<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 11/13/2016 3:35 AM, Jameson Quinn
      wrote:<br>
      <br>
      <blockquote type="cite">
        <div>What I mean is that if you take a non-election-theorist,
          present an election scenario to them, explain who won and why,
          and ask how they would strategize in the place of voter X,
          they are more likely to suggest counterproductive strategies,
          and less likely to see any strategies that actually might
          work, in Condorcet than in Bucklin-like systems.</div>
        <div><br>
        </div>
      </blockquote>
      <br>
      The strategy incentives for Condorcet voting methods vary widely. 
      Some have a random-fill incentive while others have a truncation
      incentive. Some have<br>
      a stronger or weaker incentive to equal-top rank than others, and
      some are more vulnerable to Burial than others.<br>
      <br>
      Smith//Approval has a truncation incentive like Bucklin's, only
      less strong. In addition Bucklin has an equal-top rank/rate
      incentive.  I don't see the problem.<br>
      <br>
      BTW, why does it matter if "non-election-theorists" when asked
      suggest "counter-productive strategies"?  Shouldn't we be
      encouraging sincere voting?<br>
      If they don't want to do that, why can't they just take the
      strategy advice of their favourites?<br>
      <br>
      <span class="">35: C >> A=B<br>
        33: A>B >> C<br>
        32: B>A >> C<br>
        <br>
      </span>
      <blockquote type="cite"><span class="">In Smith//approval, one
          vote alone would shift the above honest election; so the fact
          that it does not in PAR is indeed notable.</span></blockquote>
      <br>
      <span class="">I don't see why.  The example I gave just happened
        to have a close CW.  PAR seems to give an A=B tie unless (as I
        assume) it breaks tied final<br>
        scores in favour of the "leader" (A).<br>
        <br>
      </span>
      <blockquote type="cite"><span class="">In particular: in PAR,
          there is no way for the B voters to strategize such that they
          win the above election, while still ensuring that C does not
          win no matter what the A voters do.<br>
        </span></blockquote>
      <br>
      <span class="">Of course, that is why it's called a "chicken
        dilemma".  In what method <i>can</i> "</span><span class="">the
        B voters to strategize such that they win the above election,
        while still ensuring that C does not win no matter what the A
        voters do" ??<br>
        <br>
      </span>
      <blockquote type="cite">MJ passes IIA. PAR fails it, as you say,
        but passes LIIA. <br>
      </blockquote>
      <br>
      As do some Condorcet methods. It isn't one of the criteria I care
      much about.<br>
      <br>
      As I understand it, IIA can only be met by methods that fail
      Majority (like positional methods that pretend that the voters'
      ratings are on some scale independent of the<br>
      candidates).   MJ  is a variety of Median Ratings which is
      normally claimed to meet Majority. <br>
      <br>
      I would be a bit surprised if IIA can be met by a method (such as
      MJ and Bucklin) by a method that fails Irrelevant Ballots
      Independence.<br>
      <br>
      There is some rubbish about Independence of Irrelevant
      Alternatives (IIA) on Electowiki.  I'll address that in a later
      post.<br>
      <br>
      Chris Benham<br>
      <br>
      <br>
      On 11/13/2016 3:35 AM, Jameson Quinn wrote:<br>
    </div>
    <blockquote
cite="mid:CAO82iZw104HcJd==nVCqKGNDuF_z3yzLNH-WQUQVwJ2fsZDDVw@mail.gmail.com"
      type="cite">
      <div dir="ltr"><br>
        <div class="gmail_extra"><br>
          <div class="gmail_quote">2016-11-12 10:45 GMT-05:00 C.Benham <span
              dir="ltr"><<a moz-do-not-send="true"
                href="mailto:cbenham@adam.com.au" target="_blank">cbenham@adam.com.au</a>></span>:<br>
            <blockquote class="gmail_quote" style="margin:0 0 0
              .8ex;border-left:1px #ccc solid;padding-left:1ex">
              <div bgcolor="#FFFFFF" text="#000000"><span class="">
                  <div class="m_-680782387900465520moz-cite-prefix">On
                    11/12/2016 7:53 AM, Jameson Quinn wrote:<br>
                  </div>
                </span><span class="">
                  <blockquote type="cite">
                    <div dir="ltr"><br>
                      <div class="gmail_extra"><br>
                        <div class="gmail_quote">2016-11-11 12:50
                          GMT-05:00 C.Benham <span dir="ltr"><<a
                              moz-do-not-send="true"
                              href="mailto:cbenham@adam.com.au"
                              target="_blank">cbenham@adam.com.au</a>></span>:<br>
                          <blockquote class="gmail_quote"
                            style="margin:0px 0px 0px
                            0.8ex;border-left:1px solid
                            rgb(204,204,204);padding-left:1ex">
                            <div bgcolor="#FFFFFF">
                              <div
class="m_-680782387900465520m_-8048598155404746332gmail-m_-2730114550739300614moz-cite-prefix"><span
class="m_-680782387900465520m_-8048598155404746332gmail-">On 11/11/2016
                                  10:14 PM, Jameson Quinn wrote:<br>
                                  <br>
                                  <blockquote type="cite"> I think that
                                    simple PAR is close enough to FBC
                                    compliance to be an acceptable
                                    proposal.</blockquote>
                                  <br>
                                </span> I'm afraid I can't see any value
                                in "close enough" to FBC compliance. 
