A theory about this: the light uses a microprocessor, and the minimum voltage for it to run is 1V. I think that when the voltage drops below that, the designers at LiteFlux thought that although it would still work, it would become unpredictable and might even randomly rewrite the EEPROM etc. So they made the light turn off at 1V, and allowed the user as a compromise to use the light again by doing the "rest the cell for a while" trick.
It's not really battery protection, but microprocessor protection with a nice side benefit.
One way to confirm this is to pull the light apart and see what processor it uses. Anybody up for it?
Also, I've heard reports that the "battery" protection doesn't work for 14500s, as in it only cuts off at 1V, which by then the cell is damaged. Can you confirm this?
Very interesting theory rotototo. I'd love to hear from Khoo or LiteFlux to confirm/deny. Recalling earlier threads in which the challenge of reliably running a MCU torch off 1.5V power source, I think that the microprocessor runs at much more than 1V, but below that the step up can't reliably supply the voltage it needs. But then again, I'm not a torch engineer!