VoteTrustUSA - Electronic Voting Machines: Programmed for Failure?

[Wilms]

Electronic Voting Machines: Programmed for Failure?

By Joan Krawitz, Executive Director, VoteTrustUSA

March 13, 2006

Howard Stanislevic's full report, “DRE Reliability: Failure by Design?” can be downloaded here.
http://www.votetrustusa.org/pdfs/DRE_Reliability.pdf

A second report "Voting Systems Batch Test Results – Reliability," by Stanislevic and John Gideon can be downloaded here.
http://www.votetrustusa.org/pdfs/BatchTestReliability.pdf


snip

Current federal standards allow almost 10% of electronic voting machines to fail every Election Day, according to “DRE Reliability: Failure by Design?” a new report issued by the VoteTrustUSA E-Voter Education Project. The report notes that the acceptable failure rate is even higher – approaching 25% -- in a 5-day early voting period.

The report was authored by Howard Stanislevic, a network engineering consultant whose experience includes working with the Internet Engineering Task Force on Internet Protocol Performance Metrics. Stanislevic points out that the failure rate allowed for touchscreen voting machines (also known as Direct Recording Electronic or DRE) exceeds the actual failure rate of the 40-year-old lever machines still in use in New York by 44%. The Department of Justice has filed suit against New York State for failure to comply with the federal Help America Vote Act (HAVA). HAVA provides funding for states to replace lever and punch-card voting machines with more modern and accessible equipment in time for the first federal election of 2006.

A second paper, "Voting Systems Batch Test Results – Reliability," by Stanislevic and John Gideon of VoteTrustUSA and VotersUnite.org, examines the results from the recently completed "batch testing" of voting systems manufactured by Diebold, Hart Intercivic, and Sequoia Voting Systems, and puts the information from those tests, provided by the California Secretary of State’s office, into the context of the inadequate reliability standards.

The VoteTrustUSA E-Voter Education Project analysis found that federal guidelines would have allowed an Election Day machine failure or replacement every 37 seconds in Maryland, every 23 seconds in Georiga, and every 78-79 seconds in North Carolina and New Jersey.

snip

http://www.votetrustusa.org/index.php?option=com_content&task=view&id=1041&Itemid=26

[OnTheOtherHand]

1. damn, that's gruesome stuff

Would it be too much to ask that the DREs at least be required to appear to work? Sigh.

[Kurovski]

2. Not a single fricking one should be "allowed" to fail.

But then, you already knew that. Recommended.

[GuvWurld]

3. Of course they are designed to fail. Of course. Of course. Of course.

Of course. Of course. Of course.
Of course. Of course. Of course.
Of course. Of course. Of course.
Of course. Of course. Of course.
Of course. Of course. Of course.
Of course. Of course. Of course.

For how long have I been saying we don't need to bother with past statistics and data?

For how long have I emphasized that we must only address election conditions?

For how long have I been pointing out the inherent uncertainty guaranteed by current conditions?

How many other examples have I cited to describe inherent uncertainty created by means other than "elections"? See Blueprint For Peaceful Revolution.

When we hear about the "reality-based community" it is only distinguishable from the "faith-based community" (not a religious reference, but rather a way to identify those who unquestioningly accept the official lies, on faith) because the corporate-military-government-media juggernaut has intentionally blurred what is a "fact" and what is "reality."

To discover that the machines are programmed to fail is not a revelation of anything new. It is an admission of what we've known all along. This is what it feels like when cognitive dissonance melts away.

[Febble]

4. Yay.

This is the stuff.

[ktlyon]

5. sure glad my ATM machine does not have that error rate

the banking system would collapse

[eridani]

10. It wouldn't matter if it did

You have 90 days or more to fix problems relating to your ATM transactions. Also, Diebold is quite proud of its open source software for ATMs.

[ktlyon]

15. that is not my point

if they can be mistake free at an ATM they can be mistake free at the ballot box. And yes it does matter if my bank makes a mistake. If they do I have to take my time and possible expense to straighten it out, also I will be pissed.

[eridani]

17. My point is that ATMs are not mistake free

--not at all. Mistakes happen all the time. Comparing voting machines to ATMs is comparing complex equipment that is rarely used with complex equipment that is used billions of times an hour worldwide. That process shakes the bugs out, so the error rate per transaction is far lower with ATMs.

[bleever]

6. You don't need a head of foil to assume this had an effect on the variance

between the vote "totals" and the exit polls.

:)

[Wilms]

7. Poor Reliability of the machines possibly could have been a factor.

But to be fair, I don't think the author of the OP article, or the authors of the two papers cited, or even the test data referenced in the two papers, indicate significant trouble of that sort.

I seem to recall one of the DRE's had bungled a vote in the midst of a VVPAT jam. Which is real bad, but not something that would throw exit polls.

But also to be fair, the CA tests were set up as if you had a race with hundreds of voters, not thousands. Would we have found tabulation errors by having run 250,000 ballots? :shrug:

Finally, in many of the CA tests, a scary problem was repeated over and over and over again. It's not included in the Reliability spec, because it was not a machine, but a "human error", that caused it.

