[EM] voting reform effort in DENVER - PLEASE HELP

Chris Benham chrisjbenham at optusnet.com.au
Sun Jun 11 09:09:52 PDT 2006

Jan Kok wrote:

>I would like for Denver to adopt something better than plain old IRV.
Looking at the link, I was glad to see that they were proposing to 
implement IRV properly,
and not some abomination like automated Top-Two Runoff (TTR) or not 
allowing full ranking.

>1. Proportional Representation. Rob Richie prefers it, several others
>favor it, I fully support it (though I haven't studied it in any
>depth), Gary Swing (a Coloradoan who has some expertise in PR)
>supports it... Where there's a will, there's a way... By the way,
>resistance to change from incumbent council members could be reduced
>by merging districts into one common PR district when the incumbent
>doesn't run for re-election, or loses an election.
>1a. Asset Voting (which could be done with a Range Voting ballot -
>just normalize...), or Fractional Approval Asset Voting (which can be
>done with an Approval Voting style ballot).
>http://math.temple.edu/~wds/homepage/multiwin.pdf . I'll mention this
>to Rob Richie and the irv-l list as a SIMPLER way to achieve PR than
>STV. Who knows, if Rob really wants PR but has been holding back
>because he thinks STV is too complicated, maybe he will take an
>interest in some version of Asset Voting.
>1b. I just noticed this paper as well...
>http://math.temple.edu/~wds/homepage/rerange.pdf - Reweighted Range
>Voting. I'll need to read it before deciding whether to recommend it.
>2. Approval Voting. It's as simple as it gets! Many voters may not
>even notice the change. Because this is the simplest, cheapest change
>(other than reverting to straight Plurality), I think this is equal or
>better than IRV in viability.
>2a. AV+. If people object to giving an identical vote to their
>favorite and the lesser evil, we could propose AV+, which allows
>voters to mark a candidate as "Favorite", "Approved", or nothing
>(disapproved). Favorite and Approved votes are counted and reported
>separately, but are added together for determining the winner. Any
>existing voting machines can count AV+ ballots, and make sure that
>voters don't mark a candidate as both Favorite and Approved.
>2b. AV+PW ("pairwise") An Abd Lomax invention, I think. Is this
>Condorcet with equal rankings allowed and three levels (favorite,
>approved and nothing)? If so, isn't there the possibility of circular
>ties? If so, what is the completion method? Is there a simple way to
>describe the method that ordinary people can grok? This method
>probably requires the voting machines to store and send ballot images,
>so it can't be done with primitive voting machines, but it should be
>doable on voting machines that can handle IRV.
>3. IRV/AV hybrid. (Thanks to Abd Lomax for bringing this idea to my
>attention.) Use ranked ballots with equal rankings allowed. A ballot
>like A=B>C=D contributes one vote to A and one vote to B as long as
>either A or B survives, then contributes one vote to C and one vote to
>D as long as either of them survive. In each round, eliminate the
>candidate with the fewest votes. If someone sees something horribly
>wrong with this method, please say so on the EM list, before I
>embarrass myself on the irv-l list by proposing this.
>4. BTR-IRV (equivalent to Condorcet with Plurality completion). A
>candidate is "lowest ranked" on a given ballot if he is mentioned in
>last place among all surviving candidates, or he is not mentioned at
>all on the ballot. In each round, eliminate the candidate who is
>lowest ranked on the most ballots, until there is just one survivor.
>(I've stated that a bit differently than before, but I think it's
>equivalent, right? And there is a reasonably simple hand counting
>procedure: count up the number of mentions of each candidate on all
>the ballots, eliminating duplicate mentions on any ballot, and
>subtract that from the number of ballots, to obtain the number of
>ballots that don't mention each candidate. Then sort the ballots
>according to last choices. In each round, eliminate the candidate who
>is lowest ranked on the most ballots, and redistribute the ballots to
>the last surviving choices.)
>5. BTR-IRV/AV hybrid (The "mongrel" method? :-) It's tempting to think
>this is equivalent to Condorcet with Approval completion, but I'm not
>sure it is. This needs more thought... Would the counting procedure
>given above for BTR-IRV work here and do the "right thing"?
The only alternative on this list that I agree is really better than IRV 
is PR

Elections should be decided directly by the votes of voters, and as far 
as practicably possible all voters
should have equal/power weight  in this process.  So I reject Asset 
Voting. Candidates should not be
Approval Voting doesn't  meet Majority for Solid Coalitions and is 
vulnerable to disinformation.

The one you give as "IRV/AV hybrid"  is  ER-IRV(whole) and in my view is 
far worse than standard
IRV because it is much more vulnerable to "paradoxical" Pushover-like 
strategy, as I pointed out with
an an example here a while ago.

I suggest  that if  you are looking for good single-winner alternatives 
to IRV, your list should include
both versions of  Definite Majority Choice(DMC): the one that interprets 
all ranked candidates as
approved (of course allowing truncation), and the one that allows voters 
to enter an approval cutoff
so that they can rank unapproved candidates.

This has two simple definitions:
(1) Elect the CW if there is one. If not, eliminate (drop from the 
ballots) the least-approved candidate.
Repeat until there is a winner.
(2) Eliminate all candidates that are pairwise-beaten by a more-approved 
candidate. One of the
remaining candidates x will pairwise beat all the other remaining 
candidates. Elect x.

This method meets Condorcet, Clone Independence and Mono-raise.

Number 4 on your list is all wrong. As I understand it, BTR-IRV stand 
for "Bottom Two Runoff-IRV",
which at each stage eliminates the pairwise-loser of the  two remaining 
candidates with the fewest top-preferences.
It's just an attempt to smuggle a not very  good Condorcet method past 
IRV supporters.
Neither it nor the method you define is "equivalent to Condorcet with 
Plurality completion".

The method you define is  Coombs or one of the two versions of  Coombs 
(the worse one). It is far worse than
IRV. The version you give fails "Majority Favourite".  It is possible 
that a candidate with more than half the first
preference votes will be eliminated. (The other version has a majority 
stopping rule).

Chris  Benham

More information about the Election-Methods mailing list