What is Player Agency?

[Quertus]

So, we've all heard the term "Player Agency" bandied about, often in relation to the dreaded "railroad". But what does it really mean? What do you believe it means for players to have "agency"? Is it measurable? Is it something you can test for?

This thread is intended for us to express our ideas about the phrase, and for us to compare our understandings and examine our assumptions at a level of detail that may make some uncomfortable.

Viewer discretion is advised.

[Anxe]

Player Agency is the degree to which each player feels that their decisions and input have an impact on the events, outcome, and tone of the game. High Player Agency games like Dungeon World allow the players to actually build the world that their PC resides in. Low Player Agency games like Paranoia are a mockery of what a proper roleplaying campaign should be, but that's kind of the point (And even this is an overgeneralization of what Paranoia is).

[Psikerlord]

Isnt player agency mostly about the players getting to make meaningful decisions. It's a spectrum, more agency in a sandbox campaign, less in a linear adventure path, but both have some.

[PhoenixPhyre]

To me, agency (player or otherwise) is the ability to make meaningful choices.

That requires
+ Freedom: There have to be choices to be made. While watching a movie, you have no freedom to change the events. No agency. Rare that this happens in an RPG though.
+ Consequences: different choices have to have different effects. If a parent asks "do you want vanilla or chocolate ice cream" but, no matter what the child chooses, gives them vanilla, there was no agency. Only a false choice. As a corollary, you have to have the ability to choose wrongly. If every choice goes the right way, no matter what you chose, were you really making meaningful choices?
+ Knowledge: You must be able to (at least with some level of surety) be able to predict the possible consequences of actions. If every action involves a roll on a d1000 table ranging from success to "and the world explodes," there's no real agency.

[Nifft]

Player Agency = impact of player decisions on the game.


That can mean a lot of different things.

For example, one game might have players control powerful PCs whose in-game choices have consequences which resound through the setting.

Another game might have low-power PCs, but give players direct control over setting elements -- allowing players to declare truths which shape the setting, and which their PCs then leverage to some advantage.

A third game might be about low-power PCs without any special player control of the setting, but the player's in-character decisions matter a lot because the setting has no specific plotline, and the NPCs react in a reasonable way to the disruptive actions of the PCs -- some for, some against.


One way to think of Player Agency is to imagine the opposite of a Rail Road game.

[Max_Killjoy]

To me, agency (player or otherwise) is the ability to make meaningful choices.

That requires
+ Freedom: There have to be choices to be made. While watching a movie, you have no freedom to change the events. No agency. Rare that this happens in an RPG though.
+ Consequences: different choices have to have different effects. If a parent asks "do you want vanilla or chocolate ice cream" but, no matter what the child chooses, gives them vanilla, there was no agency. Only a false choice. As a corollary, you have to have the ability to choose wrongly. If every choice goes the right way, no matter what you chose, were you really making meaningful choices?
+ Knowledge: You must be able to (at least with some level of surety) be able to predict the possible consequences of actions. If every action involves a roll on a d1000 table ranging from success to "and the world explodes," there's no real agency.

That's very close to how I'd describe it... and I'd include a tie-in to verisimilitude. That is, player agency is both in part and part of the "fictional reality" (the settings, the NPCs, etc) reacting to the PCs' actions as if it and they were real.


And more controversially, my personal definition goes on to assert that the PC is the player's "interface point" with the fictional reality, the one part of that fictional reality that the player controls. This makes the player the "soul" of the player character, and it makes the PC's inner feelings, desires, thoughts, and choices sacrosanct and inviolate, with other players (GM or otherwise) only able to intrude with the individual player's explicit permission to do so. Don't mind control or mechanically social control or otherwise hijack the PC of a player who does not enjoy that aspect of the gaming endeavor.

.

[Cluedrew]

To me, agency (player or otherwise) is the ability to make meaningful choices.

