essential concept in uncovering election fraud: heteroskedasticity or "the variation in the variability".
Large packets of votes with no variation at all 133,389 all for "biden" is extreme heteroskedasticity.
the bigger the reporting center, the more opportunity for fraud.
normally in statistics, big samples mean more accurate results. the exact opposite is true here: the bigger the voting population, the more possibility of systematic fraud.
the percentages from a center counting 30 votes will be far more accurate than one reporting 300,000 votes, where there will be massive interference. it's an inverse universe and it's not hard to spot. this election stinks more than any election i've ever seen.
i first noticed this extreme heteroskedasticity in the 2016 democratic primaries, when hillary clinton beat bernie sanders in new york state. you could see the fix on the map. the whole state was for bernie, except for those large population dots.