The case for the occurence of algorithmic vote flipping

Does anyone know why the original blog was taken down?

I don't know why, but I didn't really like the article anyway. For example, they explain the X-Axis as being time-based. Also there is no talk about all the work we did to prove that the slopes are not caused by demographics.

The Judge in Orange county has a better document in his hands right now! :)
 
Several people have emailed me asking for what data is needed for a thorough analysis.

Here's what's needed from your County election clerk:

Absentee (mail-in voting):
1) How many ballots were sent out (preferably for each precinct)?
2) How many ballots came back and were counted again for each precinct? (That to find cases of "Vote Plucking")
3) What equipment is used to scan the Absentee ballots? (Sometimes the equipment is different and so are the results)
4) What is the software serial number for each machine?
5) When was the software updated?
6) Who did the work on the machines (updating the software, installing ballot configuration files)?
7) What are the precinct-level results for Absentee voting?
8) Do they have Zero-Count poll tapes of some proof of Zero Count?
9) Do they have signed poll tapes? I need them (copies if nothing else)

Early voting:
1) What equipment is used for early voting?
2) What is the software serial number for each machine?
3) When was the software updated?
4) Who did the work on the machines?
7) What are the precinct-level results for Early Voting?
8) Do they have Zero-Count poll tapes of some proof of Zero Count for Early Voting?
9) Do they have signed poll tapes for Early Voting results? I need them (copies)

Election day voting:
1) What equipment is used for the election day voting (including paper ballots if used)
2) What is the software serial number for each machine?
3) When was the software updated?
4) Who did the work on the machines?
7) What are the precinct-level results for each candidate in Election Day Voting?
8) Do they have Zero-Count poll tapes of some proof of Zero Count for Election Day Voting?
9) Do they have signed poll tapes for Election Day Voting?

Provisional ballots:
1) How many for each precinct?
2) How many were rejected for each precinct?

Other things needed:

a) Central Tabulator
1) What brand?
2) What software version?
3) Is the software version the exact same as what the State approved?
4) When was it updated?
5) Who did the work on the central tabulator?
6) The raw data BEFORE it enters the central tabulator?
7) A directory listing of all the files (with all file attributes) on the central tabulator
8) Is the Central Tabulator connected to a network of any kind? (perform netstat command)
9) How do people enter data in the Central Tabulaor? USB, DVD, PCMCIA cards?
10) Are there procedures to prevent virus infection of the Central Tabulator?
11) Is there an Anti-Virus program installed on the Central Tabulator?
12) Is there any other software on the Central Tabulator that can perform any kind of vote processing? (Shelby Co TN is one example)

b) What are the results of the 1% recount if performed?


c) How was the election data transmitted?:
1) By phone call from the precinct?
2) By e-mail?
3) By hand carried memory cards or USB cards?
4) By hand carried paper ballots?
5) By over airwaves like WiFi?

d) Have any candidates filed for re-count requests?


Please provide all data in electronic format, not scanned images or paper. I can't re-type millions of numbers. PM me if you can help.

The county is legally obliged to provide ALL that data. You may however have to pay a small fee like $35, which I had to pay in Riverside. In other counties, there was a fee to make photocopies of the data, but that's not too useful anyway because nobody wants to type in tons of numbers.

If there's fraud and you get me ALL that, I guarantee you I will find the fraud.

Thank you in advance,

RR
 
New report on election integrity for 2012:

Counting Votes 2012: A State by State Look at Election Preparedness
http://countingvotes.org/

Executive summary:
http://countingvotes.org/sites/default/files/countingvotes2012execsummary.pdf

Excerpt:

