Burlington Vermont’s 2009 Mayoral race provides a great example for visualizing the results of ranked choice elections. It’s the most consequential real-world case where a Condorcet winner beat a runoff winner (the legally elected winner), but where both beat the plurality winner.

If you like controversy, you’ve come to the right dataset.

There’s more about that election here and here. There’s an unexplained difference pertaining to four exhausted votes between the data I transferred and the official results, but the point is to showcase these new data visualization experiences, and those quirks don’t affect the outcome.

Sites associated with Indaba Application Network (IAN) are now able to present four visualizations that can help voters understand how ranked choice voting works to reflect their collective intent. Since most jurisdictions are using Instant Runoff tabulations, I’m presenting two charts — stacked bar and Sankey — and a data table to illustrate that how that method played out. Those will be followed by chart arrays presenting Borda and Condorcet results.

Here’s the rendering of the Sankey chart.

In the chart below I decided to present the stacked bar visualization separately for each round rather than with the round-by-round interactive click-through component that’s been the traditional way of presenting runoff results. This is because: 1) Printable dataziv can be served across more kinds of media, and; 2) It allows better amplification of details that might matter most in the final rounds.

The next image is a table of percentages that might be popular with the “just the facts, ma’am” numerically inclined. Note that the exhausted ballot counts are shown individually by round here, while the previous visualizations summed the exhausted counts round after round. FYI, I chose these colors… Progressive Pink, Republican Red, Democratic Blue, Independent Orange, and Green Green… to help tell the story about how the race played out.

The Borda method is generally considered to be the least reliable of the ranked choice tabulation algorithms. It’s arguably more vulnerable to gaming. Nevertheless, by leveraging a format akin to the “Customer Review” ratings on Amazon.com, the charts shown below provide a simple and familiar perspective on how candidates performed. These raw details can be exceptionally revealing. Note the relative concavity or convexity in the shape of the vote distributions. It’s evident that the plurality winner was all base with little depth, while the Borda winner (who was also the Condorcet winner) had undeniably broad appeal despite placing a lowly third in first round, first place votes.

Highlighting the Borda totals alone (here counting 6 points for each 1st choice, 5 points for 2nd, etc.) shows even more starkly the potential cost of being a polarizing candidate versus the value of being a generally acceptable one… depending on the tabulation algorithm, of course. This exposes one of the most persistent complaints about the Instant Runoff algorithm… If an electorate is deeply divided on first choices, but all voters have the same second choice, AND if that undeniably acceptable second choice happens to receive the smallest number of first choice votes, the unhappy consequence is that the candidate most likely to reflect the community’s best possible chance at consensus would wind up as the first to be eliminated.

Fortunately, this is rare. But it happens. Which is why Burlington presents such an intriguing example.

Typical Condorcet visualizations combine the results for all the candidates within a single grid, saving lots of space at the expense of easy readability. The alternative shown below breaks out each candidate’s details into separate charts, simplifying the challenge of sorting through matchups. This style of display is intended to showcase the “round robin” aspect of the Condorcet pairwise tally method by getting to the bottom of the ultimate political question, “Who crushed whom?”

There are no surprises here. The results and anomalies of this Burlington race have been picked over many times since 2009. The Instant Runoff tabulation system was no help to Republicans who may have considered the Democrat less unpalatable than the Progressive. Yes, it’s apparent from the low number of match totals that many Republican voters marked no other ranks on their ballots. But it wouldn’t have made a difference. Their candidate was the last to be eliminated and their second choice — under those rules — was of no consequence.

The outcome would most certainly have been the same with a special runoff election between the Progressive and the Republican, so at least ranked choice voting saved everyone in Burlington some time and expense. But the Republicans were furious and the Democrats were meh after this election, so the old system was restored.

As the use of ranked choice voting has become more widespread across the United States, the plurality winner has typically been the instant runoff winner… by far. So, other than inescapable post-election sore-loser complaints, there have rarely been serious reasons for controversy.

