[EM] Re: Condorcet package-wvx

[Ted Stern]

More comments on Dave's proposal for single winner elections:

On 23 Feb 2005 at 01:00 PST, Dave Ketchum wrote:
>
> Trying to define a complete Condorcet based election method:

<...justification...>

>
> How vote?  I suggest one list of candidates, with room beside each for 
> ranking.  Ranks read as 1 (top) to 9, with indecipherable ranked below 9 
> (CAN happen with paper ballots such as absentee; should be impossible with 
> electronic voting machine).  Blanks are lower as not voted:
>      A number repeated gives all with that number the same rank.
>      Skipping numbers does not disturb order of used numbers.
>      1-9, indecipherable, and blank, is 11 choices - more than likely 
> useful with even 99 candidates, but anyone seeing a problem here go solve
> it.

One more comment here.  Nine choices is a lot!  Some voters have problems with
even one -- they circle the blanks on an optical scan ballot, or mark the
blank with an X.

If you want more than 3 choices, five or six would be within their grasp.  I
like the idea of illegible and blank, adding 2 extra choices.

I would add this extra feature, default approval cutoff.

Any ranked candidates above indecipherable would be considered approved,
unless a higher approval cutoff were entered.

> 	Truncation PERMITTED: As an extreme, a voter happy with Plurality's
> service should be permitted to vote for one preferred candidate, and be 
> done with it.  We are also supporting the voter who had pains with 
> Plurality spoilers and NEEDS to rank multiple candidates.
>      Write-ins permitted:  Must be, though each voter may be restricted 
> to a single written per candidate list.
>
> Counting votes:
>      (wv) seems the appropriate choice.  If two voters rank a pair of 
> candidates (a=b) as equal, then (a>b) and (b>a) should each get one count.

This is not standard tallying of ranked ballots for Condorcet.  It creates
votes.  I maintain it won't pass the constitutional test.  A Condorcet
completion method using this pairwise tally is not using (wv), it is using
something else, now called "wvx".

If you split the vote between candidates, you are at least not creating votes,
and could pass a constitutional test.  Maybe call a completion method using
that ranking winning votes, split equal rank (wvs).  But it is still not wv.

>      The Condorcet array for each precinct, the whole district, and any 
> meaningful subset of the district, should be public.
>      Cycles CAN happen, and should be recognized as near ties.  How to 
> resolve cycles must be solved before system is ready for voting.
>      Write-ins counted:  Yes, as if one more candidate - each of which 
> can be voted for by more than one voter (could demand advance notice but, 
> right now, no notice makes sense to me).

While I think the basic idea is reasonable, you need to commit to some kind of
cycle resolution.  So what shall it be?  Ranked Pairs, BeatPath or River?  Or
something else?

Before you decide, consider how you would handle the following cases.  This
could easily happen within a short time after adoption of Condorcet.  Think of
A as "Green", B as "Democratic", C as "Republican".

       Sincere preference
       35: A>B>>C
       25: B>A>>C
       40: C>>A>B

       (1) B defects, buries A
       35: A>B>>C
       25: B>>C>A
       40: C>>A>B

       (2) C truncates, B defects, buries A
       35: A>B>>C
       25: B>>C>A
       40: C>>A=B

       (3) C truncates, B truncates
       35: A>B>>C
       25: B>>A=C
       40: C>>A=B

       (4) A changes approval cutoff as a poison pill for B
       35: A>>B>C
       25: B>>A=C
       40: C>>A=B

       (5) B and A change approval cutoff but retain sincere preference
           ordering.
       35: A>>B>C
       25: B>>A>C
       40: C>>A=B

Now consider Jobst Heitzig's Grand Compromise Proposal last fall, which is
similar to yours.

  http://lists.electorama.com/pipermail/election-methods-electorama.com/2004-October/014024.html

And recall that James Green-Armytage proposed a similar ranking, Approval
Weighted Pairwise, last June.

  http://lists.electorama.com/pipermail/election-methods-electorama.com/2004-June/013241.html

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

Frango ut patefaciam -- I break so that I may reveal