>>>>>Bit of a shame, because according to at least one analysis there was a mathematically valid justification for a recount:
>>>
>>>1) I can't find validation of their findings of 7% more votes for Trump where there are voting machines.
>>>
>>>2) Assuming it's accurate, it's still a deductive fallacy until you eliminate proper explanations for voting patterns. For example, where and why are machines used vs paper? That could provide an explanation by itself.
>>>
>>>No matter: recount is demanded and so far it suggests that the Russian hackers preferred HRC over Trump, stealing 15,000 votes for her in Milwaukee alone. Perhaps it's actually easier to hack the paper votes than the machines which is why the machines reflect a more honest match to the audit = 7% more for Trump. In any case, conspiracy theorists can breathe a sigh of relief that the Commies were thwarted in their dastardly plan. Also, popular vote protesters might consider how many votes the Ruskies stole for HRC in their own states to deny Trump the popular vote that would have prevented rioting protest.
>>
>>Hmm, Halderman's article on Medium links to Ron Rivest (the "R" in RSA), amongst others: https://medium.com/@jhalderm/want-to-know-if-the-election-was-hacked-look-at-the-ballots-c61a6113b0ba . I would have thought that if the data on which their analysis is based is public enough for them to analyze, then they should have published it - but I can't find it. Looks like I'm guilty of parroting the MSM.
>>
>>I don't know "where and why are machines used vs paper". Maybe wealthier districts can afford machines? But does that translate to a preference for HRC or Trump?
>>
>>Absent some other explanation, a variation of 7% (if it is actually 7%) on a large sample is probably statistically significant. Might be grist for some theses in years to come ;)
>
>
>I doubt very much that the distribution of votes asked via machines vs. on paper is random, so this needs to be corrected ex post facto, opening the doors to statistical corrections based on personal assumptions.
>
>Great way to earn money, if you can fudge the dice and are payed for doing just that ;-)

Based on the very limited returns of the recounts I've seen so far, you're probably right. Still, I'd be interested to find out what is causing the variation - do you have any ideas?