OMG! New twist in Vermont's "audit!" A COMPUTER will make the random selection!

[garybeck]

This keeps getting crazier by the day.

For the last year we've been pushing for audits in Vermont. Finally they said yes a couple months ago. So we immediately started asking for details. One of the most important questions we've been asking is about the selection of the audited precincts. How can we observe the selection? Will it be on TV? do we have to go somewhere to observe the selection, so we can have confidence that it is truly random?

After weeks of stalling, last week I finally got a reply from the Deputy Sec of State, stating that in fact the selection will not be random. They would, behind closed doors, decide which precincts would be audited.

I then sent out a scathing newsletter to our Vermonters for Voting Integrity group, stating that if it's not random it's meaningless. how do we know that someone doesn't know ahead of time which precincts are going to be audited? how do we know anything at all if it's not random. the brennan report clearly states they must be both random and transparent, or forget it.

So many people wrote to the SoS office and demanded to know why they're not doing the audit in a random way. They all got similar replies. "Why on earth would you think it's not random? Of course it will be random." many of these people forwarded their repliese to me and asked me what they should do. Since they seem to have stopped answering my emails, I can now only communicate through others. I guess I've been blackballed by the SoS.

So, I suggested they ask the question directly. Ask them where you have to go to watch the random selection. Tell them you want to observe the random selection and want to know where to go.

For two days people were writing to them, asking this question, and we got silence. today we are getting the answer, and it's sickening:

I think we plan to use a computer generated random selection process so there will be nothing to observe. --Kathy DeWolfe, Director of Elections of Vermont

Let's take a look at what she said. First of all she says "I THINK." Excuse me, the election is 1 week away and you don't know how you're going to select the audits? What is it we are paying you to do?

Now the real problem is that the "random" selection now appears to be done on a computer. I'm very skeptical of this. Are we to believe the SoS just happens to have a computer program she downloaded off the internet that allows her to make random selections under a variety of parameters? Although a program like this is fairly easy to write, they don't just come out of thin air, and you can't find a random selection software at your local Best Buy. Considering she isn't even sure how it's going to be done, that tells me they don't have the software yet. Where do they intend to get this program?

The skeptical part of me is saying, they could be getting this program from Diebold. How do we know they're not? And if they are, the audit is completely meaningless. Even if they're not getting it from Diebold, it is bringing computers in to solve a problem created by computers. This makes no sense at all. The whole purpose of an audit is to MANUALLY check and verify the computer counts. We don't want, we don't need, we can't have, a computer with secret software code deciding which pricincts to audit.

This is not random, this is not transparent, this is not acceptable.

The lottery winner is chosen in public view, so people can have confidence it was chosen randomly.

The military draft is chosen in public view, so people can have confidence it was chosen randomly.

The election audit NEEDS to be chosen in public view, for the same obvious reasons.

Vermont is a small state and we don't have that many precincts. Is there any logical reason on earth they can't put some numbers in a hat and pull out a few, in public view? Why do they need to go out and buy a computer program to do this? This is very suspicious!

They have changed their story on this at least 3 times. First they said we don't need audits. Then they wouldn't tell us anything. then they said it's not going to be random. then they said it is. then they said they "think" a computer is going to do the random selection. What the ___you fill in the blank___ is wrong with these people? Do they have any idea what they are doing?

Whether or not you live in Vermont, we are all affected by the results of this election in every state. You are all welcome to contact our secretary of state and express your view on this, especially if you agree the audit must be RANDOM and TRANSPARENT, and that means the selection must be made in PUBLIC VIEW and not inside some computer on some unknown computer program! Otherwise it's meaningless!

Contact Deb Markowitz, Vermont Secretary of State:
[e-mail]

[truthisfreedom]

1. i'm 100 percent with you on this. we have a similar system in MN and now

i want to know how the random selection process will work for audits. hmm...

[Old and In the Way]

2. If the list of precincts are loaded into Excel, there is a random

generator function built into the spreadsheet. Why not use that?

[Tin Man]

9. You're correct - the Excel "RandBetween" function works well for this purpose

Anybody with Microsoft Office can do this.

[garybeck]

14. but, to be *transparent* it really has to be in public view

it's much easier to allow public viewing of someone picking a number out of a hat, than someone tinkering with an excel spreadsheet. agree?

