Proposed Solution for Election Accuracy, Transparency and speed

[Land Shark]

General Idea: Multiple groups of bipartisan hand counters check each other on their subgroup of ballots. This is highly likely to remove all error in counting, but if there is still error, if the ballots have been randomized to begin with, then each hand counting group will have similar results within statistical tolerances. Finally, if hand counters are insufficient a machine scanner can be used because it too is being checked by simultaneous hand counts, eliminating the risk of substantial scanner error or tabulator hacking. If the final computer used to report results is modified, defective, or hacked, the paper trail of the counting handgroups would reveal it.

Same idea above, just in more detail:

If all the paper ballots are counted at the precinct level (either fully or to the extent time allows), then sent to central counting, shuffled or randomized in one large bunch but just marking each as an absentee, provisional or election day vote for statistical purposes, then the entire randomized group would be homogeneous.

Within the homogeneous group, you then split into various groups for counting, each of which has reps of all parties plus independent observers, and they use the counting method that is most accurate: nobody moves on until we all agree on the count.... this virtually eliminates human error and error generally so long as motivated people at cross purposes are present who are not outright crooks, but the crooks are under intense observation by others.

Separated into separate groups with several to many motivated observers in each, it will be nearly impossible for a group to cheat or make a mistake.

But IF THEY DO, since we've randomized the whole population if there are any statistically significant differences BETWEEN groups, we will know to do a recount right away, or, to be more precise, not to release any initial counting totals until double or triple checked.

The additional advantage of this approach is not just accuracy, but also speed: If the number of ballots are just too high for the number of available human counters, a machine scanner can also be used by humans to speed up PART of the counting, but even the machine scanner is being cross-checked by the groups of human hand counters.

Thus we get initial accuracy, we get all the speed we truly need, and it's done mostly with volunteers (even partisan ones) so it shouldn't be hard to staff. But in the worst case scenario, if we had to hire some of the hand counters to have enough, this is a wonderful jobs program that also teaches people about their democracy and will increase their pride in our societal commitment to accuracy and fairness. That should be a good temporary jobs program, perhaps even high school government classes would get the day after elections off so they can stay up late and know one of the answers to what they can do for the their country....

An additional possible fallback is if there are simply too many races on a ballot every now and then, the elections department releases the full counts on the HOT races (at least those expected to be close before the election) on election night or the morning thereafter, along with the PARTIAL machine scan results where machines are used. Then the hand counts and checking can be finished up later on for all the races, with the cross checks as identified above.

I think this would work, would make sense to the average citizen, and would be something the loser of the election could trust. (the most important standard, since winners always trust the system). It also has the benefit of getting rid of and being more accurate than touch screens.

[sweetheart]

1. The problem is in registration

As much as i admire your solution for counting. The area where the
cheating is going on, is in allocating crap voting technologies to
democratic precincts, too few to handle the number of voters... and
as well, in abusing the disenfranchisement lists, challenges, and
provisional ballots to deny minority voters their voting rights.

In this regard, i'm in favour of putting a video camera and a microphone
recording every single second of the registration desk at every single
polling place, that the frauds are captured on camera and actionable.
If you want to apply a simple, cheap technology to fix the real fraud,
i believe that is the simplest.

[Land Shark]

2. Good idea, but the only problem I'm trying to solve here is COUNTING

As you wisely observe, there are other issues as well. But this proposed solution should not interfere with those other solutions, it is focused on counting the vote.

[sweetheart]

4. As well you'll need some sort of machine

Undervoting, overvoting and unintelligible ballots is part of the
problem with a paper ballot. The lever machines were better at
capturing the intent of the voter, as they did not allow ballot
corruption of that sort.

We're talking several thousand ballots that your counters will be
leaving to the side in the "spoiled" area... pretty much invalidating
the accuracy of your count, in terms of the bigger picture.

Rather, i would suggest, that if an election falls i to the margin
of counting error, that a run-off election be held. That is the sane
thing to do, as is done in more first world countries.