However, as candidate counts proliferate and political ideologies fractionate, the likelihood of Burlington-style events will increase. This article is being posted just days before the Democratic Mayoral Primary in New York City, so anticipation and speculation is reaching a high. But it’s clear that if Americans want reforms that can break the polarizing cycle of two-party duopolies, systems that optimize for elevating center lane candidates offer the strongest promise to do so. Ranked choice ballots are a good first step… but not the last.

Interested in trying out a ranked choice straw poll? Check out Mayor21.com.

The music in this video is from Rain Dance, the opening track of Herbie Hancock’s Sextant, released by Columbia in 1973.

I’m hoping to brand Sygnol (AimsPoll’s API provider) with music from the whole album, using snips for lead ins, interludes, promos, and tone.

YouTube already permits me to mix and replay this music. But the Terms of Service allow YouTube to display paid ads over any posted video containing Sextant content.

A share of ad revenue collected by YouTube presumably presumably gets transferred to Herbie Hancock’s publishing agency, but I think a direct ad-hosting arrangement would significantly amplify the benefits to him and his audience.

For now, this (very early) Sygnol interface demo is live as well.

Fairvote, a leading proponent of ranked choice voting, has been running a series of events to celebrate its 25^th anniversary. It all culminates in the Champions of Democracy Awards Reception, in New York on Monday night.

I think it’s appropriate to send congratulations to Krist Novoselic, who’s been serving as Fairvote’s Chair for nearly a decade.

By taking on such a prominent role in such a venerable organization, he’s become a key figure in the effort to build a 21^st century American democracy worthy of the name. From what I’ve learned about his career, it’s a position that fits his longstanding commitments to free speech, civic education, constitutional government, and open collaboration.

Fairvote’s stated mission is to be a trailblazer for bold, nonpartisan election reform. The community attracts people who want to fix our broken democracy. And it empowers people throughout the United States who are dedicated to making repairs that endure.

The organization began to take shape in 1992 using the name Citizens for Proportional Representation. It later rebranded itself as the Center for Voting and Democracy and ultimately as Fairvote. The banner evolved to illuminate the goal.

A quarter century of persistence on behalf of any noble mission is impressive, but achieving fundamental election reform isn’t around the corner. Twenty-five years of accumulated experience means recognizing there will be thirty, forty, fifty and more. And the stakes are higher than ever.

Twenty-five years ago, when FairVote took root, the United States dominated the geopolitical stage. President George H.W. Bush believed he was forging a New World Order that would put the “rule of law” (American-flavored, of course) over the “law of the jungle.” That same year, Krist Novoselic was playing on stages all across the planet. He was in a band that upended the existing musical order by putting honesty, empathy and inclusive morality over superficial and exploitive commercialism.

In 2017 it’s obvious that the superpower hasn’t aged as well as the rock star.

These days it’s not uncommon to hear that this nation’s preeminence has faded, along with respect for the values that once made American culture so vital. We now find ourselves in a world where dictatorships and illiberalism and vicious ethno-nationalism are on the rise. More ominously, we find ourselves in a country led by people who are happy to shrug off those trends, and even welcome them.

That’s why it’s so important to celebrate people who dare to be Champions of Democracy. When we’re thirsty for positive signs, they remind us to look for democracy’s wellsprings in traditional American principles such as free association, fair elections, and responsible self-governance.  Those champions point the way with durability and resolution, and demonstrate how to be effective stewards in practice.

We deserve a modern electoral system. Reform is badly overdue. It’s necessary to admit just how badly things are broken in order to appreciate the size of the challenge. But by facing it we can also appreciate the redemptive capacity of America. Because every step toward democracy ever taken in America was badly overdue.

The glory of our history is that so many people keep daring to take those steps.  The virtue of groups like FairVote, and activists like Krist Novoselic, is that their footprints are so fresh.

Nevermind the images

Sygnol top-level candidates.

Here’s another illustration of the listicle sorting app I’m putting together. For now, I’m still making images by hand. But I’m also moving ahead on building tools for automated assembly.

