<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 11/3/2016 8:44 AM, Michael Ossipoff
      wrote:<br>
      <br>
      <blockquote type="cite">x: A<br>
        1: C=A<br>
        1: C=B<br>
        x: B<br>
        <br>
        x = any number greater than 1.   MMPO  elects C.<br>
        <br>
        <div class="m_-7655863722893086320moz-cite-prefix">(enduote)<br>
          <br>
        </div>
        1. When I asked who is wronged, you said that the A voters are
        wronged, because, rightfully, it should be an A&B tie, and
        the A voters should have a 50% probability of their favorite
        winning. </blockquote>
      <br>
      All the voters except 2 are seriously "wronged" because their
      exclusive favourite (A or B) should be elected with 50%
      probability (instead of 0%),<br>
      <br>
      (And after randomly choosing one of A and B, only one of the two
      who voted C equal-top will not have one of their equal-favourites
      elected.)<br>
      <br>
      <blockquote type="cite">Wronged? Sure. Significantly-wronged? No.
        It certainly can't be called an outrage.</blockquote>
      <br>
      That's right. Having your favourite top-voted on half of the
      millions of ballots lose to a candidate that is voted above bottom
      on only two ballots<br>
      certainly can't be called an outrage.<br>
      <br>
      <blockquote type="cite">
        <div class="m_-7655863722893086320moz-cite-prefix">Ridiculous
          result? Sure. ...caused by ridiculous voting.<br>
          <br>
        </div>
        ...in an implausible example.</blockquote>
      <br>
      The voting isn't ridiculous. The idea that the A and B voters
      should have bothered to "cooperate" against such a minor candidate
      is ridiculous.<br>
      <br>
      The example's "implausibility" is forgiven by its extreme
      simplicity.  I think of a voting method algorithm as a
      quasi-intelligent device. It is more<br>
      understandable and maybe forgiveable if the device is "confused"
      by a relatively complicated example (say with lots of candidates
      in the Smith set<br>
      and/or cycles within cycles), but anything that gets it so wrong
      in such a simple example is unforgivably "stupid" and should be
      binned.<br>
      <br>
      <blockquote type="cite">
        <div class="m_-7655863722893086320moz-cite-prefix">That being
          said, I admit that the bad-example shows that MMPO has a
          wrong-looking-result-vulnerability.<br>
          <br>
        </div>
        They go with methods that have deluxe properties. </blockquote>
      <br>
      I don't think I agree with that premise, or that MMPO has "deluxe
      properties" (presumably you mean 'deluxe strategy properties').<br>
      <br>
      I don't like random-fill incentive and consider Later-no-Harm to
      be a mixed blessing, and I'm not very impressed by "Winning Votes<br>
       strategy".<br>
      <br>
      I like that it meets FBC.<br>
      <br>
      Chris Benham<br>
      <br>
    </div>
    <blockquote
cite="mid:CAOKDY5DakmmRczHRMAW+o0kjAT+ustZj0iPaodLV7W==WwW8Yw@mail.gmail.com"
      type="cite">
      <div dir="ltr"><br>
        Chris--<br>
        <br>
        I'd said:<br>
        <br>
        <div>
          <div class="gmail_extra">
            <div class="gmail_quote">
              <blockquote class="gmail_quote" style="margin:0 0 0
                .8ex;border-left:1px #ccc solid;padding-left:1ex">
                <div bgcolor="#FFFFFF" text="#000000">
                  <div class="m_-7655863722893086320moz-cite-prefix"><span
                      class="">
                      <blockquote type="cite">
                        <div> <span class=""></span><br>
                          <span class=""> </span></div>
                      </blockquote>
                    </span><span class="">
                      <blockquote type="cite">
                        <div>For #2, it really takes something with a
                          lot of valuable & otherwise-unobtainable
                          properties, to justify replacing Approval,
                          Score or Bucklin with it.<br>
                          <br>
                        </div>
                        I suggest that only Plain MMPO qualifies.</blockquote>
                    </span></div>
                </div>
              </blockquote>
              <div><br>
              </div>
              <div class="m_-7655863722893086320moz-cite-prefix">You
                replied:<br>
                 <span class=""><br>
                </span> Say you ask people who are used to and/or like
                any of  Approval, Score,  Bucklin,  or  plurality (aka
                FPP) or Top-Two Runoff or IRV or Benham:<br>
                <br>
                (a) Would you accept a method that in a 3-candidate
                election with 20 million voters can elect a candidate
                that is voted above equal-bottom<br>
                on only two ballots?<br>
                <br>
                x: A<br>
                1: C=A<br>
                1: C=B<br>
                x: B<br>
                <br>
                x = any number greater than 1.   MMPO  elects C.<br>
                <br>
              </div>
              <div class="m_-7655863722893086320moz-cite-prefix">(enduote)<br>
                <br>
              </div>
              <div class="m_-7655863722893086320moz-cite-prefix">1. When
                I asked who is wronged, you said that the A voters are
                wronged, because, rightfully, it should be an A&B
                tie, and the A voters should have a 50% probability of
                their favorite winning. <br>
                <br>
              </div>
              <div class="m_-7655863722893086320moz-cite-prefix">My
                answer to that was & is that the A voters have
                indicated that they don't consider B any better than C.
                They're indifferent between C and the other candidate
                whom they'd like to have a 50% probability of winning.
                Wronged? Sure. Significantly-wronged? No. It certainly
                can't be called an outrage.<br>
                <br>
              </div>
              <div class="m_-7655863722893086320moz-cite-prefix">2, You
                make it sound as if what happened in that example is
                entirely the fault of the voting-system, and maybe the
                two A=C and B=C voters..<br>
                <br>
              </div>
              <div class="m_-7655863722893086320moz-cite-prefix">But it
                was done by the all A voters and B voters. In a method
                known to look only at pairwise opposition, those voters
                ensured that A & B will have as much pairwise
                opposition as anyone can have, because they refused to
                rank eachother's candidate over C.<br>
                <br>
              </div>
              <div class="m_-7655863722893086320moz-cite-prefix">Refusal
                to co-operate with a rival can be a mistake in any
                method. It's hardly surprising if it's a mistake in MMPO
                too.<br>
                <br>
              </div>
              <div class="m_-7655863722893086320moz-cite-prefix">So
                don't let the A voters & B voters convince, you when
                they whine about that result.   ...the result of their
                refusal to co-operate with eachother.<br>
                <br>
              </div>
              <div class="m_-7655863722893086320moz-cite-prefix">Think
                of it as the chicken-dilemma defection-blunder
                magnified.<br>
                <br>
              </div>
              <div class="m_-7655863722893086320moz-cite-prefix">Ridiculous
                result? Sure. ...caused by ridiculous voting.<br>
                <br>
              </div>
              <div class="m_-7655863722893086320moz-cite-prefix">...in
                an implausible example.<br>
                <br>
--------------------------------------------------------------------------------------------<br>
                <br>
              </div>
              <div class="m_-7655863722893086320moz-cite-prefix">That
                being said, I admit that the bad-example shows that MMPO
                has a wrong-looking-result-vulnerability.<br>
                <br>
              </div>
              <div class="m_-7655863722893086320moz-cite-prefix">They go
                with methods that have deluxe properties. What else is
                new.<br>
                <br>
              </div>
              <div class="m_-7655863722893086320moz-cite-prefix">The
                problem is that, of course, MMPO gets its unqualled
                properties-combination by looking only at pairwise
                opposition.<br>
                <br>
              </div>
              <div class="m_-7655863722893086320moz-cite-prefix">For one
                thing, that means that it disregards how much positive
                support a candidate has. So it was possible to write a
                bad-example in which someone wins with practically no
                positive support...without that example's result
                seriously wronging anyone.<br>
                <br>
              </div>
              <div class="m_-7655863722893086320moz-cite-prefix">Also,
                because it's _pairwise_ opposition, a candidate with two
                factions each ranking a different candidate over C, the
                method ignores the fact that C has _two_ candidates
                ranked over hir by a large number of voters, whereas A
                & B each have only one candidate ranked over them
                (by just one more voter). <br>
                <br>
                MMPO doesn't count how many candidates are ranked over
                C, or the total number of voters who rank C lower.
