I'm sorry but the flashlightreviews article is just wrong. It keeps referring to lux when it is really talking about candlepower. It is easy to confuse since lux is what his meter displays. The key point is - when talking about the lux of a light, you must specify the distance. For example a light that reads 100 lux at 1m would read 10 lux at 10 m. Lux is not the output of a light - it is a measure of how much that light illuminates the object it is shining on. Lux is the important measure for profesional lighting designers and photographers. If you are designing a lighting system for a football stadium, you don't care about the light itself - you care about evenly lighting the field to a certain specified lux level.
Consider 2 perfect lights. Both are 100 lumens. Both have a perfect square beam that is even with sharp edges. One has a 5 degree beam and the other has a 10 degree beam. If you shine both lights at a wall, with the 5 degree beam 11.5 meters away and the 10 degree beam 5.75 meters away, both lights will make identical 1 meter square 'spots' that are 100 lux (100 lumens/square meter). However the candlepower of the 5 degree beam is 4x that of the 10 degree beam since the light is concentrated in a slimmer beam that has 1/4 the cross sectional area at a given distance.
If you'll take a closer look at the entire website, so that
you understand the context in which that article was written, you'll find that he is correctly talking about lux readings. Heck, the very principles that you are using here
are the basis for his measurements. :shakehead
In fact, Doug uses the inverse square law to derive his throw ratings, which is one of the most useful techniques I've seen for objectively evaluating lights. And it's all documented on his website.
For the most part on CPF, when you see lux measurements, they are taken at 1 meter. Otherwise, the measurements are meaningless, as you point out. After a while, though, most of CPF understands that 1 meter is the commonly-accepted distance to measure at, so it's often left off.
That article assumes all measurements are taken at the same distance. I think that if you re-read it, keeping that assumption in mind, that you'll find it's correct.