Back in May 2021 on Colorado’s western slope, MAGA Mesa County Clerk and Recorder Tina Peters, on the hunt for evidence of voter fraud in the 2020 presidential election, allegedly broke election security and allowed an outside party to copy Mesa County’s voting equipment hard drive. Multiple copies of the hard drive soon proliferated on the internet and fell into the hands of various amateur election integrity sleuths.
Florida software and database engineer Jeffrey O’Donnell was one of those amateur investigators. According to his bio, his 40-year tech career includes stints at Rockwell International, Westinghouse Electric Nuclear, and U.S. Steel. He now owns and operates a small software development company that specializes in network and data security.
O’Donnell claims no prior experience in computerized voting management systems, election integrity, statistics, or political science. Nevertheless, he’ll be the featured elections expert at Shasta County District 4 Supervisor Patrick Henry’s Election Machines Town Hall event this Thursday at the Holiday Inn at 6 p.m.
The event’s social media posts have undergone several revisions. To date, the flyer below appears the latest version that includes several rules, clarifications and caveats that were not present on the earlier flyers.
A life-long fan of the late Rush Limbaugh and a fervent Trump supporter, O’Donnell was programmed by Trump to reject the 2020 election results. He began communicating his doubts online with like-minded observers on Telegraph and other social media as The Lone Raccoon of the Apocalypse, or The Lone Raccoon for short, as 2021 began.
He was intrigued by My Pillow CEO Mike Lindell’s Election Fraud Cyber Symposium, held in Sioux City, South Dakota in August 2021. At Sioux City, Lindell alleged Chinese operatives used software algorithms and other computer gimmicks to flip Trump votes to Biden votes and steal the 2020 election. It sounded plausible to the computer engineer.
Not long after the symposium, as O’Donnell would later tell the story on Lindell’s Frank TV streaming service, a friend mentioned that copies of the Mesa County voting equipment were available online. O’Donnell downloaded a copy, which featured digital ballot images, batch numbers, case vote records, and other election data. He was surprised when he quickly found one anomaly in the 2020 presidential election data and a second identical anomaly in the 2021 municipal election.
For O’Donnell, the discovery of the Mesa County anomalies was the key to Lindell’s network of election fraud promulgators. O’Donnell soon found himself working with Texas A&M emeritus computer lecturer Walter Daugherity on a report for Peters’ legal team (among other legal issues, she faces trial this fall for a slew of felonies and misdemeanors related to the election systems breach). The brash Mesa County clerk was already a Lindell favorite and thanks to Report 3, as it has come to be known, O’Donnell would soon find himself guesting on Frank TV and in Lindell’s employ as a computer analyst.
According to Report 3, the unauthorized creation of a new election database occurred during both the 2020 presidential and 2021 municipal elections, followed by the reloading of thousands of ballots. The authors note that such changes could have been implemented by Mesa County election officials but claim multiple times throughout the document that they’ve thoroughly interviewed all of the officials involved and no one recalled creating the databases or reloading the ballots.
That leaves one of two options, the report insisted. Either the databases were triggered remotely via one of the electronic management system’s external data portals or they were algorithmically triggered by a process embedded within Dominion Voting System’s software. Report No. 3 concluded a crime had been committed and recommended de-certifying Mesa County’s election results.
But there’s a big problem with Report 3. The authors didn’t interview anyone from Mesa County as they had claimed. Not a soul.
As Republican Mesa County District Attorney Daniel Rubinstein discovered after a thorough investigation, the authors of Report 3 failed to interview any of Mesa County’s election officials and employees. Daugherity told investigators he didn’t do any of the interviews and that all of the interviews were done by his co-author O’Donnell. O’Donnell refused to cooperate with the investigation and referred Rubinstein to Peters’ attorney. He did not reply to an email from A News Cafe seeking comment.
“Despite repeated claims that there was extensive questioning of the Mesa County election officials, we were not able to locate a single person who said they were interviewed by the drafters of Report 3,” Rubinstein said in his report.
As Rubinstein’s investigation of Report 3 revealed, a Mesa County official had indeed created the new election databases, elections manager Sandra Brown (Brown has since been dismissed without cause). Here’s how it worked on both election nights.
In lay terms, as batches of ballots were scanned in by an optical tabulator, digital ballot images were fed to a separate bank of computers for adjudication by election judges. During both election nights in question, the adjudication screens froze and Brown used a procedure called “reset in progress batches” to unfreeze them. In both cases, the procedures failed, so Brown stopped the adjudication sessions and started new ones, which automatically created the new adjudication databases which were subsequently found by the authors of Report 3.