                Thereby we get the wrong-looking result where C has both
                of the other candidates ranked over hir by a lot of
                people, but still beats A & B, each of which have
                just one candidate ranked over hir--by one more voter..<br>
                <br>
              </div>
              <div class="m_-7655863722893086320moz-cite-prefix">These
                are obvious inevitable results of looking only at
                pairwise opposition.<br>
                <br>
              </div>
              <div class="m_-7655863722893086320moz-cite-prefix">But,
                even with those wrong-result vulnerabilities, the
                resulting bad-example doesn't significantly wrong
                anyone.<br>
                <br>
              </div>
              <div class="m_-7655863722893086320moz-cite-prefix">Even
                though no serious wrong is done by those wrong-looking
                results, what can be done to avoid them?<br>
                <br>
              </div>
              <div class="m_-7655863722893086320moz-cite-prefix">* Rank
                some in-between compromise candidates over the worst
                one, ok? Is that too much to ask?<br>
                <br>
              </div>
              <div class="m_-7655863722893086320moz-cite-prefix">In that
                way, you rank eachother's candidate over the worst one,
                giving hir more people ranking those candidates over
                hir, to give hir more pairwise opposition than the
                candidates you like better.<br>
              </div>
              <div class="m_-7655863722893086320moz-cite-prefix"><br>
              </div>
              <div class="m_-7655863722893086320moz-cite-prefix">Set
                your rivalry aside and do that.<br>
              </div>
              <div class="m_-7655863722893086320moz-cite-prefix"><br>
                * Don't allow write-ins.  Or, if you do allow write-ins,
                make sure that you rank compromises, not just your
                favorite, so that you're all voting eachother over
                whatever unknown write-in comes along, so  that hir
                individual pairwise oppositions won't be less than those
                of the candidates you like more..<br>
                <br>
              </div>
              <div class="m_-7655863722893086320moz-cite-prefix">All the
                rank methods have strategy problems, to go with the
                deluxe properties that they bring. The more deluxe
                properties, the more kinds of wrong-looking results are
                possible. You get what you pay for.<br>
                <br>
              </div>
              <div class="m_-7655863722893086320moz-cite-prefix">So you
                deal with those special  vulnerabilities, as the price
                of the deluxe properties.<br>
                <br>
              </div>
              <div class="m_-7655863722893086320moz-cite-prefix">Michael
                Ossipoff<br>
                <br>
                <br>
                <br>
                <br>
              </div>
              <div class="m_-7655863722893086320moz-cite-prefix"><br>
                <br>
                <br>
                <br>
                <br>
                <br>
                <br>
                <br>
              </div>
              <div class="m_-7655863722893086320moz-cite-prefix"><br>
                 <br>
              </div>
              <div class="m_-7655863722893086320moz-cite-prefix"><br>
                <br>
                <br>
                <br>
              </div>
              <div class="m_-7655863722893086320moz-cite-prefix"><br>
                <br>
              </div>
              <div class="m_-7655863722893086320moz-cite-prefix"> <br>
                (b) Do you think it is fair and sensible that a voter
                who is only interested in getting their favourite
                elected can increase the chance of that<br>
                happening by fully ranking (below top) all the other
                candidates at random? <br>
                <br>
                 <br>
                <br>
                What response do you think you will get?  <br>
                <br>
                I don't live in the US, but I would think that most
                people wouldn't even believe that question (a) is
                serious, and that anyone who tries to<br>
                tell them that that example only "looks bad" would blow
                their credibility.<br>
                <br>
                Chris Benham<br>
                <br>
                <br>
                <br>
                <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">
                  <div>
                    <div class="h5">
                      <blockquote type="cite">
                        <div dir="ltr">Jameson--<br>
                          <br>
                          <div class="gmail_extra"><br>
                            <div class="gmail_quote">On Sat, Oct 29,
                              2016 at 11:16 AM, Jameson Quinn <span
                                dir="ltr"><<a moz-do-not-send="true"
                                  href="mailto:jameson.quinn@gmail.com"
                                  target="_blank">jameson.quinn@gmail.com</a>></span>
                              wrote:<br>
                              <blockquote class="gmail_quote"
                                style="margin:0px 0px 0px
                                0.8ex;border-left:1px solid
                                rgb(204,204,204);padding-left:1ex">
                                <div dir="ltr"><br>
                                  <div>....Here's my "ideal
                                    characteristics" for a political
                                    single-winner election system, more
                                    or less in descending order of
                                    importance:</div>
                                  <div>
                                    <ol>
                                      <li>FBC<br>
                                      </li>
                                    </ol>
                                  </div>
                                </div>
                              </blockquote>
                              <div><br>
                              </div>
                              <div>Yes, I consider FBC to be #!.<br>
                                 <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 dir="ltr">
                                  <div>
                                    <ol>
                                      <li><br>
                                      </li>
                                      <li>Handles center squeeze (ie,
                                        some form of weakened Condorcet
                                        guarantee that's compatible with
                                        FBC) <br>
                                      </li>
                                    </ol>
                                  </div>
                                </div>
                              </blockquote>
                              <div>But maybe the goal of electing the
                                CWs unnecessarily complicates votiing.
