Actually, the only argument is that the distribution of first digits doesn't follow a smooth curve, so some people believe that indicates irregularities. But as thoreau said, if every precinct was targeted at something like 600 voters, then there's no way that Benford's Law would give any reasonable clue whether there was some irregularity.Number 6 wrote: ↑10 Nov 2020, 11:29So, the argument is that there are anomalous numbers of people casting ballots in some precincts? I'm guessing that the person thinks that there are improbably high numbers of ballots being cast in democratic leaning precincts, and is implying ballot stuffing in Dem precincts, and/or ballots being undercounted in Rep precincts. Seems like an easy thing to check-voter lists and turnout numbers aren't exactly secret, or even difficult to come across.Highway wrote: ↑10 Nov 2020, 10:42Basically, it's taking all of the vote counts from precincts, and only looking at the first digits of those counts. So say that the precincts had the following vote counts:

175

180

245

97

382

727

Then you'd take the first digits there: 1,1,2,9,3,7 and plot a bar graph of their occurrences. So the 1s would be twice as high as all the rest. If you do that with a lot more data, you expect to see a smooth decreasing curve from 1 to 9. But as you can see, with small sample sizes, you don't get a smooth curve. Say the second precinct there had 25 more people come vote, for a total of 205. So you'd have 1,2,2,9,3,7. Now you'd have a spike at 2, where you wouldn't expect it.

All "Benford's Law" is is a rough description of the expected relationship of a lot of numbers. It's useful as a quick tool for auditing financial records, because it can show that there may be some data that doesn't look right. It's not a proof of anything, just a 'this isn't what we'd expect, let's look closer at it."

But I also take the point that the numbers are too small at the precinct level for any of this to mean anything.

Basically it's just throwing something at the wall that sounds important and trying to sow disinformation.