good questions.
There's a few reasons I lean towards the LSD NIMH over the li-ion in this application but allow me to say that it's a very close toss-up between the 2 options and you can go with either.
1. You can buy 4 NIMH LSD cells for the cost of 1 good protected 14500.
2. The runtime on either will be very similar.
3. The li-ion will be about 12.5% brighter than a LSD NIMH, not really enough to notice but worth a mention. There are some 14500/AA compatible lights on the market that gain significant output with a 14500 and are designed to use the 14500 (jetbeam has one that really screams on a 14500, 225 emitter lumen on the li-ion, 130 on a regular AA, this would be enough of a difference for me to lean towards recommending the 14500 because it would be so darn cool to have a light that small pushing a cree that hard!).
4. Protection circuitry is another failure point and is one more thing that can "go wrong." When the LSD NIMH option for a flashlight works basically "just as good" as a li-ion option, I tend to lean towards the LSD recommendation for the combined factor that the NIMH chemistry is safer, and could be considered more reliable in some ways.
-------------------------------------------
On the subject of alkalines:
Little 1 cell pocket rockets have to pull a fairly significant amount of juice (current) from the battery in order to boost the voltage up to run the LED. Say it's pushing the LED at 1.5W (estimated), it would have to pull something like 1.7W from the cell in order to makeup for efficiency losses in the boost process, Alkaline cells may say 1.5V on the package, but in reality, under a load, they deliver much less, quickly falling to below the 1.2V that most NIMH can maintain above through the majority of the run. As the voltage drops, the circuit has to pull more and more current to continue pushing the same wattage. At 1.2V, it needs 1.4166 amps to get 1.5W to the LED. AA size Alkaline cells basically fall on their face at anything over about 0.75 amp, so you can pretty much bet on a high powered compact AA light having very poor runtime when used with alkaline cells. Alkaline just has too much internal resistance and will waste most of it's available power as heat within the cell. At high loads, NIMH cells outperform them in available capacity by leaps and bounds.
The circuitry will try it's best to keep the light running full brightness on any power source you give it within the accepted voltage range, but most of those circuits start to dim when the source power drops below about 1.1V, some can stay in regulation down to around 0.9V. Not sure about the Olight. So with an alkaline cell, you'll get some constant regulated output (short lived) and then it will be all down hill from there as that voltage drops down too low...
for comparisons sake, taken from SilverFoxes battery shootouts:
A sanyo eneloop at a 1 amp discharge delivers about 2.32 watt-hours capacity.
The top performing alkaline, a duracell ultra, at 1 amp discharge, came in at 1.27 watt-hours capacity. most brand came in at around 1 watt-hour or less. And in all of the tests, a substantial amount of that energy is delivered below 1.1V, which could mean a lot of diminishing output. I've seen output/runtime charts for a number of 1xAA powered lights, and on alkaline cells, they all wind up dropping below maximum output long before the cell is used up.
Now, if you use the light in lower output modes, alkaline cells will perform better and be able to maintain those lighter loads "reasonably" well. Alternatively, the T25, with 2 cells, would only have to draw about 0.7A to drive the LED at that same hypothetical 1.5W, the amount of useful power available from the cells would rise substantially, making alkaline cells slightly more practical, and since even when the cells drop down to say, 0.9V, the combination of the 2 is till 1.8V, plenty to keep the light in regulation.
------------------------------------
Hope that clears a few things up
Eric