The case for the occurence of algorithmic vote flipping

[RonRules]

The distribution of slopes is again very "peakish", and not Gaussian.

Hey Liberty, I have a math question here. Since we are cumulating the votes on the X-Axis and this accumulation has a smoothing, centering effect, is it not reasonable to think that you will NOT get a Normal Gaussian distribution, even with nothing but honest nuns at the precinct?

 

[Liberty1789]

Hey Liberty, I have a math question here. Since we are cumulating the votes on the X-Axis and this accumulation has a smoothing, centering effect, is it not reasonable to think that you will NOT get a Normal Gaussian distribution, even with nothing but honest nuns at the precinct?

Sure. I make no claim that it SHOULD be Gaussian. Just illustrating against a well known (and mathematically very convenient!) distribution.

 

[RonRules]

Now, how did Romney's slopes look like in Louisiana in 2012 ?

The answer is: not spectacularly differently from what has been seen historically over there.

I beg to differ, Romney has a clear bias to the right.

For proper comparison, since we have evidence of tainted elections back to 1988 in the US, it would be best to do your histograms using a different county. I recommend Canada. These guys shoot straight like a curling rock.

Here is where you get Canadian data:
http://www.elections.ca/scripts/resval/ovr_41ge.asp?ddlEDRes_prov=35&lang=e

Pick the "Download compressed CSV files in ZIP format, by province or territory" option. Use another province than Ontario, I did part of that one already.

 

[dsw]

If nobody reading this thinks I have a valid point in any of what I've posted then obviously I'm wasting my time and everyone else's, and I regret that. If there is a vote flipping algorithm at work, and it covers as many races over as long a time period as many here believe (including Iowa where the count was done by hand, witnessed in public in many cases by Ron Paul supporters, with individual precinct results reported in real time), then it's one of the biggest and most important stories in our nation's history and I wish you nothing but success in making a clear and compelling case that will get some national attention.

 

[RonRules]

Looking at your CDF's VERY carefully, the four leftmost curves have colors in the chart that appear to be Robertson 88, Duke 92 Buchanan 92, and Buchanan 96, all four of which you indicated were victims of vote flipping in your previous Louisiana charts.

Likewise the "winners" on the right side are Bush 92 and Dole 96, both establishment candidates.

Please confirm that I've got the colors right.

Hey COLORBLIND people, don't try this at home!

 

[drummergirl]

If nobody reading this thinks I have a valid point in any of what I've posted then obviously I'm wasting my time and everyone else's, and I regret that. If there is a vote flipping algorithm at work, and it covers as many races over as long a time period as many here believe (including Iowa where the count was done by hand, witnessed in public in many cases by Ron Paul supporters, with individual precinct results reported in real time), then it's one of the biggest and most important stories in our nation's history and I wish you nothing but success in making a clear and compelling case that will get some national attention.

Please do not consider it a waste of time. For one thing, there are tons of lurkers who were probably asking the same questions as you. Anything that makes us clarify and refine the analysis is productive. (Even if it does drive me to drink lol)

 

[jjockers]

You can't use the math to prove that the graph shouldn't flatten out unless you FIRST prove that you haven't introduced this kind of correlation by the way you've sorted the data. If you prove the right kind of independence first, THEN (and only then) the math argument is rock solid.

Yup, absolutely correct. The flippers seem to have a very difficult time understanding this rudimentary concept. If they want to 'prove' or substantiate the claim of fraud, they first must demonstrate that precinct size is not correlated with any variables that would produce the phenomenon (such as DEM % in your graphs).

I'm just as baffled by the flipper direction, as this concept is simple and yet continues to be ignored.

 

[RonRules]

Yup, absolutely correct. The flippers seem to have a very difficult time understanding this rudimentary concept. If they want to 'prove' or substantiate the claim of fraud, they first must demonstrate that precinct size is not correlated with any variables that would produce the phenomenon (such as DEM % in your graphs).

