The case for the occurence of algorithmic vote flipping

[RonRules]

Ok, so I have been looking into R and have had a surprisingly easy time learning it. The way they deal with the data was confusing for a while, because it is very different from a traditional programming language. I have also found that there seems to be a function to do anything. I write out 15-20 lines of code to do something, and then later come back and replace it with a 1 line R function that I found.

RonRules, I think the math for computing the cumulative totals and percentages should be easy(~30 lines of R or less). The part I am having the hardest time with is rearranging the data before I do the math. If the data for TX or CA is fairly standard then it might help me to see some of the .csv files from those 2 states.

Thank you so much for taking this on. Indeed, because the data formats vary a lot, and within a state they vary, it makes the job difficult.

Nebraska's elections were just yesterday, so I jumped in and did a chart at the state level. I found that all counties report at the state level and have the exact same format. It is also tabular with columns for each candidate, making the data format nearly perfect.

If you wan to practice with a state, I recommend Nebraska. Nebraska is one of those states with sleepy little counties like Wisconsin that likely hand count ballots. It would be a useful exercise to do the entire state. If you can do one county, the rest will be very easy.

Here's the overall state. Even though Romney had no need to flip at all, he was still a flipper. That indicates that whatever trick causes this is already installed and they can't get to it to turn it off.

Data source: (Download spreadsheet)
http://electionresults.sos.ne.gov/resultsCTY.aspx?type=SW&rid=649&pty=REP&osn=100&map=CTY

I can't wait to see if Utah will be a flipper.

 

[RonRules]

I'm not sure how you are getting the X-axis on these charts. Where does count-over-time come from?

Other problem is even if this is 100% perfectly valid, I do not know what can be done or what it would change. If it won't change the outcome wouldn't I be wasting $100k of taxpayer money to demand a recount and just look like "sour grapes" damaging any future effort to gain elective office?

Something like this, if true, should probably be brought by a 3rd party. Perhaps some unknown "Citizens for Fair Elections" could step up and say "It won't change the outcome, but an anomaly demands transparency" for the sake of future electoral integrity.

I dunno, this is all new to me.

Thank you Gunny for participating here. We're all "gunning" for you.

The X-Axis does not represent time. It represents the cumulative addition of all the precincts votes from smallest to largest going from left to right on the chart. Precinct sizes vary all over the place, because of constant re-districting and for other reasons. The chart implies that you did much, much better in small precincts. Those precincts could be anywhere, but most likely in rural areas.

Did you get the impression as you campaigned that your support varied as widely as is seen on the chart? I doubt it.

I'm now using a program to make the charts that someone else wrote. I can't make changes to the graphics routine that makes the charts, but it looks like we're changing to a better system, the statistical language "R". Hosef is learning it and I have to learn it too.

As far as a recount, I agree that this is too expensive. Because the problem is nationwide, (even yesterday's Nebraska electionas you see above), we are seeking support from academia to make a definitive statement that there is a violation of basic statistics, nationwide and we demand an investigation at the attorney general or FBI level.

I sincerely believe this problem is that serious. Whether Ron is in or out, I will continue this analysis. I believe it is the most important project I have ever taken in my life.

If you know the election clerks in WAKE county and if printed ballots are still available, I recommend an informal spot check on a handful of the smallest precincts and compare that to a handful of large precincts. If the informal count shows a that your true count does not match the published count, you should be able to file a suit to get a re-count done at the county's expense.

 

[RonRules]

I just had the urge to chart the largest county in Nebraska. That's usually where we see the worse shenanigans.

Sure enough Romney was flipping at his best:

Note a slight difference when you compare the entire state chart and Douglass county. Ron Paul mostly flat-lined at the state level, but was flipped down in Douglass county. That's because we have noticed that if a candidate is below the 10% level, the flipping algorithm ignores him. We believe the reason is to prevent an opposing candidate vote count to go negative. It's happened in the past (see Volusia county FL)

This is serious folks. There's a total of 3,144 counties in the US. Romney is flipping up in about 99% of them. This just can't be, it's a mathematical impossibility.

