It feels like goalpost shifting most of the time.
@hawkeyefan or myself or somebody else will make a set of statements operating under a constrained, workable for internet discussion set of either caveats or assumptions, and then Robert will go "well we can't really discuss
that, because we need to expand to The Entirety of Play" or something.
Like, since he keeps bringing work into things: where I work we do a lot of analysis of highly complex human factor systems. We state assumptions up front that kinda level set the discussion that "we're not able to model everything here because of variation." So things like "we're using
the rules of the game as a starting point for discussion of agency because we simply cannot account for all the house rules or bit and bobs added in."
If you want to further add in "some styles of play within the same ruleset allow for greater or lesser agency because of xyz" as has been pulverized around sandbox vs non-sandbox play here, cool! You can speak from teh assumption of "given the same ruleset, player-chosen direction/priority of play within a GM provided sandbox gives more agency then AP play because..." and we're on the same page.