I'm just as baffled by the flipper direction, as this concept is simple and yet continues to be ignored.

You are welcome to make charts to prove/disprove points. I have made charts that show practically zero correlation between "Republicaness" and precinct size, and you mean to tell me that you have proven "Romneyenss" as a function of precinct size is a fact without flipping?

The big problem you're going to have is to figure out how to first remove the OBVIOUS vote flipping that has flipped over the entire 2012 primaries in favor of Romney. You first need to use historical data or other country's data (like I have) to establish a solid baseline from which you can either show or disprove that these cumulative charts should be horizontal. We will watch your statistics like a hawk and the confidence claims you make.

With sufficient data, you will inevitably come to the conclusion that if a US GOP establishment candidate is present, the chart exhibits a steady sloping in favor of that person and to the detriment of other Republicans.

The simple explanation is vote flipping. The complex, difficult to prove explanation is demographics. Generally the simplest explanations are the correct ones.

Even if you were to find an statistically discernible effect favoring Romney as a function of precinct size, please do a deep brain search and wonder why:
1) Gingrich in particular (which I allege, has similar demographics to Romney) is not affected positively, but rather suffers from this effect, while other are not at all affected. That's the key. If Romney naturally did better, the others would do worse, but we see sometimes that only one candidate is affected.
2) Why 5 other candidates are likewise affected depending on the state and their relative success against Romney.

This chart alone should end the discussion with reasonable people:

You're going to have to figure out what proportion does this Romney-specific effect contributes to all charts we have seen thus far, without reverting to silly games like measuring precincts with 7-8 Libertarian votes.

I'll grant you a couple of Romney specific effects:
1) Mormon population percentage. However, in my experience living in Utah for a couple of years. Mormons are rural dwellers, to a large extent.
2) Rich bankers in rich suburbs. However, rich people have homes on large lots in sparsely populated suburbs. The corresponding precincts are small.

I make the claim that overall, if we could magically erase vote flipping, you would see Romney do BETTER in small precincts populated by Mormons and bankers. We'll see how Utah turns out.

So meanwhile, we need more hands on deck, and we need to produce tons of data so the cherry pickers have a forest of lemon trees to look at.

The Java program is not too difficult to use, but a few minutes of hand holding will save you time. PM me if you've got problems using it.
There's been 64 downloads just this week. Somebody must be using it. Don't be shy, post your charts.
http://sourceforge.net/projects/voteanalyze/?_test=b

 

[jjockers]

You are welcome to make charts to prove/disprove points. I have made charts that show practically zero correlation between "Republicaness" and precinct size, and you mean to tell me that you have proven "Romneyenss" as a function of precinct size is a fact without flipping?

The big problem you're going to have is to figure out how to first remove the OBVIOUS vote flipping that has flipped over the entire 2012 primaries in favor of Romney. You can also use historical data or other country's data (like I have) to demonstrate the solid baseline that these cumulative charts are horizontal, unless a US GOP establishment candidate is in the race.

Even if you were to find an statistically discernible effect favoring Romney as a function of precinct size, please do a deep brain search and wonder why:
1) Gingrich in particular (which I allege, has similar demographics to Romney) is not affected positively, but rather suffers from this effect.
2) Why 5 other candidates are likewise affected
3) What proportional effect of this Romney Specific affects all charts we have seen thus far.

I'll grant you a couple of Romney specific effects:
1) Mormon population percentage. In my experience living in Utah for a couple of years. Mormons are rural dwellers, to a large extent.
2) Rich bankers in rich suburbs. Rich people have homes on large lots in sparsely populated suburbs. The corresponding precincts are small.

I make the claim that overall, if we could magically erase vote flipping, you would see Romney do BETTER in small precincts populated by Mormons and bankers. We'll see how Utah turns out.

So meanwhile, we need more hands on deck, and we need to produce tons of data so the cherry pickers have a forest of lemon trees to look at.

