Depending on the results over the next seventy-two hours, sore loser Donald Trump and his supporters will probably complain about problems with the machines that count the vote and other election security issues.  While every state is different in the exact means that they use to secure the election, most have roughly similar processes that pretty much guarantee that what Trump will be saying is pure fiction (which has never stopped him in the past and will not stop him this year).

One thing that you may hear is the vote counting machines can be hacked.  This claim is a half-truth.  The machines are computers, and, in theory, any computer can be hacked.  The people who raise this claim may even note that the internal security in the programming of the machines is weak.  They may be right about that, but that argument misses that the external security is very strong, and there are checks in the process to detect if a machine has been compromised.  The biggest external security is that the actual counting machines are not connected to the internet — either by cord or by wireless.  The results are manually uploaded from the precinct machines to the county’s machine and the results are then distributed.    Hacking these machines requires having physical access to the machine (which only a limited number of people do) or the virus has to be installed at the company that does the programming for the machine.

Even if you could get to the machine, setting up the software to manipulate the results is not easy because it has to defeat the checks.  The basic checks are:  1) a pre-election “logic” or “accuracy” test on every machine; 2) a post-election “logic” or “accuracy” test on every machine; 3) a hand count audit on random races from random precincts; and 4) the voter logs from the election.  The logic or accuracy test is using a test deck featuring a known number of votes for each candidate including various different weird ways that people can vote (overvotes, skipping races, voting a straight party ticket by marking the bubble for the party, voting a straight party ticket and then marking the bubble for that party in each race, voting a split ballot by marking the party box and then voting for the other party in individual races, etc.).  The test is whether the machine gives the known result.  If it doesn’t there is something wrong with the programming.  And here is where the difficulty for the hack comes in.  The person doing the hack does not know when the tests will be done.  If the hack takes effect immediately, the machine will fail.  The hack has to take effect after the first test and then revert back before the second test — a much more complex hack.  Even if that effort succeeds, the hack has to be on only the machines that are not used in the audit.  If the hack is on the machines used in the audit, the error will be caught, and the audit will expand to other precincts.  And lastly, the hack has to merely switch votes.  The voting logs will detect if a precinct has too many votes recorded.  In short, it is almost impossible in practice to successfully hack the counting process. Continue Reading...