Man, this is good stuff!
BOL, I'll bet the batteries that went down in the first few lights would still fire up a few of the contenders here. I'd save them in state for now and put them in some of the long runners here to eliminate the batteries as causers; just for kicks... That said, I would keep the test going unless all lights start to appear to cut out earlier than expected and the batteries are at the same voltage levels as the other ones.
Ayway (insert disclaimer here, bacause I don't know much about a few of these lights, like the liteflux, etc. and I am NOT bashing any lights!!!) one thing that I'm taking away right off the bat is that, for a "when the chips are down" light, I want the most basic twist on/off one level simple circuit light there is. Three levels, clickies, programmability, all great things and all work incredibly well, but they can and do impact reliability from time to time. Now, I am NOT saying that single level twisties don't fail, but the lights I've always had to send back for issues have been ones with extra do dads that make them nice to use lights, but don't necessarily improve the power to LED pathway. So, GO SINGLE LEVEL TWISTIES!!!!
I also think that running a test like this on my own personal stock might be the way to go to ensure the lights I have can go the distance... I've never really tried to see how far my own personal lights have gone... A good smoke test before I get one out for camping and it goes TU on me.