This 4×4 grid is a draft for the top level of the Sygnol namespace, presented in alphabetical order.  When the system goes live, you’d be able to rearrange the tiles. That interaction would power features such as playlist-making and ranked-choice polling.  You’d also be able to navigate into sortable sublists. For example, selecting OS — the flag for Oscars — would take you into a 3×3 grid of Best Picture nominees, similar to what I presented earlier.

As more content is added, grids would get much larger, and sub-channels would proliferate.

Sygnol is intended to model a new direction for civic software and liquid democracy. It’s inspired, in part, by sites such as 2008’s 10Questions and 3 American Questions from 2016. Those election-year experiments were launched with the goal of giving voters a direct way to compare candidates’ official statements on key policy issues.

10Questions’ ambitious interactive web site was hosted by the New York-based Personal Democracy Forum, heavily promoted by the New York Times, and generously funded by the Knight Foundation. Unfortunately, the candidates ignored it, just as they later ignored the 3 American Questions site. That latter project didn’t produce much more than a few videos explaining the concept, but its featured topics were very well thought out: Climate Change, Federal Debt, and Confidence in Government. So they’ll be available in the inaugural Sygnol namespace as CC, FD, and GV.

I’ll explain the other prospective topics in due course, but one I want to emphasize is ER — Election Reform. It’s central to Sygnol’s mission. An earlier ranked choice interface I built had a poll on Priorities for Improving US Elections. This version is being designed to cover many more issues at far greater depth.

Code	Title
AA	Aimspoll Agenda
BH	Big History
BI	Border Integrity
CC	Climate Change
EQ	Income Inequality
ER	Election Reform
FD	Federal Debt
FP	Foreign Policy
GV	Confidence in Government
HC	Healthcare Reform
HH	Herbie Hancock
IN	Indaba Strategy
OS	Oscar Voter
SP	Secular Progress
SY	Sygnol Namespaces

Sygnol

Oscar Best Picture Nominees 2017

I’m setting up an experiment on ranked choice polling via Twitter. It lets you rank your preferences for the upcoming Oscars. To use it, you would tweet your ranks using the list below. You can use the link to see a trailer for the film.

Code	Title	Link
mo	Moonlight	9NJj12tJzqc
ma	Manchester by the Sea	gsVoD0pTge0
li	Lion	ziOLGzKq6oo
la	La La Land	0pdqf4P9MB8
hi	Hidden Figures	5wfrDhgUMGI
he	Hell or High Water	JQoqsKoJVDw
ha	Hacksaw Ridge	s2-1hz1juBI
fe	Fences	spCxVd9ctFs
ar	Arrival	tFMo3UJ4B4g

The tweet would start with #syg to announce you’re intending to tweet some ranks. Then you would insert the flag os+ to announce you’re voting for the Oscar’s list (There will be other flags down the line). And then you’d arrange the codes for your list. Case doesn’t matter. There are no punctuation marks between the elements, just spaces.

So a valid tweet would look like this:

#syg os+ mo ma li la hi he ha fe ar

The idea is to rank your preferences, not to predict the winner.

I’m also building a tool to read Tweets based on that syntax. If you’re interested in this, and in seeing other ranked choice polls as they come out, please follow @aimspoll It would help a lot.

Ask not which one to vote for, but how to rank them all

Users of previous ranked choice straw poll sites I created are being invited by email to try AimsPoll. The message I’m sending out has a short description of what’s going on. I wanted to say more, but the initial draft got way too long. So, anyone interested in the full thought process can read it here.

Hello from Craig Simon,

Please take a moment to rank the candidates and issues for the 2016 Presidential race.

Your opinions are important.

Use the links at the bottom of this email to take the poll right away.

Or, if you want to get reacquainted first, read on.

* * *

Thanks.

Every recipient of this email probably remembers taking a ranked choice poll sometime over the last eight years.

That’s how you got on this list… You signed up at indaba.org, or choiceranker.com, or wevote.net, or americanquorum.com, or mayor2011.com, or one of the related Facebook apps I’ve built since starting this project.

Those sites generally focused on political surveys. Some of them also had big, diverse turnouts for various sports and entertainment polls. But they all shared the same database engine. And they all were motivated by the mission of “building better tools for better democracies.”

