Howdy!
It has to do with a combination of factors that all come in to play. At the end of the day, LF is better off recommending against the use of "clicky" switches to reduce the liability... Here's some components of the equation that went into that decision:
When LF did IMR bulb tests with clicky switches, they did extended continuous runs, allowing everything to get hot in the process. They managed to destroy numerous SF clicky switches in their continuous run tests. They used the IMR-M6 and IMR-M3T in these tests primarily as I understand, which both do draw more than an SF HOLA (not considering M6 bulbs)..
Many li-ion powered configurations will run longer than a CR123 powered configuration under a similar load because the larger li-ion cells have better capacity, so those continuous runs are much longer than what a stock SF configuration would deliver, which adds up to more heat buildup in the switch from resistance. Consider- A pair of IMR18650s will drive an IMR-9 or IMR-M3 for upwards of 30-35 minutes, a set of 3 CR123s would only run that same bulb for ~15-20 minutes. Technically speaking, the IMR series of lamps *can* be run on many LiCo cells as well, provided they are large enough to safely deliver the current, this was probably taken into consideration- A pair of modern high capacity LiCo 18650s will drive that IMR-9 for roughly an hour. Obviously, for overall safety, it would be best not to run a high powered compact incan for that long, as the heat would likely make the flashlight body impossible to hold, and put the cells at risk of explosion in worst case scenario, reduced useful life at best case scenario..(which is why there are "heat" warnings and continuous run limits recommended on many higher power LF bulbs as well)
One of the primary factors that went into the decision to recommend against clicky switches is likely the low internal resistance of LiMn chemistry cells and the effect that has on peak current flow when starting from a cold filament. With CR123s, a cold filament will never be able to get more to flow from those cells than slightly less than they would deliver into a dead short, which isn't all that much since CR123s have an impedance of ~0.25 ohm per cell, (~0.75ohm for a stack of 3 in series). An 18650 size LiMn chemistry cell has as little as ~0.05 ohm resistance (0.1ohm for a stack of 2 in series), which translates to potential current spikes on a cold filament many times greater than can be delivered by CR123s. There is noteworthy resistance in those clicky switches, anywhere from ~0.1-0.3 ohm depending on the condition of the switch is pretty common. Sudden failure from high current spikes seems possible to me especially when working with the larger LiMn cells (like 18650s).
----
All that said, many people will probably use clicky switches and not have any problems. I have been using the clicky on a modern E2E for a little while now with a pair of IMR16340s and the IMR-E2 lamp and have not had it fail yet. I am accepting that I am taking a risk of damaging the clicky. I try not to use it for more than a few minutes at a time.
Some configurations, like, say, a pair of IMR16340s driving an IMR-9, actually have less running current and runtime than 3 CR123s driving a P91, I don't see any reason why one shouldn't feel comfortable running that configuration with a Z59. I guarantee that Mark@LF would agree that this configuration would technically be "fine" in a private "off the record" conversation with a known educated individual, but maintain the broad sweeping general recommendation against all clicky switches speaking on behalf of LF. From the perspective of LF, much like any company, it's much better to simplify recommendations, making simple broad sweeping recommendations rather than try to have a complicated reference to determine how the products can be used. It's the job of someone like you or myself to come along and read between the lines and fill in the blanks for folks who would like to have a broader understanding but without increasing the liability for LF in any way since we are operating completely independently.
To illustrate an example of this self-preserving behavior that must be used in business... email SF and ask them if you can run 17500s in the 9PLED. More than likely, the person you talk to won't even know what a 17500 is, but regardless of whether they do or not, they will recommend only using SF brand cells in their flashlights, which insures their survivability by reducing liability for warranty work, or injury, or property damage caused by alternative configurations. If SF gives the "go-ahead" to the use of various 3rd party add-ons or modifications, then SF is opening themselves up to unwanted and un-deserved liability. LF is essentially doing the same thing here. Many of us here on CPF are well aware that a quality set of protected LiCo 17500s in that 9PLED would be both safer and more economical than CR123s.
----
I think an educated CPF member can probably make an informed decision about which configurations are suitable for use with the clicky switches and get by pretty well. Obviously, lamps with a running current that are lower than SF HOLAs, that you intend to use with cells that won't deliver radically high potential current spikes (protected LiCo instead of LiMn) and only used in short "bursts" of say, 5 minutes or less, will very likely never be the culprit in a destroyed clicky switch. Higher current lamps, like the IMR-M3T, used on IMR18650 size cells, do have a pretty good chance of damaging a clicky switch in my opinion.... Keep in mind that they are ~3.4A compared to the ~2.4A of a SF HOLA that is intended to be used with CR123s.
SF used to sell a bulb for the M4 called the N62 IIRC. It was basically an IMR-M3T/WA1111/64250 category bulb, drawing enough current to sag those poor CR123 cells down into that ~1.75V per cell range. Runtime was horrible and the safety wasn't too great with the cells. I am pretty sure the reason it was eliminated was probably the safety issue at the cells, but I also wonder how many clicky switches they blew out on those bulbs.
---
I did find a definite guaranteed way to destroy a solarforce reverse clicky...
2xIMR18650s>64275 (well over 6 amps). 2 clicks was all it took.
-Eric
