I've heard/saw some inkling of the high readings, but I find it hard to dispute Rick and Rodney's initial calibrations based on many high end commercial ANSI/NEMA FL-1 tested flashlights. My PVC lightbox is a manxbuggy1 original, and we worked together on the calibration, and even using the same exact flashlights. I'll grant you 10% is well within the tolerance, but I rarely, if ever, get a reading higher that claimed, usually lower, again, on name brand supposed FL-1 tested modern lights. I'm still using the same meter and occasionally regression test it to my old lights and still seems on. I hope I've been in the ballpark because the Q8 rating is based on my readings, and I thought djozz's readings and mine were close at the time.
The LK cells do pretty darn well, though I think the VTC5D is the current 18650/26650 king, edged out by the 30T though. I'l like to see what the 25S does, available here: https://www.18650batterystore.com/18650-p/samsung-25s.htm. The old Samsung 20R used to be the hot cell, and was hoping for a 30R or 30S for a long time, but it never happened.
I agree this light does much lower than spec'd. Seems they, like many, just make the #'s up. In theory, doubling up the FET's should do better but I never heard of those FET's before, so could be low performers. Also fairly long 20 AWG for 3 low Vf LED's will be a bottleneck, but maybe those springs are the bigger bottleneck - same here, never saw springs like that before.