All posts in the topic IRV and multiple candidate elections (Short link)
Summary
- There are 27 posts — by 16 authors — in this topic.
- Latest post made by John Wilson at Dec 18 19:18 UTC
Just read Steve Brandt's article Hand-counting ballots in instant-runoff vote called 'huge nightmare’ http://www.startribune.com/politics/local/35201754.html At the same time, my email has dates for City DFL caucuses and conventions. Caucuses held Mar 3rd, City Convention on May 16th with ward conventions in March and April. Will someone who knows IRV please explain exactly how the three seat citywide Park Board election will work? I know the 2 candidate citywide Board of Estimate and Taxation election will be similar. For citywide park board, will we vote for nine? Three positions times choice 1, choice 2, choice 3? Are the votes ranked 1-9? Or will this be 3 in choice 1, 3 in choice 2, 3 in choice 3? How will second choice votes and third choice votes be added in to get three candidates a 50%+1 vote majority? What is the basis for computing a majority? What about bullet balloting? If my goal is to help just one candidate get through (like when a friend and neighbor ran for School Board), does bullet balloting still make sense? Shawne FitzGerald Powderhorn
Shawne Fitzgerald asks about IRV procedures for multiple-seat city elections. The IRV ordinance is at http://www.municode.com/Resources/gateway.asp?pid=11490&sid=23 with the multiple-seat procedures at 167.70. Steve Brandt Reporter Star Tribune 612-673-4438 <email obscured><mailto:<email obscured>>
I've always like the idea of, and have supported the concept of IRV . . . .
until now. After watching the recount unfold, and the significant amount of
error being made just to ill in a circle, I can;t begin to imagine the
confusion that would happen on an IRV ballot, and at what point a recount
would be triggered.
I do still support runoff elections, like Georgia is experiencing. We would
be bypassing the recount right now, and getting ready to vote for the top 2
candidates instead.
Flash
Centrisity.com
On Mon, Dec 1, 2008 at 11:11 AM, Brandt, Steve <Steve.Brandt@startribune.com
In Georgia, under IRV, they wouldn't have to run this second election--IRV
would combine the two elections into one. Think of the efficiencies!
Georgia can do any danged thing it wants to do in its democratically
held elections. Come some future point, Georgia citizens may well be
saying the reverse of us.
Laura
Southeast/Como
On Dec 1, 2008, at 6:42 PM, Kelly O'Brien wrote:
> In Georgia, under IRV, they wouldn't have to run this second
> election--IRV would combine the two elections into one. Think of
> the efficiencies!
>
>
Laura Waterman Wittstock
President and CEO
Wittstock & Associates
913 19th Ave SE
Minneapolis, MN 55414
612-387-4915
blog:
cornplanter's revenge
www.laurawatermanwittstock.com
Shawne FitzGerald asks good questions about the use of Instant Runoff Voting in Minneapolis elections next year. These are just the kind of questions that voter education will address next year as the municipal election season gets underway. The City of Minneapolis and FairVote Minnesota are planning an extensive education campaign to ensure all voters understand IRV and know how to effectively complete a ranked ballot. In response to Shawne's questions: Q. Will someone who knows IRV please explain exactly how the three seat citywide Park Board election will work? I know the 2 candidate citywide Board of Estimate and Taxation election will be similar. For citywide park board, will we vote for nine? Three positions times choice 1, choice 2, choice 3? Are the votes ranked 1-9? Or will this be 3 in choice 1, 3 in choice 2, 3 in choice 3? How will second choice votes and third choice votes be added in to get three candidates a 50%+1 vote majority? What is the basis for computing a majority? A. First, a quick overview of the scope of Minneapolis 2009 municipal elections. There are a total of 25 seats to fill in 2009: . Mayor . 13 Council Members . 6 Park & Recreation Board District Commissioners (each single-winner) . 3 Park & Recreation Board At-Large Commissioners (3-seat multi-winner) . 2 Board of Estimate and Taxation At-Large Members (2-seat multi-winner) IRV (more technically known as Single Transferable Vote) allows voters to rank candidates according to their preference - 1st choice, 2nd choice, 3rd choice, etc., and, in doing so, consolidates two-rounds of voting into one (so there will be no September primary next year). The threshold for winning (representation) is 50% + 1 in single-winner elections and goes down as the number of winners to be elected within the district increases, according to the formula (1/#seats +1)+1. Candidates running for the single-seat races - mayor, council and park district - must attain a majority of votes. Candidates running for the two-seat (at-large) Board of Estimate and Taxation race must attain 33% + 1 vote and candidates running for the three-seat (at-large) Park Board race must attain 25% + 1 vote. In multi-winner elections, IRV ensures that the majority of voters elect the majority of winners while giving smaller voting groups better chance of electing candidates that represent them. Specifically how the Park Board members are elected: Six of the 9 board seats are district-based single-winner elections. Each seat is filled independently by district voters and each of the six winners needs a majority of votes to win. As a voter, you rank your choices in order of preference, up to a maximum of three rankings allowed by the existing voting equipment. Three of the seats are filled at-large. It is one (1) multi-winner race in which voters rank their preferences among the field of candidates running for the three seats. As a voter, you rank your choices - just as you do in the single-winner elections - in contrast to the "vote for up to three" instructions under the previous at-large method. The threshold for winning is 25% + 1 vote. How the vote counting works: In single-winner elections If a candidate receives a majority of first choice votes, that candidate wins. If not, the candidate with the fewest votes is eliminated and votes cast for this candidate are redistributed to remaining candidates according to second choice on each of those voters' ballots. All the ballots are counted again and if a candidate receives a majority, that candidate wins. If not, the process is repeated until one candidate has a majority of votes. In multi-winner elections After all first choices are counted, if no candidate reaches the threshold, then the candidate with the fewest votes is defeated and votes cast for this candidate are redistributed to the second choice on each ballot. If a candidate receives more votes than needed to win, then the surplus portion of each vote for that candidate is redistributed to those voters' next choices. The process repeats until all seats have been filled and ensures minimal wasted votes. A step by step explanation of how IRV works in single and multi-seat elections is on the Minneapolis Better Ballot Campaign website (the multi-seat part starts on tab 9): http://mpls.betterballotcampaign.org/node/223.A. Q. What about bullet balloting? If my goal is to help just one candidate get through (like when a friend and neighbor ran for School Board), does bullet balloting still make sense? A. No, with ranked choice voting, bullet voting and other forms of strategic voting are unnecessary and ineffective. Candidates with the greatest and broadest support win. Your lower ranked preferences can never harm your higher ranked preferences. Please see http://www.fairvotemn.org and http://mpls.betterballotcampaign.org/ for additional information and on-line FAQs or contact me at <email obscured>. Jeanne Massey FairVote MN Kingfield
How IRV tallies races involving multiple seats depends on
interpretation that has to be consistent with all seats. For something
like Bd of Est & Taxation: does IRV regard the race as being Seat 1,
Seat 2, Seat 3 or as being 3 Seats grouped. The ranking topology and
tallying are obviously different. In the past, a person would run for
a seat and not a particular seat: thus, the ballot would instruct vote
for three. How this translates into IRV structure is anything but a
minor logical problem. Suppose 10 people have filled for 3 seats: do
all 10 have an equal shot at each seat. If so, then the vote for
three, means your first vote is to be ranked among 10 people, your
second vote ranked among 10, and your third vote ranked among 10
because the voter can not know any winner - the theory of
distributions thus enables 120 different combinations. If P people are
competing for S seats, then there will be P!/[S!*(P-S)!] different
distributions. Of course, the number of choice combinations can be
reduced to manageable proportions if the limits how many people will
be allowed to file for election. One way would be to raise the filing
fee to, say, $100,000 or some such.
