Maine: Ranked Choice Voting in an Optical Scanner Environment
Updated: 24 January 2016
Short URL:

(Back to Electronic Voting < Elections < Issues < Home)

Post-2016 Election Update
This page was designed in late 2015 and early 2016, during the citizens initiative campaign for Ranked Choice Voting (RCV) in Maine. That effort eventually passed in the November 2016 election, kicking off an 18-month phase-in period.

During the campaign, the Hancock County Democratic Committee (HCDC: the most active in the state) informed the RCV campaign that it would oppose RCV on the grounds that the procedure's effective requirement for electronic tabulation of results in the absence of effective security measures puts the integrity of elections at risk -- unless the RCV campaign committed to supporting the implementation of election anti-fraud measures during the phase-in period. The RCV campaign agreed to this.

At the same time, the HCDC pushed through an election-transparency plank for the party platform at the state Democratic convention:

D. Ethical Government and Democracy
We support a democratic government that:
4. Makes the process of electronic voting transparent and fraud-proof by putting all aspects of it under public control, including the use of open-source ballot-handling software and an effective auditing process.

Personal note: The purpose of this page is not to criticize RCV, which in and of itself I believe to be an excellent concept. Instead, this is an attempt, with an apology, to put RCV in the context of electronic vote tabulation, without which it is a practical impossibility. This is an important frame of reference for those who, like me, are familiar with the security problems associated with electronically-assisted elections and the resulting potential for large-scale election theft, of which there are many proven cases. To be sure, Maine has been one of the least likely targets for election theft, at least prior to the Republican strategy of controlling State Houses and legislatures, and, unlike certain other key states, it is not known for a politicized election bureaucracy. So one could make a case for dismissing concern about e-voting in Maine. But administrations and officials change. And in a larger context, Maine's "dirigo" motto of "leading the nation" carries the implicit potential of convincing other states to use RCV, other states whose records regarding e-voting are not so pristine. In such environments, any proposal that mandates electronic voting in any form warrants careful scrutiny.

Maine History -- Maine has a history, at least in the recent past, of electing governors with less than a majority. As has been pointed out:

  • Races with more than two candidates are common in Maine and often result in winners elected by fewer than half of voters.
  • In 9 of the last 11 races for governor, candidates were elected by fewer than half of voters.
  • In 5 of those races, candidates were elected by fewer than 40% of voters.
  • None of Maine's governors have been elected to their first term by a majority of voters in the last 40 years.

This did not become a major issue, however, until Maine's experience in the years following the present governor's 39% victory in 2010. His dysfunctionally confrontational governing style and personalization of politics has polarized the electorate, threatened to paralyze the legislative process, and brought serious consideration to his impeachment. That original 39% minority victory catapulted the previously little-discussed concept of Ranked Choice Voting (RCV) into the forefront of Maine politics, and the governor's subsequent repeat-minority re-election in a second controversial 3-way race prompted the formation of an active citizens' group pushing for its adoption.

Ranked Choice Voting -- RCV allows voters to rank their preferences when three or more candidates (or issue options) are involved. If no candidate gets a majority of the voters' #1 ratings, the one with the least number of #1 ratings is eliminated, and the #2 choices of the people who supported that candidate are distributed among the corresponding candidates. This continues until one candidate has a majority. Everyone's vote gets counted once, even if it wasn't for his/her #1 choice, while at the same time an accurate picture of people's actual preferences (the #1 choices) is demonstrated.

This mechanism provides RCV with at least three major advantages:

  • It guarantees that elections with more than three candidates will be won by majority vote (though some consider it a weak majority, more of a parliamentary "coalition of the willing").
  • It eliminates the "spoiler" effect, enabling people to vote #1 for the candidate who -- even if "unelectable" -- they believe to be the best, without helping to defeat their alternative choice (whom they rate as #2).
  • It tends to reduce incivility and polarization in campaigns -- and in politics in general -- by providing an incentive for a candidate to reach out beyond his/her narrow base in order to earn a #2 rating from other voters which might eventually give him/her a majority.

The Catch -- There is a catch, however. In a traditional non-RCV race, the totals for each person are tabulated in each municipality, where it is possible and practical to do this by hand-count, the local totals then being simply reported to the state, where they are added up for a final result. In an RCV race where there is no immediate majority winner, however, tabulation can't be done on a local level. All the votes must be combined in an aggregate before the RCV elimination-and-redistribution process can commence. And at the state level, with roughly 600,000 ballots, doing the RCV process by hand would be a practical impossibility.

