I’ve been writing about this since last August. Today, my article on New York’s voter roll algorithms has been published in the peer-reviewed Journal of Information Warfare. Titled “The Caesar Cipher and Stacking the Deck in New York State Voter Rolls”, the article describes in great detail multiple well-hidden algorithms that I found in New York’s voter rolls last April.
After discovering the algorithm, I initially hesitated to put much time or resources into studying it. This is because it was clearly very complex and would take a very long time to unravel. I worried I could spend a great deal of effort and end up with nothing. In an age of rising gas and food prices, I didn’t want to spend the money and time away from my primary career chasing what might turn out to be a mirage.
Go Ad-Free, Get Exclusive Shows and Content, Go Premium Today - $1 Trial
Initially, I discovered one algorithm. Today, I call it the “Spiral” because of its characteristics. The Spiral is used to connect county voter ID (CID) numbers to State Board of Elections Identification (SBOEID) numbers in a unique way, effectively creating a third voter ID number that is both hidden from normal view and accessible to those with knowledge of the algorithm.
If a third ID number was needed for some reason, and it was thought-wise to connect the CID and SBOEID numbers to do it, the easiest way would be to simply append one to the other. For instance, if your SBOEID number is 99,967,345 and your CID number is M321,543, then the combined number is 99967345M321543. A number like that would be in the open and accessible to anyone who needed to use it.
Instead, the Spiral algorithm performs many operations to disguise the relationship between the numbers. In the example given above, the new number, which I call the “Algorithm ID” or AID, would be something like 34,674. No one looking at the SBOEID or CID numbers attached to the AID would have any way of knowing what the AID is without knowledge of the algorithm.
What is the AID? I am calling it an ID number because that is what it looks like to me. However, I would need more information to know for certain that is what it is. The AID is a serial number that represents all of the transformations performed by the Spiral algorithm. It allows, in short, the Spiral’s manipulations to be fully reversed. I added the number to the file so that I wouldn’t lose track of what the algorithm was doing. Then I realized that there was a distinct relationship between the type of numbers used by the Spiral and fixed values baked into the AID. That made it look like the AID had to be the numbers I assigned, and no others. If they were different in any way, they would be inconsistent with the Spiral. Therefore, it is plausible that creation of the AID is at least part of the intended purpose of the Spiral.
How does the Spiral work? I discuss this in more detail on my SubStack (here is a sample) but the simple version is this:
- The number space used by SBOEID numbers, comprised of the numbers 1-99,999,999 is partitioned into three bands. There is no way to discover the boundaries of these bands without exhaustive manual effort
- Within the three partitions, the one in the middle, “In-Range” is subdivided into 67 partitions. One for each of New York’s 62 counties, and 5 “buffers”. There is no easy way to discover the boundaries of these partitions without exhaustive manual effort.
- A decimal point is added to the left of all CID numbers, after separating any alpha components (like the letters C, M, and N). After the decimal point is added, the numbers are sorted in ascending order. The effect of this is to totally scramble the order of the CID numbers. This is because the number 55 comes before 56 but the number 550 comes after 56. When decimalized, 550 becomes .550, and 56 is .560, thus reversing their natural order.
- The Spiral performs a very complex series of mathematical operations on the list of in-range numbers for each county. You can read the Journal of Information Warfare article or my SubStack for details.
- The sorted list of decimalized CID numbers is cut into strips. Each of those strips is cut again and the numbers at the beginning are moved to the end.
- The scrambled list of CID numbers is mated to the scrambled list of SBOEID numbers.
- The decimal point is deleted from all CID numbers, thus hiding the fact it was ever there.
The AID is based on the final position of numbers after the Spiral has done its work. This is no small amount of obfuscation.
The Spiral does not protect any personal voter data. All voter rolls are public, except for social security numbers and driver’s license numbers. Rather than encrypt those, they are removed from the publicly available data in their entirety. To protect the public data, the relationships between voter names, addresses, birthdates, and voter history (among other things) would have to be masked. None of these are masked. None of the relationships are obscured in any way. The SBOEID and CID numbers are not masked either. If they were, an SBOEID would look something like this: XXX,XXX,324, where the X’s stand-in for masked numbers.