Jack Ferman
Kingfield Neighborhood
<email obscured>
As much I think very positively about IRV, has everyone seen what people do to
their ballots? People in Mpls, cannot even fill in ovals. After repeating to
every voter on election day, fill in your ovals in black ink completely...yall
saw what folks do. Thousands of ballots with blobs,X's- dashes, squares,
slashes, happy faces, peace signs, Griggs shorthand.....thousands....Marks all
over the ballots except where they were supposed to fill in. What are we going
to do to educate the public about how to vote by ranking preference?? I just
don't get it. Was it a fluke that there was tens of thousands of ballots with
blobs and slashes and dashes, or is this due to people just not listening to a
simple directive?
Gina Palandri
Election Judge 9/8
I'm surprised to hear that you saw so many flawed ballots when you served as an
election judge. I participated in the senate recount as an observer, and for
every 1000 or so ballots that I saw, maybe 3 or 4 were incorrectly marked. In
fact, I was impressed by how uniform and tidy those ballots were overall.
Minneapolis voters will have no more trouble filling in their bubbles on an IRV
ballot than they do now. And after the IRV education campaign that will happen
next year, they will probably do better than ever.
Dear Jeanne,
Thanks for the response. Just to clarify based on the Fair Vote website:
In the voting for the three city-wide Park Board Commissioners, will I
cast just ONE first choice vote?
Thanks,
Shawne FitzGerald
"In the voting for the three city-wide Park Board Commissioners, will I cast
just ONE first choice vote?"
Also, a follow-up to that:
If my 1st choice candidate is elected, do my 2nd and 3rd choice votes count at
all toward electing the 2nd and 3rd seats?
"I'm surprised to hear that you saw so many flawed ballots when you served as
an
election judge. I participated in the senate recount as an observer, and for
every 1000 or so ballots that I saw, maybe 3 or 4 were incorrectly marked. In
fact, I was impressed by how uniform and tidy those ballots were overall"
Sadly, I was very surprised. I saw so many more than I expected, and could not
really believe how many people cannot follow the instructions, or choose not
to.
I'd like to think that as an election judge that has served for a long time-
and hearing the instructions to tell people to please make sure to fill in your
ovals, that I saw thousands of everything...but ovals. And write-ins for
cartoon characters. I suggest alot of voter "re-education" after seeing what I
saw. Esprecially for the amount of seniors that I see in our precinct....
Gina Palandri
9/8
Shawne Fitzgerald asks: "In the voting for the three city-wide Park Board
Commissioners, will I cast just ONE first choice vote?"
Yes. You will be given a second and third place vote as well, at the very
least. When machines that can handle more than three rankings are certified
for use, we may revisit the maximum number of rankings voters will be given,
but the minimum has been set - by ordinance - at three.
Tim Bonham asks: "If my 1st choice candidate is elected, do my 2nd and 3rd
choice votes count at all toward electing the 2nd and 3rd seats?"
Yes. They will not count as whole votes - giving voters more than one vote is
what made the Duluth system unconstitutional - but you will be able to apply
whatever portion of your vote you did not use to elect your first choice to
your second choice. Think of it as change. You "spend" your whole vote on
your first choice. If lots of people choose that same candidate, you get a
bulk discount, reducing that candidate's "price." Let's say the "price" comes
down to three quarters of a vote. You can then, if you choose, apply the
quarter you have left over to your second choice.
Robin Garwood
Aide to Second Ward Council Member Cam Gordon
Member of FairVote MN Board
Cooper resident
Robin's explanation seems opaque. If there are 3 at-large seats and
the instruction is vote for three and there are 12 people running:
Smith, Jones, Martin, Anderson, Johnson, Peterson, Iverson, McNulty,
McTavish, Cavendish, Edwards and Collins; it seems a first choice
could be be Collins, Smith, & McNulty - and a second choice could be
Smith, Edwards and Johnson - and a third choice could be Martin, Jones
and Iverson. That's what my logic implies, but Robin seems to be
saying one may vote for only one name with your 2nd & 3rd choices
adjusted for some rather strange calculus. This last is a huge oops
not ever spelled out in the election approving IRV. BTW, with 12
people running for three seats, there are 220 possible different
combinations of names.
