HAVA section 301 Advisory

[FogerRox]

At the end of the PDF It speaks to Error rate or Accuracy standards

http://www.eac.gov/docs/EAC%20Advisory%2005-004%20(%204%20page%20fit%20).pdf

This is from HAVA Section 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.

This is the FEC section 3.2.1:
http://www.fec.gov/agenda/agendas2001/mtgdoc01-62/overview.htm
Human Error Rate vs. System Error Rate: In the Standards, the term “error rate” applies to errors introduced by the system, not by a voter’s action, such as the failure to mark a ballot in accordance with instructions. The updated accuracy standard is defined as a ballot position error rate. The error rate applies to specific system functions, such as recording a vote, storing a vote and consolidating votes into vote totals. Each location on a paper ballot card or electronic ballot image where a vote may be entered represents a ballot position. The Standards set two error rates:

1. Target error rate: a maximum of one error in 10,000,000 ballot positions, and

2. Testing error rate: a maximum acceptable rate in the test process of one error in 500,000 positions.

This system error rate applies to data that is entered into the system in conformance with the applicable instructions and specifications. Further research on human interface and usability issues is needed to enable the development of Standards for error rates that account for human error.
----------------------------------

HEre is the NASED list of qualifying systems:
http://www.nased.org/ITA%20Information/NASEDQualifiedVotingSystems12.03to7.05.pdf

Notice how few systems have the coveted 2002 certification?
DREs I suspect of not meeting Section 301 accuracy requirtments:

Sequoia Advantage- dates back to 1986 would you pay full price for an 18 year old computer? Handicapped accesable with VVPR for $10,000 ?
Unilect Patriot DRE
ES&S iVOtronic

In addition every DRE with only 1990 certification should be suspect.
Every vendor that gets the 2002 certs trumpets it on their web site.
If you could wouldnt you want to have the 2002 certs? If your machine could pass the tests? Of course if you knew that your machine cant pass the test- why try if you know the DRe wont pass----
HEre are the 1990 standards:
http://www.verifiedvotingfoundation.org/downloads/FEC_1990_Voting_System_Standards.pdf

[FogerRox]

1. So the FEC 3.2.1 as cited in HAVA title III, section 301 is:

http://www.fec.gov/agenda/agendas2001/mtgdoc01-62/overview.htm

1. Target error rate: a maximum of one error in 10,000,000 ballot positions, and

2. Testing error rate: a maximum acceptable rate in the test process of one error in 500,000 positions.

NOw the 1990 standards are here in Appendix F.5
http://www.verifiedvotingfoundation.org/downloads/FEC_1990_Voting_System_Standards.pdf
Desired error rate = 1 in 10,000,000
Maximum acceptable = 1 in 100,000

Further more--- in F.5=
This results in the following decision criteria:
• If the system makes one error before counting 167,753 consecutive votes correctly, it will be
rejected.
• If the system reads at least 297,589 consecutive votes correctly, it will be accepted.
• If the system correctly reads more than 167,753 votes but less than 297,589 when the first
error occurs, the testing will have to be continued until another 465,342 consecutive votes
are counted without error (a total of 762,763 with one error).
This test design replaces the horizontal axis in the time-based illustrations with the total number of
trials. Just as there was a minimum time to accept without failure, there will be a minimum data sample
size to accept without error. As a practical matter, the test is terminated if an error occurs in less than
167,753 votes. The vendor is then required to improve the system.

Review:
2002
1. Target error rate: a maximum of one error in 10,000,000 ballot positions, and

2. Testing error rate: a maximum acceptable rate in the test process of one error in 500,000 positions.


1990
Desired error rate = 1 in 10,000,000
Maximum acceptable = 1 in 100,000

So the maximum acceptable rate of error is 5 fold more------in the 2002 standards

[Kip Humphrey]

2. Test mode testing is a sham and a fraud when testing security & accuracy

Informing a system it is being tested by running in "test mode" defeats all tests for security or accuracy. Proper testing would be in real-time election mode without the system receiving any signal that a test is being performed.

[Pig Farmer]

3. Test mode testing....

....is a sham.

We got us a winner!

So how does the poor vendor or ITA make a large number of votes on a DRE other than a test mode? You think it's reasonable for someone to stand there and punch in 100,000 or whatever votes, WITH NO HUMAN ERROR?

PF

[FogerRox]

4. You have a point-- BUT -- its not completely true-- to the point

that most DREs DO NOT HAVE 2002 certification. If the process was so easily gamed most vendors would pass the Accuracy test, they cant. They havent, newer designs like the Liberty, Avante, & Accupoll
Seem to be the leading candidates for passing accuracy requirements.

Which is good for us-- this is how you decertify these machines.

One county at a time
one state at a time

[FogerRox]

5. Addotionally- NON HAVA complaint voting machines cant be used in a

Federal election. These machines are not eligible for HAVA funds.

[FogerRox]

6. In NJ

6 voting machines are certified for use--- and all 6 have not been tested to the 2002 certs.

[Ellipsis]

7. Would you Nutshell this.

Am I to understand, none, of the curent machines in use are certified for a Federal election?

[helderheid]

8. I just sent this to my list

Many read the Salt Lake Tribune and Deseret News articles, as well as the Salt Lake Tribune editorial page about California's decision to reject Diebold voting machines in their elections due in part to their 10% failure rates. As it turns out, that failure rate makes Diebold (and other voting systems with similar failure rates) not HAVA compliant! Click here to read more.

http://udpc.org/hava.htm

[FogerRox]

9. post #1- follow NASED link--

If the voting machines used in your state do not have 2002 certification,as listed on the NASED site-- it is very likely that they are not HAVA compliant.

HAVA non-complaince means they cant be used in Fed elections.. furthermore--- if a jurisdiction wants to buy new voting machines---
These new machines must be HAVA conmplaint to be eligible for HAVA money.

[Ellipsis]

10. okay.

Do you know if there is an aproved vendor list with models ?

[FogerRox]

11. The NASED link is just that

only voting machines on the NASED list--
that have 2002 certification----
Could possibly be HAVA compliant--- And yes- that list of machines with the 2002--in blue-- on the right hand side--- is a very short list-----

Which means that a very large number of voting machines are not going to meet this requirement by the deadline-

[FogerRox]

12. I have recieved more email concerning our work in NJ

on Section 301 of HAVA. This may be the single most important federal law--- to be used to eliminate the use of some DREs, nation wide--

I cannot stress the importance of this issue--

Persons can take election officials to court for violations of section 301 of HAVA-- See title IIII section 401 of HAVA.