Of coincidences and patterns: The evidence for fraud

August 18, 2004

There is a need to clarify concepts at this time. There is a lot of confusion about statistics, CNE results, coincidences, patterns and what it is that the opposition via the Coordinadora Democrática (CD) is claiming as evidence of fraud. So, I thought I would review the basic facts:

Each voting center had “mesas” or tables. Each mesa consists of volunteer “members” and witnesses, all chosen at random from voter rolls. The members handle the logistics, the witnesses are supposed to be checking that things are done properly and correctly. Both sides had witnesses and members in most tables. There were 8,142 mesas or tables in the country, some did not have the Smartmatic electronic voting machines.


Each voting center  had from one to five tables and each table could have up to three Smartmatic machines (1) to vote in. In each Smartmatic machine, people voted. Ther are two buttons with the possible answers “Yes” and “No” (2) and the voter has to select the answer of his/her choice by touching the correct button. To finish the voter confir her/his choice touching the Vote button (3) on the screen. When the voter touches the “Vote” button, the machine prints a physical ballot (paper ballot) allowing the voter to instantly see if his/her selection was recorded properly. The paper ballot must be deposited into the ballot (4)box , so that poll workers can return the ID card to the voter. The poll workers testify and guarantee that the ballot was introduced in the ballot box.



Sebastian from zonageek has actually downloaded all of the preliminary data from the CNE site and placed it in a .dat file here as an Excel file. Here you can read state by state, school by school, how many tables there were, the number of voters and the number of Si and No votes. However, this table by the CNE does not show the details of the machine vote individually.


At the end of the voting day, after polls closed, the people at each table would gather around, put the ballot box in a plastic bag with a precinct that all would sign and proceed to read how many Si and No votes there were in the Smartmatic machine readout. The Smartmatic machine would print a cover sheet (Acta) that all members and witnesses of the table would sign as confirmation that they agreed with their content. They would all take a copy, one Acta per Smartmatic machine as in the following picture, where I show three Actas from a single mesa, i.e. three Smartmatic machines:



This picture shows one of the cases of “coincidences” what is now being called a “pattern” or irregularity. In this particular case, the number of Si votes is exactly the same in all of the machines at a single table or mesa. As I described last night, such a pattern was first found in Bolivar state by J.J. Rendon, who said on TV he had detected already 24 centers with a problem like that out of the 60 voting centers in Bolivar state, a fairly large proportion.


The case in the picture is an extreme case in which all three numbers for the “Si” vote agree. In most cases, out of three Smartmatic machines, two have the same number (x) of Si votes and the third one a different one (y), but this third number is always less or equal (y<, y=x) than the one the other two had. But the third number (y) is NEVER larger (y>x) than the two identical ones.


Let’s look at what is so improbable about this:


 I) First of all, it is highly improbable that so many coincidences occur. Even if the number of voters were divided evenly between machines, abstention and different voter’s intention would vary the numbers from machine to machine sufficiently that this would be a rare occurrence.


II) If it were a matter of even distribution of voters between machines, then the percentage of equal NO votes in two machines in the same center should be equal to the coincidence of the Yes votes in two machines. No such occurrence has been reported and in all examples I have seen (Something like 15-20 of them) not once did the number of NO’s repeat.


III) But perhaps the most powerful argument is that if this were a matter of probabilities, the third number would be either higher or lower then the two numbers that coincide, much like tossing a coin, half the time above, half the time below. But as far as I know ALL of the cases that have been reported correspond to the third number being less or equal to the two identical ones.


How many cases are there?


Well, this is hard to answer at this time. J.J. Rendon has said 24 of the 60 centers in Bolivar state have the pattern. Enrique Mendoza said that he already had 500 cases documented. Nelson Rampersad said that  25% of all Smartmatic machines, more than 5,000 have irregularities, but he did not specify what these irregularities were. The Governor of Zulia state said that 33% of that state’s 931 tables had irregularities, but once again, he did not clarify if they were these type of coincidence although he did say he had consulted mathematicians and technicians who had told him this was essentially impossible. This suggests he is talking about the same pattern. This alone could be around 1,000 Smartmatic machines in 300 centers.


These are HUGE numbers that have very little probability of occurring, least of all, if they all have the same pattern in which the third number is smaller. I have not been able to confirm this.


What explains the coincidences?


Well, the theory is that they wanted to have the total number of votes be consistent with the final number of voters in each machine. Thus, the machines were somehow programmed with a table, machine by machine, center by center, so that above a certain number any additional Si vote would be assigned to the No’s. Then there were four possibilities: No machine exceeded the number, one machine exceeded the number, two machines exceeded the number or all three machines exceeded the number. Only the last two would be detectable.  Since there are also mesas or tables with one or two machines, these two would also show no significant pattern. Thus, the numbers being mentioned are quite significant and if true, likely to be present in 50% of all machines.


Can it be something else?


Sure it can, some voting machines have been released to the public only to make mistakes the first time they were tested. These machines were never tested before under field conditions, heavy voting and/or long hours of voting.


However, voting machines can be rigged and have a history of being rigged. As this article attests, e-voting is still not tamper proof and manufacturers seem to be after the money and have not made their machines full proof yet. Given the fact that the software was handed out to the CNE, that the machines were programmed and given to the CNE, who in turn gave it to the military, many things could have happened after they were first programmed by the company that made them.


However, one thing is clear, in either case; a proper audit of the printed ballots would show the problems. The CD refused to accept the audit because it felt it should have the right to choose some ballot boxes. This for two reasons, the CNE decided on the algorithm to choose the boxes, this could be as rigged as the original voting machines. The second is that the CD knows where the biggest differences between exit polls and final results are, it could pick those as a testing ground of this possible massive fraud. Thus, if the CNE did not allow for the CD choosing some boxes, The CD could not participate in a process that may become the final nail of its own coffin, unless its conditions were met.


Can the boxes be rigged?


Yes and No. It would be difficult to rig all of them, some 15,000 to 20,000 boxes with 10 million little papers in the correct proportions to match the results in each box. Then they would have to fake the signatures in each precinct. Seems hard. In fact, you could even check for fingerprints in the ballots. If real, they should all have different fingerprints, if faked, pick three at random and the fraud will be revealed.




My own personal opinion is there was fraud and if the CNE people knew about it, the sample is rigged and nothing will be detected in the audit. If none of the boxes are from the cases with coincidences it would generate new doubts about the whole process. If they did not know, tomorrow in the audit, the whole fraud will come out and everything will implode. Beyond that we will be in unchartered territory.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: