<html>
  <head>
    <meta content="text/html; charset=utf-8" http-equiv="Content-Type">
  </head>
  <body bgcolor="#FFFFFF" text="#000000">
    <div class="moz-cite-prefix">On 11/23/2016 10:59 AM, Kevin Venzke
      wrote:<br>
      <br>
      <blockquote type="cite"><span style="font-size: 16px;"
          id="yui_3_16_0_1_1479854184911_61103">Hmm, with a movable
          cutoff MDDA already violates Plurality with three candidates.
          Do you think symmetric-completion of the bottom can save it?</span></blockquote>
      <br>
      Yes.<br>
      <br>
      Something Toby Pereira wrote  (9 Nov. 2016) regarding "Irrelevant
      Ballots" got me thinking. The criterion I defined just talks about
      ballots <br>
      that plump for nobody, but there can also be a problem with
      ballots that only vote nobodies below, say, equal-top.<br>
      <br>
      Under MDDA with symmetric completion only at the bottom, adding 
      ballots like that can wash away  "majority-defeat"<br>
      disqualifications and make the result less Condorcet-consistent.<br>
      <br>
      This has led me to think of a modification to fix this problem
      that looks to be too good to be true, but so far I can't see how.<br>
      <br>
      *Voters submit rankings with an explicit approval cutoff. (I
      prefer default placement to be just below candidates ranked below
      no-one).<br>
      <br>
      On the ballots that have been symmetrically completed at the
      bottom, find the smallest set S of candidate/s that
      majority-strength pairwise<br>
      beat all the outside-S candidates.<br>
      <br>
      Disqualify the outside-S candidate/s and delete all the ballots
      that make no distinction among the inside-S candidates.<br>
      <br>
      Repeat as many times as possible.  If at the end of this process
      more than one candidate hasn't been disqualified, elect the<br>
      one of those by normal MDDA (SC).*<br>
      <br>
      I am fearful that this might fail FBC  and/or mono-raise, but I
      can't (yet) see how it does.<br>
      <br>
      45: A>B>>C<br>
      10: A=B<br>
      40: B<br>
      05: C<br>
      <br>
      100  ballots. After symmetrically completing at the bottom we get 
      A>B 47.5 - 42.5,    A>C  75-25,    B>C  95-5.<br>
      <br>
      Normal MDDA(SC)  disqualifies only C and then elects the most
      approved candidate, B.<br>
      <br>
      My suggested version first disqualifies C and then deletes the 5
      C  and  10 A=B  ballots and then disqualifies B leaving  A (the
      CW)<br>
      as the winner.<br>
      <br>
      What do you think?<br>
      <br>
      Chris Benham<br>
      <br>
      <br>
      On 11/23/2016 10:59 AM, Kevin Venzke wrote:<br>
    </div>
    <blockquote cite="mid:1434287094.44562.1479860969145@mail.yahoo.com"
      type="cite">
      <div style="color:#000; background-color:#fff;
        font-family:HelveticaNeue, Helvetica Neue, Helvetica, Arial,
        Lucida Grande, sans-serif;font-size:12px">
        <div id="yui_3_16_0_1_1479854184911_60274"><span><font size="3">Hi
              Chris,</font></span></div>
        <div class="qtdSeparateBR" id="yui_3_16_0_1_1479854184911_60273"><br>
          <br>
        </div>
        <div class="yahoo_quoted" id="yui_3_16_0_1_1479854184911_60298"
          style="display: block;">
          <div id="yui_3_16_0_1_1479854184911_60297">
            <div id="yui_3_16_0_1_1479854184911_60296">
              <div dir="ltr" id="yui_3_16_0_1_1479854184911_60295"
                style="font-family: HelveticaNeue, "Helvetica
                Neue", Helvetica, Arial, "Lucida Grande",
                sans-serif; font-size: 16px;"> <font
                  id="yui_3_16_0_1_1479854184911_60294" face="Arial"
                  size="2">
                  <hr id="yui_3_16_0_1_1479854184911_62295" size="1"> <b
                    id="yui_3_16_0_1_1479854184911_62869"><span
                      style="font-weight:bold;"
                      id="yui_3_16_0_1_1479854184911_62868">De :</span></b>
                  C.Benham <a class="moz-txt-link-rfc2396E" href="mailto:cbenham@adam.com.au"><cbenham@adam.com.au></a><br>
                  <b><span style="font-weight: bold;">À :</span></b>
