The two people, (manxbuggy and his brother) did their best and built a contraption that produced miles better results than what our eyes can do, but they simply did not have the knowledge and equipment to make it measure lumens very accurately. If we want to use the lumen unit to describe the amount of light coming out of a flashlight, I think it is a good idea to approximate the lumen as closely as we can, and maukka has the equipment (professional sphere and verified calibrated tungsten light source) to get closer to the truth than with the tube-style light measuring devices and an average over a number of flashlight manufacturer’s claims (whose measurements can hardly be called independent, and I pretty much doubt how serious their equipment and method is).
I really think that your equipment is adequate as a way to measure light output of most types of flashlight, and as long as your calibration (whatever calibration in fact) is constant over time, it is an useful way to measure differences between flashlights, but if the aim is to approximate the lumen (which is a S.I. unit), maukka’s approach is simply way more accurate.