<html><head></head><body><div class="ydp6c183f3eyahoo-style-wrap" style="font-family:Helvetica Neue, Helvetica, Arial, sans-serif;font-size:16px;"><div></div>
        <div>Hi Chris,</div><div><br></div>
        
        </div><div id="ydp7d1d2e83yahoo_quoted_0606643543" class="ydp7d1d2e83yahoo_quoted">
            <div style="font-family: Helvetica Neue, Helvetica, Arial, sans-serif; color: rgb(38, 40, 42);">
                
                <div style="font-size: 13px;">
                    Le dimanche 9 juin 2019 à 21:20:34 UTC−5, C.Benham <cbenham@adam.com.au> a écrit :
                </div>
                <div style="font-size: 13px;">>Kevin,</div><div style="font-size: 13px;">></div><div style="font-size: 13px;">>So to be clear the possible "complaint" some voters might have
      (and you think we should take seriously) is "We lied </div><div style="font-size: 13px;">>and the
      voting method</div><div style="font-size: 13px;">>(instead of somehow reading our minds) believed us".</div><div style=""><div id="ydp7d1d2e83yiv7139577953" style=""><div style=""><p style=""></p><div style=""><font size="3">A burial strategy has two scenarios that could give rise to a complaint. One is where burial succeeds. In that case the voters who complain aren't the ones who buried. The other scenario is where burial backfires. In that case it is, I guess, possible that actually *all* voters were using burial. So you may argue that they don't have a valid complaint. But implicit to my concerns is the premise that the voters are behaving rationally under the incentives of the method. If a method produces arbitrary results given rational voters then it will be hard to retain it. I think if it happens even once it will be a problem.</font></div><div style="font-size: 13px;"><br></div><span style="font-size: 13px;">
      ></span><br clear="none"><div style="font-size: 13px;">>So therefore it is good to have a less expressive ballot because
      that reduces the voter's opportunities to tell stupid </div><div style="font-size: 13px;">>lies and if
      the method</div><div style="font-size: 13px;">>is simple enough then maybe also the temptation for them to do so.</div><div style="font-size: 13px;"><br></div><div style=""><font size="3">You're making it sound as though a simpler ballot just tricks people into not lying. Expressiveness isn't the point. The reason three-slot C//A (or implicit etc.) deters burial is that there is far more risk in trying it. It is highly likely to backfire no matter what other voters do. "Low expressiveness" of the ballot doesn't guarantee this and isn't a prerequisite for it either.</font></div><div><span style="font-size: 13px;"><br></span></div><span style="font-size: 13px;">
      ></span><br clear="none"><div style="font-size: 13px;">>But I've thought of a patch to address your issue.  We could have
      a rule which says that if the winner's approval </div><div style="font-size: 13px;">>score is below
      some fixed </div><div style="font-size: 13px;">>fraction of that of the most approved candidate, then a
      second-round runoff is triggered between those two </div><div style="font-size: 13px;">>candidates. 
      What do you</div><div><span style="font-size: 13px;">>think of that?  What do you think that fraction should be?</span></div><div><span style="font-size: 13px;"><br></span></div><div><font size="3">I think there is some confusion here between what my issue is, and the voter complaints you asked about. While I think voters will be unhappy with a ruined election, ruining it is what reduces the burial incentive. If the risk outweighs the benefit then people won't do it. (That's an assumption.) This patch seems to remove the risk while leaving the benefit unchanged. </font><span style="font-size: medium;">Burial will ultimately do nothing, except to sometimes move the win from the CW to the AW. But that makes the potential gain even clearer: If the voted CW turns out to not be your candidate, you would have had a second chance at the win by voting instead to deny CW status to that candidate. If the voted CW *is* your candidate, then you're no worse off for using burial.</span></div><div><span style="font-size: medium;"><br></span></div><div><span style="font-size: medium;">Kevin</span></div><div><span style="font-size: medium;"> </span></div>
      <br clear="none">
    <p></p>
    <div class="ydp7d1d2e83yiv7139577953yqt1633905774" id="ydp7d1d2e83yiv7139577953yqt46183" style="font-size: 13px;"><div class="ydp7d1d2e83yiv7139577953moz-cite-prefix">On 10/06/2019 9:57 am, Kevin Venzke
      wrote:<br clear="none">
    </div>
    <blockquote type="cite">
      </blockquote></div></div><div class="ydp7d1d2e83yiv7139577953yqt1633905774" id="ydp7d1d2e83yiv7139577953yqt09108" style="font-size: 13px;"><div><div class="ydp7d1d2e83yiv7139577953ydp8f615ceyahoo-style-wrap" style="font-family:Helvetica Neue, Helvetica, Arial, sans-serif;font-size:16px;">
        <div>Hi Chris,</div>
        <div><br clear="none">
        </div>
        <div>
          <div>
            <div>
              <blockquote type="cite" style="margin:0px;padding:8px;color:rgb(38, 40, 42);font-family:Helvetica Neue, Helvetica, Arial, sans-serif;">>>I don't think
                it's ideal if burying X under Y (both disapproved) can
                only backfire when Y is made the CW.
