Optical Scan is Computerized Vote Counting

[Wilms]

No risk-based audits? No recounts??

Then your vote may not be worth the paper it was written on.

[Stevepol]

1. IMO, without audits, optiscans are a bigger risk than touchscreens.

They're safer to use to steal elections because they seem to be safer since they have paper. But unless the paper is audited, that paper is absolutely meaningless.

[robinlynne]

2. oh yeah.

[elocs]

3. They use optical scan where I live and they have hand recounts.

It's very easy to see the black line which connects the arrows by the candidate. I feel very confident with it. There is no absolute, foolproof system. It's all about who counts the votes and how the votes are counted.

[Wilms]

4. Hand recounts? Very good.

Now. How does one go about getting one of those hand recounts?

[elocs]

5. In our last election they had hand recounts where somebody on won by 2 votes.

It was a mayoral race. The loser did not have a problem with the result, he just wanted it to be clear that the winner won so that would not hang over him. If I remember correctly the winner's margin increased by 4 votes or so which is not many, but twice the number he had originally won by out of not a lot of votes cast (small city).

[Wilms]

10. I was wondering more what the triggers are.

Recounts aren't as available as you might want in many parts of the country.

WI does have audits. Unfortunately, they aren't risk-based, like NJ's audits.

[eridani]

21. Triggers ought to be statistical, based on the total number of voters in a precinct

--(or other designated district) and the % difference between the candidates in the audited race. The American Statistical Association is very willing to help out with this.

[Wilms]

26. Like I asked, "I was wondering more what the triggers are."

Specifically, the race margin under which a recount would be triggered even if there wasn't an audit.

[eridani]

27. That would also be a matter of statistical calculation

Decide what level of confidence you want, note the number of votes in the district and the percentage difference between the candidates. Then the statistician gives you a number.

[Wilms]

28. You missed my question, but I feel S.A.F.E. with your answer!

:D

A few things:

First, it's not at all clear to me that WI is using a risk-based method such as the SAFE method. If it is, please inform me.

Second, I was asking what the recount trigger is for WI because a fixed percentage audit may or may not provide a high enough confidence level...but it wouldn't matter in cases where a recount (100% Hand Count) was triggered by the margin.

Linked to that is a conversation I had with an auditing enthusiast. There I wondered what states had recount laws that are margin triggered...and wondered what the trigger level was. A small point, but I considered that less auditing needs to be done if the trigger margin for a recount is wide, and that might help sell it to expense weary lawmakers.

I'm glad to know, however, that you know about statistical (risk-based) auditing. It's promotion was part of the reason I made that round-about OP.

[eridani]

31. Might have gotten into the wrong subthread here

I was thinking about Minnesota's law.

[OnTheOtherHand]

32. WI triggers -- recounts and audits

Wilms, I think most people have been using "triggers" to refer to the circumstances under which a post-election audit is escalated. It isn't exactly a technical term, and you were clear about what you meant.

Here is Citizens for Election Integrity/Minnesota's summary of the WI audit law. It isn't risk-based, and it's a "cold audit" that happens after certification and really isn't designed to fix incorrect election outcomes.

Here is a state PDF that walks through WI's recount procedures. As I read this, there is no automatic recount no matter how close the race -- a candidate always must petition for the recount -- but the recount is free to the candidate if the margin is 0.5% or less (or under 10 votes in small races). I found this link courtesy of Fair Elections Wisconsin.

ETA: In case this wasn't obvious from the way I answered, I have no first-hand knowledge of WI election law.

[Wilms]

33. Thanks, OTOH. This is the info I had hoped elocs was familiar with.

Perhaps not. And perhaps that's why we haven't heard back.

In addition to bringing up recounts, auditing, risk-based issues, and triggers, I was trying to get some to realize that without a solid protocol the "at least I vote on a paper ballot" meme is amongst the most dangerous statements made on behalf of reform.

This I sense is the result of an anti-DRE fundamentalism that blinded itself, and others, to the limitations and dangers Optical Scan.

Needs corrections, I think.

[OnTheOtherHand]

34. yes, the paper ballot is just a foot in the door

I think that's huge -- I don't agree with folks who see the op-scans as a decoy -- but it sure isn't enough.

[Wilms]

35. Well, let's OTOH that, if you don't mind.

We both get into enough trouble with wording things. ;)

I take that we both agree an unaudited Optical Scan is not much of an advance over DREs.

