Joel [Minnesota] on Nostr: Interestingly enough, there is a slightly increased probability in clustering when ...

Interestingly enough, there is a slightly increased probability in clustering when working preceint by preceint. 16,000 votes is ~3% of the total expected, ~600k, single vote ballots. Having a single box all be this type for the same candidate is statistically improbable and would require fraud. Grok estimated it at (0.008)^16000 chance - less than 1x10^-300. I think the number is small enough that the limit is zero.