Thus, RCV is unavoidably accompanied by electronic vote tabulation. And for those who understand the inherent lack of integrity of today's electronic voting systems, whether using an electronic touch-screen or optical scanning of paper ballots, this presents a problem. RCV institutionalizes electronic vote tabulation. If we are to abandon forever the possibility of a return to the hand-counting method that served us well for the previous two centuries, and which is still used in many countries, we should at the very least accompany RCV with provisions for an effective audit the results.

This page describes the voting-technology situation in Maine, the issues related to the optical scanners that have increasingly replaced hand counting in the state, and how all this impacts on the proposal to institute RCV.

The Essentials of the RCV/OptScan Trade-Off Problem
These selected items from the resources found elsewhere on this page (each indicated there by a *)
constitute a "handout" -- either separately or as a single PDF document -- useful for an overview of the issue:

  1. Alternatives to the present "Plurality" voting method
    1. Ranked Choice Voting -- RCV-Maine's short explanatory video
    2. Approval Voting -- "vote for everyone who isn't unnacceptable"

  2. The election-integrity problem of voting machines --
    This are more fully covered on the main e-voting page, but are summarized here because there would be no problem with RCV elections if this were not a problem.
    1. The basic problem
      1. The systems are certified by vendor-supported test labs.
      2. The occasional independent studies can't -- by definition -- prove there are no flaws; they can only point out the flaws they manage to find. Nevertheless, they almost always turn up serious security issues
      3. These flaws are in the software/firmware/hardware that has been certified at the federal and state levels, but are frequently not addressed.
      4. The software and firmware are tightly-held trade secrets, and the few instances of analyzed source code have been shocking from a security standpoint.

    2. * How to Steal an Optical Scanner Election
      This describes general technique by which a simple programming change on a memory card can untraceably alter the outcome of an election. Whether such a change can actually be made in a given situation is the product of many variables.
    3. Code Red: Computerized Election Theft and the New American Century, Election 2016 edition, by Jonathan Simon
      I consider this book, with its clear question-and-answer format and detailed statistical analysis, to be the clearest answer to why an essentially liberal country -- as evidenced by the large number of ballot initiatives nationwide that would normally be considered "liberal" passing with large majorities -- elect conservative people who are dead set against those initiatives.
      1. The book  (from Amazon or Maine interlibrary loan) and a couple of good Simon blog entries here and here
      2. Excellent interviews on KPFA's Guns and Butter, covering most of Simon's salient points -- 2016, 2014
      3. * Summary of 11 basic points

    4. Stephen Spoonamore interviews
      Spoonamore is an expert on computer fraud, working with the international financial industry to minimize and track down credit card fraud.
      1. * 7-minute in-a-nutshell -- this is my audio (mp3) distillation of the points in the 2006 multipart YouTube interview below that are relevant to Maine. He concentrates on Diebold, but the basic problem Diebold epitomized back then still applies to modern systems. There is an irreduceable background "noise" of credit-card fraud of 2.5% of all transactions, so Spoonamore asserts that a 1-2% handcount of randomly selected votes would be necessary to spot any fraud.
      2. Youtube Oct 2006 interview (part 1 of 6)
      3. Youtube Sep 2008 interview (part 1 of 10)

  3. * Democratic Denial
    The Democratic Party, despite being on the receiving end of virtually every instance of electronic election theft in history, has universally turned a deaf ear to the presentation of evidence of this theft.
  4. * Can RCV be consistent with trustable elections?
    RCV analysis by definition cannot be done on a municipal level, and doing it by hand-count at the state level would be a practical impossibility. Which leaves machine-counted voting. Is there any way in which that non-transparent process can be made trustable?
  5. * Auditing an RCV election -- Two possibilities
    Although a hand-count audit of a statewide RCV election with 700K votes is a practical impossibility, it may be possible, through a somewhat complex process, to do a random-subset audit. To the extent that random selections may be subject to manipulation, an alternative method is to do an independent parallel scan/tabulation/analysis using the free open-source Transparent Election Verification System developed in Humbold County, CA.
  6. The Maine situation --
    Maine still has some towns that do hand-counting, but has recently converted all its optical scanners to ES&S DS200 precinct scanners (see below)
    1. * Maine's current election-methodology situation -- a big-picture summary based on a conversation with a Maine election official
    2. Which Maine municipalities use which? -- DS200 optical scanners vs. hand-counts
      1. * Table   (derived from the spreadsheet, below)
      2.    Spreadsheet   from the Maine Bureau of Corporations, Elections and Commissions

    3.    RFP #201208364: Lease of Ballot Scanning and Tabulating System (2012) (53pp)
      1. Usability criteria -- worth 30 points (pp 26-40)
        "The following questions (B.1-B.75) deal with equipment and software necessary for the system."
      2. Security & Accuracy criteria -- worth 15 points (pp 41-45)
        "The following questions (C.1-C.27) deal with security and accuracy of the proposed system to protect the integrity of Maine's elections."

