Player agency is often argued about on these boards. This post will not attempt to reconcile such arguments. It will enunciate the definition of agency that I use and the reasons I do so.
Agency defined
Agency in games is the product of inviolable rules which the players know and can rely on to achieve known goals.
I’ll look at each of these elements in turn – but it’s important to note that these elements don’t stand in isolation. They work in unison. They must all be present to create agency – the absence of even one of them is fatal.
Removing the elements of agency
If we set up a chess board but now I tell you that the pieces will move in ways I decide from moment to moment, you can no longer move. You can only suggest moves you’d like to make and see if I approve. You can't take my pieces, you can only ask permission to take my pieces. You can't checkmate me unless I choose to be checkmated. All you can do is ask my permission and I can accede or deny - thereby forcing you to ask permission for something else. And I can always deny permission. You’re no longer playing chess – you’re an advisor, suggesting ways for me to enjoy my own game. My game world, my rules, my stories.
Similarly, if we set up a chess board and start to play and I play d4 and you go d5 and then I play c4 and you accept the gambit and go to play dxc4 and I tell you that the realistic result is that you lose your d-pawn, play becomes futile. If the rules of chess are mutable – based on my conception of realism or aesthetic preference or whim – rather than reliable, again, you can no longer play chess. You can thrash around moving pieces, hoping I don't change the rules. But you have ceded your agency over what happens in the game.
If we set up a chess board and now I tell you that the goal of chess is no longer to checkmate the king, you can still make moves. It's just that making moves is worthless. Is the goal to be the first to take all your opponents pawns, or both their knights, or to lose all your pieces? Or to have more pieces than the opponent on the b-file on move 7? Who knows? Without knowing, there is no game - you can move pieces for no purpose and nothing more.
Interestingly, in all of these cases you can still pretend to play chess. You can cos-play - dutifully asking permission and moving ‘your’ pieces as if you have agency – when in fact you have none: I make all the meaningful decisions. This describes the vast majority of rpg play – which is no surprise, since we’ve imported traditional rpg processes into chess.
RPGs
Most RPG play I’ve seen fails to include at least two of the four building blocks of agency. Significant amounts of play I’ve seen, and seen described, simply doesn’t feature any of them at all. For example, most RPGs fail to:
RPGs and goals
One critical area of divergence between RPGs and most other games are that many RPGs don’t clearly state the goals of play. Chess defines the goal – checkmate the opponent or run out their clock. Football sets goals as the, well, goal. In tennis it’s points, to win games, to win sets.
RPGs very rarely define the goal of the game. Instead, it invites participants to create characters, and it is the goals of the characters which define the goals of the game. However, this creates an opportunity for GMs to exercise all the agency - while pretending the players have it – by controlling the goal of the game.
It’s not enough to say the goal of the game is ‘fun’. This can be the goal of any game – tennis or chess. But the goal of play in each of these games is different. One may have fun playing, but in playing you are trying to achieve a different goal – checkmate or clock the opponent, score points. Those are the goals of the game.
It isn’t enough to say the goal in D&D is to level up. This is like saying the goal of chess is to improve your rating. Without a goal for a game of chess (checkmate, clocking the opponent) there is no way to win a game such that you improve your rating. D&D is the same – levelling up is a measure of your success at achieving goals. It is a tortology to also claim that levelling up is the goal. You can only level up through instances of play – so what is the goal of each instance of play?
A character in an RPGs need someone to author their goal. The person who authors the goal for a character is the person who exercises agency over that character.
Dungeon-crawling, with no GM agenda, and no ‘plot’ or ‘story’, can easily have player-authored goals – I’m going to leave this dungeon alive and with as much treasure as I can carry. I’m going to own a +2 sword or die trying. Ironically, or perhaps tellingly, it’s easiest without a GM – I did so nearly 40 years ago with a few friends and the random tables at the back of the AD&D DMG, and more recently with the roll and write game Four Against Darkness and boardgame League of Dungeoneers.
Jump forward a few decades from random-table D&D and you can play games such as Burning Wheel and Apocalypse World, which (using different techniques) prompt the players to author the goals for their characters and then defines the role of the GM to make achieving the player-authored goals for each character exciting and difficult.
However, I see little traditional play where the goal has been authored by a player. There’s very little where the goals are even known by the players – instead they are withheld in the name of ‘mystery’ or ‘tension’ or ‘realism’.
What often happens is the breadcrumb trail – endless instructions to go here and do that, and then there and do this, and then being attacked by this thing, and ambushed by that thing. Fetch this, kill that, uhoh there’s a skeleton army on the way. You must destroy it. Fairly soon you don’t need the overt instructions anymore – it’s implicit.
