An interesting point - there’s no other games where people argue about player agency.
That's because no serious card player argues that five hundred has as much agency as bridge; or no serious traditional game player argues that backgammon has as much agency as chess or go.
Leaving aside the ability to measure or intuit the spin of the wheel, I assume that few casino-goers argue that roulette has as much agency as blackjack.
In these well-known games, the contrast between "lighter" games that involve more luck and less decision-making, and "serious" games where player decisions play a bigger role in determining outcomes, is relatively obvious.
For some reason, there are RPGers who want to argue that players exercise agency over the game although it is someone else's decisions that are determining outcomes. That's the prima facie puzzle. The solution to it is that they are treating the GM's decision-making as if it were a non-agentic mechanism.
Classic D&D has many techniques - including the implicit obligation on the GM to stick to their notes - to try and render the GM a non-agentic mechanism. Lewis Pulsipher was a great advocate of this in his White Dwarf articles in the late 70s.
It seems obvious to me that a GM who departs from those techniques, and who makes decisions about outcomes based on what "makes sense", is an
extremely agentic decision-maker, quite different from the classic model. That many RPGers deny this difference seems to me deeply bound up with play traditions that go back to at least the mid-80s.