I think you got it, between this and the follow-up. I've been trying to add something meaningful to it but besides weird corner case notes (for example, in a less challenge based game the ability to choose wrongly becomes less important) I have very little to say. Good description, you could probably apply it to non-role-playing games as well.

[Anxe]

I'm curious whether people define Player Agency by actual events or just feelings. Using the ice cream example. Say a person has never had vanilla or chocolate and doesn't know the difference. They choose vanilla and are given chocolate. The person eats what they ordered and are satisfied.

Same thing occurs in roleplaying. DM designs an adventure for goblins and tells the players they can fight the kobolds or the goblins. The PCs choose the kobolds. The DM uses the same adventure and because kobolds and goblins are almost the same the DM doesn't change anything. The players are satisfied and feel like their choice mattered. They influenced the narrative (kobolds instead of goblins), if not the actual mechanics of what they interacted with in the game.

[Nifft]

I'm curious whether people define Player Agency by actual events or just feelings.

Actual events.

It should be possible for a disinterested 3rd party to read through a game later and determine the degree of player agency.

Same thing occurs in roleplaying. DM designs an adventure for goblins and tells the players they can fight the kobolds or the goblins. The PCs choose the kobolds. The DM uses the same adventure and because kobolds and goblins are almost the same the DM doesn't change anything. The players are satisfied and feel like their choice mattered. They influenced the narrative (kobolds instead of goblins), if not the actual mechanics of what they interacted with in the game.

That's just illusionism.

It's why multiple-choice questions are lower-agency indicators than free-form or open questions.

[Scripten]

That's basically the idea of the Quantum Ogre, which is not nearly as sneaky as some DMs seem to think and is a violation of player trust. If you want to give your players choices, then those choices should matter.

[GreatKaiserNui]

I can't wait for Darth Ultron to crash his way into this thread like Kool-aid man and then ramble his way through a subject he has no comprehension of.

/S

[Glorthindel]

That's basically the idea of the Quantum Ogre, which is not nearly as sneaky as some DMs seem to think and is a violation of player trust.

Like anything, this is the point where the idea of agency jumps the shark and goes screaming off in to the woods with its underpants on its head.

Agency is a tool, not a goal.

The goal is to give the players a fun, exciting, and engaging game. No other goal matters. The role of "Agency" is to facilitate that, because players tend to have more fun, and be more engaged when their actions and ideas matter (and I say tend, because that is not an absolute rule, some players just want to watch a story unfold). Setting "Agency" as the goal is putting the cart before the horse - providing agency at all cost serves no benefit if the players are not having fun, and aren't engaged.

It should be possible for a disinterested 3rd party to read through a game later and determine the degree of player agency.

I am not interested in the opinions of third party campaign-monitors who are going to go through my DM notes and mark me out of 10 on "Agency" or any other false goal. The only people that matter to me are the people around the table. If they are happy, I have succeeded at what I aimed to do. If they have not, I have failed. If they leave the table beaming and chatting excitedly I do not care how much or how little agency I provided, because that is not the purpose of us sitting down together.

For that reason, I strongly dispute the collective sniffing being done at illusionary agency. It is just as an acceptable tool as real agency if it achieves the goal of helping the players enjoy the session. And the fact is, illusionary agency has its place. Barring the mythical "perfect DM" who can create intriguing fun plots and events on the fly (and they do exist, but expecting every DM to be that is setting yourself up for disappointment), a DM has a finite amount of time to create material for his session. I have no issue with invoking the quantum ogre if the parties decisions cause them to sidestep a fun, involved encounter that I know they would enjoy. Especially if they will never realise that their decision should have caused them to miss the encounter. To me giving the players that enjoyment is more important that slavishly adhering to the god of agency. Claiming it is "a violation of trust" is outlandishly hyperbolic - the players "trust" you to give them a fun gaming session, not to adhere to some carved-in-stone code of DM behavoir that places agency up on a pedestal above all other things. Yes, agency often does contribute to giving them that fun session, but there are times when it can form a straight-jacket that hinders giving them the very best session they could experience, and then, it should be sacrificed without hesitation.