Even when this doesn’t happen the secondary presumption is that the players will choose from a menu of goals pre-scripted by the GM. These are still the GMs goals.
D&D doesn’t ask you to create characters with stuff they need to achieve. It asks you to create blank slates waiting to serve the GM. The GM gives you a hook and you are expected to bite. This is the ceding of your agency – of the infinite things your character could want to achieve and choose to pursue, you are expected to do the things and go to the places and talk to the people that the GM has written this adventure about. D&D gives primacy to the GMs world and the GMs prep. It makes the players, through their characters, subservient to both.
In games featuring player agency, the starting point is the creation of characters with their own purpose and their own goals. I may have been outcast for my sorcery, but I will become King. I will avenge my murdered wife, taken from me by an apparition which appears only during a lunar eclipse. The world is secondary. Only once we know the characters, can we understand the world we need to create to bring these exact characters to life. When the players have agency, the world serves the characters, not the other way around.
But my players have agency!
Assuming the truth of a conclusion isn’t persuasive. Agency in chess, in tennis, in poker, in bridge, in Monopoly, comes from them having inviolable rules, which are known and which allow the player to reliably achieve a known goal.
The fact that a roleplaying game is played in the imagination doesn’t change the requirements for agency. We can play chess without a board and pieces, purely in the imagination too. In the absence of inviolable rules which the players know and can rely on to achieve known goals imaginary chess falls apart exactly the same as over the board chess.
The imaginary component of RPGs does, however, provide additional cover for claims of agency in games which feature very little.
Common arguments used to argue that players have agency include:
1: Pseudo-resolution processes
When I talk about resolution which players can rely on that means that the outcome is codified into the rules. Taking a piece in chess is an outcome I can rely on. But moving your piece and asking me, the GM, to tell you what happens is completely devoid of reliability. It reliably produces the outcomes the GM chooses – and that is all. Roll a dice and the GM describes what happens gives the same player agency. None. Of course, it is also beloved of illusionist / railroading GMs.
Much of D&D functions in this way – roll a dice, add a number from your character sheet, and I tell you what happens.
Systems like Burning Wheel and Apocalypse World have codified outcomes. But – and this is critically important – those codified outcomes work in tandem with player-set goals. The elements of agency don’t work in isolation.
So, a common argument is to try and focus in on a moment of resolution – such as a die roll and resulting narration - in a game with agency and a game without. Frequently they look quite similar. GMs from games dictated entirely by them say ‘Aha! it’s all the same’. What they elide is that when the GM has already set the goal – your characters are all tasked with killing A or solving B or discovering C – the resolution process is immaterial. If your goal has been set for you, play is all about doing what you’re told, not what you want.
A game can feature a player-authored goal such as ‘get rich’. But if that game doesn’t feature codified mechanics for acquiring wealth (Monopoly, Acquire, 1830 all feature such mechanics) there’s no agency. It’s the combination of goals and mechanics that are vital. Try playing Monopoly where another player (let’s call them the GM) has complete control of the game’s money supply, asset availability and asset prices – and come back and tell me how much agency you have.
2: ‘Real life’
People keen to maintain the illusion of agency in their roleplaying games often try to equate gameplay agency with 'real life' agency.
Real life has no person who chooses the weather. It has no person who chose to create mountain ranges or oceans. No person who decides every building in your local town, or everyone who lives within a 10-mile radius of you, or how close the nearest policeman is right now. These things are not determined by individual agency - even though in some cases they may be the accretion of dozens, or thousands, or billions of human decisions.
For these things to be included in a game someone has to decide them. Making stuff up is not analogous to the real-world’s deterministic factors. Nothing about who makes stuff up makes any of it more realistic. Denial of your agency to make stuff up is nothing to do with 'realism' - it's simply denial of your agency to make stuff up. The primacy of GM world-building and GM prep is given cover by claims of ‘realism’ and ‘consistency’. Both realism and consistency are entirely possible (and, in my experience, overwhelmingly likely) in games featuring player agency.
3: ‘Character’ Agency
This goes back to the chess board with the mutable rules and hidden goals. Just because the chess pieces move doesn’t mean you have agency. I can let you move pieces and still retain all the agency by controlling the property of the pieces and the goals of the game. Claims that characters doing things in the fiction are, in themselves, evidence of agency are smoke and mirrors.
4: Perfect Information
Some people take the definition 'inviolable rules which the players know and can rely on to achieve known goals' and insert their own strawman of 'perfect information' or 'perfect situational awareness'.