Jack Ferman
Kingfield Neighborhood
<email obscured>
By Jack's reasoning, instead of voting for three candidates we'd be voting for
three slates. Robin's explanation makes sense to me. If a candidate needs 25%
+1 vote to win a park board at-large seat, and pulls in 33% of the vote, that
"extra" 8% will be redistributed to the 2nd choices on those ballots. As with
single-seat elections, that method ensure majority support for the winning
candidates.
On Tue, Dec 9, 2008 at 9:33 PM, Kelly O'Brien <kellyobrien@visi.com> wrote:
> By Jack's reasoning, instead of voting for three candidates we'd be voting
for three slates. Robin's explanation makes sense to me. If a candidate needs
25% +1 vote to win a park board at-large seat, and pulls in 33% of the vote,
that "extra" 8% will be redistributed to the 2nd choices on those ballots. As
with single-seat elections, that method ensure majority support for the winning
candidates.
who chooses what 8% will be redistributed? This makes no sense to me.
the first 8% counted could have different 2nd choices than the last
8%
John Harris
webber-camden
Before all these mathematical gyrations begin, I think the first question to
answer is, does IRV even apply to multiple seat races?
In response to Steve Kotvis' question re will ranked voting be used for the multi-winner elections: Yes, it will be used to fill the 3 at-large Park Board and 2 at-large Board of Estimate and Taxation seats. In response to John Harris' question re *whose* votes are redistributed: The surplus portion of EACH vote for the winning candidate with surplus voters is redistributed to those voters' next choices to ensure that everyone's vote is counted equally. To review how the process works: 1) Each voter casts a single vote, ranking candidates in order of preference. 2) Count all first place votes for each candidate. 3) Determine the vote threshold to win: (votes/seats + 1) + 1. For example: Number of seats to be filled Votes needed to win For 1 person (mayor, city council) 1/2 + 1 (majority) For two people 1/3 + 1 For three people 1/4 + 1 4) Elect any candidate reaching the winning threshold. If all seats are filled, the election is over. If any seats are open, a runoff round of counting occurs. 5) In a multi-winner election, if any candidate exceeds the winning threshold, reallocate the surplus votes *proportionally* from that candidate to the next choices on each of the ballots cast for that candidate. Surplus votes are reallocated at a *fraction of its value* to the ballot's second choice (or next choice still in the running). For example, if the winning threshold is 1,000 and the candidate gets 2,000 votes, reallocate half of every ballot. 6) Return to Step 3. 7) If there are no more candidates over the threshold to redistribute surplus votes from, start at the bottom eliminating last place candidates (they have lost the election). The votes from these candidates are redistributed (as whole votes) to the next choice on each ballot. 8) Stop at any time when all seats are filled. Each voter casts a single vote which can transfer (based on the rankings) during steps of the runoff to ensure that no votes are "wasted" and that voters win their fair share of representation, i.e., the majority of voters elects the majority of seats and smaller voting blocks have a better chance of gaining representation. The more seats there are to fill, the more proportional the outcome. If a voter ranks just a first choice and stops ranking, and that first choice wins a seat, then the voter is represented the way he or she voted. If a voter has a preference regarding who fills the additional seats, then he/she should continue to rank candidates so that his/her "share" of the vote beyond what their first choice needed to win (the surplus) can be used to help elect the additional winners. I like Robin Garwood's dollar analogy: "Think of it as change. You "spend" your whole vote on your first choice. If lots of people choose that same candidate, you get a bulk discount, reducing that candidate's "price." Let's say the "price" comes down to three quarters of a vote. You can then, if you choose, apply the quarter you have left over to your second choice." *Single Transferable Vote* ensures that if a candidate has more than enough votes to win, part of each vote will go to help elect additional candidates. Or if a candidate has low support and cannot win, the voters who supported that candidate can have their vote count for another candidate who has a better chance of winning. The order in which candidates benefit from transferred votes is determined by the voters' rankings. For additional information, please see How IRV Works at": http://mpls.betterballotcampaign.org/node/225 (multi-seat elections begin on tab 9) Jeanne Massey FairVote MN Kingfield
On Wed, Dec 10, 2008 at 9:23 AM, Jeanne Massey <jkmassey@earthlink.net> wrote:
> In response to John Harris' question re *whose* votes are redistributed:
> The surplus portion of EACH vote for the winning candidate with surplus
> voters is redistributed to those voters' next choices to ensure that
> everyone's vote is counted equally.