[Floret]

I'm curious whether people define Player Agency by actual events or just feelings.

Both, I think. Or rather, actual agency by events; percieved agency (Which is the more important one) by feelings, but... well. (Agency defined as "player control over what happens; ability for player decision to matter")

See, one of the things I personally highly value in RPGs, that makes me feel like I have agency, is rolling the dice. Not having dice rolled, but rolling them myself. One classic GM tactic (Rolling perception rolls and the like for the players) feels utterly bizzarre to me, and if done by a GM I at least feel uneasy. Heck, I even really, really dislike systems like DnD or L5R, where for an attack, the attacking faction rolls on a static value, and if the GM rolled well enough you are hit. (Though I have no problem with my players hitting me with their rolls; and also dislike the slug of active defense systems. Some problems I have.)

Technically, whether I roll or the GM does it for me, changes nothing. Nothing about the outcome is different. If I, as a player, rolled on a defense value instead of the GM on the same attack value, nothing would be different. But it feels differently. When the dice are in my hand, I feel like I am in control.

Does it violate player agency if the GMs planning or the Adventure Path or w/e writes in one possible solution (though others would theoretically be possible), and the players instinctively choose that one? I think you'll be hard pressed to find someone arguing it does; even though, looking through the results, they followed the "rails" exactly. So it can't be following a pre-planned path precisely that is the problem.*
The thing about the quantum ogre is: If it is done in a way the player's don't notice, they feel perfectly fine and like their agency is kept intact. The problem is, of course, players tend to be the creative sort and come up with all sorts of things - and you can't quantum ogre everything without someone noticing.

It's a difficult question. Everyone has their own preferences, and limits - I know I have (and have seen) weird ones. I've had players ask for input on what their characters think; I've consulted the dice for the question of "does my character give in to this temptation" and would do it again.
I also happen to believe that player agency is not the be-all and end-all that it is made out to be sometimes, and can be violated, if agreed upon by both sides (In session 0, by knowing GM styles beforehands, etc.). I think percieved agency is indeed the important part, to have it in the places you want it and agreed upon it being; but that percieved agency is nigh-impossible to uphold without actual agency being given.

(*One game of Shadowrun had me get a bit nervous, because the player's actions were diverging so far from the expected, I'd have to really, really think around if I wanted to keep what run should be next (Though in a pinch I would have changed my planning). Up until the last minutes of the run, where the mages' player suddenly comes up with the perfect follow-up - that lead directly into the original plans I had. Out of nowhere. I would never have suggested that myself, I didn't get the idea he came up with.
So... effectively the players re-joined the railroad, out of their own free will, without any prompting from myself. Player agency can be a weird creature sometimes.)

Also, Glorthindel's post comes pretty close to some of my own feelings on the matter.

[NichG]

On the one hand, I'm all for keeping separate an understanding of what tools can be brought to bear to make a good campaign, versus particular philosophical preferences surrounding which of those things are intrinsically 'good'. On the other hand, there is a danger in going all-in on the illusionism path, which is that generally speaking DMs can easily overestimate how well their illusions will hold up. Experienced players will have at the very least a good subconscious sense about which choices they have to be careful about and which choices correspond to 'what color would you like your armor to be?' - and that can give away the game very easily.

The kobold/goblin example in particular is a really dangerous example for the DM, because even if they haven't specified anything about kobolds and goblins in their setting yet, its a sure thing that many of the players at the table will have encountered kobolds and goblins in fiction and in other campaigns, and will have some kind of mental image associated with what makes them different. So when the DM pulls a quantum ogre and makes goblins into crafty trappers who prefer to lure their prey into elaborate death-traps, at worst the jig is up and at best the DM has still introduced a small amount of dissonance that will make the players feel less confident that they get what's going on - something that can build up to the game as a whole feeling incoherent if it persists.

It's okay to use this device, but use it with care.