On Election Day, Nov. 6, the stakes will be high. A number of critical races will be very close, and some might be decided by very few votes. At the same time, it is highly likely that voting systems will fail in multiple places across the country.i In fact, in every national election in the past decade, computerized voting systems have failed – machines haven’t started, machines have failed in the middle of voting,ii memory cards couldn’t be read be read,iii votes were mistalliediv or lost.v
Our elections are so complex, with so many different jurisdictions and varying technologies, that problems are inevitable. And, as the technology used for elections has become more complicated, the opportunity for error has substantially increased.
This report reviews how prepared each state is to ensure that every eligible voter can vote, and that every vote is counted as cast. Because we cannot predict where machines will fail during the upcoming national election, every state should be as prepared as possible for system failures.
The Verified Voting Foundation, the Rutgers Law School Constitutional Litigation Clinic and Common Cause surveyed states’ voting equipment and ranked the states according to their preparedness. The rankings are based on how states compare to a set of best practices already being used in some places.
The report ranks states from worst to best (inadequate, needs improvement, generally good, good and excellent) in these five areas of evaluation:
1) Does the state require paper ballots or records of every state? When computer failures or human errors cause machines to miscount, election officials can use the original ballots to determine correct totals. Additionally, paper ballots or records can be used to audit machine counts to determine if outcomes are correct.
2) Does the state have adequate contingency plans at each polling place in the event of machine failure? Machine repair should occur quickly and emergency paper ballots should be made available.
3) Does the state protect military and overseas voters by ensuring that marked ballots are not cast online? Voting system experts at the National Institute of Standards and Technology and cyber security experts at the Department of Homeland Security warn that even state-of-the-art online voting technology lacks adequate security and privacy protections. Ballots cast over the Internet can be subject to alteration and voters may lose the right to a secret ballot.
4) Has the state instituted a post-election audit that can determine whether the electronically reported outcomes are correct?

Listed below are examples of past machine failures and how they impacted various elections:
 Following a June 2009 election, officials in Pennington County, South Dakota, discovered a software malfunction that added thousands of non-existent votes to the county totals.vi
 In a municipal election in Palm Beach County, Florida, in March 2012, a problem with election management software allotted votes to the wrong candidate and the wrong contest. The official results were only changed after a court-sanctioned public hand count of the votes.vii
 In the 2008 Republican presidential primary in Horry County, South Carolina, touch screen voting machines in 80 percent of the precincts temporarily failed, and when precincts ran out of paper ballots, voters could not cast ballots in their home precinct.
 In a test-run for an online election in the September 2010 Washington, D.C., primary, a hacker team was able to change all of the votes to “elect” their own candidates. The online voting system was days away from being launched in a real election for use by overseas and military voters. After the incident, the Internet voting system was canceled.viii
Similar vote-counting errors may go undetected during the 2012 elections unless the mistake is so large and obvious – like the software malfunction in South Dakota – that it can’t be ignored, or the state has adopted procedures – like the post-election audit done in Florida – as recommended in this report.
 
statechartmap.jpg
 
I'd like your opinion about this analysis methodology:

I read this Russian election fraud paper. It is recent and published this year.

http://samarcandanalytics.com/elect...results for the 2011 Russian elections v2.pdf

They have a pretty incriminating chart here, which relates voter turnout with candidate (party) success.
RussianElectionFVoterTurnout.png

"This plot shows a large number of polling stations which have near perfect turnout. Of ninety-five thousand polling stations, almost seven thousand reported a turnout above 97%. The support for United Russia has a very strong dependence on turnout, while a simple model would predict that these two should be largely independent. Similarly, the support for the Communist Party declines proportionally with increasing turnout ."

This method is now referred to as the "Shpilkin" method, named after one of the three authors on the paper.

So, I decided to see how our Golden Boy Mitt Romney does with these types of charts, specifically in Orange County, CA; That's where the trial is! :)

I charted individual points for each candidate's results in each precinct as a function of turnout %.
2012_CA_OrangeCountyPresPrimariesRepubCertFVoterTurnoutCast.png


Discovery 1)
The spreadsheet shows a strong slope linear least-squares fit benefiting Romney. I did not get what I expected. My rationale was that if a candidate gets his voters excited about voting (like Ron Paul), you would expect that the greater the turnout %, the higher his own percentage as a function of turnout %. (Remember I'm just referring to his portion of the turnout %) Again the X-Axis is turnout % and not precinct size. A specific precinct size could be anywhere on the chart. You could have a precinct with 5 votes, but a high turnout 100 %, 5 / 5 for example and it will be at the rightmost point on the chart.

