<div dir="auto">To give credit where credit is due, I want to thank Richard Lung for getting us thinking along the inclusion/exclusion axis corresponding to near-first/near-last ballot ranks ... That contributed to my confidence in the lottery/anti-lottery idea as a means of de-cloning Borda, Copeland. and Kemeny-Young, which has panned out so beautifully.<div dir="auto"><br></div><div dir="auto">Thanks, Richard!!!</div></div><br><div class="gmail_quote"><div dir="ltr" class="gmail_attr">El vie., 14 de ene. de 2022 8:18 a. m., Forest Simmons <<a href="mailto:forest.simmons21@gmail.com">forest.simmons21@gmail.com</a>> escribió:<br></div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div dir="auto"><div>Two equivalent ways of decloning Borda ... both relying on a combination of the benchmark and anti-benchmark lotteries BML and ABML.<div dir="auto"><br></div><div dir="auto">First method:</div><div dir="auto"><br></div><div dir="auto">Ballot B contributes to candidate X's score the sum of BML(j) [summed over candidates j ranked below X] minus the sum of ABML(k) [summed over candidates k ranked above X].</div><div dir="auto"><br></div><div dir="auto">Think of it this way ... the importance of candidate X to the voter of ballot B is the estimated strength of the candidates she likes less than X, minus the estimated weakness of the candidates she likes better:</div><div dir="auto"><br></div><div dir="auto">[If the ones you like better are weak and the ones you dislike in comparison are strong, you should support X.]</div><div dir="auto"><br></div><div dir="auto">[Also this is a natural DSV kind of way to specify an approval cutoff]<br><div dir="auto"><br></div><div dir="auto">Method 2.</div><div dir="auto"><br></div><div dir="auto">The score for X is the difference in the BML expectation of X's pairwise matrix row and the ABML expectation of X's pairwise matrix column.</div><div dir="auto"><br></div><div dir="auto">The former is a source of pride for X ...the weighted average pairwise scores of X over the candidates, where the weights are estimates of the strengths of those candidates.</div><div dir="auto"><br></div><div dir="auto">The latter expectation is a source of chagrin for X ... the weighted average of the scores against X by the other candidates, where the weights are estimates of the weakness of those candidates... if you are soundly trounced by a weak candidate, you should feel chagrin, if not shame!</div><div dir="auto"><br></div><div dir="auto">This gives another idea for a kind of MinMax/MaxMin method:</div><div dir="auto"><br></div><div dir="auto">Let M(X) be the Max value of the product of ABML(k)*P(k,X), where P(k, X) is the percentage of ballots on which k is ranked ahead of X.</div><div dir="auto"><br></div><div dir="auto">Similarly, let m(X) be the min value of P(X, j)*BML(j), where P(X,j) is the percentage of ballots ranking X above j.</div><div dir="auto"><br></div><div dir="auto">Elect argmin(M(X)-m(X)).</div><div dir="auto"><br></div><div dir="auto">[Maybe this is another key to fpC-fpA]</div><div dir="auto"><br></div><div dir="auto">Note the strong symmetry of this method, also.</div><div dir="auto"><br></div><div dir="auto">All in all, I'm pretty excited about the usefulness of the anti-benchmark lottery in conjunction with the benchmark lottery.</div><div dir="auto"><br></div><div dir="auto">If BML(X) is an estimate of X's viability, then ABML(X) is an estimate of X's weakness.</div><div dir="auto"><br></div><div dir="auto">Of course, the BML is not the only clone-free lottery ...Jobst's MaxParC, the Martin Harper Lottery, the Random Approval Lottery, the Nash Lottery and its "Ultimate Lottery" generalization ...all of these and others have their corresponding anti-lottery counterparts ... just reverse the ballots to get them.</div><div dir="auto"><br></div><div dir="auto">Why didn't we think of this before? It was staring us right in the face! </div><div dir="auto"><br></div><div dir="auto">What other obvious ideas are still invisible to us?</div><div dir="auto"><br></div><div dir="auto">-Forest</div></div><br><br><div class="gmail_quote"><div dir="ltr" class="gmail_attr">El vie., 14 de ene. de 2022 6:24 a. m., Forest Simmons <<a href="mailto:forest.simmons21@gmail.com" target="_blank" rel="noreferrer">forest.simmons21@gmail.com</a>> escribió:<br></div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div dir="auto"><div>Kristofer, <div dir="auto"><br></div><div dir="auto">I have a feeling you might like this method:</div><div dir="auto"><br></div><div dir="auto">For simplicity assume complete rankings.</div><div dir="auto"><br></div><div dir="auto">For each candidate X, let fp(X) and lp(X), respectively, be the total number of first place votes and last place votes, respectively, of the candidates that are pairwise defeated by X, and pairwise defeat X, respectively. </div><div dir="auto"><br></div><div dir="auto">Elect argmax(fp(X)-lp(X))</div><div dir="auto"><br></div><div dir="auto">What do you think?</div><div dir="auto"><br></div><div dir="auto">-Forest</div><div dir="auto"><br></div><div dir="auto"><br></div><br><br><div class="gmail_quote"><div dir="ltr" class="gmail_attr">El vie., 14 de ene. de 2022 2:11 a. m., Kristofer Munsterhjelm <<a href="mailto:km_elmet@t-online.de" rel="noreferrer noreferrer" target="_blank">km_elmet@t-online.de</a>> escribió:<br></div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">On 14.01.2022 11:01, Kristofer Munsterhjelm wrote:<br>
> On 14.01.2022 04:05, Forest Simmons wrote:<br>
>> I've always admired the basic idea of Kemeny-Young, namely a cost metric<br>
>> for conversion of one ranking into another ... simply the number of<br>
>> transpositions (elementary swaps) required.<br>
>><br>
>> The deal-breaker draw-back of the Kemeny cost metric is that it is clone<br>
>> dependent ... if it were independent of clones the " cost" of<br>
>> transposing AB to BA would be the same as the net cost of transposing<br>
>> the rank order of respective clones of A and B:<br>
>><br>
>> the order a1a2a3b1b2b3<br>
>> to the order b3b2b1a3a2a1<br>
> <br>
> Of course, I have to restate here that Ranked Pairs is such a method:<br>
> the metric is just leximax. (ordering a is better than ordering b if the<br>
> greatest pairwise victory consistent with a is higher than the greatest<br>
> pairwise victory consistent with b, with tiebreaks for second greatest,<br>
> third greatest, etc.) And it's cloneproof.<br>
<br>
Here's a thought: let parameterized p-Kemeny be: take the lp norm of the<br>
vector of all pairwise victories consistent with the output ordering.<br>
Break ties by gradually lowering p, i.e. l(p-epsilon), further ties by<br>
l(p-2 epsilon), etc.<br>
<br>
Then Kemeny is l_1 and RP is the limit as p approaches infinity, l_inf.<br>
I think? Are any other values of p interesting?<br>
<br>
-km<br>
</blockquote></div></div></div>
</blockquote></div></div></div>
</blockquote></div>