[Old and In the Way]

17. Yeah, it's a no-brainer....it is ridulous to even consider a random#

generator on a computer when there are plenty of analog ways that will assure public confidence.

[Tin Man]

22. Agreed - A bingo cage & ping-pong balls would work too, and

...and provide better transparency.

[garybeck]

3. Wait there's more! They just called me!

Just got a call from the Deputy Secretary of State! I guess I'm not completely blackballed after all. I wonder if the call I made to my State Representative had something to do with it!?

They intend to download a free program "from Google" to make the "random" selection! They don't have it yet, but they're sure they can find one in the next week.

there you have it... a free program, best case written by a highschool student as a project, and worst case written by Diebold themselves, will be making the SECRET-RANDOM selection.

Thank God for Google. I guess all is well now and I can go back to my easy chair and remote control.

[meuniermr]

27. I'd look skeptically at this;

Google has been hijacked by the CIA; According to a former CIA employee who goes by the name "Robert D. Steele", who is the top non-fiction reviewer on Amazon. This accounts for their censorship in China and of what videos are in their top 100.

[garybeck]

28. Google....

yeah, I don't trust them... one interesting thing is that the founders of Google are funding one of the largest manufacturing plants for solar panels ever built. They claim to have a new technology that will plummet the cost of solar panels and make them cheaper than conventional utility power. check it out at nanosolar.com

[emlev]

4. In San Diego they use a Diebold random [sic] number generator

to choose the precincts for audit.

Google Diebold and "random number generator" and you'll find out more about this. Don't know if the other e-"voting" vendors have them.

It's yet another outrage.

[garybeck]

5. I'll BET that's what they're doing. n/t

[garybeck]

6. couldn't find anything on Google. if you have anything on that San Diego

thing please let me know. I'd love to go directly to them and ask if that's what they're using.

[emlev]

7. Very interesting. There used to be a lot there. Now nothing.

How does one go about having something disappear from Google? I mean, just killing a site doesn't do it, does it?

I'll see what I can find you, Gary.

[Pacifist Patriot]

8. Fox meet Hen House

[robinlynne]

10. Here in LA, last election the registrar let an election protection person pick

the random precincts using a random table of numbers, after poeple got into exactly the same argument. It simplifies things. no doubt about how they got chosen.

[robinlynne]

11. found this about how random selection should be made. hope it helps.

The following are essential to a meaningful 1% Hand Tally:

1. The random selection must be made by a transparent method, such as
10-sided dice or numbered ping pong balls selected from a tumbler
(methods used in Contra Costa and Alameda Counties). Using
vendor-supplied computer generated random numbers is absolutely
unacceptable, as the results could be known in advance to a vendor
insider.

2. The units sampled (DREs by precinct; DREs+Optiscanners by precinct;
Optiscanners by precinct; boxes of mixed-precinct absentee ballots)
must be the same as the units by which the statement of vote is
recorded.

If the statement of vote is traditionally reported by precinct, showing
votes cast in the polling place and absentee, that statement must be
augmented by a full report using the units sampled in the 1% manual
tally. The county must agree to prepare and make public the full
information by units sampled, in order for the 1% manual tally to be
useable as a check on the electronic statement of vote.

If these two conditions are not met, the 1% manual tally becomes
laborious window dressing that is worse than meaningless, as it will
give a false aura of validity to the election results.

We need to work with our Registrars of Voters before election day to be
sure these two conditions will be met.

[garybeck]

12. thanks, what is the source of that? n/t

[robinlynne]

15. email forum from which I'm not supposed to quote. an EP person.

[robinlynne]

16. Gary, if you really need a name as a reference, i'll write him a note and ask for his

permission. let me know.
RL

[garybeck]

20. well put it this way....

if I use it, it would be to send to the SoS. So it would have to be someone with credentials. If I say "some guy on the internet wrote this" they won't pay much attention.

thanks

[robinlynne]

23. don't know about the credentials. But I just sent an email asking if I can put

you in touch with him (her).

[garybeck]

13. UPDATE #2: My letter to the Secretary of State's office

See the update above; they called me to tell me it will be a computer who decides which precincts will be counted. They also gave me some details about the date of the audit. the "random" selection will take place on 11/10 and the audit will take place on 11/16. So there you go... if there's any funny business found by the audit, chances are the "winners" will already have been flown to Hastert's office and sworn in, so it would take a lot of undoing to change any results. Don't ask me why they can't audit on election night, as half of our precnicts hand count their ballots anyways. I expressed my disappointment on the phone to him, that having a computer select the precincts is not transparent and it is not acceptable, and I followed up with this letter to him:

--------------------

Hi Bill,


Thanks for the phone call. Please allow me to explain why a computer generated selection is not acceptable.