The Java program is not too difficult to use, but a few minutes of hand holding will save you time. PM me if you've got problems using it.
There's been 64 downloads just this week. Somebody must be using it. Don't be shy, post your charts.
http://sourceforge.net/projects/voteanalyze/?_test=b

I've made and distributed a comprehensive tool for SC that shows dynamic charts and analysis for most of the key demographics. It's posted and linked all over these threads. I've posted the conclusion a dozen times: 2/3 of the phenomenon can be explained by county-level demographics in SC (remainder by precinct size). However, no one has attempted to perform any similar deep-dive type demographic analysis (except da32130).

The next step should be focused on finding a state that has precinct level demographics available. Precinct "demographics" should/could include Income/wealth, average age, % female, % white, % mormon, % GOP vote in 08, pop density, dollars donated per candidate, campaign visit (yes/no), and anything else that might explain the phenomenon.

*edit* Here you go. The only posted comprehensive Demographic analysis thus far: http://dl.dropbox.com/u/63868969/RPF/SC_2012_Primary_Analysis.xlsx

Note that the Exhibit_Pivot worksheet will allow you to create pretty much any demographic breaks you want. Very easy to use to see how income, age, etc creates some of the phenomenon in SC.

 

[affa]

I've made and distributed a comprehensive tool for SC that shows dynamic charts and analysis for most of the key demographics. It's posted and linked all over these threads. I've posted the conclusion a dozen times: 2/3 of the phenomenon can be explained by county-level demographics in SC (remainder by precinct size). However, no one has attempted to perform any similar deep-dive type demographic analysis (except da32130).

The next step should be focused on finding a state that has precinct level demographics available. Precinct "demographics" should/could include Income/wealth, average age, % female, % white, % mormon, % GOP vote in 08, pop density, dollars donated per candidate, campaign visit (yes/no), and anything else that might explain the phenomenon.

*edit* Here you go. The only posted comprehensive Demographic analysis thus far: http://dl.dropbox.com/u/63868969/RPF/SC_2012_Primary_Analysis.xlsx

Note that the Exhibit_Pivot worksheet will allow you to create pretty much any demographic breaks you want. Very easy to use to see how income, age, etc creates some of the phenomenon in SC.

A couple questions/comments:

1) I don't understand how you can evaluate precincts based on 'county level' demographics. I know from other posts you concede it's unfortunate you don't have better data, but I simply don't understand how you can even begin to evaluate a precinct level phenomenon with county level data to begin with. Just like you can't figure out anything about a specific state with country level demographics - that is, can you establish anything of value regarding Alaskan voting habits if I only hand you overall United States demographics? Perhaps I'm just misunderstanding you, so please clarify... because I view the county/precinct divide as similar, just on a different scale, given how large some counties are.

2) If there IS flipping going on, any factor the flippers decide to flip on could very easily correlate with demographics. Correct? I mean, it's not hard to believe that areas with large voter turnout, say, are different demographics than areas with low voter turnout. So, for example, if the flippers flipped on voter turnout (which I'm not saying they do), any flip would be somewhat 'hidden' within the demographics that define 'high voter turnout' areas. The same is potentially true of -any- factor the flippers flip on, including precinct total vote tally. This doesn't negate your study, I'm just pointing out why this is so difficult to both prove and disprove.

3) The reason many of us haven't dug into 'precinct level' demographics is because so far, I, at least, haven't found them. I've scoured on zip codes, housing/realty sites, and other general indicators, but none so far divvy up any area (understandably) into the political precincts. I've actually done hours upon hours of examining previous elections at the precinct level to try to eek out demographic data, and come to the conclusion that I don't think 2008 is even remotely representative of common/established voting habits of a precinct (or even county, really). This is looking at both precinct votes in primaries, general elections, senate, governor, and house of rep for precincts pre AND post 2008). And that's assuming 2008 data is pure, an assumption I'm not sure is true in the first place.