seriouly, I'm a college educated guy and I'm trying really hard to
understand this but it is still not clicking. Now you introduced half
votes and quarter votes...
I get the single winner bit. This multi-seat race needs some help, IMHO.
John Harris
webber-camden
eagerly awaiting more attempts to explain the process
John Harris wrote:
> On Wed, Dec 10, 2008 at 9:23 AM, Jeanne Massey <jkmassey@earthlink.net>
wrote:
>> In response to John Harris' question re *whose* votes are redistributed:
>> The surplus portion of EACH vote for the winning candidate with surplus
>> voters is redistributed to those voters' next choices to ensure that
>> everyone's vote is counted equally.
>
> seriouly, I'm a college educated guy and I'm trying really hard to
> understand this but it is still not clicking. Now you introduced half
> votes and quarter votes...
>
> I get the single winner bit. This multi-seat race needs some help, IMHO.
Well, this system allegedly works just fine in Cambridge, a town
that probably has more Ph.D.s per capita than any other place
in the world. Minneapolis should provide an interesting test of
whether normal people can make sense of this.
Tim mentions Cambridge, and that this works just fine there.
Can anyone find a sample ballot from there for a multi seat election to use as
an example for us on the forum.
I am supportive of IRV, and cant wait. I have confidence in it working and
working well.
The math involved for the multi seat elections is a bit complicated, but that
is where computers and programs come in handy.
We will have the paper ballots as a backup.
I am curious to see what our ballots may look like for next years city
elections, anxious actually.
Ron Leurquin
Nokomis East
Leurquin, Ronald wrote: > Tim mentions Cambridge, and that this works just fine there. Maybe it does, and maybe it doesn't; I don't know. But, I do know that Cambridge, home to Harvard and MIT, has a very educated population and and an exceptional number of PhDs, while Minneapolis is still working on ensuring that an adequate number of students graduate from high school. And, maybe IRV is a good thing. Again, I don't know. But, I am pretty sure that the claim that IRV is really very simple is, charitably, a gross exaggeration. > Can anyone find a sample ballot from there for a multi seat > election to use as an example for us on the forum. Better yet, here are two sample election results. First is the result for Minneapolis 2005 mayoral election: two candidates, three numbers (including write-ins): 01 = R. T. RYBAK 43,198 02 = PETER MCLAUGHLIN 25,807 03 = WRITE-IN 1,268 You don't need a PhD to figure out who won and how they did it. Second, here is the result for the 2007 Cambridge city council elections. Sixteen candidates (plus write-in and "exhausted"). A chart with 291 entries, some of them negative. And, in the end, every winning candidate received exactly the same number of votes. <http://www.rwinters.com/elections/election2007/2007official.pdf> A voting process that requires nearly 300 entries to explain the results sure doesn't seem simple to me. But, I think that Jeanne Massey did a pretty good job of convincing us that IRV is anything but simple. Contrast her explanation with "the person with the most votes wins".
Please pardon the length of my post. I'm responding to Mr. Salo's
questions and comments.