                                Maybe someday, there won't be a bottom
                                set, for most voters, and, with honest
                                elections & honest, open media, it
                                will be clear who's the CWs. But now,
                                Approval's best strategy doesn't require
                                that. Brams pointed out that Approval's
                                results can be better than the CWs.<br>
                                <br>
                              </div>
                              <div>But, when it's desirable to elect the
                                CWs, and it isn't obvious who's the CWs,
                                then the wv methods, and methods with
                                wv-like strategy (such as MMPO), are the
                                methods that make it easiest to protect
                                the CWs.<br>
                                <br>
                              </div>
                              <div>Also, it should be noted that you
                                can't be sure how often there will be a
                                CWs, under different and better
                                conditions.<br>
                                 <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 dir="ltr">
                                  <div>
                                    <ol>
                                      <li><br>
                                      </li>
                                      <li>Relatively simple to explain<br>
                                      </li>
                                    </ol>
                                  </div>
                                </div>
                              </blockquote>
                              <div><br>
                              </div>
                              <div>Approval & Score are easy to
                                explain. I've had conversations in which
                                only Approval & Score were accepted
                                as being plain & un-elaborate enough
                                to be acceptable..<br>
                                <br>
                              </div>
                              <div>You know that SARA & XA are
                                complicated and not easy to explain.
                                I've tried explaining them.<br>
                                <br>
                              </div>
                              <div>MMPO?:<br>
                                <br>
                              </div>
                              <div>"The winner is the candidate who has
                                fewest people voting someone else over
                                hir."<br>
                                <br>
                              </div>
                              <div>{...some same other candidate) <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 dir="ltr">
                                  <div>
                                    <ol>
                                      <li><br>
                                      </li>
                                      <li>Minimal strategic burden</li>
                                    </ol>
                                  </div>
                                </div>
                              </blockquote>
                              <div><br>
                              </div>
                              <div>We often hear it said that Approval
                                has a large strategic burden, but, for
                                most people, with our current
                                candidate-lineup, there's nothing
                                difficult about it: Approve (only) the
                                progressive candidates. That's optimal
                                for most people.<br>
                                <br>
                              </div>
                              <div>And if the time comes when, for most
                                people, there isn't a bottom-set, then
                                that will be a happy circumstance, in
                                which it doesn't matter terribly anyway,
                                which candidate wins.<br>
                                <br>
                              </div>
                              <div>In such conditions, with honest
                                elections and honest, open media, it
                                will likely be pretty obvious who's the
                                CWs, and, in the absence of
                                tep-set/bottom-set, the best strategy
                                will be to just approve down to the CWs.<br>
                                <br>
                              </div>
                              <div>And, if there were no bottom-set, and
                                if it happened that it _wasn't_ clear
                                who the CWs was, and it was 0-info, then
                                it would just be a matter of approving
                                down to the expected winning
                                merit-level. Maybe, under those
                                conditions, that would be the
                                candidate-mean.  Or maybe (as now)  the
                                estimated mean merit of what voters want
                                (which can be estimated by the
                                candidates' merit-midrange, if the
                                election is 1D.  ....but it might not
                                remain 1D under different and better
                                conditions).<br>
                                <br>
                              </div>
                              <div>..but I repeat that, with only your
                                former top-set remaining as winnable
                                candidates, it won't make as much
                                difference which one wins anyway.<br>
                                <br>
                              </div>
                              <div>People often consider Approval voting
                                more difficult than it is.<br>
                                <br>
                                 <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 dir="ltr">
                                  <div>
                                    <ol>
                                      <li>Summable (ideally O(N), no
                                        worse than O(N²) in practice,
                                        though I might accept some
                                        special pleading for the use of
                                        prior polling to reduce to
                                        O(N²).)<br>
                                      </li>
                                      <li>Handles CD, or at least, CD
                                        offensive strategies don't in
                                        practice mess up the center
                                        squeeze properties.<br>
                                      </li>
                                    </ol>
                                  </div>
                                </div>
                              </blockquote>
                              <div><br>
                              </div>
                              <div>MMPO meets Weak CD.<br>
                                 <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 dir="ltr">
                                  <div>
                                    <ol>
                                      <li><br>
                                      </li>
                                      <li>Some arguable track record<br>
                                      </li>
                                    </ol>
                                  </div>
                                </div>
                              </blockquote>
                              <div><br>
                              </div>
                              <div>Of course at EM we discuss two
                                completely separate and different method
                                proposal choices:<br>
                                <br>
                              </div>
                              <div>1. Proposals to electorates who have
                                only had Plurality.<br>
                                <br>
                                2. Later proposals to replace one better
                                voting system with another.<br>
                                <br>
                              </div>
                              <div>If we're talking about #1, then there
                                are only a few to choose from:<br>
                                <br>
                              </div>
                              <div>Approval, Score, and Bucklin.<br>
                                <br>
                                <div><br>
                                </div>
                                I suggest that all 3 of those should be
                                offered to initiative-proposal
                                committees, and that the public should
                                be polled, or consulted in
                                "focus-groups", regarding which of those
                                3 methods they'd  accept<br>
                                <br>
                                <br>
                              </div>
                              <div>For #2, it really takes something
                                with a lot of valuable &
                                otherwise-unobtainable properties, to
                                justify replacing Approval, Score or
                                Bucklin with it.<br>
                                <br>
                              </div>
                              <div>I suggest that only Plain MMPO
                                qualifies.<br>
                                <br>
                              </div>
                              <div><br>
                              </div>
                              <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 dir="ltr">
                                  <div>
                                    <ol>
                                      <li><br>
                                      </li>
                                    </ol>
                                    <div> Approval does well on 1,3,4,
                                      and 6, is OK on 2, and bad on 4
                                      and 5.</div>
                                  </div>
                                </div>
                              </blockquote>
                              <div><br>
                                <br>
                              </div>
                              <div>No. Contrary to what we so often
                                hear, Approval doesn't have a high
                                strategic burden, as I discussed above.
                                <br>
                                <br>
                              </div>
                              <div>Other than MMPO's CD, improvements
                                over Approval by more complicated
                                methods are illusory.<br>
                                <br>
                              </div>
                              <div>And your standard #5 was summability
                                and count-complexity. Approval is
                                precinct-summable, and its count is the
                                easiest and least computation-intensive,
                                among voting-systems.<br>
                                <br>
                              </div>
                              <div>Michael Ossipoff<br>
                                <br>
                              </div>
                              <br>
                            </div>
                            <br>
                          </div>
                        </div>
                        <br>
                        <br>
                        <p color="#000000" align="left"><br>
                        </p>
                      </blockquote>
                      <p><br>
                      </p>
                    </div>
                  </div>
                </div>
                <br>
                ----<br>
                Election-Methods mailing list - see <a
                  moz-do-not-send="true" href="http://electorama.com/em"
                  rel="noreferrer" target="_blank">http://electorama.com/em</a>
                for list info<br>
                <br>
              </blockquote>
            </div>
            <br>
          </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: 2016.0.7859 / Virus Database: 4664/13333 - Release
        Date: 11/02/16</p>
    </blockquote>
    <p><br>
    </p>
  </body>
</html>