 

[jjockers]

A couple questions:
1) I don't understand how you can evaluate precincts based on 'county level' demographics. I know from other posts you concede it's unfortunate you don't have better data, but I simply don't understand how you can even begin to evaluate a precinct level phenomenon with county level data to begin with. Just like you can't figure out anything about a specific state with country level demographics.
2) If there IS flipping going on, any factor the flippers decide to flip on could very easily correlate with demographics. Correct? I mean, it's not hard to believe that areas with large voter turnout, say, are different demographics than areas with low voter turnout. So, for example, if the flippers flipped on voter turnout (which I'm not saying they do), any flip would be somewhat 'hidden' within the demographics that define 'high voter turnout' areas. The same is potentially true of -any- factor the flippers flip on, including precinct total vote tally. This doesn't negate your study, I'm just pointing out why this is so difficult to both prove and disprove.
3) The reason many of us haven't dug into 'precinct level' demographics is because so far, I, at least, haven't found them. I've scoured on zip codes, housing/realty sites, and other general indicators, but none so far divvy up any area (understandably) into the political precincts.

1) - it's the best data we have, and it actually works decently to explain the phenomenon in SC. It's not a bad approximation, just not ideal. How it works: precincts are assigned their county demographics. Median Income, % White, % Female and % Over 65 at the county level are decent predictors of Romney vote % in a precinct. Even precinct level data is an approximation -- what we really want is demographic info for only those that voted
2) - Yes, I agree, that could be the case. We've discussed this earlier. The regression analysis is essential information, regardless. It's the first thing an outsider would request. Disregarding demographics on the basis of the point you make is a mistake, as others will then disregard the analysis in its entirety. Once the key factors are identified, then it would be time to question whether the proposed flipper depends on these factors.
3) Same here, plus time constraints. I'm best suited analyzing data, not collecting it. I've got to believe that some state has detailed precinct demographic info, and with the manpower we have investigating this phenomenon, someone will volunteer to take on the task. Is it possible that the campaign has voter info at the precinct level that could be of some help?

 

[Liberty1789]

The four leftmost curves have colors in the chart that appear to be Robertson 88, Duke 92 Buchanan 92, and Buchanan 96.

Likewise the "winners" on the right side are Bush 92 and Dole 96, both establishment candidates.

Please confirm that I've got the colors right.

100% right!

 

[RonRules]

2) If there IS flipping going on, any factor the flippers decide to flip on could very easily correlate with demographics. Correct? I mean, it's not hard to believe that areas with large voter turnout, say, are different demographics than areas with low voter turnout. So, for example, if the flippers flipped on voter turnout (which I'm not saying they do), any flip would be somewhat 'hidden' within the demographics that define 'high voter turnout' areas. The same is potentially true of -any- factor the flippers flip on, including precinct total vote tally. This doesn't negate your study, I'm just pointing out why this is so difficult to both prove and disprove.

Point of information!

I am a flipper, which means that I believe that votes are flipped solely in favor of Romney in 2012 through some fraudulent method that we have not yet discovered.

The people you refer to in the above paragraph are "anti-flippers". If people believe that demographics, large voter turnouts, religion, etc., are the probable cause for the observed effect, then we should refer to them as "anti-flippers" not "flippers".

This is confusing enough, we don't want anybody to flip out.

May I suggest that we call them "anti-flippers" or "aflippers", much like we have Theists and Atheists.

If you don't like that, we could call them "flip deniers" or "flipniers".

"jjockers" appears to be a hybrid flipper. He believes that demographics explains part of the flip. Maybe he could be described as a "part-flipper", "Agnostic flipper", or "Agflipper"

Other possibilities for those that don't believe that flipping happens though fraud could be termed as: "contraflippers", "dissaflippers", "negaflippers" (maybe not this one), "refutaflippers", "rejectaflipppers" or the lame "nonflippers"

This is important guys/gals. We're soon going to make the news in a big way and these words will become part of the American Lexicon.

 

[affa]

Point of information!

I am a flipper which means that I believe that votes are flipped solely in favor of Romney through some fraudulent method that we have not yet discovered.

The people you refer to in the above paragraph are "anti-flippers". If people believe that demographics, large voter turnouts, religion, etc., are the probable cause for the observed effect, then we should refer to them as "anti-flippers" not "flippers".

This is confusing enough, we don't want anybody to flip out.

May I suggest that we call them "anti-flippers" "aflippers", much like we have Theists and Atheists.

If you don't like that, we could call them "flip deniers" or "flipniers".

"jjockers" appears to be a hybrid flipper. He believes that demographics explains part of the flip. Maybe he could be described as a "part-flipper", "Agnostic flipper", or "Agflipper"

Other possibilities are "contraflipper", "dissaflipper", "negaflipper" (maybe not this one), "refutaflipper", "rejectaflippper" or the lame "nonflipper"

This is important guys/gals. We're soon going to make the news in a big way these words will become part of the American Lexicon.

When I use the term 'flipper', I'm speaking about the person committing the flip. The flipper flipped votes. Provide a different word for me to use, and I'll give it a whirl.

 

[Liberty1789]

If nobody reading this thinks I have a valid point in any of what I've posted then obviously I'm wasting my time and everyone else's, and I regret that.

You are not wasting anyone's time. I see your sober mathematical feedback/judgement as one of the best we've had so far. Some of your points really make me think long and hard, for which I am very grateful.

 

[40oz]

When I use the term 'flipper', I'm speaking about the person committing the flip. The flipper flipped votes. Provide a different word for me to use, and I'll give it a whirl.

trusters and fencers?

 

[RonRules]

When I use the term 'flipper', I'm speaking about the person committing the flip. The flipper flipped votes. Provide a different word for me to use, and I'll give it a whirl.

I think you should refer to those people as "regressionists".

Use in a sentence: "The regressionist flipped Romney .3%, as a function of Grecian Formula usage in the precinct."

There's a Wiki for everything:
http://en.wikipedia.org/wiki/Grecian_Formula

Hey, check this out:
"The product is 98% clear liquid, 0.29-0.34% lead acetate and a small amount of sulfur."

No wonder Santorum and Romney seem so retarded. It's the lead poisoning stupid!

Gingrich is pretty smart. Ron's the smartest and you won't see a drop of that stuff used on his beautiful grey hair.

 

[Scott_in_PA]

Post #133 -> #137 LOL

Maybe this info will help.

http://www2.census.gov/census_2010/04-Summary_File_1/
http://www2.census.gov/census_2010/05-Summary_File_2/

I'm lurking and learning.

 

[JVParkour]

I'll be a pancakeflipper. The correlation to # of waffle houses in a precinct is clearly the determining factor.

Seriously though, this is interesting, and regardless of the end result it makes me realize that I should brush up on my stats!

 

[dr.k.research]

David Terr, Raytheon, and Algorithmic Flipping

Note: this was written in 2008, in an attempt to pursue election fraud against Dr. Paul. Author: C. Ingram, RE: an indivdual named David Terr:

Wolfram Research is a firm which makes its own specialized software, known as Mathematica. This software is used for cryptography. Terr’s specialty revolves around the development of algorithms for various mathematical calculations.

Raytheon is a monstrous military company. Depending on government contracts for income this company lobbies extensively for an active military. It is the company’s goal to create or maintain wars. With Dr. Paul’s plan this company would no longer be subsidized. Thus, it is attempting to undermine the doctor’s election bid. It can be no coincidence that the Raytheon-sponsored polling site, USAElectionpolls.com, gives its own political endorsement, while disparaging Dr. Paul. By electing these warmongers, Raytheon knows, there will be perpetual war and, therefore, endless profits. Thus, using its own proprietary mathematical software, Raytheon and its agents creates its own versions of polls to defraud voters and their bid for a Paul presidency.

Full material:

http://healthfreedom2012.com/HFblog/wp-admin/post.php?post=205&action=edit