Auditing The Vote on Opscans In NH

[BeFree]

It should concern every voter to make a proper distinction about whether the totals from voting procedures are as accurate as can be.

Looking back at the last few elections we see that the use of the new-fangled electronic voting machines results leave us with a lot of questions.

Take the machines that wiped out thousands of votes in NC in 2004. Because of that NC now has a paper trail mandate so that votes are never again totally lost.

We know that votes cast on electronic black boxes can be destroyed.

We saw it again in Florida just last year, when 18,000 votes in one race magically disappeared, with no paper trail. People's votes were completely destroyed. And because of that, Florida too has moved to a paper trail.

So now that we have moved on to a paper back-up system we should all applaud. I do.

But if we don't look at the paper trail, how will we ever know if the machines are counting correctly? We must take the next step and do a thorough double check of the totals delivered almost instantaneously by the machines, and do it within a time frame that meets the public's expectations for quick results.

Easily done with the optical scan ballots; not so easy with the cash register type of paper trail.

With optical scan ballots we run the ballots through the machines and establish the machine's first count.

Then, by hand, officials pick one race and divide the ballots into separate piles of each candidate.

Then we run each individual pile through the scanner, writing down the number of ballots counted by the machine from each individual pile.

The totals we have now should closely match the counts of the first machine count. If the counts are similar, a double checking has taken place and should result in few, if any, questions.

Trust is established and verified all within a matter of minutes, or we have found the machines to have made a terrible first mistake.

Anything less, frankly, would be anti-democratic and all too dependent on proven to fail machines.

[Warren Stupidity]

1. You have to randomly sample the ballots.

for example by randomly selecting sufficient precincts for a hand recount in order to have the appropriate confidence level that the sample reflects the entire population. One machine isn't sufficient, and the selection process has to be both random and done after the election so that nobody can know which precincts will get audited before hand. The ballots and the equipment have to be under evidence chain of custody control from the start of the election until the post audit certification so that there is less possibility of tampering to begin with. For example, NH allowed technicians to access tabulators during the election, and that is completely wrong. If a machine malfunctions either it is replaced with a spare machine that is under the same custody rules or that precinct gets hand counted.

[riqster]

3. Very good builds-thanks for giving people the details. nt

[BeFree]

4. No randomizations

Just an immediate and full recount done in the minutes after the first count.

Just take the main race on the ballot, make as many piles as there are candidates in that race and run each pile full of each candidate through the counter again.

Oh, run the machine not in election mode, but 'TEST' mode.

[Warren Stupidity]

6. I'm ok with that too.

If I understand you correctly you want every machine run through a test run after a hand count (required to sort the ballots correctly.) I can't quite figure out what we would need the machines for though. The sort is a hand count minus recording the number in each pile.

How would you know if the 'test' mode works the same as the 'election' mode? If I were rigging the system my test mode would work just fine.

[BeFree]

8. Almost

A true hand count would take time, some say too much time.

What I propose is that the piles are not counted by the humans doing the separation.... just piles made. Then let the machine do the actual counting. Done in minutes, not hours as hand counting takes.

See, if the machine is counting the ballots correctly, (and the humans making the piles did so correctly) the machine readout will show the total votes for each single candidate all by its lonesome. No confusion whatsoever.

The reason it needs to be in test mode is that that mode can be used over and over again and is not likely to have hacked software controlling that counting.

[riqster]

11. A true hand count was done as phase one of your process, it seems to me.

By sorting the ballots by candidate.

[BeFree]

12. Well, sort off

But just sort of. No tabulation of the number of ballots is done by humans. No count, just sorting into piles. Then the machine counts each individaul pile, real fast. Fast enough that a recount, or audit, is accomplished fast enough for everyone.

I'm outta here. Over and out, Cya.

[riqster]

13. From where I sit, it could easily be shifted to a hand count

...if we take the machines out of the audit entirely, it will clear the air far more effectively. The counts will still be recorded on memory cards with the same programming that they had on Election Day, after all.

That being said, there's a lot about your proposal that has merit. Fresh thinking is needed.

[riqster]

2. Well said!

What's wrong with double-checking? Hell, I do it with my checkbook ewvery time I balance it, and that has impact on far fewer people than does this.

[BeFree]

5. Exactly

As it stands we don't have a systematic 'double-check' of the machines, even tho there is ample proof, evidence, and common sense that the machines fail.

And yet I read here and there that people still trust the machines without question. WTF?

[riqster]

7. People are wack.

They don't want the government to take their taxes, because they don't trust the government with their money. At the same time, they blindly send their kids off to Iraq to be killed, because they trust Bush when he tells them it's a noble cause.

So many ways we see this, in so many places. It seems the higher the stakes, the less inclined Amercans are to give a shit.

[Warren Stupidity]

9. They blindly send somebody else's kids.

Not that I am in favor of the draft, quite the opposite, but it is simply a fact that if we had a draft as broad as the Vietnam era draft there would be a whole lot less blind trust. Anyhow, something like 70% of the population agrees the war is crap, they are divided (and confused) about what to do about the crap that it is.

The people who are wack are that 20-30% who fit your stereotype: the ones still supporting war while thinking they are against 'big gummint'. I've always wondered how many of those dead enders actually have kids over in Iraq keeping Bush's legacy intact until 2009.

[riqster]

10. True. My post was overly broad.

The important bit was the last sentence, which applies to a lot more people WRT Election Integrity, including the Progressive establishment like Daily Kos et al.