                <div>>></div>
              </blockquote>
              <span style="color:rgb(38, 40, 42);font-family:Helvetica Neue, Helvetica, Arial, sans-serif;">>Why
                is that?  </span></div>
            <div><span style="color:rgb(38, 40, 42);font-family:Helvetica Neue, Helvetica, Arial, sans-serif;"><br clear="none">
              </span></div>
            <div><span style="color:rgb(38, 40, 42);font-family:Helvetica Neue, Helvetica, Arial, sans-serif;">Because I
                think if voters decide to attempt to prevent another
                candidate from being CW, via insincerity, there should
                be risks to doing that. Of course there is already some
                risk. But if you "knew" that a given candidate had no
                chance of being CW then there would be nothing to lose
                in using that candidate in a burial strategy.</span></div>
            <div><span style="color:rgb(38, 40, 42);font-family:Helvetica Neue, Helvetica, Arial, sans-serif;"><br clear="none">
              </span></div>
            <div><span style="color:rgb(38, 40, 42);font-family:Helvetica Neue, Helvetica, Arial, sans-serif;">>The
                post-election complaint (by any of the voters) would be
                .. what?</span></div>
            <div><br clear="none">
            </div>
            <div>For either a successful burial strategy, or one that
              backfires and elects an arbitrary candidate, I think the
              possible complaints are clear. Maybe someone would argue
              that a backfiring strategy proves the method's incentives
              are just fine. But that wouldn't be how I see it. I think
              if, in actual practice, it ever happens that voters
              calculate that a strategy is worthwhile, and it completely
              backfires to the point that everyone would like the
              results discarded, then that method will probably get
              repealed.</div>
            <div><br clear="none">
            </div>
            <div>><br clear="none" style="color:rgb(38, 40, 42);font-family:Helvetica Neue, Helvetica, Arial, sans-serif;">
              <span style="color:rgb(38, 40, 42);font-family:Helvetica Neue, Helvetica, Arial, sans-serif;">>If
                you don't allow voters to rank among their unapproved
                candidates then arguably you are not even trying to
                elect the sincere CW.</span><br clear="none" style="color:rgb(38,                 40, 42);font-family:Helvetica Neue, Helvetica, Arial, sans-serif;">
              <span style="color:rgb(38, 40, 42);font-family:Helvetica Neue, Helvetica, Arial, sans-serif;">>Instead
                you are just modifying Approval to make it a lot more
                Condorcet-ish.  </span></div>
            <div><br clear="none">
            </div>
            <div>Not an unfair statement. If you require voters to have
              that much expressiveness then you can't use implicit.</div>
            <div><br clear="none">
            </div>
            <div>To me, the motivation for three-slot C//A(implicit) is
              partly about burial, partly about method simplicity,
              partly about ballot simplicity. C//A(explicit) retains 1
              of 3. (Arguably slightly less for the Smith version.)
              Possibly it has its own merits, but they will largely be
              different ones.</div>
            <div><br clear="none" style="color:rgb(38, 40, 42);font-family:Helvetica Neue, Helvetica, Arial, sans-serif;">
              ><br clear="none" style="color:rgb(38, 40, 42);font-family:Helvetica Neue, Helvetica, Arial, sans-serif;">
              <span style="color:rgb(38, 40, 42);font-family:Helvetica Neue, Helvetica, Arial, sans-serif;">>A lot
                of voters like relatively expressive ballots. I think
                that is one of the reasons why Approval seems to be a
                lot less popular than IRV.</span><br clear="none" style="color:rgb(38, 40, 42);font-family:Helvetica Neue, Helvetica, Arial, sans-serif;">
            </div>
          </div>
          <div><span style="color:rgb(38, 40, 42);font-family:Helvetica Neue, Helvetica, Arial, sans-serif;"><br clear="none">
            </span></div>
          <div><span style="color:rgb(38, 40, 42);font-family:Helvetica Neue, Helvetica, Arial, sans-serif;">I have no
              *inherent* complaints about the ballot format of explicit
              approval plus full ranking.</span></div>
          <div><span style="color:rgb(38, 40, 42);font-family:Helvetica Neue, Helvetica, Arial, sans-serif;"><br clear="none">
            </span></div>
          <div><span style="color:rgb(38, 40, 42);font-family:Helvetica Neue, Helvetica, Arial, sans-serif;">Kevin</span></div>
          <br clear="none">
        </div>
        <div><br clear="none">
        </div>
        <div><br clear="none">
        </div>
        <div><br clear="none">
        </div>
      </div>
      <div class="ydp7d1d2e83yiv7139577953ydpb6a32dd9yahoo_quoted" id="ydp7d1d2e83yiv7139577953ydpb6a32dd9yahoo_quoted_0445501564">
        <div style="font-family:'Helvetica Neue', Helvetica, Arial, sans-serif;font-size:13px;color:#26282a;">
          <div> Le jeudi 6 juin 2019 à 21:03:19 UTC−5, C.Benham
            <a shape="rect" class="ydp7d1d2e83yiv7139577953moz-txt-link-rfc2396E" href="mailto:cbenham@adam.com.au" rel="nofollow" target="_blank"><cbenham@adam.com.au></a> a écrit : </div>
          <div><br clear="none">
