Statistical methods for detecting election irregularities from the numerical results cannot hope to find problems with many of these factors. They are best suited to detect issues with the counts that might result from fraud, manipulation, or error. They generally work by assuming that “fair” and accurate election results would follow a particular hypothesised statistical model, then looking for evidence that the results in question do not follow that model. That is, they assume that the results are correct and that they were generated by a known, understood statistical process, absent evidence to the contrary.