That said. This is a gross misapplication of Benford's law. Benford's law is valid in cases where there is no expectation of structure in the underlying data set and when the range of numbers spans orders of magnitude. If instead you have things like voter precincts of a range of sized (i.e. 500-5000 voters) rather than randomly sized voter precincts, then both criteria for Benford's law to be valid are violated.
Given how quickly Benford's law was trotted out and then discredited it almost seemed to me to be intentional chaff to invalidate claims of vote fraud.
Please don't include this discredited test for vote fraud in this graphic. It has zero validity, will be shredded by anyone with a mathematical background, and undercuts all the evidence.
Great graphics.
That said. This is a gross misapplication of Benford's law. Benford's law is valid in cases where there is no expectation of structure in the underlying data set and when the range of numbers spans orders of magnitude. If instead you have things like voter precincts of a range of sized (i.e. 500-5000 voters) rather than randomly sized voter precincts, then both criteria for Benford's law to be valid are violated.
Given how quickly Benford's law was trotted out and then discredited it almost seemed to me to be intentional chaff to invalidate claims of vote fraud.
Please don't include this discredited test for vote fraud in this graphic. It has zero validity, will be shredded by anyone with a mathematical background, and undercuts all the evidence.