1/ @TXAG , himself currently under indictment for securities fraud, and apparently also under investigation by my former agency for public corruption, is suing to have other states' citizens' votes thrown out in a bizarre filing laden with impressive-looking bits like this:

2/ That does sound really specific and dramatic. But, having spent a lot of time examining election numbers, I have to wonder if Mr. Paxton is that bad at math - or just a really bold con artist. Because no "common statistical test" he alludes to says anything of the sort.

3/ Paxton's "expert" (Charles J.
Cicchetti) pulls some really absurd statistical shenanigans. Essentially, he argues that the time between 2016 and 2020 having been so short, the statistical distribution of voters should be the same (this is dumb). He then tests the 2020 counts

4/ and finds - amazingly - that people didn't vote the same way! In other words, he uses a year when Trump won, compares it to one where Trump lost, and then says "If all those people voted the same way, there's 0% chance Trump lost!"

5/ In Ciccetti's filing, there are no formulas, but he focuses in on Georgia in some detail. He points out that at 3:10AM on November 4 (why this dramatic moment in time, I don't know) with about 268,204 votes remaining, Biden was behind by 57,346 votes but *somehow* led by

6/ 14,000 on 11/18. Impossible, he declares!
Wrong!
If we split the remaining votes as they were in 2016, this would be correct, but samples of the mail-in ballots showed they were going around 2-1 for Biden (66.66667%). Ciccetti points out vote counts are binomial distributions

7/ but this is a skewed one. In fact, after taking several samples of the remaining ballots & finding 2/3rds D, Georgia could have been called for Biden RIGHT THEN. Because Biden only needed to win ~61% of the remaining votes to have a majority. And more importantly, the odds of

8/ Biden winning are the SUM of the probabilities for all of the ways he could win - not just the specific 14,000 vote lead. Put simply, Biden needed less than 163K of the remaining votes to win, but ANY number from 163K to ALL of the remaining votes wins it for him.

9/ If the remaining mail-in votes from the heavily democratic counties had only split 61% in favor of Biden, he STILL wins 99.95% of the time. Ciccetti is pulling a slight-of-hand here - claiming the remaining votes should be treated as a normal distribution around the 2016 mean.

10/ This misdirection goes to the heart of the question:

Is this bad math or fraud?

Republicans spent a HUGE amount of time prior to the election concocting a "mail-in-ballot fraud" narrative. @ProjectLincoln tried to warn everyone about it.

The GOP, knowing their "base" was

11/ less afraid of COVID (especially *because* they tried to cast it as a hoax), and certain that Democrats (shown by surveys to be more afraid of COVID) would choose to mail in ballots, developed a plan.

I believe the GOP conspired in a "scheme and artifice to defraud" the US.

12/ This scheme and artifice to defraud included the following components:

- Spreading rumors of "massive mail-in-ballot fraud," including using officials of the US Gov - in violation of the Hatch Act - to falsely state that foreign powers were manipulating mailed ballots.

13/ - Directing the US Postmaster General to disable the USPS's ability to process mail-in ballots - especially in heavily democratic areas

- Producing and transmitting by wire and mail in interstate commerce false statements in furtherance of this scheme

14/ including: knowing false declarations by purported poll watchers, unsubstantiated allegations about Dominion, a US company that manufactures voting machines, and false testimony by "expert" witnesses.

Once the votes were counted and Biden emerged as the clear winner,

15/ the conspirators began a campaign of intimidation, using US Government employees in their official positions including the POTUS in an attempt to coerce, threaten, and extort election officials into changing the outcome of the election.

16/ This campaign culminated in Texas AG Ken Paxton filing an affidavit with the Supreme Court which he knew to be false and/or misleading, including incorrect statistical calculations and false claims of hacked voting machines - despite multiple recounts of the PAPER

17/ ballots confirming - and even extending - President-Elect Biden's lead.

That last bit is important - as @C_C_Krebs pointed out, paper can't be hacked. Spooky claims of stolen thumbdrives and hacked voting machines are both unsupported and irrelevant.

My conclusion? Fraud

Note: if anyone is thinking of writing an amicus brief, I'm in