People attempting to elevate false claims about election fraud have lots of ways to try to make their arguments seem more robust. There are centuries of chicanery to draw upon, after all, and some tried-and-true techniques that simply need to be upgraded for the modern era.

For example, many of the fraud claims rely heavily on intentionally confusing or complicated jargon. They cite what is presented as “statistical analysis,” involving various manipulations of numbers in Excel spreadsheets until they can be framed as nefarious. (These analyses are somehow never reinforced with proven examples of the rampant illegal voting they claim to reveal.) And there are various appeals to authority intended to show just how serious this all is: claims about lawsuits and imminent court cases and the sorts of hinted ramifications that never appear to manifest.

This week we were gifted with a particularly wonderful example both of how dishonest or deluded actors try to reinforce their claims about fraud and how they cannot be reasoned away from them. The website Gateway Pundit — long a storehouse of unfounded conspiracy claims and surreal assertions — has upped the ante in its claims that the 2020 presidential election was stolen from Donald Trump. It will pay you $10,000 if you can explain the anomalies that it claims to have found.

Clearly the folks at Gateway Pundit must be pretty confident in this, right? Why else would they put up that kind of money?

The answer is that they are putting up that money to intimate precisely that sort of confidence. And that no one will ever receive the reward because their claims are divorced from reality, to the extent that this has already been explained, repeatedly, but they refuse to accept it.

But, hey. On the off chance that rationality might for once defeat conspiracy theorizing, here’s my bid for the $10,000. It can be paid to The Washington Post’s “Helping Hand” charitable program.

The heart of the purported mystery centers on what the site calls a “drop and roll” process, a name that links two different data questions because calling it “drop and roll” makes it sound more nefarious. I feel it’s useful to put a fine point on this, so I will: This is all very ridiculous.

The well-explicated inexplicable mystery the site hopes to have explained is presented in a video produced with the now-common aesthetic of the amateur Internet sleuth. Here, for example, are the claims made about the results in four states.

The “drop” refers to the addition of large batches of ballots at different points in the vote counting. The “roll” refers to what the video claims are statistically “impossible” margins in vote batches that occurred at the same time. The implication is that the big increases in votes (“drops”) are then slipped into the totals in smaller chunks (“rolls”).

But this, by itself, makes no sense. It’s conflating two totally different things: the addition of large numbers of ballots from large counties and cities with the public presentation of the results coming in. Think about it. If the vote totals show an addition of a few hundred thousand ballots, giving Biden a wide edge, what does that have to do with purportedly simultaneous incremental additions given to him in a feed of vote totals? These unidentified masterminds throwing the election come out and say “Biden gains a net of 400,000 votes” or whatever, and then slip those results into the totals in batches of 100 at a time? What is the point?

Anyway, the “drop” has been explained ad nauseam. I’ve explained it before, looking, for example, at Wisconsin. In some states, at some points, big places with lots of voters reported big batches of results. It’s not in the least bit complicated, and there’s nothing about those results that deviates from past elections. In other words, a big dump of votes from Milwaukee might shift the count, but it’s not like Milwaukee didn’t heavily favor the Democrat in 2016.

What’s honestly amazing about Gateway Pundit’s approach to all this is that it even acknowledges that this has been debunked but then pretends that it has somehow unbunked things.

This is an actual thing an actual person typed and apparently thought was convincing:

“When the fact-checker finally addresses the issue of the hundreds of thousands of votes dropping early in the morning after Election night they state that there were many more absentee ballots due to COVID. We really don’t know that. There was no study that we know of where they prove this argument. They don’t cite their source. Maybe there were more votes due to cheating by Democrats, isn’t that just as plausible?”

No? It’s idiotic?

I mean, this is unfalsifiable. If you say to this person, There were a lot of absentee ballots because of the pandemic, and they say, Oh but were there?????, you just sort of have to shrug. Okay, man.

Then there’s the “roll,” which, impressively, is dumber.

It centers on lists of data like this one.

That same fact-check of the site’s ridiculous claims points to a different version of data that gets at the same thing.

These, the site explained in November, are incremental updates from the New York Times’ feed of results following the election.

“One oddity in the file noted immediately is that the results for votes are not in whole integers (e.g. 1, 2, 3…)," the site writes. “All of the entries have fractional amounts. This makes no sense since ballots do not come in fractions in the US. Each vote equals one vote.”

This is presented as proof that something weird is going on instead of what it actually shows: These are obviously not raw vote results.

The way election night data works is that media outlets like The Post and the Times subscribe to services that pull in updated information from states and counties. In 2020, both papers used Edison Media Research. That company provides not only results but also expectations for how many votes will be cast overall. It’s what’s called modeled data, using what we know to figure out what’s likely to remain.

So let’s say it’s 12:01 p.m. and the model thinks that about 3 million votes will be cast, 51 percent of them for Biden. Then a precinct reports its totals, and it had lower-than-expected turnout. At 12:02, the model might then decrease the overall number of expected votes and shift the expected outcome of the race in Trump’s direction. If you subtract the expected number of votes for Biden at 12:02 from the total at 12:01 and compare it with Trump’s, you might think that there was a reduction in votes or that Trump was being unfairly punished. But the reality is that you have no idea what you’re looking at.

It’s possible that the data shown above represents something else, but the fact that it includes decimals indicates that this is something other than actual data. And, therefore, that this is something derived from reality, not reality itself. (This, too, was unbunked unsuccessfully by Gateway Pundit.)

In short, the explanation of the “drop and roll” is that the “drop” was “vote counting” and the “roll” was “something Gateway Pundit made up to imply nefarious activity.

Looking forward to seeing that check.

Jeremy Bowers contributed to this report.