I guess what I mean by player agency is the sense that the player's decisions are allowed to affect and drive play.
Rules that are followed rather than vetoed are one way of increasing that agency.
But I don't think that rules should be followed 'despite what the fiction allows' or 'even when it doesn't make sense'. I don't even recognise that dilemma, barring edge cases or very poor rulesets. As I said earlier in the thread, I think that it's trivially easy for a GM to find an in-universe justification for denying a player's action. And I think it's often just as easy for a creative player to find an imaginative reason they should allow it. Ultimately any GM is either looking for reasons to say yes or looking for reasons to say no. I prefer to say yes unless there really is a very good reason to say no. I don't find that 'the GM came up with a reason to say no based on imagined factors within the gameworld' to be any more realistic or plausible than 'the player came up with a reason to say no based on the same'.
I don't think increased player agency and narrativism are necessarily synonyms but they do naturally fit together. One can increase player agency in any game without spilling over into narr though. I think it's still a label or dial that has some value. I can imagine gamist or sim play that also features a lot of player agency but it's in the service of reflecting/realising the setting or creatively solving obstacles or the like. Maybe the higher ranges of player agency are inherently narr or narr-adjacent though.
I wanted to circle back to this earlier comment.
I appreciate the thoughtful response, and I understand your position.
I think that where we differ is that I don't find "player agency" to be a useful descriptor
between systems. Especially when the usage is completely different
within one sphere (say, people who play D&D) than it is within another sphere (say, people who are playing BiTD).
There are a number of reasons for this, but the easiest one to point to is the same one we keep seeing in these conversations. Some games have very tightly integrated rules and processes of play that (for example) authorize the authoring of fiction and are designed specifically so that "the player's decisions are allowed to affect and drive play." More importantly, they almost always have a tight control of setting and genre (BiTD is a great game for heists in Doskvol, but it is not such a great system if the players want to fictionally author a portal to Victorian England and engage in some Wuthering Heights shenanigans) that allows for that.
On the other hand, there are games that, for a number of reasons, are much more agnostic about ... well, everything. From assumed setting, to assumed genre, to even questions about the way to DM responds. In those cases, it is easy to point to examples of dysfunctional gaming and say, "D&D has low player agency because there was this time that a DM once said no to something that was completely reasonable!"
Which is true- and that can suck. But that's really just a conversation about how varied the styles of play are within D&D, and also the collateral conversation of how D&D is often the entry point for new gamers- and a lot of new DMs have trouble learning the game.
IME, the major difference is these different games is that it still fairly rare for people to pick and play the indie games as their first game, and if people are playing it "badly" or "incorrectly" they will often bounce right off of it (or be told, since it is a tightly integrated ruleset and process of play, that they are, in fact, 'doing it wrong,' which is more difficult to do with D&D). Perhaps we will some day reach a point when these games are widespread enough among newbies, and have been played for so long, that we will get to hear regular horror stories. That's not a bad thing, by the way!