And that problem was bad Ballot Definition Settings!! And we'll assume they were programmed in error but we know it's feasible to make an error on purpose when setting up a real election.

So I'm glad you mentioned this.

[GuvWurld]

8. Please school me on ballot definition settings (eom)

[Wilms]

12. Beware the "Straight-Party Option"

It matters not whether your state allows it. It's probably available on the GUI (the screen, in this case, the ballot programmer works off) for most of the voting systems systems in use.

How to STEAL an Election -- for Real!

by Bill Bored

Nov-14-05

snip

Here is how you rig an election, even though they apparently had voter-verified paper ballots -- and it wasn't even Diebold!

Read it, link to it, email it, kick this thread, nominate it, and DO SOMETHING ABOUT IT!

HELP END THE MADNESS!

snip

http://www.democraticunderground.com/discuss/duboard.php?az=show_topic&forum=203&topic_id=401378

[Febble]

11. I would also agree

that it would have an effect on the variance.

What worries me far more is that it implies, as so many other things also imply, that the people responsible for your elections think a ballpark estimate is good enough for democracy. It isn't. You don't want an estimate of who won. You want the leaders your citizens voted for.

[Amaryllis]

9. check this out on error rates allowed by HAVA:

HAVA 301 A (5) Error rates.--The error rate of the voting system in
counting ballots (determined by taking into account only those
errors which are attributable to the voting system and not
attributable to an act of the voter) shall comply with the error
rate standards established under section 3.2.1 of the voting
systems standards issued by the Federal Election Commission
which are in effect on the date of the enactment of this Act.

I don't know what "the error rate standards established under section 3.2.1 of the voting systems standards issued by the Federal Election Commission " are but I know they are way less than 10%

[Wilms]

13. I did that, too.

Here, the applicable standard is Reliability, not Error Rates.

Volume 1, Performance Standards
Sect. 3 - Hardware
3.4.3: Reliability

The reliability of voting system devices shall be measured as Mean Time Between Failures (MTBF) for the system submitted for testing. MBTF is defined as the value of the ratio of operating time to the number of failures which have occurred in the specified time interval. A typical system operations scenario consists of approximately 45 hours of equipment operation, consisting of 30 hours of equipment set-up and readiness testing and 15 hours of elections operations. For the purpose of demonstrating compliance with this requirement, a failure is defined as any event which results in either the:

a. Loss of one or more functions; or
b. Degradation of performance such that the device is unable to perform its intended function for longer than 10 seconds.

The MTBF demonstrated during qualification testing shall be at least 163 hours.

But the Stanislevic paper, also notes...

It’s important to note that by definition, any MTBF spec is only an average.

While the above standard is repeated verbatim (except for some of the typos) in Sect. 4.3.3 of the latest (2005) version, the actual National Certification Test Design Criteria mandated by the EAC in Volume II, Appendix C, Sect. C.4 of the 2002 VVSG actually allows 4 failures after 409 hours of testing (MTBF of 102.5 hours), while the same section of the 2005 VVSG allows 6 failures after 466 hours (MTBF of 77.8 hours).

http://www.votetrustusa.org/pdfs/DRE_Reliability.pdf


-on edit-

That last snip is bugging me.

While (and despite) the MTBF is an awfully lax 163 hours, the "National Certification Test Design Criteria" they use would actually allow 102.5 hours now, under the 2002 standards, and then allow 77.8 hours to qualify to the 2005 standard! :grr:

Why are we moving backward? :banghead:


So everybody, all together now! :applause:

The Election Assistance Commission’s Voluntary Voting Systems Guidelines (VVSG) are developed by a

Technical Guidelines Development Committee (TGDC)

http://www.eac.gov/tgdc.asp

[Wilms]

14. So remember. Don't confuse the Reliability Spec with Error Rate

As long as we're clear on that let's go off topic to answer the question.

For testing purposes, the acceptable error rate is defined using two parameters: the desired error rate to be achieved, and the maximum error rate that should be accepted by the test process.

For each processing function indicated above, the system shall achieve a target error rate of no more than one in 10,000,000 ballot positions, with a maximum acceptable error rate in the test process of one in 500,000 ballot positions.

http://www.eac.gov/election_resources/v1/v1s3.doc

[Bill Bored]

16. The real question is:

Does the error rate spec even apply during the failures?

Let's say 10% of your machines do fail on Election Day.
Do these machines have to meet the error rate spec during the failures?
And how long are the failures?
If they're less than 10 seconds, they aren't even counted as failures.
How many votes could be corrupted by a computer in that amount of time?
How will the voters become aware of this?
How will the software behave during a hardware failure?

This reliability issue really makes it seem like we're dealing with a proverbial house of cards, doesn't it?

[Wilms]

18. More credit where it's due...

It has been brought to my attention that erstwhile DUer, MajorFlaw, was very helpful in the preparation of the "DRE Reliability..." paper, for which he deserves our full appreciation.

:applause:

[hedda_foil]

19. Kick for MajorFlaw!

And a promotion to Colonel for a great job!

[autorank]

20. Hey you have no "major" flaw. Thank you. You're a great guy.

What a pleasure to hear this. KR

[Amaryllis]

21. kick