[Land Shark]

6. Unreasonable to expect perfection

Every car or voting system has SOME problems, we need to choose the one with the fewest problems that is also 1. transparent and 2. verifiable and 3. not secret, nonobservable or of high expense to audit on the practical level.

The MarkSense machine is a touch screen but only for the purpose of helping to create an extremely clear mark on the paper ballot, which is the real record. Some of those machines might help the elderly but those that are confident of their ability to do it by hand could also do so.

Watch out: machines are very expensive and become bottlenecks like Ohio unless there is a HUGE absentee program, in which case these expensive toys are not TRULY handling the vote anyway. Thus, the only way the machines are affordable is if WE'RE NOT USING THEM in any big way. Thus, they're ideal for ADA compliance and voter assistance, but if people can grab and vote their paper ballot up against a wall, they need not wait for a booth and then there need not be ungodly lines.

Alternatively we can go to vote by mail, having similar flexibility as to where and when one votes. This could facilitate vote buying, but if that was a significant risk we would see it already in Oregon (where 100% is vote by mail) or Washington state (where 65% is vote by mail through the liberal absentee program)

[sweetheart]

8. ... hence the principal of a runoff election

Rather than worry that a poll will be carried by the margin of error,
why not simply declare a runoff poll when the result is within that
margin... then everyone is happy.

The postal ballot has its own foibles, including spoilage, lost ballots
and, unless registered mail is used, there can never be the presumption
that the ballots all make it in. Including the postal margin of error,
postal balloting MUST trigger a runoff election within the margin
of error. Even Ukraine has runoff elections. It is not rocket science
and basically prevents the fraudulent conditions you're trying to
prevent using bookkeeping accounting standards in counting.

[CanOfWhoopAss]

16. Lever machines had faults too.

Someone else could probably explain it better but... Some of the lever machines where rigged so that when the lever was pulled the wheels that punched the ballot where shaved. The lever moves and everything appears to work but the ballot records nothing.

[skids]

3. Interesting idea.

Bringing more audit into the counting process from the start is good in general.

One thought I had for those counties that refuse to go to 100% hand-count from the start is this: in voting day, each voter has a one in 10 (or something) random chance of voting on a paper ballot. The paper ballots are compared statistically to the machine ballots and if the results are outside the MOE for a ward/city, a mandatory hand count is to be done before any results are reported to the SoS.

[Land Shark]

5. that's another way to help with the idea above, thanks! n/t

[understandinglife]

7. Yes, it will work and work well. However, you invoke a problem that...

....does not ever need to exist -- a computer in the tally process.

No computer is needed anywhere in the process. None.

Tally's are done in a bound ledger at the precinct. That bound ledger is escorted to the county seat where another bound ledger is used to record the precinct results. That bound ledger and all the precinct ledgers are escorted to the State SoS office responsible for tallying. All county results are entered into a bound ledger. A telephone call is then made to appropriate media to annonce results. All proceedings are a/v recorded. No computers; all acts supervised by people.

And, as a separate issue, the format for the ballot should be paper and pen.

And, I think it essential to separate Federal elections and votes on amendments from the day that State and local elections and other State/local level ballot measures are considered.

[Zan_of_Texas]

9. UL: Agree.

Why is everyone intent on inventing a brand new system, when Canada and some rural areas on the US and some European countries do paper, with paper counting?

I like the bound ledger book, by the way.

If the count is accurate in the precinct, and agreed by all hands, it would diminish any need for a recount.

The fly in the ointment remains absentee, military and overseas ballots, where I think real fraud can still be committed. But, at least the paper ballot vote on ONE DAY, holiday, can get most of the votes visible and transparent.

And, yes, if we tighten up the count, the registration fraud and shenanigans will increase (like they could get worse than Ohio?!). So, it's like a long balloon filled with water -- the weakest point will burst the whole thing. Gotta make there be NO WEAK POINTS.

