Yeah. I don't understand people so lacking in curiosity that they take a story seriously even when it says outright, "we know math doesn't work this way, but we also know you're too dumb to notice our conclusions are based on just that broken math."

ETA seems to have moved on to Professor Mebane's model. Have you seen it? He's a respected academic in this area, but other statistic academics can't make sense of it. I'm converging on my own assessment but am interested in other people's opinion. websites.umich.edu/~wmebane/PA2...

He'd be more convincing if he could write grammatically correct English, but the basic idea of putting in several suspect assumptions into a model to predict whether people would vote for Harris or Trump is about the same as reading tea leaves.

I actually believe Mebane is actually saying that the application of his model on PA revealed "false positives". It's why he drew a parallel with the German parliamentary election. But no one amplifying the "250K fraudulent votes" ever gets to this sentence.

Whatever he thinks he sees in his teacup, he's measuring his initial assumptions multiplied by patterns in the voting data. It's meaningless. I could believe voter suppression could play a part in the election outcome, but it would have to be proved based on the mechanisms used to do it.

This post was before the 3-state version was replaced, but conclusion remains......they thought they would get justification for a recount and got an academic exercise to misuse bsky.app/profile/r5-t...