          </div>
          <div><br clear="none">
          </div>
          <div>
            <div id="ydp7d1d2e83yiv7139577953ydpb6a32dd9yiv8583323751">
              <div>
                <p>Kevin,<br clear="none">
                </p>
                <blockquote type="cite">Specifically should "positional
                  dominance" have the same meaning whether or not the
                  method has approval in it?</blockquote>
                <br clear="none">
                If the voters all choose to approve all the candidates
                they rank, then yes.  (For a while I was wrongly
                assuming that Forest's suggested<br clear="none">
                default approval was for all ranked-above-bottom
                candidates, but then I noticed that he specified that it
                was only for top voted candidates).<br clear="none">
                <br clear="none">
                One of my tired examples:<br clear="none">
                <br clear="none">
                25: A>B<br clear="none">
                26: B>C<br clear="none">
                23: C>A<br clear="none">
                26: C<br clear="none">
                <br clear="none">
                Assuming all the ranked candidates are approved, C is by
                far the most approved and the most top-voted candidate.
                <br clear="none">
                Normal Winning Votes (and your idea 2 in this example)
                elect B.<br clear="none">
                <br clear="none">
                <blockquote type="cite">I will go easy on these methods
                  over failing MD, because it happens when some of the
                  majority don't approve their common candidate.</blockquote>
                <br clear="none">
                For me this this type of ballot avoids the Minimal
                Defense versus Chicken Dilemma dilemma, rendering those
                criteria inapplicable.<br clear="none">
                <p>48: A<br clear="none">
                  27: B>C<br clear="none">
                  25: C<br clear="none">
                  <br clear="none">
                  The problem has been that we don't know whether the
                  B>C voters are thinking "I am ranking C because
                  above all I don't want that evil A<br clear="none">
                  to win" or  "My C>A preference isn't all that
                  strong, and I think that my favourite could well be
                  the sincere CW, and if  C's supporters rank<br clear="none">
                  B above A then B has a good chance to win. But if they
                  if they create a cycle by truncating I'm not having
                  them steal it".<br clear="none">
                  <br clear="none">
                  With the voters able to express explicit approval we
                  no longer have to guess which it is.<br clear="none">
                  <br clear="none">
                </p>
                <blockquote type="cite"> I don't think it's ideal if
                  burying X under Y (both disapproved) can only backfire
                  when Y is made the CW.
                  <div><br clear="none">
                  </div>
                </blockquote>
                Why is that?  The post-election complaint (by any of the
                voters) would be .. what?<br clear="none">
                <br clear="none">
                If you don't allow voters to rank among their unapproved
                candidates then arguably you are not even trying to
                elect the sincere CW.<br clear="none">
                Instead you are just modifying Approval to make it a lot
                more Condorcet-ish.  <br clear="none">
                <br clear="none">
                A lot of voters like relatively expressive ballots. I
                think that is one of the reasons why Approval seems to
                be a lot less popular than IRV.<br clear="none">
                <br clear="none">
                Chris Benham<br clear="none">
                <br clear="none">
                <div class="ydp7d1d2e83yiv7139577953ydpb6a32dd9yiv8583323751yqt8593146214" id="ydp7d1d2e83yiv7139577953ydpb6a32dd9yiv8583323751yqt96283">
                  <div class="ydp7d1d2e83yiv7139577953ydpb6a32dd9yiv8583323751moz-cite-prefix">On
                    6/06/2019 5:34 pm, Kevin Venzke wrote:<br clear="none">
                  </div>
                  <blockquote type="cite"> </blockquote>
                </div>
              </div>
              <div class="ydp7d1d2e83yiv7139577953ydpb6a32dd9yiv8583323751yqt8593146214" id="ydp7d1d2e83yiv7139577953ydpb6a32dd9yiv8583323751yqt32286">
                <div>
                  <div class="ydp7d1d2e83yiv7139577953ydpb6a32dd9yiv8583323751ydpe4c7db39yahoo-style-wrap" style="font-family:Helvetica Neue, Helvetica, Arial, sans-serif;font-size:16px;">
                    <div>Hi Chris,</div>
                    <div><br clear="none">
                    </div>
                    <div>I've been short on time so I don't actually
                      have much thought on any of the methods, even my
                      own.</div>
                    <div><br clear="none">
                    </div>
                    <div>I suppose Idea 2 is the same as
                      Schwartz-limited MinMax(WV) if nobody submits
                      disapproved rankings. I'm not sure if it makes
                      sense to reject the method over that. Specifically