As the Brennan Report states clearly, the audit must be both random and transparent. If the selection is made in the privacy of your office on a computer it is not transparent, and when it's not transparent, we have no way of knowing if it is truly random. In previous emails to me you indicated it is not going to be random at all, so now that you are saying it is going to be random, we have even more reason to desire real transparency in the random selection. The winner of lotteries is chosen in public view for a reason - so people can have confidence it is truly random. Military drafts are chosen in public view for the same reason. The same reasoning applies to our election audits.


Another important reason we can't approve the use of software to make the random selection is, in San Diego a similar situation existed. An investigation found that the election officials actually used a random number generator *given to them by Diebold.* I don't think I have to explain why this would be totally unacceptable and would make the audit into a farce. You can't use software written by Diebold to check the accuracy of Diebold machines. Obvious, right? Please tell me that's not what you're planning on doing here in Vermont.


Vermonters for Voting Integrity stands firm on our request for a truly random and transparent audit. We are glad there is going to be an audit, but if the selection is going to be made behind closed doors, it is unacceptable. The experts are on our side on this, if you just read the summary recommendations of the Brennan Report.


As I said on the phone, I would like nothing more than to tell our organization that a good effective audit is going to take place. I would like nothing more than to tell our members they should write to your office and thank you for doing this, and they should stop pestering you about the details of the audit. But I cannot do this considering the most vital point of the audit - the random selection - is still in serious question. Please reconsider your decision to use a computer program to select the precincts to audit. The benefit of having the public's confidence on your side and the praise you will get from our group if you agree to have a transparent audit will surely override any benefits of using a computer program (to be honest, I have a hard time finding any benefits anyway. What exactly are they?). As it stands, the audit is not transparent and we must continue to press for transparency.


I await the day when your office announces a truly transparent and random audit, with a statistically significant sample size (we still have no idea what the sample size will be, so please let me know if you have that information). On that day I will let everyone know they should thank you and your office for helping ensure our elections have validity and integrity.


Regards,
Gary Beckwith

[Febble]

18. It is of course dead easy

to randomise lists using Excel - I do it regularly - but you are absolutely right that it is unacceptable, because it would also be easy to rig.

The thing is, is that random is only half the recipe - the other is unpredictability, so you have to ensure that NO-ONE has run the algorithm before the precincts are announced, and that NO-ONE has time to rig the votes between the announcement and the audit.

Excellent letter.

[garybeck]

19. uh oh.

we have 6 days between the selection and the audit. during that time the ballots will be transferred from the polling places to a the state capitol, where they will for several days.

I don't understand why the audit can't be on election day.

[Febble]

26. Well, that's why audit protocols

have to be so picky, and why custody of the ballots between polling place and count/recount/audit has to be absolutely ensured. In the UK we have all kinds of semi-ceremonial ways of achieving this. And we don't even do random audits, if there is a question over the count, we hand-count the whole damn thing again, straight away. (I know, I know....)

But seriously, the US doesn't seem to have the first clue about how to conduct elections! They don't (in many places) seem to be much more accurate or secure than a show of hands.

OK, I'm preaching to the choir. What do you do? Exactly what you are doing, I suppose, refusing to put up with it. If people running elections don't know how to run elections then they should damn well find out, a lot more than a week before the election.

Good luck!

[Bill Bored]

21. Why not just let LHS do it on their GEMS computer? LOL! nt

[garybeck]

25. GEMs....

that's still the elephant in the closet for me. i've been so focused on the opscans, the memory cards, and secret vote counting inside the scanners, I have not learned hardly anything about how our totals are tabulated. that's coming for me next i guess.

[mhatrw]

24. This is so fucked up. First, there's no such thing as a truly random

number generator. Second, because of the potential for computer selection bias, this is almost as bad as no audit at all.

Use dice. Use a bingo wheel. Flip a bunch of coins. Put your hand into a bucket and pull out a slip of paper. Just about ANYTHING is more truly random and harder to rig than using a computer.