As a point of personal privilege, I would like to offer an observation rooted in nothing more than my own biases and inclinations. That is this: Were I being sued for more than $1.3 billion, I would either try very, very hard to find evidence that would allow me to win the lawsuit or I would very, very rapidly try to reach a settlement. But then, I am not Mike Lindell, MyPillow CEO and game-theory enthusiast.

That latter description comes from Lindell’s response to Dominion Voting Systems’s massive defamation suit against him: “Lindell admits he loved game theory.” Lindell also admits that “he’s a numbers guy” and that he is “good at math.” These admissions, such as they are, are part of a point-by-point response from Lindell’s attorneys to Dominion’s February suit (in which they seek the aforementioned $1.3 billion). Dominion cites Lindell as describing himself as a numbers guy and, in his response, Lindell agrees — though he disagrees with most of the assertions Dominion makes that bolster their argument against him.

Dominion’s play is to establish that Lindell knows enough about numbers that his insistences about rampant voter fraud are clearly insincere and that his claims that Dominion allowed the 2020 presidential election to be stolen were, in fact, efforts to gin up attention so that he could sell more pillows.

“Lindell pushed this false narrative even though there was no evidence to support it because he had learned from his knowledge of game theory and prior experience … that lying about Dominion would be good for MyPillow’s bottom line,” the February suit states, “and would lead to other benefits, including, for example, Trump’s anticipated endorsement for Lindell’s run for governor of Minnesota.”

Maybe. Or maybe Lindell sincerely believes that he has proved that Dominion stole the election. Maybe, as I’ve argued in the past, he’s been convinced that he has the goods when he very obviously doesn’t and exists in so protected a bubble of reality that he presses forward with that conviction despite having precisely zero evidence to bolster his claim. He is the emperor and his fraud claims are his clothes and here he comes, parading into court.

But let me be specific. At the heart of Lindell’s response to Dominion is his assertion that Dominion machines were vulnerable to vote manipulation and, further, that votes were manipulated using the machines. I am not a lawyer, so I will leave it to others to determine whether proving this claim would actually settle the legal dispute. But I am myself decent at math and pretty good at assessing statistical analysis so I can say with certainty that Lindell’s response very much does not prove that the election was stolen.

We can categorize his claims into three groups: claims about vote-switching, statistical analysis about the results and claims about proof of international hacking.

Vote-switching. Lindell’s countersuit makes a couple of claims about votes having been switched in the 2020 election. Here’s one:

“For example, in Bibb County, Trump was reported to have 29,391 votes at 9:11 pm EST while simultaneously former Biden was reported to have 17,218 votes. A minute later at the next update, these vote numbers switched, with Trump now having 17,218 votes and Biden having 29,391 — a 12,173-vote switch in Biden’s favor. YouTube — owned by Google, Inc. — removed this news video after this switch was revealed.”

This is in Georgia, where the ballots were audited and recounted after the election multiple times — meaning that any such “switch” would either have been detected or would have involved changing actual ballots. But it didn’t happen anyway. An article from the right-wing Epoch Times linked in the suit explains that the claim is based on “time-series election data,” almost certainly meaning reported results from a news organization like the Associated Press. If anything happened, it’s likely that the AP switched results and then corrected its error, something that happens occasionally. When Donald Trump made a similar claim, the county flatly denied anything happened.

The suit makes another vote-switching claim centered on Michigan.

“Litigation involving Dominion’s voting machines in Antrim County, Michigan, initiated after approximately 6,000 votes were discovered to have been wrongly switched between Presidential candidates — ostensibly due to a so-called ‘glitch’ — proved Dominion’s machines could be manipulated and hacked to generate this ‘glitch.’ ”

Yeah, no. This is one of the oldest conspiracy theories out there and one debunked over and over, including by Republican legislators in the state. The short version of the story is that the ballot was changed but the machines that read the ballots weren’t, so the tabulation didn’t properly align vote choice with candidate. It’s as though you were passing out personalized gifts to a class of kids who are in line alphabetically but you weren’t told that Kevin joined the class. Every kid from Kevin on is going to get the wrong gift.

What’s remarkable about the claim in the suit, though, is that this error is not just an error but purported proof that the machines could be “manipulated and hacked.” Or — bear with me here — it shows that the machines are subject to human error.

Statistical analysis about the votes. At another point in Lindell’s response, there’s an extended explanation of analysis conducted by a guy named Douglas Frank. I am infatuated with this claim because it combines so many of the hallmarks of misleading statistical nonsense, from technical-sounding lingo to complete irrelevance from the point trying to be made. I’ve written about it more than once.

Again, the short version: Frank took data from various counties in several states and looked at turnout by age. He then averaged the turnout across those counties, dropped it into Excel and generated a polynomial trend line. That’s the technical-sounding bit, but it’s literally a few clicks in the application. Then he compared that line to all the results in each county and, lo and behold! the age distribution in those counties correlates to the line he created. He calls that line the “key” and implies it’s somehow a formula used to generate vote results.

It is hard to type with hands clenched into fists from annoyance, but I’ll try. First of all, there’s a consistent pattern to turnout by age in every election anywhere: Older people vote more. So yeah, a line that shows older people voting more is going to correlate to the results nearly everywhere. Nevermind that the line he generates uses data from the state itself, which, as I analogized previously, is like trying to predict how fast Usain Bolt ran in a number of past races by averaging a number of past Usain Bolt race results. Your result is going to be close!

All of this is beside the broader point: This isn’t proof of fraud either. It’s just some nonsense that someone who’s actually good at math should know better than to treat as serious.

Proof of international hacking. This summer, Lindell hosted an event in South Dakota at which he promised to show concrete evidence of Internet traffic in which votes were manipulated. When the event rolled around, the same thing happened that happens whenever a con artist is asked to put up: There was a glitch that prevented the evidence from being shown. We did learn at that point, at least, that the purported source of Lindell’s information is someone notorious for past claims that didn’t pan out.

But Lindell’s response nonetheless pushes forward, claiming that an attachment shows “a subset of twenty documented successful hacks through the election management system in the states of Michigan, Pennsylvania, Georgia, Wisconsin, and Arizona resulting in a total 555,864 votes switched from President Trump to candidate Vice President Biden in the 2020 general election.”

Actually, what that attachment shows is a table listing IP addresses for computers resolved to various locations around the world (China, Iran, Germany, Brazil) accompanied by terse descriptors of the alleged fraud. “CHANGED,” one column reads, with cells underneath labeled as “TRUMP: DOWN.” Then in another column, “VOTES,” the tally that was purportedly changed.

It’s basically as though you took this —

Hacker country
Votes stolen
From who
Hacker code

— and used it to assert that Trump won the popular vote. I mean, the tally of votes purportedly stolen by the various IP addresses don’t even add up to the number of votes stolen per county in the same document. Hackers in Brazil dropped Trump’s total by 13,449, but that total isn’t comparable to any of the county-level totals that were purported affected. It’s just gibberish.

Here’s the thing. If you have digital evidence that someone stole votes, show the actual packet of information in which that occurred. You don’t need to summarize it; this is a major lawsuit that could cost hundreds of millions of dollars. Show the actual theft, not the functional equivalent of a police report (if police reports were written by dudes who were told about crimes by their roommates)!

This, more than anything, is the amazing part of Lindell’s response. There’s nothing new here, at all. It’s the same claims he’s been making for months and that he’s for months been completely unable to substantiate. Maybe that’s good enough for the folks who tune in to his bespoke TV network, but it’s not going to convince a judge and it’s certainly not going to worry Dominion.

The day of reckoning for Lindell’s claims is getting closer and closer and he has done nothing to prepare.