Monday, June 18, 2012

Disappearing Ink

This is a horrible development for open democracy, but as a clever innovation it is worth thinking about.  News out of Egypt is that supporters of one candidate (name undisclosed to keep politics out of this discussion) handed different kinds of pens to different voters.  If the voting booth attendant thought a voter was going to vote for the attendant's preferred candidate, the voter was given a real pen.  If the attendant thought the voter was going to vote for another candidate, the voter was given a pen with disappearing ink, so the vote wouldn't count.

There are a few performance issues for this innovation that make this harder than it seems at first glance.

1.  First, you need the booth attendant to be reliable.  If even one of them either intentionally (double agent or feels guilty) or unintentionally (over a beer, bragging to friends about his cleverness) reveals the plan you are done for.  This isn't the kind of thing that wouldn't spread pretty quickly.

2.  Then you need the pen exteriors to be so similar that the voter can't tell they are getting a different pen than other voters.  But the attendant needs to be able to tell the difference.  Kind of like using marked cards in poker, it needs to be subtle.

3.  Next, you need the attendant to be able to tell who is planning to vote for which candidate.  This doesn't have to be 100%, but it has to be reliable enough to impact the final vote count.  If they are wearing a big pin on their shirt or chanting slogans then it is easy.  But can you tell through demographics?  And how much training would the attendant need?  You are not going to get the nation's keenest observers to be voting booth attendants across the country. 

4. If you don't want any voters to notice, the ink has to last longer than any of them sticks around in the voting booth.  This is a behavioral question. Most people leave quickly, but someone who is confused might not.  Or maybe someone who is physically infirm might need a few minutes to rest.  What are all the other use cases for sticking around?

5. And then if you don't want any of the counters to notice, the ink has to disappear faster than the quickest path from voting booth to counting station.  This is more of a process mapping problem.  What is the path?  Do completed ballots go to counting stations only at the end of the day?  Or at periodic intervals such as when the ballot boxes fill up?  Even if you know the official policy, are there in-the-field exceptions you have to worry about?

6. Finally, you have to figure out how to make ink that reliably fits this window. Even one screw up could get you caught.  So you need to find an ink formulation that has real Six Sigma performance in terms of accuracy and precision. 

Well, they got caught so one of these things failed.  But it might have lasted long enough to have worked.  We may never know.