That experience means you’re already familiar with the concept of sorting your preferences 1st, 2nd, 3rd, 4th, and so on. It also means you have some sense of the system’s simplicity, and perhaps even an appreciation of its power.

* * *

The poll I’m inviting you to take today uses the highly regarded SurveyGizmo platform. Big players in the online opinion research industry have started offering ranked choice systems, and SurveyGizmo is one of them. That’s great news. It proves the growing popularity of the approach.

Unfortunately, there’s a tradeoff. Robust as it is, SurveyGizmo doesn’t include important features you might remember from the sites I built, such as interactive media, social authentication, and real time tabulations.

Someday it might be possible to have the best of both worlds. For the time being, however, what’s most important is that the upcoming 2016 campaign opens the door for another chance to leverage the power of ranked choice polling.

The challenge for this cycle is to transform the results into action. That requires getting people who matter — politicians, pundits, and plutocrats, too — to take notice. And achieving that will depend, first of all, on understanding what AimsPoll is for.

* * *

AimsPoll’s surveys are intended to have predictive power…  but not like the demographically weighted, margin of error sampled, voter forecasts you’re used to seeing. Ironically, for each traditional survey that gets a prediction right, many more prove to be notoriously wrong.

As a wise man once said, the best way to predict the future is to invent it.

Imagine an interactive petition that enables diverse groups of people to coalesce on the fly as they articulate their common resolve, compelling responses from the targets of their message. The idea isn’t as futuristic or utopian as it might first appear. In fact, it’s only a few steps more advanced than the straw polls that Republican Party conservatives put on so frequently.

Why have those straw polls become so increasingly popular? Much of the answer has to do with fundraising and promotion, of course. But there’s more. As it turns out, straw polls promise the discipline and winnowing effects of organized public reflection. They fortify a group’s sense of its own coherence. They demonstrate concentrations of local power. They make news. They transform the participants into a force to be reckoned with.

Applying ranked choice technology to straw polls adds two benefits. First, visually interactive data collection tools let users express a fuller range of opinions in a very short period of time. Second, advanced tabulation algorithms reveal where views will coalesce most effectively.

Those virtues pay off most highly fragmented, multi-candidate races, like the current GOP contest. For that matter, we should keep in mind that the Democratic contest isn’t over till it’s over.

Why should members of a thoughtful community settle for a “frontrunner” who can only muster a 13% plurality? Antique ballot technology isn’t up to the challenge of discovering their fundamental alignments or representing their most broadly shared views.

Modern techniques raise smarter questions about overall attitudes toward the candidates. Who has the greatest depth of support?  Who triumphs most often in head-to-head challenges? Who would win the most thorough possible runoff?

Ranked-choice ballots give fairer results by surfacing for more meaningful details about voter sentiments. That’s why your participation in AimsPoll truly matters.

The long-term goal of this project is to have candidates create video and essay responses for each of the top ranked issues. When that happens, primary and caucus voters will be able to use AimsPoll to access side-by-side comparisons on the issues they care about. So, how do we make that happen?

For better or worse, we live in a horse race obsessed media environment. Discussions of issues are neglected  by journalists who above all crave news about who’s gaining and who’s falling behind. That hunger for a contest can be exploited. This is because the results of ranked choice polls look like horse races on steroids. Journalists can mine them for rich stories about the trials and tribulations of candidates jockeying through the pack.

The trick is to steer journalists into linking their stories about horse races with stories about issues. Here’s how. If a handful of contenders can be enlisted into competing in a “Rank the best candidate videos on issues x, y and z” contest, the others will face pressure to follow.

As we know, just a few dozen votes in a prestigious straw poll ahead of the Iowa Caucuses can mean life or death to a candidate’s prospects. When everything is at stake, the jockey needs to look for all potential openings up ahead while guarding against moves from behind and from the flanks. Each candidate realizes that participating in a debate is risky, but that skipping it is a sure way to lose it.

