I've been to two major events so far, Thunder in the Mountains 2016 and COPE 2016, and I've been offering free testing for Violet wand output and for gapped electrodes like body contacts.
The function of the safety gap in such electrodes is to prevent the low voltage from the wall plug from conducting through the person. The gap needs to be small enough so that when the high voltage fires across the gap the energy lost in the gap is minimal, but large enough that the peak line voltage (which could be as high as 325V on a 230V line or 170V on a 120V line) can't sustain the spark when the high voltage is not present.
This is the same reason that fuses have voltage ratings. A 32V automotive fuse is not guaranteed to be able to stop the current in a circuit with higher voltage like a 120V AC line. Arcing may allow current to continue to flow through the blown fuse!
I've tested maybe 20 body contacts at these two events, and two of those were defective, with the safety gap shorted. That is not a comforting number. This is better odds than "Russian Roulette". but not much better.
I have not had the opportunity to dissect those failed pieces, which would have been interesting. I can easily predict a couple of failure mechanisms though:
1: Carbon tracking across a plastic insulator. When a flush cut wire end is used as the gap, the spark will end up running over the surface of the plastic. The heat and ozone from the spark will carbonize the plastic over time, turning it into a conductor. This is similar to the carbon tracking that happens on spark plug boots, wires and distributor caps.
2: Errant wires. Any wire projecting out past the plastic could over time, with flexing of the body contact cable, bend slightly and short to the outer contact.
3: Other manufacturing defect or use driven failure mechanism.
There are a number of vendors making these devices, and I can't possibly know all the techniques that they might have used to create the spark gap. No matter what the failure mechanism, the effect is the same, the safety gap no longer works as it should but the worst part is that the user has no idea this has happened. A safety barrier has been removed without your knowledge.
What do you do about it: Test your equipment. The required equipment (a 1000V "megger") isn't expensive. $20-$30 A group could buy one and let the members use it at munches and play parties. The "megger" builds up 1000V and places that across the test leads. You connect one lead to each end of your body contact or other gapped electrode, and the meter reads out what the resistance is. It should read "1 " which indicates open circuit. A lower reading indicates that current is leaking through, probably due to carbon tracking, or maybe contamination like dirt or moisture inside the gap assembly.
I developed my "Clearview Spark Gap" because of these issues, The gap material is borosilicate glass which simply will not develop carbon tracking, and the gap spacing is determined by the diameter of the spheres, which won't change over time. It simply won't fail in a way that would be dangerous.