As such, I argue, it IS in effect a decoy. At least, I'd be willing to hear out someone who claimed as much since there are many, who in not adequately assessing the situation, assume the paper ballot is THE means and only means necessary to an end.

Blinded, decoys, etc. You get the pixels.

[OnTheOtherHand]

36. well, I think unaudited op-scan is heinously dumb

"Not much of an advance" is not the phrase I would choose, although I certainly see your point. I think Florida is a lot better off with unaudited (or practically unaudited) op-scan than it was with DREs, because (1) it is much closer to getting meaningful audits than if it still had the DREs, and (2) meaningful recounts also are possible. But I still think unaudited op-scan is heinously dumb. So if we disagree, it's a pretty darn subtle disagreement.

For anyone who thinks that paper ballots are sufficient in themselves, op-scan does serve as a decoy. (Note that you specified "unaudited" op-scan, and I didn't.)

[Wilms]

37. Well hrumffffff!

This calls for a 200 post thread where I bash you repeatedly for being a vanilla ice cream enthusiast.

That. Or I'm taking my marbles and going home.

:stickstoungueout:

LOL!

[eridani]

20. Pass a law like Minnesota's n/t

[gristy]

6. Paper ballots are the way to go

If optical scan makes counting them easier, then I'm all for it. :thumbsup:

[RC]

7. How do you feel about accuracy?

[gristy]

8. Pardon me for answering a question with a question

But what is your concern(s) about the accuracy of paper ballots?

[Wilms]

12. It may mean accuracy of the count OR the outcome.

Ultimately, the latter is more important. And it's actually unverified in most elections.

Risk-based audits seem the way to go when using computers to count votes.

[RC]

16. It's the optical scan that I have a problem with.

Optical scanners are computers and being computer can be programed to spit out whatever results the private owners want.
Other countries still hand count paper ballots. To ensure honest in our elections, hand count the paper ballots.

[kster]

17. In America we work along some people want audits some people want to

do a full hand count of the optical paper ballots before the ballots leave the neighborhood.

We are Americans and we can do both audits and full hand counts.

I will not deny them their audit so long as they do not deny us our full hand count at the polling place.

Its the American way! Working along with each other.

:)

[eridani]

23. No, do both. One method is a check on the other

It's like paleontologists who count tree rings by hand and also send samples back to the lab for carbon dating. Agreement between two different methods reinforces the validity of the dating.

[eridani]

22. For single contests, hand count is the gold standard

When tabulating three or more races, opscan pulls ahead.

[diva77]

40. Please read up on the "Hursti Hack" which was conducted in Leon County, FL

http://en.wikipedia.org/wiki/Hursti_Hack


The optiscans are easily hackable. They make unaccountable counting easier, not accurate counting.

From the actual report, posted at http://www.blackboxvoting.org/BBVreport.pdf

SECURITY ALERT: July 4, 2005
Critical Security Issues with Diebold Optical Scan Design

Executive Summary
The findings of this study indicate that the architecture of the Diebold Precinct-Based Optical Scan
1.94w voting system inherently supports the alteration of its basic functionality, and thus the alteration
of the produced results each time an election is prepared.
The fundamental design of the Diebold Precinct-Based Optical Scan 1.94w system (AV OS) includes
the optical scan machine, with an embedded system containing firmware, and the removable media
(memory card), which should contain only the ballot box, the ballot design and the race definitions, but
also contains a living thing – an executable program which acts on the vote data. Changing this
executable program on the memory card can change the way the optical scan machine functions and the
way the votes are reported. The system won’t work without this program on the memory card.
Whereas we would expect to see vote data in a sealed, passive environment, this system places votes
into an open active environment.
With this architecture, every time an election is conducted it is necessary to reinstall part of the
functionality into the Optical Scan system via memory card, making it possible to introduce program
functions (either authorized or unauthorized), either wholesale or in a targeted manner, with no way to
verify that the certified or even standard functionality is maintained from one voting machine to the
next.
snip

[rurallib]

9. As long as they are backed by random (and I do mean random)

audits - they should be ok. Every voting rights group I know is OK with Opti-scan with random audits and open source code.

[Wilms]

11. Make that risk-based and I'm happier.

Then there's less to worry about with the pedigree of the source code.

As far as open source goes, it's not in use anywhere and I'm not surprised.

[rurallib]

14. I believe California is going to open source.

As for the risk- based - there just needs to be some randomness so only certain precincts are not the ones always selected.

