HR 550: Thank Tank. . . . . .. Percent audit? 2 stage Audit?

[FogerRox]

Ok I may not be the one to talk about audits percents and such. But what the heck.

It seems like alot of folks want a statewide percent Audit, say 2%, but in small jurisdictions or races, a higher Audit rate is needed. DO I have this right?

Does that mean that a county wide race would need, say 4%, Town races 5%?

Questions:

-In a large sample- statewide- is 2% capable?

-TOwn council races in my area might have 5k or 6k votes cast. In those size races what is needed % wise?

-2 stage Audits, we are talking, say 2% statewide, plus what?

I think Febble posted up that sort of info, I will search for it later.

Have at it........

[Peace Patriot]

1. Given the lack of voter confidence, the utter non-transparency of the

vote "tabulation" (done with 'TRADE SECRET,' PROPRIETARY programming code, owned and controlled by rightwing Bushite corporations), the huge insecurity and hackability problems with these machines, the unreliable results, and everything else, I want a 100% audit for the next two federal elections. Then we can talk about what to do next.

A 2% audit of the current system is absurdly inadequate. A 2% automatic audit of a system with real paper ballots and in which secret source code has been banned in both the voting machines and the central tabulators might suffice, if certain other conditions were met, including a ban on partisan vendors, an easier, less expensive system for requesting recounts, tougher recounts where recounts are requested, and public access to all ballots. Right now, the voting system is designed like a nuclear weapons facility, with the "enemy" being the voters and citizens of this country.

I realize that HR 550 is a compromise that we may have to stomach, on premise that incremental election transparency is better than no transparency. The bill appears to be designed to pass in a Congress controlled by Bush junta criminals. If passed, it would give us some means of monitoring '06 results--for instance, a real paper ballot, and access to the "trade secret" codes (unless that provision goes soft, as I suspect that it will, and gets knocked out)--as opposed to the current situation, with virtually no audit/recount controls. I would hope that, if, somehow, huge voter turnout manages to overwhelm the machine advantage to Bushites and warmongers this November, that a much better bill would be introduced: one that repeals HAVA altogether, as the piece of shit, corrupting, porkbarrel, fascist, Bushite legislation that it is; or one that cleans house in the elections industry and banishes all corporate privatization of our elections (which will rid us of the more corrupt election officials). Most of the dangers in electronic voting--and all the corruption around it--are created by the PRIVATIZATION of our elections. Get rid of that, and you will have honest government officials and true representatives of the people designing a real voting system with adequate controls--and we can all contribute to that discussion.

The problem--in essence, the loss our sovereignty as a people, which we exercise through voting--is not likely going to be solved by fiat of this Congress. Their predecessor, the Anthrax Congress, took away our right to vote; this Congress is not likely to give it back. And I am immensely distrustful of any federal solutions, even ones that sound good--for fear of last minute amendments that could make things worse, loopholes, and poison pills, such as a nationalized voter database (try to get back onto a voter list that you have been purged from, if you have to deal with Bushite bureaucrats in Washington DC, for instance)--and I especially fear this Congress. IF HR 550 could be passed as is, or if it could be passed with a stronger audit provision, fine. But what are the chances of that? Somewhat better, I'd say, than when Tom Delay was running the House. But not a whole lot better. The Democrats don't have a great record on holding firm, and some of them are themselves corrupt on electronic voting and other gov't electronics issues. (--the ban on secret software will be the first to go, since, if passed, it will likely mean the corporations will pull out of the "market").

Anyway, if Holt is going to try to push this in THIS Congress, I'll be holding my breath the whole time. More likely, it is intended to make the Democrats look good to an aroused grass roots, who are fed up with their past silence on Bushites corporations "counting" all our votes behind a veil of secrecy. What I mean is that the Dems have no intention of banning secret software--and thus corporations--from our election system, and it probably doesn't matter how weak the audit provision is, because it will not be passed this year, and they know it.

