Reply #19: I would guess that a hacker or hacker would not have "one" method...

[Sancho]

Over the last 3 or 4 election cycles, if we hypothesized the hackers point of view; it is possible that an opportunity for vote switching would make more sense in precincts where registrations were close, and the purpose was to turn race (like Coleman - Franken).

Dropping votes (undervotes) would be more applicable if the precinct was majority one party, but the goal was to shave the state or district total as you would see in Miami on a statewide election.

That is why detection by statistical SEM would need precinct level data on a variety of situation in order to spot abnormalities across many possible combination. If the "hacker" formula was preset to a maximum SE in order to prevent obvious problems, other methods would be more sensitive.