[Wilms]

15. Why do you think CA will go "open source"? OVC? Or a vendor?

Randomness IS crucial. But there are a few other concerns less well known. Sample size in particular. Have a look at this paper. It'll show why a fixed percentage audit is not a risk-based one, even if it's random.

Percentage-based versus SAFE Vote Tabulation Auditing: A Graphic Comparison

February 9th, 2008

This page describes and links to two versions of a paper on percentage-based versus statistical power-based vote tabulation auditing by John McCarthy, Howard Stanislevic, Mark Lindeman, Arlene Ash, Vittorio Addona, and Mary Batcher. The paper is intended for anyone interested in vote tabulation auditing, from concerned citizens to policy-makers and from people with minimal mathematical skills to professional statisticians.

The latest, longer version of our paper, entitled "Percentage-based versus SAFE Vote Tabulation Auditing: A Graphic Comparison," can be found at SAFE Auditing - Revised November 2, 2007 Version - PDF

A shorter version of the paper was published in the February, 2008, issue of The American Statistician, one of the flagship publications of the American Statistical Association (see http://www.ingentaconnect.com/content/asa/tas/2008/00000062/00000001/art00002 ) A .pdf version is also available on this web site at TAS 'Percentage-Based versus Statistical-Power-Based Vote Tabulation Audits' - PDF

http://www.verifiedvotingfoundation.org/article.php?id=6483

[Wilms]

13. HUGE New Diebold "Accuvote" Optical Scan Vulnerabilties Uncovered by University of CT

"if the AV-OS election reporting printouts are the sole means of reporting the election results (as it is the case in fact in many jurisdictions) then one can write quite complex malicious reporting functionalities that get triggered in specific cases (when e.g., the number of votes of a certain candidate are below a certain percentage) and perform arbitrary vote transfers between the candidates."

http://www.democraticunderground.com/discuss/duboard.php?az=show_mesg&forum=203&topic_id=499058&mesg_id=499058

[kster]

18. New meme optical scan/Hand Counted paper ballots at the polling place and risk-based audits

Its the American way and everyone gets what they want. We can do all three.

Sound good Wilms?

[eridani]

19. Audits are essential, and Minnesota has proven that they can be done.

Not that I don't feel like personally smashing their SecState's head against a brick wall a few times when he says crap about how happy he is that their audit validated his trust in his equipment. No, you fawking idiot, DISTRUST is the default position with respect to any piece of complex equipment, or even simple equipment. That's why you can trust most scientific results--constant performance auditing is absolutely mandatory.

[kster]

24. I agree you should be able to audit, and we should be able to

Hand count all the optical scan paper ballots at the polling place at the close of election, its America, and its the way we roll. :)

[eridani]

25. Except that 70% of voters vote by mail

But don't let reality interfere with your plans.

[kster]

29. Even Scott McClellan knows when the gig is up

and you should to, the people need to take back the control of the ballot counts, by hand counting the ballots in their neighborhood/polling place at the close of election and you damn well know this.

No more machine counts, ITS JUST ASININE FOR CHRIST SAKE!

But be my guest keep talking your bullshit.

:mad:

[eridani]

30. Fine. Good luck in moving to WA state and advocating enabling legislation

--to abolish mail-in voting. You talk like you have a magic wand that can force everyone to become a poll voter. I don't believe in such things myself. And ranting about something that is never going to happen is not really activism; it's just a waste of time

[Stevepol]

38. How about another Rec here just to get wider viewing?

[BeFree]

39. Correct

There be some who wish it weren't so, but it is. Computers are totally relied upon when using an opscan, and given how hard it is to get any audits or recounts, it is surprising that anyone here supports the use of opscans. Then again, that support may not be a surprise if they want votes to continue to go to republicans.

[kster]

41. Audits are camouflage, for WE MUST TAKE YOUR PAPER BALLOTS

away from your neighborhood/polling place in order to do the audit.

Audit and or Hand Count the Paper Ballots BEFORE the ballots are allowed to leave the neighborhood.


Anything less is just F*CK*NG ASININE.

But sure as I'm sitting here THEY WILL TRY AND SELL audits as election protection. BULLSHITT!


Go Befree..........

[BeFree]

42. Write on kster

Why do they keep on harping about audits when the real problem is with the computers to start with? Why not attack the computers head on? Do they really believe computers are reliable?

Even the staunchest proponents of the way 2004 turned out is now quite aware of the fallibility of the computers. Of course they still try to sell us on the idea of keeping the damn things. Crazy? Or what?