                  Michael Ossipoff <a class="moz-txt-link-rfc2396E" href="mailto:email9648742@gmail.com"><email9648742@gmail.com></a> <br>
                  <b id="yui_3_16_0_1_1479854184911_62872"><span
                      style="font-weight: bold;"
                      id="yui_3_16_0_1_1479854184911_62871">Cc :</span></b>
                  EM <a class="moz-txt-link-rfc2396E" href="mailto:election-methods@lists.electorama.com"><election-methods@lists.electorama.com></a>;
                  Forest Simmons <a class="moz-txt-link-rfc2396E" href="mailto:fsimmons@pcc.edu"><fsimmons@pcc.edu></a><br>
                  <b><span style="font-weight: bold;">Envoyé le :</span></b>
                  Mardi 22 novembre 2016 10h51<br>
                  <b><span style="font-weight: bold;">Objet :</span></b>
                  Re: [EM] Trying to have CD, protect strong top-set,
                  and protect middle candidates too<br>
                </font> </div>
              <div class="y_msg_container"
                id="yui_3_16_0_1_1479854184911_60302"><br>
                <div id="yiv4410041191">
                  <div id="yui_3_16_0_1_1479854184911_60301">
                    <div class="yiv4410041191moz-cite-prefix"
                      id="yui_3_16_0_1_1479854184911_60300"><span
                        style="font-size: 16px;"
                        id="yui_3_16_0_1_1479854184911_60845">On
                        11/22/2016 9:25 AM, Michael Ossipoff wrote:</span><br
                        clear="none">
                      <br id="yui_3_16_0_1_1479854184911_60832"
                        clear="none">
                      <span style="font-size: 16px;"
                        id="yui_3_16_0_1_1479854184911_60843">>>With
                      </span><font id="yui_3_16_0_1_1479854184911_60876"
                        size="3">MDDTR, if your plump for X makes hir
                        lose, it's because you added a ballot. It has
                        nothing whatsoever to do with the fact that the
                        new ballot plumped for X.<br>
                        >>Your ballot made X lose in spite of the
                        fact that it was a plump for X, not because it
                        was a plump for X.<br>
                        >>But in IRV, when you make X lose by
                        raising hir from last place to 1st place, that
                        raising of X was the only thing that you did,
                        and it is the reason why X lost.</font></div>
                    <div class="yiv4410041191moz-cite-prefix"
                      id="yui_3_16_0_1_1479854184911_60300"> <font
                        id="yui_3_16_0_1_1479854184911_61028" size="3">><br
                          clear="none">
                        <span id="yui_3_16_0_1_1479854184911_60840">>That
                          "distinction" is meaningless and completely
                          useless.  The idea that adding a ballot is
                          "something you did" that rates a mention is
                          ridiculous.</span></font></div>
                    <div class="yiv4410041191moz-cite-prefix"
                      id="yui_3_16_0_1_1479854184911_60300"
                      style="font-family: HelveticaNeue, "Helvetica
                      Neue", Helvetica, Arial, "Lucida
                      Grande", sans-serif;"><font size="3"><br>
                      </font></div>
                    <div class="yiv4410041191moz-cite-prefix"
                      id="yui_3_16_0_1_1479854184911_60300"
                      style="font-family: HelveticaNeue, "Helvetica
                      Neue", Helvetica, Arial, "Lucida
                      Grande", sans-serif;"><font
                        id="yui_3_16_0_1_1479854184911_60874" size="3">I'm
                        not sure about this specific example but I think
                        this kind of distinction could be a useful
                        defense. For IRV I might argue that a mono-raise
                        failure happens not just from raising the winner
                        but also <i
                          id="yui_3_16_0_1_1479854184911_62253">lowering</i> some
                        other, incidental candidate. The reason