Based on the general enthusiasm for Ron Paul I expected to see a positive slope for Paul (but the same overall low results that he got in Orange County, CA).

Instead I got a very strong positive slope for Romney. I am wondering what would drive that. It appears like ballot stuffing in the higher turnout precincts.

Discovery 2)
Normally a certain number of vote are "cast", but a lesser number of votes are actually "counted" because some ballots are rejected. That's normal and fine.
Have a look at this chart, which uses the actual votes counted (as opposed to votes cast):
2012_CA_OrangeCountyPresPrimariesRepubCertFVoterTurnoutCounted.png


Note that the slope is higher (0.9462) in the second chart, benefiting Romney and affecting Ron Paul. Romney has a higher slope when you chart the "counted" ballots. In other words, the more ballots are thrown out, the better Romney does. This could indicate that ballots for other candidates were possibly rendered invalid and thrown out. That activity is more effective in larger precincts.

Also, Orange county rejected an astounding 11,972 votes out of 234,396 cast. I don't know if that's normal, but it seems like a lot.

Can you repeat this type of analysis in your state/county to confirm my findings here?

Comments?
 
Last edited:
It sounds like the same folks are fixing both elections.


I'd like your opinion about this analysis methodology:

I read this Russian election fraud paper. It is recent and published this year.

http://samarcandanalytics.com/elect...results for the 2011 Russian elections v2.pdf

They have a pretty incriminating chart here, which relates voter turnout with candidate (party) success.
RussianElectionFVoterTurnout.png

"This plot shows a large number of polling stations which have near perfect turnout. Of ninety-five thousand polling stations, almost seven thousand reported a turnout above 97%. The support for United Russia has a very strong dependence on turnout, while a simple model would predict that these two should be largely independent. Similarly, the support for the Communist Party declines proportionally with increasing turnout ."

This method is now referred to as the "Shpilkin" method, named after one of the three authors on the paper.

So, I decided to see how our Golden Boy Mitt Romney does with these types of charts, specifically in Orange County, CA; That's where the trial is! :)

I charted individual points for each candidate's results in each precinct as a function of turnout %.
2012_CA_OrangeCountyPresPrimariesRepubCertFVoterTurnoutCast.png


Discovery 1)
The spreadsheet shows a strong slope linear least-squares fit benefiting Romney. I did not get what I expected. My rationale was that if a candidate gets his voters excited about voting (like Ron Paul), you would expect that the greater the turnout %, the higher his own percentage as a function of turnout %. (Remember I'm just referring to his portion of the turnout %) Again the X-Axis is turnout % and not precinct size. A specific precinct size could be anywhere on the chart. You could have a precinct with 5 votes, but a high turnout 100 %, 5 / 5 for example and it will be at the rightmost point on the chart.

Based on the general enthusiasm for Ron Paul I expected to see a positive slope for Paul (but the same overall low results that he got in Orange County, CA).

Instead I got a very strong positive slope for Romney. I am wondering what would drive that. It appears like ballot stuffing in the higher turnout precincts.

Discovery 2)
Normally a certain number of vote are "cast", but a lesser number of votes are actually "counted" because some ballots are rejected. That's normal and fine.
Have a look at this chart, which uses the actual votes counted (as opposed to votes cast):
2012_CA_OrangeCountyPresPrimariesRepubCertFVoterTurnoutCounted.png


Note that the slope is higher (0.9462) in the second chart, benefiting Romney and affecting Ron Paul. Romney has a higher slope when you chart the "counted" ballots. In other words, the more ballots are thrown out, the better Romney does. This could indicate that ballots for other candidates were possibly rendered invalid and thrown out. That activity is more effective in larger precincts.

Also, Orange county rejected an astounding 11,972 votes out of 234,396 cast. I don't know if that's normal, but it seems like a lot.

Can you repeat this type of analysis in your state/county to confirm my findings here?

Comments?
 
RonRules, you have exceeded your message capacity. I was asking if a primary vote is sufficient for your analysis. We just had our primary yesterday and I was going to contact my county clerk to obtain the information you specified earlier in the thread.
 