[Amaryllis]

17. See this thread for more on hand counting and Canadian system:

http://www.democraticunderground.com/discuss/duboard.php?az=show_topic&forum=203&topic_id=303880

The issue with hand counting is the very elaborate ballots we have, and ballots with splits. Andy keeps talking about this also; Land Shark, do you this working for ballots with splits like you have in King County?
My eyes glaze over with some of the technicalities.

[Land Shark]

18. Thanks Amaryllis, see this thread..

http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=203x315368

If we start out from all paper, if complexity truly becomes an issue and ballot simplification NOT an option, then we use scanners to scan a whole bunch and simultaneous hand counters are checking the scanners, and if any of the machine groups or hand counting groups come up with discrepant totals, there is an effective "recount" before the vote is even announced (but candidates would still have the right to a recount)

We need to have a position to compromise from (if necessary) and the thread above gives more details of what could be an ingenious fall back position that also has the benefit of giving elections officials a fig leaf for some of them technology purchases.

[Amaryllis]

19. I did read it! Good ideas. Oregon requires each county to randomly select

one precinct to conduct a hand count of 1% of the ballots to check against figures the Diebold tabulators are reporting. Step in right direction; far from sufficient, wouldn't you say? I know Kip of 51 Capital march was insistent during the NH recount that it was imperative to recount entire counties, not just precincts within counties to determine tabulator fraud.

We are all very nervous about the Diebold tabulators.

[Land Shark]

20. Sampling is suspect when catastrophic error possible

For example, we wouldn't sample 1% or 4% of the people entering airports as a method of callig the other 96% or 99% "safe".

So the question is, what is "catastrophic error" in elections? Every vote is sacred, at least to the person casting it, but others will argue for a lower standard. A hundred votes? Can 99 people be ignored, when the amount someone loses by is significant as well (though no where near as significant as winning in our winner-take-all system). Or is it only a catastrophic error if it changes the results of the election? (that's an extreme position, but a common one, in my opinion)

[Amaryllis]

21. A position we certainly heard often in 2004 election. "Not enough to

change the outcome." How many times did we hear that? Applied to so many things. Irrelevant, in my opinion. Votes not counting only matters if it's enough to change the election result?

[Wilms]

10. kick n/t

[Land Shark]

11. I agree with paper ballot approaches, but...

But having worked in a legislature and voted on a board, I'd be ready to trot out the option of optical scanners BUT ONLY if checked by plenty of hand counting AND ONLY if absolutely needed for speed or compromise (like elections officials needing a fig leaf so as not to lose their investment in elections technology)

[marions ghost]

12. Your plan calls for counting and surveillance by

groups of citizens at all stages of the process which seems the only way to go -- after what we have seen in these recent elections.

I'm OK with scanners used as you specify, with simultaneous hand counts, along with an equally transparent way of protecting and checking central tabulators. Optical scanning has been around for a long time and has a proven track record for reliability. It's an example of appropriate technology.

Your idea --hand-counting AS the vote is going on -- makes hand-counting an integral part of the system, while using simple machine counts as an auxilary tool. The use of hand-counting WHILE the voting is going on would help reduce the need for expensive recounts. You get more accuracy initially as you say.

But how can we get the legislators and everybody to consider an option like this when they are so gung-ho about touchscreens "with paper trails" as the solution? If your goal is to have "voter-verified paper ballots" anyway--WHY put touchscreens in there between the voter and the paper? There are just as many issues about mistakes with touchscreens as with paper ballots. After what we've seen in the last election, it's hard to imagine anyone defending them.

[Land Shark]

14. Washington state close to going to all-paper

Because of many factors, including becoming aware of the cost and dealing with the faux-objections to paper (no system is perfect, but electronic just adds more problems without solving any except once in a while it's faster)

But what's more important than election integrity and why? Are corporate secrets more important? Secret or invisible computer counts?

[Amaryllis]

22. All depends on who you ask and whether they are profiting from e-voting.

[marions ghost]

25. $5 Million for one county in WA to buy DREs!

This is the cost of hardware only. For 30% of the voters. Good grief. HOW CAN COUNTIES LIKE SNOHOMISH ARGUE DREs ARE COST-EFFECTIVE?!?