    4. * Why a town clerk likes voting machines --
      from a conversation with the Southwest Harbor town clerk

  7. DS200 and ES&S issues --
    Though many of these are of older systems, the testing mechanisms are still in place that failed to find the flaws mentioned.
    1. DS200 OptScan hackability, December 2016
      During the 2016 Wisconsin recount process, it was discovered that DS200's have a wireless modem capability, meaning that they are intended to connect to the internet, something that has been denied by ES&S. Whether this is the case with the machines currently in stock in Maine isn't known at this time.
    2. ES&S OptScan hackability 2008
      Video:  Hacking Democracy: An In Depth Analysis of the ES&S Voting Systems

      (not to be confused with the must-watch HBO movie Hacking Democracy)
      A presentation at the last "HOPE" conference in 2008 by a University of Pennsylvania hacking team who did the ES&S part of a detailed 2007 study (commissioned by the Ohio Secretary of State) of the security of various voting systems. The team found exploitable security vulnerabilities in almost every component of the ES&S touch-screen and optical-scan hardware and Unity software systems of the time. Some of these flaws could allow a single malicious voter or poll worker to alter countywide election results, possibly without detection. These were older systems, but the point is that neither the official certifiers nor the professional security company hired for the study found all of these problems, so the basic concern still applies with newer systems.
      1. Full presentation (59 min)
      2. OptScan overview
      3. OptScan analysis

    3. Security issues --
      PDF of the following five examples of articles on DS200 and earlier ES&S optical scan systems
      1. Federal agency finds defects in ballot scanners (12/23/2011)
      2. E-voting machine freezes, misreads votes, U.S. agency says (1/6/2012)
      3. We told you so: Newfangled voting machine screwed up (NYC, 5/14/2012)
      4. Paper Ballot Op-Scan Systems in FL, WI, NY, OH Confirmed to Overheat, Mistally 70% of Votes (5/14/2012)
      5. WI Paper Ballot Scanners Failed to Count 1000s of Votes in 'Citizens United' Ballot Referendum (11/19/2014)
      6. Verified Voting's DS200 Security Concerns (May 2012?)

    4. U.S. Election Assistance Commission (EAC) finds trouble -- In 2011, the EAC was forced by revelations in a Cleveland newspaper to acknowledge serious problems with the "Unity" software system used on the DS200, after the EAC had already certified it. Despite this, the EAC refused to de-certify it. ES&S withdrew its certification request and later submitted another system, which also had problems. (More info can be found the above articles; see also copies of the EAC documents referenced here.)

  8. Appendix: The Election Systems and Software (ES&S) optical scanners
    1. DS200 -- a light-duty ("precinct") scanner used in many of Maine's municipalities
    2. DS850i -- a heavy-duty batch scanner for use at the state level. It is not yet used in Maine, but would be required for RCV. (local copy of original)

Apology -- Because the above considerations are being offered quite late in the RCV game, I offer my sincere apologies to those who have in good faith put in a tremendous amount of effort in bringing RCV to the attention of Maine voters and providing them with the opportunity to choose it. In the early part of the campaign, it had seemed like a motherhood-and-apple-pie issue to me as well.

When I saw an excellent hands-on demonstration of it in Bar Harbor on September 22, 2015, it seemed to make sense, until I realized that all ballots really do have to be included in the final calculation. In a small-scale one-municipality election such as that demonstration, checking any suspicious machine tallies by hand would be straightforward. But for a statewide election (700,000+ votes), or even a Congressional-District election, it began to look like a hand count would be difficult, if not impossible. As I looked into it further, that proved to be the case.

I've raised the issue of electronic election theft at many political meetings -- often giving out copies of the Code Red book to activists and candidates -- but have been mostly met with silence or "conspiracy theory" dismissal. It is only now, with the possibility of elections in which fraud cannot be detected, that people are beginning to recognize the problem. Had this realization developed earlier, I believe more serious thought would have been given to the possibility that RCV might be a dangerously short-sighted choice for our Maine gubernatorial dilemma, at least until such time as we take the steps necessary to resolve the electronic-election integrity problem.