Anyhow, in the end this sort of thing has very little to do with real agency, because choices made without an understanding of their significance don't really constitute a form of agency. The feeling of agency comes from in essence seeing that one is able to shape the future according to one's will in the present. If there's a choice where you don't know what the futures you're choosing between will be like, that produces feelings of uncertainty or fear or unease rather than feelings of agency. It's like the difference between striving and succeeding - agency is looking back and seeing that indeed you were able to exert some measure of control successfully. The quantum ogre is an illusion that only works when you can't know the difference, so its basically got nothing to do with it.

[Frozen_Feet]

Player agency straightforwardly refers to the number of influential actions a player can take in the game.

In a roleplaying game where the player is focused on a single character, this is primarily a matter of scenario design: place the character in a scenario where the character has few possible actions, and you have low player agency. Place the character in a scenario where it has more possible actions, and you have more player agency.

If a player can control multiple characters or directly change the scenario, those up the number of possible actions and hence increase player agency. But it's wrong to think that games which allow such things are automatically higher agency. For determining agency, you are interested in quantity and impact of possible actions, not their abstract typing. For example, a player being allowed to decide whether it's day or night is not a sign of great player agency if that decision 1) does not impact the game or 2) is one of few, or only, possible action(s) available.

Player feelings have nothing to do with this. Player perception of agency is a separate issue and may differ wildly from what it actually is. The perception of agency may be influenced by player's lack of intelligence, player's lack of imagination, biases and pressures from other players, and other metagame concerns. On the extreme end, we can have players who are rules-wise omnipotent (infinite agency), such as a GM, being made to feel they only have one possible route of action due to peer pressure from other players.

I'll get to how you can test for agency when I have more time.

[Max_Killjoy]

Like anything, this is the point where the idea of agency jumps the shark and goes screaming off in to the woods with its underpants on its head.

Agency is a tool, not a goal.

The goal is to give the players a fun, exciting, and engaging game. No other goal matters. The role of "Agency" is to facilitate that, because players tend to have more fun, and be more engaged when their actions and ideas matter (and I say tend, because that is not an absolute rule, some players just want to watch a story unfold). Setting "Agency" as the goal is putting the cart before the horse - providing agency at all cost serves no benefit if the players are not having fun, and aren't engaged.

Speak for yourself.

I want my choices as a player to actually matter, not just appear to matter, even if the outcome is less "fun".

And if nothing my character does can avoid that pre-planned encounter with that ogre, then the game isn't fun anyway.

[Cluedrew]

Agency is a tool, not a goal.

The goal is to give the players a fun, exciting, and engaging game. No other goal matters.

I don't disagree but at the same time: name me a game that is fun, exciting and engaging game that does not involve meaningful choices.

Just yesterday I was reading about Chess 2, which was designed to introduce more agency into Chess, because at the grandmaster level there was not enough agency in the early and late game. It was too much pattern memorization and simple execution at that level. Apparently, I'm not a chess grandmaster and I avoided studying it to avoid that problem. Now I don't actually know if Chess 2 is any more or less fun than Chess, I haven't played it yet. But I thought it was an interesting example of a very different game that illustrates my point. Also I'm trying to figure out which if I'm supposed to capitalize Chess or not.

As for creating an illusion, you can do it for a time. But maintaining it always hits the same problem. The players are as about as intelligent as the GM and soon the illusion will break. Which is not to say every choice made in a game has to have deep & meaningful consequences, I think the have to be scattered in for the game to be fun. Or at least fun as a game, it could be fun as a movie otherwise but viewer agency is not the same thing, if it is a thing at all.

[PhoenixPhyre]

That's very close to how I'd describe it... and I'd include a tie-in to verisimilitude. That is, player agency is both in part and part of the "fictional reality" (the settings, the NPCs, etc) reacting to the PCs' actions as if it and they were real.