I think we're in for a long, difficult slog through every state/local jurisdiction in the country, to try to restore our right to vote. I think we should face that reality, and keep at it. HR 550 is something of a poison pill itself, as to that: it legitimizes HAVA and electronic voting, and strengthens centralization, including the Bushite-controlled Election Assistance Commission. What would we think of this bill if it did NOT have the ban on secret code (which, in any case, may or may not apply to the entire voting system)? What would it look like then? It would give us a paper ballot with very inadequate auditing (ergo, a fairly useless, cosmetic control), and a more entrenched, private electronic voting industry.

Beware!

[FogerRox]

2. For instance...

Does that mean that a county wide race would need, say 4%, Town races 5%?

Was my OP that meandering?

[OnTheOtherHand]

3. I will try to address your first question

(Oh, I will quickly address your second: no, not so meandering.)

Umm, look, it sorta depends on what you are trying to accomplish. One way of simplifying the mere math, that works pretty well for most purposes, is to reason this way: if machine failures/hacks/whatever are distributed in some geography in fraction "p" of all precincts, a random audit of "x" precincts in that geography should have at least a 1 - (1-p)^x probability of finding at least one of them. (The hypergeometric thing is more precise, but this is a darn fine approximation for many purposes.) For instance, suppose that 10% of machines are messed up (p = 0.1). If you randomly audit 1 machine, there is a 10% probability that you will select one of the messed-up ones (and just for argument, let's assume that if it is, there is a 100% probability that the audit will discover the fact!). If you audit 2 machines, there is a 19% chance (that's 1 - 0.9^2 = 1 - 0.81). If you audit 10 machines, there is about a 65% probability (or higher, depending on the total number of precincts). 20 machines: 88%. 30 machines: 96%. 40 machines: 98.5%. All these probabilities are a few points higher if there are only 100 total precincts (and a few points may be huge to you -- for instance, 98% is a lot better than 96%).

If the proportion of messed-up precincts goes down, all the probabilities go down, too. That proportion matters a lot more than the size of the jurisdiction. 40 random precincts out of 10,000 should work almost as well as 40 precincts out of 200 (45 out of 10,000 would be a bit better) -- but the percentages are very different.

OK, so I'm not going to answer your question (grin). The percentage depends a lot on the total number of precincts. But the number of precincts that ought to be sampled doesn't vary as much, until the total number of precincts gets much smaller. Most "towns" I know have a very small number of precincts. Some counties in Iowa turned out to have very small numbers of precincts. In those cases, 4% or 5% may not get you anywhere near the assurance you want. If a House district has 600 precincts, then, well, the formula works pretty well, so take the value of "p" that you want to protect against and dial up the "x" until you reach a level of risk that you are willing to live with.

Or not. I can help with the math, but I can't tell people what to fight for -- I mean, I could, but I urge people never to take anyone else's word about what to fight for. The math says that people's intuitions are likely to lead them astray about the percentages: a small percentage can be pretty damn informative for a state (except maybe the tiny ones), while a much larger percentage (or some other approach altogether?) is required to yield similar confidence levels for smaller geographies.

[sadjonny]

4. elect-voting data, and networking - the math equals 100% failure

If you can not VALIDATE the data, you have 100% failure. 
PERIOD.
No point of having elections, using electronics, digitized
data and networks with 100% failure.

It will be cheaper for our country to place all these
electronic boxes under a steam-roller, and use paper with a
hand count.

LOOK AT WHAT IT HAS ALREADY COST YOU!!!
THE PRESIDENT OF THE UNITED STATES WASN'T ELECTED!
THEREFORE YOU NO LONGER HAVE THE RIGHT TO VOTE.

All this talk about math is utter fucking bullshit, chaos and
confusion.  

You all are deluding yourselves, if you think that
electronics, digitized data and networks belong in our
elections.
(Go ahead lie to yourselves) 

It's an improper (domestic terrororist) use of technology.  

Find out who allowed it to permeate your local election
systems, and remove both the technology, and those people.

[OnTheOtherHand]

5. ok, that's an opinion

Several opinions, in fact. But there is an empirical claim here that hasn't been well supported. AFAICT, Bush was elected. I would be more popular around here if I stopped saying so, but them's the breaks.

As for who thinks that electronics belong in our elections, hey, whatever.

