Interesting Olight S10R low battery behavior

So while I wait for my awesome Raw Copper S1 to arrive I decided to revisit some battery issues I had with my very nice S10R. Like many folks, shortly after the low battery led comes on, with the stock protected Olight RCR123, the light goes off totally and typically won't come back on. I always assumed that the (largely useless) warning comes on right before the protection circuitry trips. More evidence is that I usually have to take out the battery and put it in a charger because the charging dock won't usually start charging after this happens.

With a primary CR123a, the light usually indicates "low battery" right away in the High setting. After a while the light goes out but will come on in all modes and run for quite some time before eventually getting quite dim. Eventually the light indicates "low battery" in all modes but still operates in all modes. Conclusion, the low battery lamp is only relevant (though minimally so) with LiIon batteries.

I tried an experiment. I took a "crap" (orange wrapper, no name 350mA) IMR 16340 and charged it fully. I then put it in the light and let her rip on high. After about "350mA" worth of use on hi (~16-17 minutes) the low battery light came on. Then a curious thing happened. A few minutes later the light stepped down to "medium" and the low battery light went off. It wouldn't cycle back to high and continued to step down until the main light was flashing. It never went out altogether and still functioned on low and moonlight.

I let the battery rest for 10 minutes and measured voltage. Result: 3.05V. Pretty abused but not near the 2.75V point when I consider the battery worth disposing of. It took a charge with no apparent ill results (I may still recycle it given the low capacity and the fact that I have no recollection where I got it from).

I then took a "crap" (old as dirt Tenergy) LifePO4 16340 and put it in the the light fully charged. Here, the light immediately illuminated the "low battery" light (like a primary) and ran for a respectable amount of time (~20 minutes) before shutting down. Came back just like a primary and ran for a good deal more time before visibly dimming. I let it rest for 10 minutes. Result: 3.20V.

Not entirely sure what to make of this. Does the Olight have some special battery protection circuitry for ICR / IMR batteries? Should I retry the experiment with a "crap" Ultrifire blue wrapper (LiCO chemistry, unprotected, claims 1000mA...Hah!)????

Thoughts...

I think S10R battery indicator starts to glow red whenever it measures the cell voltage under load being lower than some value (3.0 - 3.6 volts maybe?). If a crappy, old, worn out cell has high internal resistance, its voltage will drop below that indicator light treshold when high mode is used that pulls a lot of current from the cell. After shutting down the light, the open circuit voltage of the cell immediately jumps back to a higher value because current is not being pulled anymore and internal resistance is not causing a voltage drop. So I guess extremely crappy cells can't even be fully discharged, at least in high mode, because the light will shut down due to not getting enough voltage because of the high internal cell resistance.
When I tried S10R with an almost empty li-ion, I could change modes normally by holding the button but the high mode just looked equally bright with medium. I couldn't detect a brightness increase between the two modes but I noticed that it stayed 2 seconds in that brightness instead of 1 second, which indicates that the both modes were still there.