And more controversially, my personal definition goes on to assert that the PC is the player's "interface point" with the fictional reality, the one part of that fictional reality that the player controls. This makes the player the "soul" of the player character, and it makes the PC's inner feelings, desires, thoughts, and choices sacrosanct and inviolate, with other players (GM or otherwise) only able to intrude with the individual player's explicit permission to do so. Don't mind control or mechanically social control or otherwise hijack the PC of a player who does not enjoy that aspect of the gaming endeavor.

.

That part goes to knowledge--being able to predict in character what will happen. Now of course agency is not unlimited (no more so than it is in real life). You can give away your agency. If your character takes actions that have consequences such as dying, you don't get to say "Nope, that didn't happen" or "you're railroading me." If mind control is a thing in the setting, it can happen to PCs. I'm the nice type who tends to spare feelings, so I don't use it much, but it's still a thing that could happen, potentially.

I think you got it, between this and the follow-up. I've been trying to add something meaningful to it but besides weird corner case notes (for example, in a less challenge based game the ability to choose wrongly becomes less important) I have very little to say. Good description, you could probably apply it to non-role-playing games as well.

It's actually almost directly pulled from a discussion of moral agency (what's often called free will) in real life. Agency is agency. Same prerequisites, slightly different expressions.

I'm curious whether people define Player Agency by actual events or just feelings. Using the ice cream example. Say a person has never had vanilla or chocolate and doesn't know the difference. They choose vanilla and are given chocolate. The person eats what they ordered and are satisfied.

Same thing occurs in roleplaying. DM designs an adventure for goblins and tells the players they can fight the kobolds or the goblins. The PCs choose the kobolds. The DM uses the same adventure and because kobolds and goblins are almost the same the DM doesn't change anything. The players are satisfied and feel like their choice mattered. They influenced the narrative (kobolds instead of goblins), if not the actual mechanics of what they interacted with in the game.

This speaks to knowledge and consequences. If the consequences of the two choices are the same (down to the details), then it's a false choice. Making a choice without knowing the difference between the two is the same as flipping a coin. Either explain the difference (or, in the goblin/kobold case, the lack of difference) or don't give a choice at all.

Speak for yourself.

I want my choices as a player to actually matter, not just appear to matter, even if the outcome is less "fun".

And if nothing my character does can avoid that pre-planned encounter with that ogre, then the game isn't fun anyway.

I mostly agree, but sometimes, due to previous choices and their consequences, your character may have given up the agency required to avoid the ogre. That is, the consequences of previous actions have locked you into a path that bottlenecks on that ogre, and that's not a bad thing. Now if they're wildly different paths (going south through the forest or north through the mountains) and have the same encounter (not just with generic ogres who could have easily been on both paths, but with the exact same ogre in the exact same map, and you chose one path because there wasn't an ogre on it, then you have a problem.

For me, the problem with quantum ogres is the denial of consequences. If the party made other choices that are unconnected to said ogres, the presence or absence is irrelevant as to agency. If, however, they made choices specifically designed to avoid those ogres, and made them well, then dropping the ogres in regardless better have a darn good explanation (like teleporting ogres with a grudge, maybe. And that's still pushing it.).

[Cozzer]

+ Freedom: There have to be choices to be made. While watching a movie, you have no freedom to change the events. No agency. Rare that this happens in an RPG though.
+ Consequences: different choices have to have different effects. If a parent asks "do you want vanilla or chocolate ice cream" but, no matter what the child chooses, gives them vanilla, there was no agency. Only a false choice. As a corollary, you have to have the ability to choose wrongly. If every choice goes the right way, no matter what you chose, were you really making meaningful choices?
+ Knowledge: You must be able to (at least with some level of surety) be able to predict the possible consequences of actions. If every action involves a roll on a d1000 table ranging from success to "and the world explodes," there's no real agency.

This, with an emphasis on "knowledge" since it's the part that's often left out. If you put two identical paths, I follow the first one and something bad happens, you can't say "well you could have avoided that by taking the other identical path, so it's not true that you had no agency!".

