We are taking the experience of gaming many times and many tables. After we know what rules got used, both official and house rules, we are attempting the measure the complexity of the games. Is this not practical and constructive?
Insofar as we don't actually have the data, and no realistic way to gather it properly, no. We actually have rulebooks that don't depend on gathering data from statistically relevant samples of people.
Every mass published RPG has thousands of pages of material beyond the core mechanics.
Define "mass published".
I think saying *EVERY* one has *THOUSANDS* of pages of material is nowhere near accurate - either a poorly considered statement, or unreasonable levels of hyperbole. Unless the only games that qualify as "mass published" are D&D, Pathfinder, and White Wolf, I suppose.... but then "every" is really "three", and your point isn't so strong.
It's simply not possible to use every book for every game. "Throwing out large chunks of it" is literally the only way to play any professionally made RPG.
First it is "mass published" now it is "professionally". Please settle on one definition of what games we may or may not consider here. Otherwise, intentionally or not, the goalposts are going to tend to move. If you feel a need to limit what games we consider, let us lay that out early.
It sounds like you want a very academic discussion of complexity using a specific set of game texts. I don't think that's possible, simply because that's not how RPGs are played. Discussion of complexity is only valid if we include real life experience.
And it sounds like you are somehow of the opinion that the real life experience is significantly different from what's in the text. I am of the opinion that if the rules of a game are complicated, leaving out some of the most egregious parts still leaves you with a game that's complicated in play.
I don't think, "the game as written is complicated, but in the majority of play, it turns out not to be complicated," is a real scenario.