[EM] Does this implement AERLO without recounts?

[Ted Stern]

I've been pondering how to implement something like Mike Ossipoff's AERLO
idea without requiring a recount, the object being to discourage bullet-voting 
and burying tactics and encourage sincere ranking.

Let's assume there is a ranked-choice ballot for N candidates allowing equal
ranking.  The voter can select an approval cutoff above the lowest rank if
desired, otherwise the approval cutoff is the rank above the lowest candidate
ranking.

In addition to the standard Condorcet grid (Grid_0) an additional N grids
would be updated.

For each ballot, 

  The voter's ranked choices would be counted two ways:
  Ranking_1) The default ranking entered by the voter.
  Ranking_2) Promote all approved candidates to equal first place rank.

  Loop over i = 1 to N:
  For the i-th Grid:

    - if candidate_i is in the set of candidates approved by the voter,
      update Grid_i with Ranking_1.

    - otherwise update Grid_i with Ranking_2.

After all ranked votes are tallied, if there is a circular tie in Grid_0 that
leads to a victory for candidate_i using some robust Condorcet completion
method (e.g, wv-CSSD), use Grid_i instead with the same method.

I'm not sure the result is exactly the same as stated here:

,----[ http://article.gmane.org/gmane.politics.election-methods%3A4393 ]
| For methods that aren't pairwise-count methods, this is what AERLO is:
| 
| The voter may designate a line in hir ranking, to indicate that if no one 
| above that line wins, then that voter wants to promote to 1st place all of 
| hir above-line candidtates, and have the ballots counted again.
| 
| This is done for all ballots having an automatic equal ranking line, and, 
| when that has been done, there's only one recount, after the parageraph 
| before this has been carried out for all rankings having an automatic equal 
| ranking line with no winners above it.
| 
| (For pairwise-count methods, such as Condorcet, there's an additional 
| requirement for the promotion: It only takes place for a particular ranking 
| if there's a cycle with members above and below that ranking's AERLO line. 
| That avoids unnecessary promotions in those methods).
| 
| Yes, as I was saying in an earlier posting, ERBucklin(whole) meets FBC, but 
| not Strong FBC. To meet Strong FBC, that method needs AERLO.
| 
| For rank methods, AERLO makes FBC compliance into Strong FBC compliance, and 
| makes WDSC compliance into SDSC compliance.
| 
| Mike Ossipoff
`----

But it appears to meet the original intention -- "don't flatten my rankings
unless there's a circular tie and none of my approved candidates win."

Comments?

Ted
-- 
Send real replies to
	ted stern at u dot washington dot edu

Frango ut patefaciam -- I break that I may reveal