No remote triggers were required. No Chinese algorithms. Dominion Voting Systems said Brown’s first actions to use “reset in progress batches” were correct but she should have called customer service before stopping the adjudication sessions. Chalk it up to human error. No harm, no foul.
Rubinstein was blunt in his summary:
“This investigation is being closed with no finding of probable cause that a crime was committed by any person. There appear to be anomalies in the election logs, caused by the intentional actions of Sandra Brown, elections manager of the Mesa County Clerk’s Office. No evidence exists that Ms. Brown had any nefarious criminal motive in those actions, but rather appears to have been trouble-shooting problems in the flow of the adjudication process during the elections.
“These actions were verified to have been done by her through video evidence, corroboration of records, audit of randomly selected ballot images, interviews with witnesses and experts, and recreation of the certain scenarios using a test election environment and prove that the conclusions of Report 3 are incorrect claims. At this time, no evidence suggests that these actions negatively impacted the election.”
In the real world of report-writing, in academia, law, science, journalism, and other fields involving investigatory work, being exposed as a fabricator generally results in the end of your career. Make no mistake, the authors of Report 3 lied when they claimed they’d interviewed all the pertinent Mesa County elections personnel. The validity of their entire report hinges on that lie. They didn’t interview anybody, and Daugherity put the blame on O’Donnell.
That should have been the last we ever heard from Jeffrey O’Donnell on the topic of election integrity.
Instead, in the MAGAverse, O’Donnell has been rewarded. He’s made numerous appearances on Frank TV and other conspiracy outlets like Steve Bannon’s War Room. Last summer, Lindell invited O’Donnell to review the work of former CIA contractor and noted con-man Dennis Montgomery. Montgomery claims he helped develop a supercomputer called Hammer and software called Scorecard that can manipulate voting data. He maintains the technology has been hijacked by the Chinese.
Montgomery provided the source material for Lindell’s Election Fraud Cyber Symposium in 2021. The symposium blew up in Lindell’s face after a cyber forensics expert revealed the source material was mainly text documents filled with random gibberish, not packets of data that proved the Chinese rigged the 2020 election as promised by Lindell. Lindell was ordered to pay the forensic expert $5 million earlier this year for winning “The Prove Mike Wrong” contest.
That hasn’t stopped Lindell from continuing to claim he has the election fraud goods on China, and last summer O’Donnell privately certified the data given to Lindell by Montgomery.
“I have seen the unicorn,” O’Donnell said at Lindell’s Moment of Truth rally in Springfield Illinois last August. Presumably, he means he’s seen Hammer and Scorecard, although Lindell has yet to provide a single scrap of evidence to the public proving their existence.
On an episode of Frank TV last November O’Donnell was a little too honest about Lindell’s invite to review Montgomery’s material.
“What if I’m the guy who has to tell Mike Lindell that all of this stuff isn’t real?” O’Donnell recalled asking his wife. “I’m doomed.”
One thing’s for certain: Telling Lindell the truth isn’t going to get you invited back on Frank TV. Of course, he’s seen the unicorn! On The Lone Raccoon website, O’Donnell offers a detailed defense of Report 3 that baffles the reader with technical details and fails to mention the authors didn’t interview anybody for their report even though they claim they did.
Instead of acknowledging the errors he made in Report 3, O’Donnell has doubled down on them. He continues to ignore the fact that the anomalies he uncovered in Mesa County’s voting records were human errors and instead poses fictitious remote triggering devices and secret algorithms that have created a nationwide epidemic he calls the “Mesa pattern of election fraud.”
Last year, O’Donnell set up at least two websites dedicated to collecting crowd-sourced election data from volunteers,
Polaris Recount and Ordros Analytics. Both websites encourage users to request documents from their county election departments and post them online. Polaris Recount collects digital ballot images, which aren’t available in all jurisdictions, including California. Ordos Analytics offers instructions on contacting local election departments to request the more easily obtainable cast vote records, which are then posted at votedatabase.com.
Thanks to these efforts, registrars of voters across the country have been deluged with public records requests that have interfered with workflow even as election officials continue to receive death threats from Trump supporters.
Lindel and O’Donnell have been pushing listeners to request “cast vote records,” referred to as CVRs, which at their simplest are a text list for all the ballots received in an election before the official canvas. For example, downloading the Shasta County cast vote record at votedatabase.com yields an Excel spreadsheet listing each of the 94,084 votes cast in 2020 as a line item spread across a dozen presidential candidates, the House of Representatives, and the state Assembly and Senate races.
