<div dir="auto">I think that if you were looking at the success rates of unilateral burial options along with the basic burial defense of truncation below sincere CW ... then the difference between Winning Votes and Margins would be more striking (still in favor of wv).</div><br><div class="gmail_quote"><div dir="ltr" class="gmail_attr">On Wed, Sep 20, 2023, 2:12 PM Michael Ossipoff <<a href="mailto:email9648742@gmail.com">email9648742@gmail.com</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div dir="auto">Thanks for the experiment. I hope that Tideman’s organization won’t be promoting the old margins version of RP.  …unless just for maybe choosing between pizza-toppings or movies.</div><div><br><div class="gmail_quote"><div dir="ltr" class="gmail_attr">On Wed, Sep 20, 2023 at 07:30 Colin Champion <<a href="mailto:colin.champion@routemaster.app" target="_blank" rel="noreferrer">colin.champion@routemaster.app</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
  
    
  
  <div>
    <font face="Helvetica, Arial, sans-serif">Well, I coded up
      Minimax(WV) for my own evaluation. Rather to my disgust (and
      contrary to Darlington) I find that it does indeed outperform
      Margins for truncated sincere ballots. I ran a large number of
      trials with 10001 voters under a spatial model, 8 candidates being
      truncated to 4. Minimax (margins)=83.35% correct,
      minimax(wv)=84.09%. Other methods which outperformed standard
      minimax in the simulation include Approval Sorted Margins (in an
      ordinal version suggested by Ted Stern): 84.14%; Black: 84.22%;
      Smith,Borda: 84.22%. <br>
         Hastily written and unreliable code, not to be trusted. <br>
            CJC<br>
    </font></div><div><br>
    <div>On 18/09/2023 22:30, Michael Ossipoff
      wrote:<br>
    </div>
    <blockquote type="cite">
      
      <div dir="auto">Hi Colin—</div>
      <div dir="auto"><br>
      </div>
      <div dir="auto">Yes, margins beats wv for social-utility under
        sincere voting when there’s that vanishingly rare natural
        top-cycle.</div>
      <div dir="auto"><br>
      </div>
      <div dir="auto">But Darlington & Tideman evidently aren’t
        considering resistance to offensive strategy, which is a much
        bigger threat than natural top-cycles.</div>
      <div dir="auto"><br>
      </div>
      <div dir="auto">Protecting the CW from offensive strategy is more
        important than SU in natural top-cycles.</div>
      <div dir="auto"><br>
      </div>
      <div dir="auto">…& is better for SU.</div>
      <div><br>
        <div class="gmail_quote">
          <div dir="ltr" class="gmail_attr">On Mon, Sep 18, 2023 at
            07:03 Colin Champion <<a href="mailto:colin.champion@routemaster.app" target="_blank" rel="noreferrer">colin.champion@routemaster.app</a>>
            wrote:<br>
          </div>
          <blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
            <div> <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 href="https://arxiv.org/abs/1807.01366" target="_blank" rel="noreferrer">https://arxiv.org/abs/1807.01366</a><br>
              </font></div>
            <div><br>
              <div>On 18/09/2023 07:57, Michael Ossipoff wrote:<br>
              </div>
              <blockquote type="cite">
                <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" target="_blank" rel="noreferrer">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" target="_blank" rel="noreferrer">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" rel="noreferrer">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" rel="noreferrer">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 noreferrer" target="_blank">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 noreferrer" target="_blank">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 noreferrer" target="_blank">https://electorama.com/em</a>
                                  for list info<br>
                                </blockquote>
                              </div>
                            </div>
                          </div>
                        </blockquote>
                      </div>
                    </div>
                  </div>
                </div>
                <br>
                <fieldset></fieldset>
                <pre>----
Election-Methods mailing list - see <a href="https://electorama.com/em" target="_blank" rel="noreferrer">https://electorama.com/em</a> for list info
</pre>
              </blockquote>
              <br>
            </div>
            ----<br>
            Election-Methods mailing list - see <a href="https://electorama.com/em" rel="noreferrer noreferrer" target="_blank">https://electorama.com/em</a>
            for list info<br>
          </blockquote>
        </div>
      </div>
    </blockquote>
    <br>
  </div>

----<br>
Election-Methods mailing list - see <a href="https://electorama.com/em" rel="noreferrer noreferrer" target="_blank">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">https://electorama.com/em</a> for list info<br>
</blockquote></div>