[sadjonny]

8. It is PHYSICS, not MY OPINION!

Several opinions, in fact. But there is an empirical claim here that hasn't been well supported. AFAICT, Bush was elected. I would be more popular around here if I stopped saying so, but them's the breaks.

Oh so what you are telling me is that the SUPREME COURT elects our president, not the voters, and not the electorate.


As for who thinks that electronics belong in our elections, hey, whatever.

My friend,

a.) you do not understand physics. *especially Electro Magnetic physics*
b.) you do not understand electronics. r2 a bulb ain't enough.
c.) you are not an electronics engineer. *wading through technical data manuals or desigining chips*
d.) your are not a programmer.
e.) you probably couldn't program a pic controller. let alone understand an ASIC.
f.) you are not a systems administrator.
g.) you have not worked in a NOC.
h.) you have not hacked telco lines.
i.) you have not cracked a server.

Okay I make some assumptions here. But I bet I am right.

If I am right. Perhaps you need to step back. Or do you have an agenda I wonder?

[Amaryllis]

6. What;s the purpose ? Have to define what you are trying to find out

first. See this:

http://www.democraticunderground.com/discuss/duboard.php?az=show_mesg&forum=203&topic_id=422833&mesg_id=422833

THERE ARE DIFFERENT TYPES OF AUDITS FOR DIFFERENT PURPOSES. The purpose of the Holt audit is, in his own words, to "ensure proper functioning of a voting system by spot-checking its tally against the voter-verified paper records." NOTE: NOT TO DETERMINE THE WINNER OF THE ELECTION, or to verify the results of the election. This type of audit will not do this, but this was never its purpose. A different type of sampling is required to verify the outcome of the election.

Holt says this is the purpose of a recount: "A recount seeks to determine the actual results of an election."

That being said, there can, and must be, audits specifically designed to determine the outcome of the election, which may be more aptly called a statewide verification protocol. THIS IS NOT THE PURPOSE OF the Holt audit, and this distinction must be made to avoid talking at cross purposes with each other. Obviously, we must not rely only on recounts to determine the actual results of an election (we all know how far we get with recounts in the current poltical climate); we must have an audit process that serves to determine the accuracy of the election results.

I am not arguing for or against the Holt audit but making the point that we must distinguish the type of audit, what its limitations are, and what its purpose is BEFORE we start arguing the merits of any particular protocol and its effectiveness if we are to make any headway in this discussion. An audit protocol cannot be evaluated without consideration of the purpose and what one is trying to determine.

[FogerRox]

7. I want to:

"ensure proper functioning of a voting system by spot-checking its tally against the voter-verified paper records."

Im thinking about the 2 stage potential of HR550...

2% statewide, plus.....

[BeFree]

9. Amaryllis

Good points. You keep at it. There needs to be an understanding of what it is needing to be accomplished before IT can be accomplished.

My two cents: What can the citizens do? Supposing 550 is not passed this year, (it probably won't be passed) what can we do this year?

Could we demand, as citizens, a look at the paper trail? Several places do not allow such a thing, some do.

Where we can have a look at the trail, we should begin now to establish citizen committees organized with the sole intent of auditing the machine produced count.

The first step would be to decide what the audit intends to accomplish. My idea is that an audit would be designed to back up the machine count; to prove the machines counted correctly.

What election official would not welcome a citizen committee to prove the machines are honest? <grin>

So, my question to auditing experts is: How would one design an audit to prove the machines counted correctly?

[Amaryllis]

10. I am no expert on audit protocols; I just noticed that a lot of the

misunderstanding on these audit debates comes from people arguing over protocols without defining the purpose of the audit. So I keep asking people to define what they want to determine when the subject comes up.

As far as what to do...is there an ER reform group where you live? If you don't know, PM me and I can find out the one closest to you.

[BeFree]

11. Is there an audit expert about?

Seems like there have been a few audit experts in here. Any still here that can get us up to speed?

Where I live...I am the ER group. Small town and all that. However, it would be nice to see a few groups in a list for the new folks comin' round. Is there a site with links to the scattered groups?