 

[RonRules]

For those new to this thread, here's how an honest election should look like:

Here's an example of the Calfornia propositions that were voted in 2008. That's how is should be:

 

[RonRules]

I'm wondering if they're going to remove Ron Paul from the Utah ballots, because he's not "actively campaigning" there. I would not be surprised.

 

[romancito]

---

 

[romancito]

---

 

[RonRules]

I thought I ran into Mexican or Spanish elections when I saw this:

Departamento de Elecciones de Contra Costa
Elección Especial de Voto por Correspondencia de mayo 8, 2012

http://results.enr.clarityelections.com/CA/Contra_Costa/37859/80844/sp/reports.html

Nope, welcome to California.

We have ballots in 7-8 languages, including Korean, Japanese, Chinese and Vietnamese. I wonder if SOE software got a contract to translate their website to all these languages as well.

Oh also, California has a DEFICIT of 16 Billion.

The DEBT is 361 Billion. I don't know if that includes the debt for each county. I know Riverside has 4.3 Bil in debt.
http://www.usdebtclock.org/state-debt-clocks/state-of-california-debt-clock.html

 

[RonRules]

One case in my mind that seems clear vote flipping is the case in West Virginia where Keith Judd got approx. 40 percent of the primary vote against Obama.

Call me an idiot but I cannot see how 40 percent of people, regardless of how they may feel about Obama voted for an unknown quantity such as a name that represents an inmate in a Texas prison. I just don't see it. And I also don't see how well after Santorum and Gingrich were known to have withdrawn from a race they can gain approximately 20 percent jointly. I just can't accept it as natural.

I had trouble finding data from West Virginia. I'll gladly analyze it for you. If you can find me data at the precinct level that has a format similar to this, it's then a 10 minute job to make a chart and post it.

This is the data format that's easiest to analyse:

 

[RonRules]

And I also don't see how well after Santorum and Gingrich were known to have withdrawn from a race they can gain approximately 20 percent jointly. I just can't accept it as natural.

If you look through this thread I show that when Cain, Bachman, Perry and Huntsman quit, the very next election got them a minuscule portion of the vote. (all less than 0.07%)

On the other hand Santorum (a full two weeks after he quit on April 10) managed to get 18.4% in PA and as high as 26% in some counties. That could not be explained though absentee ballots because the results were similar to walk-ins on election day. I actually suspect that the ballot configuration file switched Paul for Santorum.

 

[romancito]

---

 

[RonRules]

I am sorry that I can't get the data you requested from me. First I am not proficient in that kind of research. But second, I do not operate anymore under a scientific paradigm, while I was raised as a scientist, like we all are, I believe science is dead - social sciences. I believe in intuition. I know when I am being cheated and 40 percent for Judd sounds like a big Turkish rug over my eyes. But I can appreciate what you are doing and value you as a non-scientific actor. Your art is spectacular.

Romantico, your post was so interesting that I dropped everything to do West Virginia. Your intuition needs to be studied scientifically because you have discovered some REALLY interesting elections.

I only have "State-Level" data, but it's interesting enough.

On the evening of the West Virginia elections, I had done a chart for the Republicans and all I could observe was "meh". Here it is again:

Now check out the Democrat elections:

There's DEFINITELY something amiss here. Possibly some serious ballot stuffing in some counties, but I really don't know.

If someone here is from WV, please ask your individual counties for precinct level data. Thanks.

 

[RonRules]

It's often informative to chart the straight precincts size on the X-Axis (non-cumulative).

The Republican chart may indicate some ballot stuffing in some precincts:

The Obama/Judd chart is also weird. Judd exceeded Obama in several counties:

Are you from West Virginia?

 

[romancito]

---

 

[RonRules]

Welcome to Mingo county West Virginia, http://www.mingocountywv.com/ (They don't have an election section on their website. That's too bad) where Keith Judd got 60.2% of the vote against the incumbent president, a guy named Barack Obama.


