Remember when a few thousand votes came in overnight from Michigan, and they were ALL for Biden? It's so statistically unlikely that even if all the votes were meant to be for Biden it's more likely that a voter circled the wrong bubble or a counter accidentally counted one ballot incorrectly (i.e. false positive) and Trump would've gotten a vote that way.
Has anybody figured out what the actual probability of that happening is?
I'm no math guy, but it seems like you could take the voter registration percentages in those counties and calculate a probability distribution of heads/tales on a weighted coin to approximate it.
There's a lot of mathematics about the false positive paradox. I don't know what the exact false positive rate for filling out a form is, but if a person misfills a ballot in 1 out of 45,000 occurences or more often, this is considered statistically significant at the .95 level (5% or less). At the .90 percent level it's up to 1 in 56,000
aka the ballot printer
Remember when a few thousand votes came in overnight from Michigan, and they were ALL for Biden? It's so statistically unlikely that even if all the votes were meant to be for Biden it's more likely that a voter circled the wrong bubble or a counter accidentally counted one ballot incorrectly (i.e. false positive) and Trump would've gotten a vote that way.
Has anybody figured out what the actual probability of that happening is?
I'm no math guy, but it seems like you could take the voter registration percentages in those counties and calculate a probability distribution of heads/tales on a weighted coin to approximate it.
There's a lot of mathematics about the false positive paradox. I don't know what the exact false positive rate for filling out a form is, but if a person misfills a ballot in 1 out of 45,000 occurences or more often, this is considered statistically significant at the .95 level (5% or less). At the .90 percent level it's up to 1 in 56,000