Timothy J. Salo wrote:
<[…] here are two sample election results. First is
the result for Minneapolis 2005 mayoral election:
two candidates, three numbers (including write-ins):
<01 = R. T. RYBAK 43,198
<02 = PETER MCLAUGHLIN 25,807
<03 = WRITE-IN 1,268
<You don't need a PhD to figure out who won and how they did it.>
= = = = =
[KB response:] This is the runoff election for a single-winner race (Mayor).
If the mayoral race had been held using IRV, there would have been no
low-turnout primary and all the candidates would have been on the ballot in
November when turnout is high. If a candidate received a majority of the
first-choice votes, s/he would be the winner with no need to have a second
round of counting. All voters could vote sincerely for their preferred
candidate as #1 without fear of "helping" elect a candidate they didn't like.
Earlier I wrote:
> And, maybe IRV is a good thing. Again, I don't know. But,
> I am pretty sure that the claim that IRV is really very
> simple is, charitably, a gross exaggeration.>
Ken Bearman was apparently trying to refute my claim when
Perhaps it's a matter of some of us think Tim Penny was the best choice for
governor and Tim Pawlenty got in.
Roger Moe would have been a better choice than Tim Pawlenty had I been given
the chance to make my second choice vote known in that election.
I for one want to vote for what I think is the best candidate, regardless of
party affiliation, and not have to fear that the Devil will get elected.
Ron Leurquin
Nokomis East
Once the voters "get it," the citywide Park Board races will be easy to
understand. Forget 2nd and 3rd choices, the key is that you have ONE
vote.
If you join a bloc that can get 25% plus one vote, your bloc will have
one seat. On the first round. Guaranteed. If your bloc has more than
25% - say 65% - then you want to break your bloc into two parts BEFORE
the election - this way you win two of the three seats.
I plan to cast my vote for a bloc that I know will win a 25% plus one
plurality. I want my vote to count and I want it to count for a winner.
Only three blocs can win. Citywide Park Board seats have no reason to
appeal citywide - there is every incentive to appeal to a narrowly
defined large plurality. All kinds of blocs are possible - political
party, interest group, geographically-based.
If you look at the citywide race through the prism of political parties,
then the DFL should get 2 seats and a coalition of Republicans and
Independents can win the 3rd seat. I think there is still enough
Republican strength in the City to do this - and they are a very
organized party. They will also pick up votes from independents who
have had no alternative but to vote DFL or Green in this race. This
bloc needs an attractive candidate.
The DFL needs to undertake massive voter education because the rules
have changed. The Democrats also need to split their vote in an
organized fashion - and there are actually a couple of ways to do this.
The DFL index is reliably 60%-70% in the City so if this can be divided
in half (as in vote for one) - the Democrats can win two seats. They
have the advantage of two strong incumbents. The DFL should not endorse
a third candidate in the citywide race. They would also need to get
their majority to quit ticket-splitting with the Greens. The DFL needs
two blocs of 25% + 1 to retain two seats.
These blocs already exist and the voters are used to the party system.
So, it these blocs get organized, the citywide Park Board races should
be decided on the first ballot. It won't matter if 20 other candidates
file and 2nd and 3rd round choices will not be counted. The election is
over when three blocs win.
Over the long run, when voters better understand how the system works,
the blocs will shift to geographic bases. Look at the park district
vote totals from the last election. The three Chain of Lakes districts
cast well over half of all total votes in the district park board
races. If we got organized and split our votes (like Chain of Lakes
West and Chain of Lakes South), we could pick up two of the three
citywide Park Board seats. The Chain of Lakes folks would then have a
majority on the Park Board.
This could be organized with a Chain of Lakes PAC and a Chain of Lakes
screening committee. But even with no organization, I think this will
happen anyway because people tend to vote their self-interest.
Shawne FitzGerald
Powderhorn
(From the tiny corner that is in a Chain of Lakes district)
Google "Douglas J. Amy" for more info on IRV and other proportional
representation systems. Personally I like "closed list" better.