                      should "positional dominance" have the same
                      meaning whether or not the method has approval in
                      it? As a comparison, I will go easy on these
                      methods over failing MD, because it happens when
                      some of the majority don't approve their common
                      candidate.</div>
                    <div><br clear="none">
                    </div>
                    <div>I would have liked to simplify Idea 2, but
                      actually Forest's eventual proposal wasn't all
                      that simple either. As I wrote, if you add "elect
                      a CW if there is one" it can become much simpler,
                      so that it isn't really distinct from Idea 1. I
                      actually tried pretty hard to present three
                      "Ideas" in that post, but kept having that
                      problem.</div>
                    <div><br clear="none">
                    </div>
                    <div>I posted those ideas because I thought Forest
                      posed an interesting challenge, and I thought I
                      perceived that he was trying to fix a problem with
                      CD. That said, I am not a fan of
                      Smith//Approval(explicit). If all these methods
                      are basically the same then I probably won't end
                      up liking any of them. I don't think it's ideal if
                      burying X under Y (both disapproved) can only
                      backfire when Y is made the CW.</div>
                    <div><br clear="none">
                    </div>
                    <div>Kevin</div>
                    <div><br clear="none">
                    </div>
                    <div><br clear="none">
                    </div>
                  </div>
                  <div class="ydp7d1d2e83yiv7139577953ydpb6a32dd9yiv8583323751ydp50d5cd69yahoo_quoted" id="ydp7d1d2e83yiv7139577953ydpb6a32dd9yiv8583323751ydp50d5cd69yahoo_quoted_0454840046">
                    <div style="font-family:'Helvetica Neue', Helvetica, Arial, sans-serif;font-size:13px;color:#26282a;">
                      <div> Le mercredi 5 juin 2019 à 21:26:23 UTC−5,
                        C.Benham <a shape="rect" class="ydp7d1d2e83yiv7139577953ydpb6a32dd9yiv8583323751moz-txt-link-rfc2396E" href="mailto:cbenham@adam.com.au" rel="nofollow" target="_blank"><cbenham@adam.com.au></a>
                        a écrit : </div>
                      <div><br clear="none">
                      </div>
                      <div>Kevin,<br clear="none">
                      </div>
                      <div>
                        <div id="ydp7d1d2e83yiv7139577953ydpb6a32dd9yiv8583323751ydp50d5cd69yiv9085021920">
                          <div>
                            <p>I didn't comment earlier on your "idea
                              2".  <br clear="none">
                              <br clear="none">
                              If there no "disapproved rankings" (i.e.
                              if the voters all approve the candidates
                              they rank above bottom),<br clear="none">
                              then your suggested method is simply
                              normal  Winning Votes, which I don't like
                              because the winner can<br clear="none">
                              be uncovered and positionally dominant or
                              pairwise-beaten and positionally dominated
                              by a single other<br clear="none">
                              candidate.<br clear="none">
                              <br clear="none">
                              On top of that I don't think it really
                              fills the bill as "simple".  Approval
                              Margins (using Sort or Smith//MinMax<br clear="none">
                              or equivalent or almost equivalent
                              algorithm) would be no more complex and in
                              my opinion would be better.<br clear="none">
                              <br clear="none">
                              I would also prefer the still more simple
                              Smith//Approval.<br clear="none">
                              <br clear="none">
                              What did you think of my suggestion for a
                              way to implement your idea 1?  </p>
                            <div class="ydp7d1d2e83yiv7139577953ydpb6a32dd9yiv8583323751ydp50d5cd69yiv9085021920yqt3873327189" id="ydp7d1d2e83yiv7139577953ydpb6a32dd9yiv8583323751ydp50d5cd69yiv9085021920yqtfd25173"><br clear="none">
                              Chris <br clear="none">
                            </div>
                            <div class="ydp7d1d2e83yiv7139577953ydpb6a32dd9yiv8583323751ydp50d5cd69yiv9085021920yqt3873327189" id="ydp7d1d2e83yiv7139577953ydpb6a32dd9yiv8583323751ydp50d5cd69yiv9085021920yqtfd71007">
                              <div class="ydp7d1d2e83yiv7139577953ydpb6a32dd9yiv8583323751ydp50d5cd69yiv9085021920moz-forward-container"><br></div></div></div></div>
                      </div>
                    </div>
                  </div>
                </div>
              </div>
            </div>
          </div>
        </div>
      </div>
    
  </div></div></div></div>
            </div>
        </div></body></html>