The Spiral does not protect non-public information. It can’t. SSN and driver’s license data is likely available somewhere to people with authority to access it. That data, however, isn’t hidden or protected any better than it already is by adding a third ID to what is already attached to voter records.
Databases are usually optimized for quicker search and retrieval. A common method is called a “B-Tree”. A simple way to describe it is this: assume you want record number 47 out of 100 records. What is the fastest way to find it? In the old days, a search might have started with the number 1, then proceeded upward until it reached 47. In other words, 47 separate checks of the database to find the record. A B-Tree asks, “Is the record in the first half or the second half of the database? It is in the first 50 records, so the last 50 are now removed from consideration. Now it asks, is it in the first half or the second half of the remaining records? It is in the second half, so records 1-25 are eliminated. It continues like this until, after 5 steps, it has found the record. Now, this “tree” is stored in an index that makes it easier to find later.
The Spiral divides the database based on orders of magnitude. That makes the database far less efficient for searches if that is what the Spiral was intended to be used for. Think of it this way, the Spiral creates the following groups of numbers, 1-9, 10-99, 100-999, 1,000-99,999, and 100,000-999,999. That would be how the search space is divided. If your target number was 563,222, it is buried in the middle of 900,000 numbers. That’s as close as the Spiral gets you to the number you want. After that, you would have to use another method to get to the number you wanted.
The Spiral is unquestionably present in New York’s voter rolls. I have seen three versions of the database and it is in every one of them. A researcher in North Carolina contacted me to say she’d made a FOIL request and received the database to check my results. Using Schenectady County’s in-range numbers from a version of the database I’ve never seen, she was able to find the Spiral. The algorithm is there. That is not in question. It is also extremely well-hidden. That isn’t in question either.
The question is: why is it there? Or, what does it do? Originally, I thought it might be a flag for illegal records. I no longer believe that. At the moment, I favor the idea that the algorithm scrambles the records in such a way that the AID, or something like it, can be used to extract records of interest without attracting the attention of anyone who doesn’t know about the algorithm. That, by the way, includes every county commissioner I’ve spoken with. It was a surprise to all but one, who didn’t want to admit he was surprised (I think he was).
There are other algorithms in the rolls besides the Spiral. I call them Metronome, Tartan, and Shingle. I have found another in New Jersey that seems to do the same thing but in a different way.
The algorithms have meaning. I know this because I can use some algorithms to predict voter status, cloned records, and deleted SBOEID numbers. None of that should be possible based on ID numbers but it is.
The opinions expressed by contributors and/or content partners are their own and do not necessarily reflect the views of Red Voice Media. Contact us for guidelines on submitting your own commentary.
Join the Discussion
COMMENTS POLICY: We have no tolerance for messages of violence, racism, vulgarity, obscenity or other such discourteous behavior. Thank you for contributing to a respectful and useful online dialogue.
We need to go back to old voting machines that added votes mechanically. Votes were entered by flipping a lever of the chosen candidate and the number increased like an old odometer. The numbers were recorded on the back of the machine and the final count was documented at the end of the night. There was no way it could be tampered with. When you exited the voter machine you had to open the curtain by pushing the main bar to the right and at the same time votes were calculated.
Any voting machines hooked up to the internet is nothing more than voter fraud just waiting to happen. It’s not who votes, it’s who counts the votes. I remember Pelosi telling Biden to stay in the basement and not worry about anything.
Electronic voting and counting machines need to be banned!
We only need one election day, in-person voting with few exceptions, VOTER ID, and numbered ballots according to registered number of voters for each district, TYPICALLY PICKED UP AT VOTING LOCATION, numbered ballots not associated with individual voters, or parties. With mandatory paper ballots hand counted locally and state-wide with video verification.
This is ominous news.