I think one of the most common problems of GMs is how they believe that giving more informations to the players would make the game less interesting. In 99% of cases, a hard choice where the characters have most of the facts is way more fun and more interesting than a plot twist where the facts change choices already taken. This is, I believe, one of the biggest differences between standard narratives and RPGs.

[Actana]

To me, agency (player or otherwise) is the ability to make meaningful choices.

That requires
+ Freedom: There have to be choices to be made. While watching a movie, you have no freedom to change the events. No agency. Rare that this happens in an RPG though.
+ Consequences: different choices have to have different effects. If a parent asks "do you want vanilla or chocolate ice cream" but, no matter what the child chooses, gives them vanilla, there was no agency. Only a false choice. As a corollary, you have to have the ability to choose wrongly. If every choice goes the right way, no matter what you chose, were you really making meaningful choices?
+ Knowledge: You must be able to (at least with some level of surety) be able to predict the possible consequences of actions. If every action involves a roll on a d1000 table ranging from success to "and the world explodes," there's no real agency.

Going to second, third or whatever-th this post. To me, putting the idea of player agency concisely more or less results in the definition of "the ability of the players to make informed, meaningful choices".

If the players don't have the ability to make choices, there's no agency: the players are being told what they're doing.

If the choices aren't meaningful; if the results are the same whatever the players do, there's no agency: the players are simply making choices that don't matter at all, and everything is based on the whims of the GM.

If the players aren't able to make informed choices; if they have no idea what happens with any choice, there's no real meaningful choice to be had: the players need to be able to have some manner of discerning what choice they should take. It's okay to present the players with a situation where they have no knowledge of what might happen, mind you. It's just that you should at the same time always have some ways of acquiring that knowledge, even if those ways might fail. But at the same time if you force a choice where they can't predict the results, you can't blame them for anything bad that happens - indeed, you are to blame as the GM for that bad happening.


Whether these choices are in-character or out-of-character depends on the situation, as both can contribute towards having agency but in different circumstances. Character building choices are almost always out-of-character. Taking actions in-game is usually both. I don't really think there are many examples where you can take a purely in-character action, as the game's structure bleeds both of them into it. Not because metagaming or anything, but simply because players cannot 100% immerse themselves into their characters. There's always a part of the player driving actions as the player, which is perfectly fine. Some people prefer higher OOC activity, others higher IC. Same with games. But in all of those situations, the ability to make informed meaningful choices is what defines player agency for me at least.

[PhoenixPhyre]

This, with an emphasis on "knowledge" since it's the part that's often left out. If you put two identical paths, I follow the first one and something bad happens, you can't say "well you could have avoided that by taking the other identical path, so it's not true that you had no agency!".

I think one of the most common problems of GMs is how they believe that giving more informations to the players would make the game less interesting. In 99% of cases, a hard choice where the characters have most of the facts is way more fun and more interesting than a plot twist where the facts change choices already taken. This is, I believe, one of the biggest differences between standard narratives and RPGs.

This I 100% agree with. I tend to err on the side of giving more information. For example, I had two campaign arc options planned for a group. They're pretty much exclusive--either you go there and deal with those, or you go the other place and deal with the other things. So I talked to them OOC and gave them a brief rundown on the types of things they'd encounter and the basic terrain:

1) You'll be in the jungle, mainly facing beasts and wild tribes, leading up to dragons and ancient technology.

OR

2) You'll be in the forest/ancient ruins, dealing with orcs leading to undead leading to demons.

That way, they can make informed choices that are based on their character and on their own personal desires. After all, the point is to have fun.

[Cozzer]

Yeah, in most cases you don't even need to go OOC, I think. It's usually enough if the GM doesn't try to replicate these cool "everything you thought you knew was wrong!" moments they saw in their favorite books, movies or videogames. It's a pitfall even pretty good GMs tend to fall into, in my experience, because these moments are actually pretty cool and it would be great to replicate them. It's just that the RPG medium has very different strengths.