This strategy builds on a lesson from Political Science 101. “Politics is about managed competition.” So manage it. Create venues that bake in attractions for signal over noise. And make things easier for “low information voters” who wait until the last few days before an election to look into who’s running and where they stand on the issues.

That’s why your opinions are so important for AimsPoll to succeed. There’s one poll for Republicans and another for Democrats. The first section lets you rank your favorite candidates in order of preference. The second section allows you to rank the issues you want to hear the candidates talk about. About 2,000 completed surveys are needed to reach the threshold at which we expect politicians and pundits will begin to pay attention.

Without the right user interface, lots of candidates could require lots of clicks before all the preferences are put in order.

Here’s a handy formula that expresses a key virtue of ranked-choice interfaces over head-to-head pairwise evaluations… N(N-1)/2.

Head-to-head interfaces are typical of “Whose hotter?” sites such as candobetter.com. The point is to simplify the task of evaluating a long list of options. Users view a randomly selected pair of images from a large set, and then select one. The choice is recorded and another pair of images is displayed. It’s fine for idle entertainment since it’s so easy. Of course, the typical user gets bored and quits long before all the options are exhausted. Why? N(N-1)/2.

The reality of N(N-1)/2 boils down to this: As the number of candidates to be evaluated increases by 1, the possible number of pairwise evaluations grows at an increasingly higher number. The growth is less than half the rate of an exponential increase, but significant nonetheless. The underlying math is simple and straightforward, as Larry Bowen demonstrates, with significant implications for interface design.

One of the most advanced head-to-head crowdsourcing applications currently online is AllOurIdeas.org, a Google-funded project led by Princeton-based Sociologist Matthew Salganik. It’s a serious, text-oriented approach toward what Salganik calls “Bottom-Up Social Data Collection.” It’s been used to help small groups design their web sites and to help US citizens discuss national priorities. But even a site as sophisticated as that can’t escape the fundamental challenge of N(N-1)/2.

With just two choices per page, hot-or-not sites and AllOurIdeas benefit from having lots of room to display big, engaging pictures or crisp easy-to-read text boxes. On a per page basis, that space-saving approach leaves little room for confusion. But problems arise as the list of choices increases. It taxes respondent patience and undermines the accuracy of results.

Imagine a comparison of just 6 candidates in a hot-or-not styled contest. According to N(N-1)/2, the various names would be seen a total of 30 times. The voter or survey taker would have to click through 15 pages of pairs before a full ordering of preferences is possible. Even worse, someone could easily enter contradictory rankings, for example choosing A over B, B over C, and C over A. This “rock, paper, scissors” problem is known to mathematicians as intransitivity. People might fail to remember how prior choices were ranked, presuming they even bother to try. Or they might change their minds as they get cranky in the middle of a long survey session.

Things gets worse as lists get longer. Not exponentially worse, but enough to ruin the experience. Contests such as the 2012 GOP Presidential debates, which typically presented 8 candidates, would require 28 pages. A 12 candidate reality show such as a season-starting American Idol would require 66. A top 25 list such as college football’s Bowl Championship Series would require 300.

The bottom line is that a hot-or-not interface might be ideal if the only two choices are, say RCV versus hot-or-not. It might also be appropriate if the site owner’s main interest is to run up page views. But it’s a very different story when serious multi-candidate evaluations are called for, accurate responses are important, and an efficient user experience is desired.

An earlier version of my article is at my other blog… Comparing Ranked-Choice and Pairwise Evaluation User Experiences.

Note from April 8, 2015. I spent several days with SurveyMonkey before discovering and falling in awe wwih SurveyGizmo. Unfortunately, Gizmo doesn’t have Monekey’s very nice question element styling feature, which I was using to put candidate pictures in the sortable list object. But so many other features are so good, the scales tip heavily. I’ll keep the Monkey poll as is, and see what happens when my month of subscription expires. Only five votes came in for this one. I was building a more ambitious survey with Monkey, but abandoend it once I saw the alternative. Some key pieces from it were used to build my first Gizmo survey, Department of Oops.

“Ask not which one to vote for, but how to rank them all.”

Move your favorites to the top, and the ones you most oppose to the bottom.

This is the long list of candidates who have made credible moves to seek the Democratic or Republican Party nomination for President. It also includes several who wouldn’t merit inclusion on a strictly pared list due to lack of serious logistical efforts or plain statement of intent. Elizabeth Warren has consistently said she is not running, but is here because of the active movement to draft her.

This poll uses the Survey Monkey service. It is excellent in many ways, but has some shortcomings. Most notable is that it forces every candidate to be ranked. That inability to leave candidates off is not a showstopper. The best way to use this is to put your favorites at the very top, and the ones you oppose the most at the very bottom. Leave those you’re most unsure of somewhere in the middle. Keep in mind that the results will be tabulated using algorithms by which voting down a candidate you strongly oppose is as important as voting up your true preferences and the others you would accept.

Follow @aimspoll on Twitter and join the AimsPoll page on Facebook to get announcements of the results.

During the process of collecting Twitter metrics on the candidates, I started adding in comparisons of several voter-oriented online groups. It’s intriguing to see which folks put a lot of effort into tweeting, despite having relatively small numbers of followers. That’s why the traction metric is useful. Going for volume without enough traction risks wheel spinning.

Organization	Handle	Followers	Tweets	Volume	Traction
NationalPopularVote	@NatlPopularVote	57	35	0	1.63
ReThink Democracy	@ReThinkDemocrcy	902	5710	5	0.16
Votenet	@votenet	1081	524	1	2.06
Votetocracy	@votetocracy	1143	37300	43	0.03
DemocracyLab	@DemocracyLab	1339	144	0	9.30
Vote411	@VOTE411	1815	4919	9	0.37
Intellitics	@intellitics	1893	4122	8	0.46
iCitizen	@iCitizen	2274	7271	17	0.31
WhyTuesday	@WhyTuesday	3424	7264	25	0.47
Fairvote	@fairvote	3672	7647	28	0.48
VoteSmart	@VoteSmart	4820	4301	21	1.12
The GovLab	@TheGovLab	6620	1859	12	3.56
League of Women Voters	@LWV	9304	17000	158	0.55
CrowdPac	@Crowdpac	9464	1233	12	7.68
Nat. Inst. on Civic Discourse	@NICDInstitute	9498	2310	22	4.11
Center for Civic Media	@civicMIT	10200	6146	63	1.66
WolfPAC	@WolfPAChq	13400	4280	57	3.13
Impact Democracy	@ImpactDemocracy	14200	8177	116	1.74
Citizens United	@Citizens_United	14300	9677	138	1.48
OpenCongress	@OpenCongress	15400	3737	58	4.12
Common Cause	@CommonCause	18300	13300	243	1.38
New America Foundation	@NewAmerica	21800	17000	371	1.28
Ballotpedia	@ballotpedia	22000	12100	266	1.82
Peterson Foundation	@pgpfoundation	24100	1517	37	15.89
National Urban League	@NatUrbanLeague	37100	8880	329	4.18
American Enterprise Institute	@AEI	43900	31200	1370	1.41
Center for American Progress	@amprog	48400	15300	741	3.16
FactCheck.org	@factcheckdotorg	68800	3818	263	18.02
NAACP	@NAACP	93400	11400	1065	8.19
No Labels	@NoLabelsOrg	129000	12300	1587	10.49
Moveon	@MoveOn	132000	14200	1874	9.30
NRA News	@nranews	145000	22700	3292	6.39
National Rifle Association	@NRA	266000	6265	1666	42.46
RNC	@GOP	333000	21000	6993	15.86
Lawrence Lessig	@lessig	334000	8741	2919	38.21
DNC	@TheDemocrats	364000	14000	5096	26.00
Heritage Foundation	@Heritage	486000	46400	22550	10.47
Greenpeace	@Greenpeace	1290000	24200	31218	53.31

Selected Public Interest Organizations, March 2013

With luck, AimsPoll will be evolving into something much more than just a Twitter feed. This video provides a look at the kinds of technology being put together.