                        mono-raise failures are offensive is that
                        supposedly the voter has done nothing but aid
                        the preexisting winner. But at least in IRV i</font><span
                        style="font-size: 16px;"
                        id="yui_3_16_0_1_1479854184911_60873">t is not
                        so clear as that.<br clear="none">
                        <br clear="none">
                        <blockquote type="cite"
                          id="yui_3_16_0_1_1479854184911_60311"><span
                            class="yiv4410041191"> </span></blockquote>
                        >Regarding MDDA,  symmetrically completing
                        the ballots only at the bottom and having a
                        moveable approval cutoff fixes its failures of
                        Mono-add-Plump<br clear="none">
                        >and Plurality and Irrelevant Ballots
                        Independence and in my opinion makes it a
                        good/acceptable method.</span></div>
                    <div class="yiv4410041191moz-cite-prefix"
                      id="yui_3_16_0_1_1479854184911_60300"
                      style="font-family: HelveticaNeue, "Helvetica
                      Neue", Helvetica, Arial, "Lucida
                      Grande", sans-serif;"><span style="font-size:
                        16px;"><br>
                      </span></div>
                    <div class="yiv4410041191moz-cite-prefix"
                      id="yui_3_16_0_1_1479854184911_60300"
                      style="font-family: HelveticaNeue, "Helvetica
                      Neue", Helvetica, Arial, "Lucida
                      Grande", sans-serif;" dir="ltr"><span
                        style="font-size: 16px;"
                        id="yui_3_16_0_1_1479854184911_61103">Hmm, with
                        a movable cutoff MDDA already violates Plurality
                        with three candidates. Do you think
                        symmetric-completion of the bottom can save it?</span></div>
                    <div class="yiv4410041191moz-cite-prefix"
                      id="yui_3_16_0_1_1479854184911_60300"
                      style="font-family: HelveticaNeue, "Helvetica
                      Neue", Helvetica, Arial, "Lucida
                      Grande", sans-serif;" dir="ltr"><span
                        style="font-size: 16px;"><br>
                      </span></div>
                    <div class="yiv4410041191moz-cite-prefix"
                      id="yui_3_16_0_1_1479854184911_60300"
                      style="font-family: HelveticaNeue, "Helvetica
                      Neue", Helvetica, Arial, "Lucida
                      Grande", sans-serif;" dir="ltr"><span
                        style="font-size: 16px;"
                        id="yui_3_16_0_1_1479854184911_62843">For
                        whatever interest it may be, I calculated the
                        "DNA" for the method you describe and got the
                        exact same 343-digit code as for ICA(explicit).
                        That's the first time I've hit a method I
                        already had...<br clear="none">
                        <br clear="none">
                        Chris Benham<br clear="none">
                      </span></div>
                    <div class="yiv4410041191moz-cite-prefix"
                      id="yui_3_16_0_1_1479854184911_60300"
                      style="font-family: HelveticaNeue, "Helvetica
                      Neue", Helvetica, Arial, "Lucida
                      Grande", sans-serif;"><span style="font-size:
                        16px;"><br>
                      </span></div>
                  </div>
                </div>
              </div>
            </div>
          </div>
        </div>
      </div>
      <p class="" avgcert""="" color="#000000" align="left">No virus
        found in this message.<br>
        Checked by AVG - <a moz-do-not-send="true"
          href="http://www.avg.com">www.avg.com</a><br>
        Version: 2016.0.7924 / Virus Database: 4664/13460 - Release
        Date: 11/22/16</p>
    </blockquote>
    <p><br>
    </p>
  </body>
</html>