) seems better for public elections,
                          due to its better thwarting & deterrence
                          of offensive strategy.</div>
                        <div dir="auto"><br>
                        </div>
                        <div dir="auto">Yes, MinMax doesn’t meet the
                          luxury cosmetic look-good criteria that RP
                          meets. </div>
                        <div dir="auto"><br>
                        </div>
                        <div dir="auto">But for one thing, I remind you
                          that natural ( sincere) top-cycles are
                          vanishingly-rare.</div>
                      </blockquote>
                    </div>
                  </div>
                  <div dir="auto"><br>
                  </div>
                  <div dir="auto">This is the same conclusión I have
                    come around to. </div>
                  <div dir="auto"><br>
                  </div>
                  <div dir="auto">And methods that break a three member
                    top cycle at the weakest link tend to reward the
                    burier faction.</div>
                </div>
                <div dir="auto">
                  <div dir="auto">
                    <div class="gmail_quote">
                      <blockquote class="gmail_quote" style="margin:0px
                        0px 0px
0.8ex;border-left-width:1px;border-left-style:solid;padding-left:1ex;border-left-color:rgb(204,204,204)">
                        <div dir="auto"><br>
                        </div>
                        <div dir="auto">So do you want to have less
                          strategy-protection, in order for the result
                          to maybe look better in a vanishingly rare
                          natural top/cycle?</div>
                        <div dir="auto"><br>
                        </div>
                        <div dir="auto">…& how bad is a violation of
                          Condorcet-Loser anyway.  “Beaten by all the
                          other alternatives” sounds like some kind of
                          unanimity, but of course it isn’t. It isn’t
                          like a Pareto-violation. I remind you that the
                          MinMax winner has fewer voters preferring some
                          particular candidate over him than anyone else
                          does.</div>
                        <div dir="auto"><br>
                        </div>
                        <div dir="auto">Clone-Criterion violation? How
                          bad that really in MinMax, especially when
                          we’re talking about a vanishingly rare natural
                          top-cycle?</div>
                        <div dir="auto"><br>
                        </div>
                        <div dir="auto">RP(margins) for a completely
                          honest electorate.</div>
                        <div dir="auto"><br>
                        </div>
                        <div dir="auto">MinMax(wv) for public elections.</div>
                        <div dir="auto"><br>
                        </div>
                        <div dir="auto">..& about a primary to
                          reduce the candidates to 5: Forget the
                          primary. If you think people will have trouble
                          rank-ordering lots of candidates, I remind you
                          that, to vote among them in a primary, they’d
                          still have to examine & choose among the
                          initial many candidates.</div>
                        <div dir="auto"><br>
                        </div>
                        <div dir="auto">…harder than ranking only the
                          ones you know & regard as deserving &
                          definitely in your accepts& preferred set.</div>
                        <div dir="auto"><br>
                          <div class="gmail_quote" dir="auto">
                            <div dir="ltr" class="gmail_attr">On Wed,
                              Sep 13, 2023 at 00:18 Colin Champion <<a
href="mailto:colin.champion@routemaster.app" rel="noreferrer"
                                target="_blank" moz-do-not-send="true">colin.champion@routemaster.app</a>>
                              wrote:<br>
                            </div>
                            <blockquote class="gmail_quote"
                              style="margin:0px 0px 0px
0.8ex;border-left-width:1px;border-left-style:solid;padding-left:1ex;border-left-color:rgb(204,204,204)">I
                              notice that RP is the only election method
                              mentioned by name in the <br>
                              Virginia agenda.<br>
                              <br>
                              A while ago I ran some simulations on
                              elections with truncated ballots. <br>
                              Something I noticed was that the presence
                              of RP in the list of methods <br>
                              made the software unacceptably slow. I
                              didn't look into the cause, but <br>
                              there's a natural explanation, which is
                              the fact that RP is known to be <br>
                              NP-complete when it deals correctly with
                              tied margins, i.e. by <br>
                              exhausting over all their permutations.
                              Presumably if some candidates <br>
                              are unpopular and ballots are extensively
                              truncated, then tied margins <br>
                              are much likelier than with complete
                              ballots.<br>
                              <br>
                              I gather that practical implementations of
                              RP choose a random <br>
                              permutation rather than exhausting. This
                              seems to me to bring a danger. <br>
                              The presence of a few vanity candidates
                              (truncated off almost all <br>
                              ballots) may lead to ties, and this may
                              lead to a comfortable winner <br>
                              looking as though he owes his victory to a
                              coin-toss. Obviously this <br>
                              undermines the legitimacy of his win.<br>
                              <br>
                              CJC<br>
                              ----<br>
                              Election-Methods mailing list - see <a
                                href="https://electorama.com/em"
                                rel="noreferrer noreferrer"
                                target="_blank" moz-do-not-send="true">https://electorama.com/em</a>
                              for list info<br>
                            </blockquote>
                          </div>
                        </div>
                        ----<br>
                        Election-Methods mailing list - see <a
                          href="https://electorama.com/em"
                          rel="noreferrer noreferrer" target="_blank"
                          moz-do-not-send="true">https://electorama.com/em</a>
                        for list info<br>
                      </blockquote>
                    </div>
                  </div>
                </div>
              </blockquote>
            </div>
          </div>
        </div>
      </div>
      <br>
      <fieldset class="mimeAttachmentHeader"></fieldset>
      <pre class="moz-quote-pre" wrap="">----
Election-Methods mailing list - see <a class="moz-txt-link-freetext" href="https://electorama.com/em">https://electorama.com/em</a> for list info
</pre>
    </blockquote>
    <br>
  </body>
</html>