                                The point of FBC is to give an absolute
                                guarantee to (possibly uninformed<br>
                                and not strategically savvy) 
                                greater-evil fearing voters.</div>
                            </div>
                          </blockquote>
                          <div><br>
                          </div>
                          <div>Yes. The guarantee you can give is "as
                            long as the world is somewhere in this
                            restricted domain — that is, essentially, as
                            long as there are no Condorcet cycles and
                            each voter naturally rejects at least one of
                            the 3 frontrunners — this method meets FBC".
                            This is much broader than any guarantee you
                            could give for a typical non-FBC method. For
                            instance, with IRV, the best you could say
                            would be "as long as your favorite is
                            eliminated early or wins overall, you don't
                            have to betray them", which unlike PAR's
                            guarantee is not something which could ever
                            be generally true about all real elections
                            for all factions.<br>
                            <br>
                          </div>
                        </div>
                      </div>
                    </div>
                  </blockquote>
                </span> C: I have in mind voters who are inclined to
                Compromise, and so it's <i> absolute guarantee</i> or
                it's nothing.   Smith//Approval also has a much lower
                Compromise incentive<br>
                than does IRV  (which in turn has a much much lower
                Compromise incentive then FPP).<span class=""><br>
                  <br>
                  <br>
                  <blockquote type="cite">
                    <div dir="ltr">
                      <div class="gmail_extra">
                        <div class="gmail_quote">
                          <div><br>
                          </div>
                          <blockquote class="gmail_quote"
                            style="margin:0px 0px 0px
                            0.8ex;border-left:1px solid
                            rgb(204,204,204);padding-left:1ex">
                            <div bgcolor="#FFFFFF">
                              <div
class="m_-680782387900465520m_-8048598155404746332gmail-m_-2730114550739300614moz-cite-prefix"><span
class="m_-680782387900465520m_-8048598155404746332gmail-"><br>
                                  <br>
                                  <blockquote type="cite">It elects the
                                    "correct" winner in a chicken
                                    dilemma scenario,
                                    naive/honest/strategyless
                                    ballots, without a "slippery slope"
                                    (though of course, this is no longer
                                    a strong Nash equilibrium). </blockquote>
                                  <br>
                                </span> How do you have a "chicken
                                dilemma scenario" with
                                "naive/honest/strategyless ballots" ?<br>
                                <br>
                                35: C >> A=B<br>
                                33: A>B >> C<br>
                                32: B >> A=C  (sincere is B>A
                                >> C)<br>
                                <br>
                                In this CD scenario your method elects
                                B  in violation of the CD criterion.<br>
                              </div>
                            </div>
                          </blockquote>
                          <div><br>
                          </div>
                          <div>You're suggesting that the sincere
                            preferences are  <br>
                          </div>
                          <div><br>
                            35: C >> A=B<br>
                            33: A>B >> C<br>
                            32: B>A >> C<br>
                            <br>
                            <br>
                          </div>
                        </div>
                      </div>
                    </div>
                  </blockquote>
                </span> C:  I'm not "suggesting". I'm stating.<span
                  class=""><br>
                  <blockquote type="cite">
                    <div dir="ltr">
                      <div class="gmail_extra">
                        <div class="gmail_quote">
                          <div><br>
                          </div>
                          <div>If you are 1 of the B>A>>C
                            voters considering whether to strategically
                            vote B>>A=C, you have no strong
                            motivation to do so, because your vote alone
                            is not enough to shift the winner to B. This
                            is what I mean by "no slippery slope".<br>
                            <br>
                            <br>
                          </div>
                        </div>
                      </div>
                    </div>
                  </blockquote>
                </span> C: One "vote alone" is very rarely enough to do
                anything, so I suppose no-one has a "strong motivation"
                to vote.</div>
            </blockquote>
            <div><br>
            </div>
            <div>In Smith//approval, one vote alone would shift the
              above honest election; so the fact that it does not in PAR
              is indeed notable.</div>
            <div><br>
            </div>
            <div>In particular: in PAR, there is no way for the B voters
              to strategize such that they win the above election, while
              still ensuring that C does not win no matter what the A
              voters do. This "safe" strategizing is grease on the
              slippery slope.</div>
            <div><br>
            </div>
            <blockquote class="gmail_quote" style="margin:0 0 0
              .8ex;border-left:1px #ccc solid;padding-left:1ex">
              <div bgcolor="#FFFFFF" text="#000000"><span class=""><br>
                  <blockquote type="cite">
                    <div dir="ltr">
                      <div class="gmail_extra">
                        <div class="gmail_quote">
                          <div><br>
                          </div>
                          <div><br>
                          </div>
                          <div>I believe that in the election you gave,
                            there is no way to tell what the sincere
                            preferences are. <br>
                            <br>
                            <br>
                          </div>
                        </div>
                      </div>
                    </div>
                  </blockquote>
                  <br>
                </span> C: From just the information on the ballots, of
                course not (like any election).<span class=""><br>
                  <blockquote type="cite">
                    <div dir="ltr">
                      <div class="gmail_extra">
                        <div class="gmail_quote">
                          <div><br>
                            Perhaps the B voters are strategically
                            truncating A; perhaps the C voters are
                            strategically truncating B. So the "correct
                            winner" could be either A or B, but is
                            almost certainly not C. <br>
                          </div>
                        </div>
                      </div>
                    </div>
                  </blockquote>
                  <br>
                </span> C: By "correct winner" I assume you mean the
                sincere CW. But there is reason to assume there is one.
