The Voting Machine Company Counters With a Strongly Worded Response But Fails to Answer the Critical Question: Do Their Machines Count Votes Accurately?
By John Gideon on 10/20/2008, 3:29pm PT  

Guest Blogged by John Gideon, VotersUnite.Org

As I briefly reported in Friday night's "Daily Voting News" a New Jersey Superior Court Judge has ruled that a court-ordered Princeton University report critical of the state's Sequoia Advantage DRE (touch-screen) voting machines could be released to the public. The only stipulation was that four paragraphs and a number of appendices were to be redacted.

The 158-page report [PDF] was released publicly on Friday afternoon warning, among other things, that:

  • The lost votes during New Jersey's Super Tuesday elections "were caused by two different programming errors on the part of Sequoia"
  • "New Jersey should not use any version of the AVC Advantage that it has not actually examined with the assistance of skilled computer-security experts."
  • "The AVC Advantage’s susceptibility to installation of a fraudulent vote-counting program is far more than an imperfection: it is a fatal flaw."
  • "The AVC Advantage is too insecure to use in New Jersey."

Before the Princeton report was even released, however, Sequoia Voting Systems issued a press statement [PDF] and a scathing response to the Princeton report. The Sequoia response, all 19 pages, is a strongly worded attack on the Princeton computer scientists and their motives, but fails to respond at all to, perhaps, the most crucial point in the devastating Princeton report...

One paragraph is typical of the language of Sequoia's response report:

In particular, we take the academics to task for their inflammatory tone - using "steal" over 100 times, their editorializing on the wonders of paper ballots and optical scanning, their numerous factual errors and cases of intellectual dishonesty (several of which we will highlight), and their inappropriate and uninformed extrapolations to versions of the AVC Advantage that were not studied.

But while Sequoia was busy throwing darts at the Princeton research team, culled from computer scientists and security experts from around the world, and calling their motives into question, the voting machine company seems to have forgotten to mention something in their response. Something very important.

As previously reported by The BRAD BLOG (here and here) the reason for the original lawsuit, the court case, the inspection and resulting report was that at least 38 Sequoia voting machines used in the February 2008 primary election in 8 counties had significant anomalies. In Section VI of the Princeton Report they explain [emphasis added]:

Certain county clerks, and others, noticed inconsistencies in the printed paper results reports from New Jerseys' Presidential Primary election of February 5, 2008. We have found that these were caused by two distinct design flaws or programming errors in the AVC Advantage voting machine. As a consequence of these flaws, voters were disenfranchised.

Sequoia's response goes through the Princeton report section by section, sub-section by sub-section but when they get to Section VI --- and specifically sub-section 56 where the Princeton team discusses the specific failures of the machines to count the voters votes accurately --- Sequoia is silent. No comments.

The Help America Vote Act of 2002 (HAVA), Section 301(a)(5), requires that the vote-counting error rate of each voting system used in federal elections must comply with the 2002 Voluntary Voting Systems Standards. It does not matter whether the system, as pointed out by Sequoia several times in their response, was tested and certified to the 1990 standards or the 2002 standards. Federal law requires that for operation in an election, the allowable error rate is 1 error in 10,000,000 ballot lines (0.00001%).

It appears that Sequoia has no arguments to respond to the failure of their voting system to accurately count votes; a failure that is a violation of federal law.