[NichG]

The neat thing with information is that it can be presented in a way that makes the player feel as if they're actually seizing agency rather than being given agency - associated with those moments of figuring out what the levers are which can move the world. This is something that works best alongside very open-ended situations, otherwise it risks being a 'guess the puzzle to advance' type of thing which has very little actual agency.

[PhoenixPhyre]

The neat thing with information is that it can be presented in a way that makes the player feel as if they're actually seizing agency rather than being given agency - associated with those moments of figuring out what the levers are which can move the world. This is something that works best alongside very open-ended situations, otherwise it risks being a 'guess the puzzle to advance' type of thing which has very little actual agency.

Agreed. My main group is very good with this, and the format allows them to decide to make sweeping changes. For them, I don't really plan campaign arcs much in advance. I know what their short-term goals are (because they've told me) and what they'll need (or are planning to do) to accomplish them. Beyond that? Depends on where they are.

They've done lots of things that totally surprised me, both session-to-session and more large-scale. I didn't expect them to try to found a fantasy UN, for example. That shaped a huge chunk of the campaign.

[Scripten]

Like anything, this is the point where the idea of agency jumps the shark and goes screaming off in to the woods with its underpants on its head.

Is it, though?

Keep in mind that the post I was replying to (in retrospect I should have quoted it) was talking about the players choosing between an adventure with goblins and an adventure with kobolds. The players have no knowledge of what either of these creatures are, which means that their agency is limited by lack of knowledge. Once they have done their kobold adventure, perhaps they later decide to go on a goblin adventure or they otherwise find out that goblins are the exact same as kobolds. Don't you think they would feel cheated?

At best, the DM could present goblins as something different during that later encounter, but in that case the choice actually affected how goblins exist in that DM's setting.

I am not interested in the opinions of third party campaign-monitors who are going to go through my DM notes and mark me out of 10 on "Agency" or any other false goal.

The point is that playing a smoke and mirrors game with your players only works for so long. If the players make choices that should give them one experience, then giving them a different experience, in spite of the confines of the setting's internal logic, breaks the consistency of that setting. PheonixPyre already broke this down quite well.

If you have two hallways and the players have seen that one leads to an ogre and another leads to a dragon, then they should encounter an ogre down one hall and a dragon down another, essentially. If you only prepared an ogre and can't come up with a dragon encounter on the fly, then don't present that illusory choice.

Claiming it is "a violation of trust" is outlandishly hyperbolic - the players "trust" you to give them a fun gaming session, not to adhere to some carved-in-stone code of DM behavoir that places agency up on a pedestal above all other things. Yes, agency often does contribute to giving them that fun session, but there are times when it can form a straight-jacket that hinders giving them the very best session they could experience, and then, it should be sacrificed without hesitation.

Can you present me with an example of where limiting agency leads to a situation that is more fun than allowing for it? Remember that providing players agency is more than just giving the players two identical doors and pretending there is an actual choice being made.

[Lord Torath]

Speak for yourself.

I want my choices as a player to actually matter, not just appear to matter, even if the outcome is less "fun".

And if nothing my character does can avoid that pre-planned encounter with that ogre, then the game isn't fun anyway.

This comes into the different definitions of "fun." In games you play in, it's very important for you to have a large amount of agency. If denied that, you're not going to have fun, and you're not going to go home talking about what a great session you just had. Anyone DMing for you needs to keep that in mind. In other groups, "fun" might not hinge so much on agency, but perhaps on having an exciting encounter. In both cases, a "fun" session is the goal, but the players may have different definitions of "fun".

As for the last part, I find the quantum ogre* works fairly well, as long as once the ogre waveform has collapsed, it stays there (where "wave form collapse" occurs the first time the party discovers information - first-hand or otherwise - that fixes its location). The ruins are in whatever direction the PCs start out in, until they talk to a scout who tells them the ruins lie to the south.

* For ogres of various species, quantities, and geological arrangements.

