[EM] Fwd: Re: Fwd: [uscv-voting_activists] Coinkie-dink? Aspenmayor-1273, Seat 1-1273 & Seat 2- 1273 ???]

[Terry Bouricius]

I contacted the administrator of the Aspen IRV election tally to clarify 
some things about the results reporting (like why winners appear to have 
the same vote totals). I am pasting his reply below. Anyone who wants to 
understand the details of the tally can read this (otherwise, it's not 
particularly interesting).

Terry Bouricius

--------------------------
Very quickly:

1.  There were no invalid votes for mayor.  See
http://www.aspenpitkin.com/pdfs/depts/38/electionresults09.pdf , which
you get to by following the May 2009 Election Results link from middle
of page http://www.aspenpitkin.com/depts/38/

2.  I believe the city is releasing the data to people who ask for it in
person at city hall.  I don't know if you can get it via a public
records request from afar.

3.  The confusion about exhausted and invalid ballots stems from using
ChoicePlus Pro, which was written to follow Cambridge's conventions on
terminology, threshold calculations, etc.  To implement Aspen's rule
with ChoicePlus Pro, we had to insert a dummy code 0 for a skipped
ranking.  Candidate 0 is then excluded from the count.  Under Cambridge
rules, these ballots are all reported as valid ballots, which was true
in the mayor's race, but there were 23 ballots invalid under Aspen's
rules for council.

The threshold calculation by ChoicePlus Pro is based on Cambridge
interpretation of valid ballots (hence 2544 for both races), which leads
to the 1273 threshold, and under Cambridge rules, the counts continue
until a candidate is elected with exactly the threshold, which is why
all winning candidates are listed as having 1273 votes in the final
round -- ChoicePlus Pro declares them elected when they hit 1273.  Under
Aspen rules, the election is over when you are down to 2 candidates, so
the final html pages are not applicable to Aspen.

If you look at
http://www.aspenpitkin.com/pdfs/depts/38/electionresults09.pdf , you'll
see Ireland elected with 1273, which in fact is a coincidence, since
there were 16 blank ballots in that race, meaning the threshold in that
race really was (2544-16)/2+1=1265.  In the other two races, the winning
candidates 1233 and 1073 votes in the final round.

As for exhausted ballots in the council race, you start out with 34
undervotes and 23 invalid votes.   There are 86 actual exhausted
ballots, which are people who did not rank any of the 4 advancing
candidates.  If you ranked an eliminated candidate first but an
advancing candidate second, your ballot counted for the advancing
candidate in the first round of the council IRV tally.  It did not
exhaust because your first choice was an eliminated candidate.

----- Original Message ----- 
From: "Kathy Dopp" <kathy.dopp at gmail.com>
To: "EM" <election-methods at lists.electorama.com>
Sent: Tuesday, May 12, 2009 2:36 PM
Subject: [EM] Fwd: Re: Fwd: [uscv-voting_activists] Coinkie-dink? 
Aspenmayor-1273, Seat 1-1273 & Seat 2- 1273 ???]


Some possible rebuttals to Terry's last post.


---------- Forwarded message ----------
From: Chris Telesca <cjtelesca at earthlink.net>
Date: Tue, May 12, 2009 at 6:14 AM
Subject: Re: Bouricius [Fwd: Re: [EM] Fwd: [uscv-voting_activists]
Coinkie-dink? Aspen mayor-1273,Seat 1-1273 & Seat 2- 1273 ???]
To:
Cc: Joyce McCloy <jmc27106 at earthlink.net>, Kathy Dopp 
<kathy.dopp at gmail.com>




Joyce McCloy wrote:
>
>
> -------- Original Message --------
> Subject: Re: [EM] Fwd: [uscv-voting_activists] Coinkie-dink? Aspen 
> mayor-1273,Seat 1-1273 & Seat 2- 1273 ???
> Date: Mon, 11 May 2009 17:36:04 -0600
> From: Kathy Dopp <kathy.dopp at gmail.com>
> Reply-To: kathy.dopp at gmail.com
> To: Terry Bouricius <terryb at burlingtontelecom.net>
>
>
>
> On Mon, May 11, 2009 at 5:21 PM, Terry Bouricius
> <terryb at burlingtontelecom.net> wrote:
>>
>> Kathy,
>>
>> As I understand it (and I am not directly involved, so may have this
>> imperfectly)
>> 1. In Aspen there were zero invalid ballots...The "spoiled" ballot
>> estimation was made by an election official referring to several voters
>> who made mistakes, but then automatically had their ballots bounced 
>> back
>> by the voting machine for correction. In the end, I believe there were 
>> no
>> invalid IRV ballots.

Well we'd really have to look at the raw voting data before making
that assumption.
>>
>> 2. I believe the city is going to, or has, posted all of the ballot
>> rankings on the Internet (as do San Francisco and Burlington) so that 
>> now
>> that the random sample audit has confirmed that the scanners recorded 
>> the
>> ballots correctly, anybody who wishes (including you) can double check 
>> the
>> IRV tally on any spreadsheet program.

No they have not done that. ALl they have done is posted the results
of the rounds.

>>
>> 3. The city did use a non-conventional majoritarian vote-tally rule for
>> the two seat council election (first and second choices counted
>> simultaneously), but the mayoral election was standard.

You bet your ass it was non-conventional. My head hurt trying to read
and comprehend it.

What it looks like they did was state up front that they would only
have the top 4 candidates in the IRV round. So any votes for the other
candidates might have automatically been considered "exhausted"
resulting in the high exhausted ballot pile numbers.

But where it gets real interesting is that the threshold for victory
in all three races was exactly the same - and that is the exact number
of votes that all three candidates got. The two council elections
were even more interesting - by what should have been the last round
with only two candidates standing, none of them had enough votes to
cross the threshold. So instead of merely declaring the candidate
with the most votes the winner, they turned around and checked the
ballots for the lowest vote getter of the last two standing to see if
there were any second or third choice votes for the top vote getter -
then see if there were enough votes there to make the leader a winner.
In those two cases, there were just enough votes to put the candidate
RIGHT ON the minimum number of votes to win the election.
It looks like they have a process so complicated and hard to
understand that 3 elections settled with exactly the same number of
votes also had three votes pulled out of someone's ass to make it look
like IRV is as easy as "1-2-3"!

Chris Telesca

>>
>> Terry Bouricius
----
Election-Methods mailing list - see http://electorama.com/em for list info