Interview
--snip--
WB: Please breakdown the cost requirements of having DRE's in Snohomish County.

PL: The hardware cost $5 million. One million was reimbursed by a Bush administration federal grant under the Help America Vote Act (HAVA). Now some people think that HAVA is really helping people vote, in a way that shouldn’t be, but again if the machines malfunction it matters little whether it is malfunction or intentional cheating for purposes of the integrity of the election. Such intent does matter a lot of course, for certain political and possible criminal charges, but those go beyond the main scope of our concern which is the integrity of our democracy.

For the $5 million, the county received just about 950 or so machines, which ended up processing only 32% of the county’s vote. Thus, in a way the only way Snohomish county, even with $5 million dollars can afford to have touch screens is because we DON’T have touch screens being used for the majority of voting.

There is also a software licensing fee of $30,000 per year that is payable. The above cost quotes do not include the cost of the service contract, personnel to service the machines, or the cost to run the elections themselves."

whole interview at:
http://washblog.typepad.com/main/2005/02/activism_at_it...

[Wilms]

26. This is another thing that gets me.

Maybe I'm wrong, but it seems the Low-tech equipment we want would save $ so that we can buy enough disabilities-compliant machines that are secure, if expensive and fragile.

I don't want to put words in Dodd's mouth, but it seems he wants lots more machines than necessary to serve the disabilities-community.

That's not economical.

[Land Shark]

15. How to get rid of touchscreens "with paper trails"

Recognize that "with paper trails" as a political demand is the fruit of some of "OUR" efforts (with "our" defined broadly as all election transparency and verification activists). So, although it's true we don't own the idea, we could make a deal about WITHDRAWING the idea by those who have changed their mind. NOt all of us have, but a lot of us have. At the very least we can emphasize that it is not a solution but a compromise.

An under emphasized issue is machines as expensive bottlenecks. That can also be unequally distributed, whereas a paper ballot can be voted against a brick wall with a black pen if the voter prefers speed over booth privacy. This raises the specter of huge lines for electronic machines because you simply can't grab your "ballot" and vote it against a wall or on your lap..... and also raises the specter of huge costs to serve the entire voting population unless there's low turnout or a huge absentee or early voting program. Legislators tend to listen to those kinds of concerns as well.

Although you may not want to expressly make this point, the overall message needs to show (prove) that technology, while it is often progress and we love it, is not progress here in elections at all because we have more important values, trust, transparency, verification, and simplicity so that the loser of the election can abide by the result. Elections offficials have unfortunately understood the convenience aspect of reporting results all too well but forgotten about the other important roles of elections, and that the only perspective that matters is that of the loser (Just ask WA state Republicans clamoring for a revote, and they'll agree)

[eridani]

27. That bottleneck issue is really important

Saw it as a poll worker with optically ccanned ballots. If the blue privacy booths were full, people just wandered off into the school library stacks, where there were plenty of flat surfaces on the shelves. Can't happen with a touchscreen, or even a punch card system.

[Wilms]

13. Nice thread ya got here. I think I'll kick it. n/t

[BeFree]

23. Finally, good work, Land Shark

Been hoping someone would put together a thread on the why's and wherefore's of using Paper Ballots. Let this be a lesson to anyone still contemplating remaining with the e-vote systems.

Bravo!

[libertypirate]

24. Complicated solutions breed complicated results.

There are two distinct reasons systems get fucked up incompetence and on purpose. The on purpose generally speaking is to offer some unfair advantage to some group, a cheat. Well the best way to beat complication is to make the thing so simple a two year old can count it.

I like the stupid fingerprint on both the ballot and the poll book, makes it personal.

My New favorite...

Anyone found attempting to thwart the will of the people should receive a minimum of 10 years in jail for each count. Anyone found attempting to disenfranchise individuals from being able to vote should receive 15 years for each count. Anyone found altering a ballot, machine, or any voting, totaling, or counting equipment should receive 30 years for each count.

lp