I'd guess the person was using high capacity NiMH where the charge is here today, gone tomorrow. There are some lights that do drain. The Fenix TK40 comes to mind. I'd guess any drain from mode memory would be of the order of micro- or nano-amps so not worth worrying about. Protection circuits in lithium batteries do cause drain, but again, not a lot.
Thinking about this, in most torches the switch is mechanical and when it is off, there is no path for current to flow, so it cannot drain cells as the switch is what completes the circuit and allows current to flow. Some electronically switched ones doubtless do drain cells, but most such things are not aimed at us budget enthusiasts. Doubt anyone here is in the market for multi-hundred dollar custom lights.
Thanks, I kind of thought that the mechanical switch in off means there is no path for energy flow. There's a fairly expensive (~$35.00) AA flashlight, I forgot the brand, which is made for long runtime, and it has a "soft" electronic switch, and they say that it does slowly drain the battery. I doubt that's the case with the Ultrafire C3 SS.
Here's another question though: How does the Akoray K-106 maintain its mode programming during a battery change? As far as I can tell mine does retain its settings.
This is just a guess, but these days I'd have thought a tiny micro-controller with a few bytes of flash memory integrated into the controller. This would probably be the simplest. Certainly my K-106 has never reverted to factory mode despite having been left for over a week with no cell in it.
I suppose it could be DRAM and a capacitor to keep it alive but I'd guess that refreshing DRAM would soon eat the charge - certainly less than a week. SRAM used to be horribly expensive, but used little or no power and a few bytes of it in a microcontroller is not going to add much to the price.
These are guesses - not keen to rip apart my K-106 to see what components are on the driver board, as I sorta prefer it working.
I have a Casio "Data bank" watch that stores phone numbers, and it is supposed to have a "Memory Protect" function that keeps your contacts intact when the battery is changed. Must be something similar to that.
A watch probably needs only a few minutes of protection and since it already has some form of RAM in order to store phone numbers, it probably just uses a capacitor to cover for when the battery is absent. A clinically dead CR2032 can probably keep memory going for a rather long time if it doesn't have to drive the watch display as well - I seem to remember nanoamps being mentioned in this context.
The only parasitic light that i have heard about is the Nitecore D10......dont know at to what extreme the drain is since the piston is always making contact with the brass ring. Too me, when the light is off the circuit is broken, no power going in at all.