[EM] Honest equal-rank/truncation?

[Toby Pereira]

 If you determinize COWPEA it would lose independence of clones (cloned candidates would struggle to get elected), so it could give quite disproportional results because of that. Off the top of my head, not sure about IIB and ULC.
    On Saturday, 18 June 2022, 17:33:40 BST, Kristofer Munsterhjelm <km_elmet at t-online.de> wrote:  
 
 On 18.06.2022 15:43, Toby Pereira wrote:
> As an aside, "Random Approval" is the single-winner version of COWPEA
> Lottery. https://electowiki.org/wiki/COWPEA_Lottery
> <https://electowiki.org/wiki/COWPEA_Lottery>

A possible way to determinize a lottery is to take the mode (the council
elected most often). That generally preserves monotonicity. Would it
preserve IIB and ULC?

I once thought about a semiproportional method that goes: pick a random
ballot, elect the candidate on it, eliminate this candidate from every
other ballot, and repeat. The determinized version is not PR, but it
gets closer the more seats there are. Possibly the determinized COWPEA
method also loses its proportionality and becomes only semiproportional,
but I haven't checked.

(I wonder if the determinization of a DAC/DSC version would be
Droop-proportional... something like "take random ballot,
round-eliminate the uneliminated candidate ranked last, repeat taking
random ballots until only one candidate remains, then elect and
eliminate (for good) this candidate, then restore the round-uneliminated
candidates and repeat".)

-km
  
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.electorama.com/pipermail/election-methods-electorama.com/attachments/20220618/77c22201/attachment.html>