The problem is that Benford's law isn't applicable to the question at hand. The reasoning is simple. Assume a large group of average sized precincts. You have 700 votes, in a binary choice election. Benford's law applied to candidate #1 says that 30% of his vote counts should start with the number 1. So basically those are vote counts from 100-199 (we'll exclude very small numbers). Then, necessarily, the vote count of candidate 2 needs to be 700-C1, or in these 30% of the time, it will be between 501 and 600. So 30% of the precincts have Candidate 2 with vote counts starting with 5, which violates the principle of Benford's law that states this should only be 8%.It is interesting how people that do not believe that the election was stolen put ZERO stock in that article. Even "intelligent" people think that Biden somehow fell outside of the theory.
This kind of mathematical proof is known as reductio ad absurdum.
Data scientists queue up to explain why claims of irregular vote tallies don't hold water
William Mebane is THE expert on Benford. It isn't a law that applies everywhere, it applies to certain kinds of data sets, but not these. I realize the fact that I am saying this makes you guys dubious, but math is math.