Sam says you should read this
This blog was created with the BlogFile software, written by Samuel Levy.

You can find Sam on Google + and LinkedIn.

On voting machines and tin-foil hats

Americans are having an election today. You may have heard. I'm not American, but I've still heard about it (for months). The last few elections have been rife with accusations of vote-fixing, cheating, and "hanging chads". This year appears to be no different.

A video showed up today on YouTube showing a voter using an electronic voting machine, selecting Obama and getting Romney. Apparently this type of thing has happened before, prompting The Simpsons to parody pretty much this same problem. Breitbart has also (although without any more than anecdotal evidence) claimed that the same thing has been happening to Romney voters, getting their votes changed to Obama.

This has set off the conspiracy brigade.

Now I want to make one thing very clear: it is very, very unlikely that this is any form of deliberate vote tampering. If it is, then it is the most incompetent vote tampering ever.

I fail to understand who would attempt to change someone's vote, and then show the voter. If a machine was going to deliberately change your vote, then it would do it silently, and not show you that it had done so.

I'm not saying that this issue is specifically the result of bad calibration, or poorly written software. All I'm saying is that the chances of this being actual intentional voter fraud are so tiny, that I could only imagine it being perpetrated by a cartoon villain who was too busy twirling their moustache to notice that they had completely ballsed up the execution.

So everyone, put your tin-foil hats to the side for a second, and focus on the real problem - there is obviously a bug in the software or the hardware which should never have gotten into production. If anyone should be blamed, it should be the company that makes the machines - not the candidates or parties they represent.

Comments have been locked for this post.

The fear is that machines are 1) inaccurate across the board, giving a 50/50 chance to Obama is actually statistically ahead in the state, or 2)rigged in such a way that perhaps the visual vote shift is simply an artifact of deeper software issues, or 3)that this is in fact an anomoly, but diverts attention from the fact that the software patches that the GOP leaders decided to install on the machines are not innocent. There is strong evidence that machines and votes were tampered with in Ohio in 2004, which led to a Bush victory. This following the 2000 election in which the supreme court basically stole the election from Gore (who was later found to have won Florida after all). With this record, it doesn't seem a stretch at all for the Teaparty people in Ohio (Husted) would find a way to shave votes in key districts.

I fear your apparent naivete.


Oh, yes! And the fact the the GOP has spent the last several months trying to pass voter suppresion laws, clearly in violation of the spirit of the Constitution if not the letter, does not inspire confidence. Nor do the long lines in swing states make me feel at ease--many of of those lines seem to be aggravated by GOP leaders actions or inactions.

Samuel Levy

I'm not saying that there isn't a problem (in fact, I think that there is a real problem there). I'm just saying that I don't think the problem is deliberate voter fraud. It seems too sloppy for that.

In Australia, I believe that there are two states which have incorporated electronic voting* - one state uses a machine like this to record the vote, but actually prints it to a physical piece of paper which is then counted manually. The other allows people to vote with paper, then electronically tallies the votes. In both cases, there's a physical paper-trail.

* I might be wrong - my state is still pencil and paper, and manual counts. I've only been told about the electronic voting practices anecdotally


If that visible a problem could turn up in production, then imagine what a mess the hidden internals are like? And this is for an application that one would hope might come under greater-than-usual scrutiny! Did the customer not review the source code? Do any sort of failure analysis? Or did they just take it on trust that the devices worked as advertised?