In “Fingerprints of Fraud: Evidence and Analysis of a Multi-State Conspiracy to Defraud the 2020 General Election,” a 281-page report released free of charge in May, O’Donnell claims the cast vote records he has collected can be used to detect patterns of election fraud. He specifically examines mail-in ballots because he believes they were “used as a critical attack vector” in the election during the pandemic. Here’s the chart for the 2020 presidential election in Shasta County, which O’Donnell claims exhibits the characteristics of a “Mesa pattern of election fraud:”
The vertical axis is the percentage of the vote, the horizontal axis is the number of votes. The red cone that extends to the right represents the upper and lower limits for the percentage of votes Donald Trump was expected to get, based on the number of votes he ultimately received and derived by “running 100,000 random elections of 500,000 ballots each.” The blue line is the number of mail-in votes for Donald Trump.
Here’s where the garbage starts. O’Donnell makes a convoluted argument that the mail-in votes for Shasta County and other counties that exhibit the Mesa pattern of fraud come in totally at random, like a coin flip but adjusted, in Shasta County’s case to 65 percent, the amount Trump won by here. Therefore, the blue line representing the number of mail-in votes should quickly jump straight up to 65 percent, running horizontally between the upper and lower limits, instead of gradually climbing diagonally toward the right.
Hence, the Mesa pattern of fraud is the No. 1 sign of a corrupt election in the United States according to O’Donnell.
“To fairly judge a county’s cast vote record mail-in results,” O’Donnell writes, “I first test the results to see if they meet the randomness assumption. The assumption is that mail-in ballot results contained in the cast vote records are randomized by the processing of the ballots themselves. As mail-in ballots are randomly requested, randomly sent out, randomly filled out, randomly returned or delivered by the voter, and not presorted by the county upon receipt, they become naturally shuffled and mixed.”
What’s that they say about assuming things?
Responding to the increasing number of case vote records requested during the past year and their misuse by amateur analysts, Max Hailperin, a retired computer science professor who has researched election technology, told Votebeat Arizona’s Rachel Leinang that amateur election sleuths that think they’re going to prove fraud with case vote records are mistaken.
“Some who are seeking these records believe they will be able to find fraud based on the sequence in which ballots were scanned, Hailperin said. For example, they would flag whether a large number of consecutive ballots came in for one candidate. But, in the case of ballots cast by mail, the order in which ballots are scanned doesn’t always correlate with when the ballots arrived. And citing ballot sequences as proof of fraud is a flawed premise: Different people, with different political affiliations, tend to vote using different methods because of their trust in early voting or other factors.”
Dan Wallach, a computer science professor at Rice University who studies election security told Leinang the same thing in regard to using cast vote records to prove fraud.
“They’re not going to tell the story that you want them to tell,” he said. “If the story you want to tell is that, clearly, this was fraudulent, you’re not gonna get that.”
Nevertheless, that’s exactly what O’Donnell is doing, according to Justin Grimmer, a professor of political science at Stanford and senior fellow at the Hoover Institution. Grimmer journeyed to Shasta County last September to debunk Dr. Douglas Frank, another Lindell associate making similar election fraud claims, at a board of supervisors meeting. He plans to attend Patrick Jones’s Town Hall this Thursday.
“He [O’Donnell] wants to say he knows the order in which the votes were counted,” Grimmer said, calling it a college freshman statistics error. He has glanced at “Fingerprints of Fraud.” “There are moments of laughter. He absolutely doesn’t know what he’s doing.”
Grimmer suggested a simple explanation for the blue line’s diagonal slope: Democratic voters were primed to use mail-in voting in 2020; meanwhile Trump demonized mail-in voting from the very beginning of his campaign. It makes sense that Trump’s mail-in votes would be lighter in the beginning and heavier as the voting deadline approached.
Grimmer is impressed with the amount of crowd-sourced election data O’Donnell has collected, even if he’s not exactly thrilled with his research. So far, Polaris Recount has examined the digital ballot images of just one county, tiny Harding County, New Mexico. “Citizen adjudicators” looked at 504 digital ballots and found six errors, three in favor of Trump and three in favor of Biden.
“If this anomaly exists in all Dominion Systems, then hundreds or even thousands of falsified votes could exist in larger counties,” O’Donnell writes. “In addition, it raises the possibility that totally blank ballots could be run through with votes being assigned to them.”
“They decided to create fraud symmetrically,” Grimmer said chuckling at the Trump-Biden tie and the attempt to spin up a small county’s results into something that sounds threatening. Then he grew more serious. All it takes to convince the election deniers is to say, “Mistakes were made.”
Why do Dr. Frank and O’Donnell do what they do? I asked Grimmer. What’s in it for them?
“Dr. Frank is super clear he’s not in this for the money,” Grimmer said, noting that Frank claims he only gets travel expenses for his efforts. It’s not clear how much Lindell is paying O’Donnell if anything. “My intuition is they are not grifting. They are true believers or, if I had to guess, maybe it’s just their way of getting attention.”