RonRules, you have exceeded your message capacity. I was asking if a primary vote is sufficient for your analysis. We just had our primary yesterday and I was going to contact my county clerk to obtain the information you specified earlier in the thread.

Before you see your county clerk, let's see if you've got a flipper first. If you are referring to Bentivolio race, that was definitely a flipper, even if we won.

In order to analyze, I need precinct-level data with each candidate in columns. I'll give you a chart and detailed statistics.

Here's Bentivolio's chart (most of the results in):
2012_MI_11thCongressionalDistrictcsv-2.png


Here are the statistics:

Code:
Correlations:		
State	kerry_bentivolio	nancy_cassis_(writein)
2012_MI_11thCongressionalDistrict.csv Correlation	-0.680934125	0.680934084
		
Standard Deviations:		
State	kerry_bentivolio	nancy_cassis_(writein)
2012_MI_11thCongressionalDistrict.csv Std Deviation	0.028625278	0.028625279
		
Percentage gained/lost:		
State	kerry_bentivolio	nancy_cassis_(writein)
2012_MI_11thCongressionalDistrict.csv Percentage Gained	-6.717169285	6.717172265
AllStates		
		
Votes won by each candidate		
State	kerry_bentivolio	nancy_cassis_(writein)
2012_MI_11thCongressionalDistrict.csv Votes Earned	20663	9646
		
Votes gained/lost:		
State	kerry_bentivolio	nancy_cassis_(writein)
2012_MI_11thCongressionalDistrict.csv Votes Gained	-2036	2036
 
Last edited:
I'm in Missouri. Where can obtain precinct level data? I assume I have to contact the election authority in my county

http://election.sccmo.org/election/index.php?option=com_content&view=article&id=27

The link just shows the final results. As you know I can't chart that.

Sometimes, the Secretary of State has the precinct-level results, but each county definitely does have it. Make sure they don't give you PDF images that would need to get retyped. A PDF file that is text is OK, provided it's set up as columsn.

Because I'm so busy with a technical paper, I can't collect the data right now.

First start with the Secretary of State. They may have the precinct-level data. If not, you have to check each county's website and download it. If it's not there call them and they will email it to you.

Put everything in one spreadsheet like this:
Code:
Precinct	Kerry_Bentivolio	Nancy_Cassis_(WriteIn)
Commerce 1	96	44
Commerce 2	113	48
Commerce 3	115	61
Commerce 4	115	54
Commerce 5	159	88
Commerce 6	94	40
Commerce 7	172	109
Commerce 8	157	80
Commerce 9	168	66
Commerce 10	121	54
Commerce 11	96	55
Commerce 12	56	36
Commerce 13	100	71
Commerce 14	83	24
Commerce 15	101	63
Highland 1	292	158
Highland 2	255	93
Highland 3	140	90
Highland 4	269	138
Highland 5	214	128
Highland 6	261	145
Highland 7	123	48
Highland 8	178	67

The columns did not come out right in the post. Make sure the votes are underneath the candidate name.
 
Last edited:
Just for grins I thought I'd check Ran Paul's election in 2010:

Here's the Primary:
2010_KY_EntireStateUSSenatorPrimaryElectioncsv.png


Here's the General:
2010_KY_EntireStateUSSenatorGeneralElectioncsv.png


The general's last twitch on the curve is because Rand had somewhat different results in the two larges counties (Fayette and Jefferson). It could be flipping, but it's inconclusive.
 
Last edited:
I realize this is an old topic but it is one that gnaws at me.

The following is a post I made for another thread but with the addition of the following chart of the Republican Party Flipping. I'm thinking the central tabulator in this case may have a name. I am curious as to what they have to say for themselves.


MattMcdonald_RepPartyFlip.jpg





If you take the Matt McDonald video as a vote count representation...


at about 3:20 or so the vote was...


8 for Ron Paul

7 for Santorum

5 Romney

2 Undecided

After the head cheeses had their way with the tally it somehow equaled;

2 for Ron Paul

9 for Romney

5 for Santorum


So you tell me how many people really voted for Ron Paul?

How many would have if the party wasn't kicking its own contestants XX XXX XXXX right out of the starting gate?

moleman.jpg



 
Back
Top