I made a relevant or let's say interesting observation:
When the light is 'on' and you interrupt the circuit (e.g. by unscrewing the tailcap), the light turns 'off' (obviously!), and when you screw the tailcap back on, the light continues to be 'off'.
However, when the light is 'on' and you interrupt the circuit only for a fraction of a second (up to 0.5sec or so), the light turns off during that short time (obviously!) but then returns to the 'light on' state, i.e. the light does not stay turned off.
For example, a shock like bumping the magnetic tailcap on a metallic surface will cause the battery inside to jump at the (+)contact disc and lose momentarily electrical contact. Instead of the light turning off and staying off, we witness a singular light "flicker". One can reproduce this behaviour with a paper clip experiment: remove the tailcap, short the tail contact with a paper clip, turn the light on, interrupt the tail contact for a millisecond, and you'll see a flicker instead of continued black darkness, but interrupt the tail contact for 1 second (or longer) and you'll see continued black darkness and you need to press the switch to turn the light back on.
I use the tailmagnet's function often. So whenever i magnestick the light to an attracting surface (with a little shock), i would witness the singular flicker. That made me think and conduct that paper clip experiment.