>>12334411You're mixing up election forecasting experts with statisticians. Election forecasting has weak predictive power, everyone knows it's a borderline grift. It's also a different field from statistical theory, which is where you get statements about the probability of a thing happening multiple times.
You say you don't have all the facts, which is plainly evident, but what makes you think the things you DO know adequately cover the situation you are in? Modelling ballot counting as roughly equivalent to flipping a coin is nonsensical. Ballots are collected inhomogeneously from multiple precincts with large variance in their chances of any individual voting for a given candidate, counted in a nonrandom order, and reported in batches that are also influenced by many different factors that have a real influence on the time series of votes counted. I don't even know all the factors, because knowing them all would take serious expertise in the details of EVERY STATE's ballot collection and counting process.
Given all that, presupposing that any apparent "anomaly" in the time series is indicative of fraud based on incomplete information about the ballot counting process is an accusation with very little weight behind it, that only serves to confirm the biases of anyone who believed there was going to be fraud in the first place.
If you want to prove electoral fraud, stick to identifying singular cases. They're better evidence than anything you can find by looking at graphs. Election forensics on a statistical level isn't something you can pick up and apply competently within a week after the election.