From Politico:

Get to know Keith Judd with these 10 fun facts:
1. Better known as prisoner number 11593-051.
2. Currently resides in a low-security prison in Texas — Federal Correctional Institution Texarkana — where he’s serving a 210-month sentence for extortion connected to making threats at the University of New Mexico in 1999. His projected release date is June. 24, 2013.
3. Says he has run for president in every election since 1996.
4. Finished third in the 2008 Idaho Democratic primary with 734 votes, or 1.7 percent.
5. Got on the ballot in West Virginia by paying a $2,500 filling fee and sending in a notarized certification of announcement.
6. Among the many things on his resume: “Member, Federation of Super Heroes, 1976-1982,” “Recording Musician/Writer/Producer, Nadine’s Music, Hollywood, California, 1968-1998” and “Agent/Individual Contractor, New
York Society of Reproductive Medicine.”
7. Favorite athlete: Gary Skidmore, a pro bowler. Favorite actor: Gene Hackman. As for his favorite food, he writes, “I forgot.”
8. Lists his religion as Rastafarian-Christian.
9. Mozart ranks as the person he’d most want to meet. “He was cool,” Judd writes.
10. Favorite president: Richard Nixon. “He got us out of Vietnam, and began world peace with China and the Soviets.”

 

[RonRules]

Wow. If I am not wrong I read in these graphics specially in the Democrat graphics that people behave entire different from different sectors, i.e. 'precincts.' Wow. '

Well, you got it. That's the whole point of this exercise. Within a state there should not be such drastic fluctuation in support for an incumbent president.

That's why I'm here.

 

[romancito]

---

 

[RonRules]

In Grant county West Virginia, Mitt Romney has a vertical spike on the chart. Here are the results:

Party Votes Percentage State Total
MITT ROMNEY Republican 1,403 78.16% 72,430
RICK SANTORUM Republican 149 8.30% 12,589
NEWT GINGRICH Republican 114 6.35% 6,499
RON PAUL Republican 109 6.07% 11,569
CHARLES "Buddy" ROEMER Republican 20 1.11% 1,048

Does Mitt Romney gather so much love in Grant County WV to get 78% of the vote?

Ron Paul 6.07%, really, REALLY?

 

[affa]

I'm not sure how you are getting the X-axis on these charts. Where does count-over-time come from?

time is not a factor in these charts.

rather, the vast majority of charts in this thread, unless label'd otherwise, are cumulative votes sorted by precinct size.

that is, you first sort all precincts by number of votes cast, then chart 'cumulative votes'.

we believe, based on historical analysis of past elections as well as study of international elections, that this should 'flat line' fairly quickly. however, in many counties in many states across America, instead of flatlining we're seeing what appears to be a distinct 'shift' of votes, usually from one specific candidate to another, with the benefactor being Romney (and occasionally McCain in 2008).

It is our belief we've found a way to identify vote flipping -- that is, say, taking 10% of candidate A's votes and giving them to candidate D. The specifics differ slightly by state, but generally, it seems that whoever is doing this ignores the smallest of precincts in a region (reduces chance of detection significantly for several complementary reasons) and begins doing it once vote totals cross a specific threshold.

The common 'debunk' is that the sudden shift of votes to Romney is based on 'demographics', but a lot of time has been spent researching this and demographics do not seem to be a significant factor. Nor does population density or rural/suburban/city divides. In fact, we see the same anomaly within 'affected' counties whether the difference in votes between precincts are only a handful or counted in the thousands... which lends credence to it being a algorithmic function.

 

[Aden]

I'm not sure how you are getting the X-axis on these charts. Where does count-over-time come from?

Other problem is even if this is 100% perfectly valid, I do not know what can be done or what it would change. If it won't change the outcome wouldn't I be wasting $100k of taxpayer money to demand a recount and just look like "sour grapes" damaging any future effort to gain elective office?

Something like this, if true, should probably be brought by a 3rd party. Perhaps some unknown "Citizens for Fair Elections" could step up and say "It won't change the outcome, but an anomaly demands transparency" for the sake of future electoral integrity.

I dunno, this is all new to me.

The X-axis is not votes over time. The X-axis represents precincts arranged by voter size. So precincts with 5, 10, 20 voters would be on the far left, and precincts with hundreds of voters on the far right. What the charts show is that votes were stolen from you in larger precincts. They choose larger precincts because they are less likely to get caught.