                And if the B voters are actively Burying C, it could be
                C.<span class=""><br>
                  <br>
                  <blockquote type="cite">
                    <div dir="ltr">
                      <div class="gmail_extra">
                        <div class="gmail_quote">
                          <div>The "CD criterion" requires the system to
                            elect C, merely to punish the B voters; I
                            think that's perverse, because, among other
                            things, it means that a system does badly
                            with center squeeze, allowing the C faction
                            to strategize and win.<br>
                            <br>
                          </div>
                        </div>
                      </div>
                    </div>
                  </blockquote>
                </span> C: No, it merely says "not B".  But CD +
                Plurality say that it must be C.<span class=""><br>
                  <blockquote type="cite">
                    <div dir="ltr">
                      <div class="gmail_extra">
                        <div class="gmail_quote">
                          <div> </div>
                          <blockquote class="gmail_quote"
                            style="margin:0px 0px 0px
                            0.8ex;border-left:1px solid
                            rgb(204,204,204);padding-left:1ex">
                            <div bgcolor="#FFFFFF">
                              <div
class="m_-680782387900465520m_-8048598155404746332gmail-m_-2730114550739300614moz-cite-prefix">
                                <br>
                                Since you are apparently now content to
                                do without FBC  compliance  and you
                                imply that electing the CW is a good
                                thing,<br>
                                why don't you advocate a method that
                                meets the Condorcet criterion?<br>
                                <br>
                                What is wrong with Smith//Approval?  Or
                                Forest's nearly equivalent Max Covered
                                Approval? <br>
                              </div>
                            </div>
                          </blockquote>
                          <div><br>
                          </div>
                          <div>Largely, it's because I think that
                            Condorcet systems are strategically
                            counterintuitive, and hard to present
                            results in. I think that will lead to more
                            strategy than a system like PAR. That's
                            because PAR can make guarantees that
                            Condorcet systems can't.<br>
                            <br>
                          </div>
                        </div>
                      </div>
                    </div>
                  </blockquote>
                </span> C: Such as?  What exactly does "strategically
                counter-intuitive" mean?  An example?</div>
            </blockquote>
            <div><br>
            </div>
            <div>What I mean is that if you take a
              non-election-theorist, present an election scenario to
              them, explain who won and why, and ask how they would
              strategize in the place of voter X, they are more likely
              to suggest counterproductive strategies, and less likely
              to see any strategies that actually might work, in
              Condorcet than in Bucklin-like systems.</div>
            <div><br>
            </div>
            <blockquote class="gmail_quote" style="margin:0 0 0
              .8ex;border-left:1px #ccc solid;padding-left:1ex">
              <div bgcolor="#FFFFFF" text="#000000"><span class=""><br>
                  <blockquote type="cite">
                    <div dir="ltr">
                      <div class="gmail_extra">
                        <div class="gmail_quote">
                          <div><br>
                          </div>
                          <div>In a system like MJ or Score, you can
                            give a number to each candidate, based on
                            their own ratings alone, and the higher
                            number wins. That is an easy way to get
                            monotonicity, FBC, and IIA.</div>
                          <div><br>
                          </div>
                          <div>In Condorcet, no candidate has any number
                            except in relation to all other candidates.
                            That's good for passing the Condorcet
                            criterion (obviously) but it breaks FBC and
                            IIA.<br>
                            <br>
                          </div>
                        </div>
                      </div>
                    </div>
                  </blockquote>
                </span> C:  Your method and MJ fail IIA.</div>
            </blockquote>
            <div><br>
            </div>
            <div>MJ passes IIA. PAR fails it, as you say, but passes
              LIIA. </div>
          </div>
          <br>
        </div>
      </div>
      <p class="" avgcert""="" color="#000000" align="left"><br>
      </p>
    </blockquote>
    <p>
      <blockquote type="cite">
        <p>Prefer Accept Reject (PAR) voting works as follows: </p>
        <ol>
          <li><b>Voters can Prefer, Accept, or Reject each candidate.</b>
            Blanks count as "Reject" if no rival is explicitly rejected;
            otherwise, blank is "Accept".</li>
          <li><b>Candidates with at least 25% Prefer, and no more than
              50% reject, are "viable"</b>. The most-preferred viable
            candidate (if any) is the leader.</li>
          <li> Each "prefer" is worth 1 point. For viable candidates,
            each "accept" on a ballot which doesn't prefer the leader is
            also worth 1 point. <b>Most points wins.</b></li>
        </ol>
      </blockquote>
      <br>
    </p>
  </body>
</html>