[PhoenixPhyre]

As for the last part, I find the quantum ogre* works fairly well, as long as once the ogre waveform has collapsed, it stays there (where "wave form collapse" occurs the first time the party discovers information - first-hand or otherwise - that fixes its location). The ruins are in whatever direction the PCs start out in, until they talk to a scout who tells them the ruins lie to the south.

* For ogres of various species, quantities, and geological arrangements.

Possibly rephrase this as "nothing is fixed until the information comes out in play?"

I tend to know that (say) there are ogres and dragons out there. I then reveal (if the players ask about local threats) that fact. When they take steps to avoid one of the two (and thereby accept the other), then the waveform collapses. I may not have fixed which one was north and which one was south, but once they decide "we're going south to avoid the ogres which we believe (based on information gathered) are to the north," then putting the ogres to the south would be a breach of agency in the absence of external forces such as a) a lying informant, b) ogres that are hunting the party explicitly and know of their movements, or something else like that. And any such thing would have had to have been telegraphed way in advance and very clearly.

[Lord Torath]

Possibly rephrase this as "nothing is fixed until the information comes out in play?"

I tend to know that (say) there are ogres and dragons out there. I then reveal (if the players ask about local threats) that fact. When they take steps to avoid one of the two (and thereby accept the other), then the waveform collapses. I may not have fixed which one was north and which one was south, but once they decide "we're going south to avoid the ogres which we believe (based on information gathered) are to the north," then putting the ogres to the south would be a breach of agency in the absence of external forces such as a) a lying informant, b) ogres that are hunting the party explicitly and know of their movements, or something else like that. And any such thing would have had to have been telegraphed way in advance and very clearly.

That sums it up very nicely! That's also the idea I always got from the Quantum Ogre article.

[Cozzer]

I don't know, I think it works for the "first act" of an adventure, where the characters are still more or less aimingless and you need to throw a few interesting things their way. But after that, every important thing that happens needs to be at least in part the consequence of something that has happened before, so the chances to use the quantum ogre strategy become fewer and fewer.

I mean, if you compare the campaign to a TV series, the quantum ogre is appliable only in the first three/four minutes of each episode (and, I don't know, maybe during the whole first episode of a season... you get the idea). You can say that whatever city the party travelled to is the city where they learn of the existence of a certain villain by happening to find one of his victims, but after that you can't say whatever city they travel to is the city where the villain's lair is located.

All of this is all IMO, of course, but the quantum ogre used more than halfway through a plot makes the whole thing feel like very railroady.

[Segev]

To me, agency (player or otherwise) is the ability to make meaningful choices.

That requires
+ Freedom: There have to be choices to be made. While watching a movie, you have no freedom to change the events. No agency. Rare that this happens in an RPG though.
+ Consequences: different choices have to have different effects. If a parent asks "do you want vanilla or chocolate ice cream" but, no matter what the child chooses, gives them vanilla, there was no agency. Only a false choice. As a corollary, you have to have the ability to choose wrongly. If every choice goes the right way, no matter what you chose, were you really making meaningful choices?
+ Knowledge: You must be able to (at least with some level of surety) be able to predict the possible consequences of actions. If every action involves a roll on a d1000 table ranging from success to "and the world explodes," there's no real agency.

This is a pretty good definition.

Agency is all about your ability to meaningfully impact the setting based on your character's capabilities. It's what separates a tabletop game with a GM from a cRPG: whatever the players can think of to have their characters do within their conceptual capabilities, they can attempt. No waist-high invisible walls, no "but thou must!" interactions, no only-two-choices-when-the-player-sees-a-third-option... they just try things. And the world reacts sensibly.

Lack of agency arises when the world is constructed to deny it (there is only one path forward and everything else is pre-designed to have literally no way through), the world alters to deny it (no matter what you try other than the approved track, it fails, or the world retcons to make it have failed), or you lack information to make meaningful choices. That last can happen in a game with agency, too, but you have to exercise your agency first to identify the ways to get information to make meaningful choices about other actions.