Nothing in the definition precludes games from featuring hidden information. Poker, for example, is predicated on exploiting hidden information for advantage. So is Bridge. Both are considered games with player agency under the definition. In both cases, the rules are known and agreed and abided by. The rules of Poker can be relied on by players to achieve the goal of winning hands - either by bidding more than others are willing to or by having the best hand at showdown. Likewise, you win at Bridge by making bids about how many tricks you can win in a hand and then scoring points for doing so (or losing points if you can't).
Indeed, hidden cards are essential in poker and bridge – without that they are solvable mathematical exercises. But in both cases the rules are not hidden or negatable, the reliability of outcomes cannot be undermined by any participants and the goals are clear for each player.
Aargh! This guy is attacking my playstyle!
You might think that. I’m not. I’ve played and run many years of low-to-zero player agency RPGs and had plenty of good times doing so. Observations over 40 years suggest agency isn’t a particularly valued currency in RPGs – the vast majority of players are perfectly willing to cede it in return for subjective experiences such as mystery or immersion.
I’ve also played many high player agency RPGs and had plenty of fun doing so. They have an entirely different feeling of creativity and collaboration which generate their own mystery and their own immersion.
However, there is a very clear difference between the two – and if I want one, I’m going to bounce off the other really hard.
The problem isn’t the playstyle – the presence or lack of agency. However - the claims about it, the illusion that the players’ actions matter in games where any or all of the goals of play, permissible moves, processes of play or results are held by the GM, are untenable.
If player agency matters to you, keep your eyes open for inviolable rules which you know and can rely on to achieve your goals. Watch for GMs with opaque processes, resolution which doesn’t give you any actual say in your character’s outcomes, GMs that expect you to agree to them setting your character’s goals for you.
Agency defined
Agency in games is the product of inviolable rules which the players know and can rely on to achieve known goals.
I’ll look at each of these elements in turn – but it’s important to note that these elements don’t stand in isolation. They work in unison. They must all be present to create agency – the absence of even one of them is fatal.
Removing the elements of agency
If we set up a chess board but now I tell you that the pieces will move in ways I decide from moment to moment, you can no longer move. You can only suggest moves you’d like to make and see if I approve. You can't take my pieces, you can only ask permission to take my pieces. You can't checkmate me unless I choose to be checkmated. All you can do is ask my permission and I can accede or deny - thereby forcing you to ask permission for something else. And I can always deny permission. You’re no longer playing chess – you’re an advisor, suggesting ways for me to enjoy my own game. My game world, my rules, my stories.
Similarly, if we set up a chess board and start to play and I play d4 and you go d5 and then I play c4 and you accept the gambit and go to play dxc4 and I tell you that the realistic result is that you lose your d-pawn, play becomes futile. If the rules of chess are mutable – based on my conception of realism or aesthetic preference or whim – rather than reliable, again, you can no longer play chess. You can thrash around moving pieces, hoping I don't change the rules. But you have ceded your agency over what happens in the game.
If we set up a chess board and now I tell you that the goal of chess is no longer to checkmate the king, you can still make moves. It's just that making moves is worthless. Is the goal to be the first to take all your opponents pawns, or both their knights, or to lose all your pieces? Or to have more pieces than the opponent on the b-file on move 7? Who knows? Without knowing, there is no game - you can move pieces for no purpose and nothing more.
Interestingly, in all of these cases you can still pretend to play chess. You can cos-play - dutifully asking permission and moving ‘your’ pieces as if you have agency – when in fact you have none: I make all the meaningful decisions. This describes the vast majority of rpg play – which is no surprise, since we’ve imported traditional rpg processes into chess.
RPGs
Most RPG play I’ve seen fails to include at least two of the four building blocks of agency. Significant amounts of play I’ve seen, and seen described, simply doesn’t feature any of them at all. For example, most RPGs fail to:
- Establish that the GM is no more privileged to break rules than anyone else
- Provide transparency to the players about how their actions will be resolved
- Provide play processes which resolve the conflicts created by the game
- Provide transparent goals for players
- They do not treat rules as inviolable (for the GM)
- There is no reliability in resolution for key elements of gameplay – instead the assumption is the GM can at any time interpose their own agency to resolve situations
- There is an assumption that the GM creates ad-hoc resolution processes – such as ‘I’ve decided what is realistic here’ or ‘I’ve decided on two possible outcomes and will roll a dice’.
- They assume the GM sets goals in secret
RPGs and goals
One critical area of divergence between RPGs and most other games are that many RPGs don’t clearly state the goals of play. Chess defines the goal – checkmate the opponent or run out their clock. Football sets goals as the, well, goal. In tennis it’s points, to win games, to win sets.
RPGs very rarely define the goal of the game. Instead, it invites participants to create characters, and it is the goals of the characters which define the goals of the game. However, this creates an opportunity for GMs to exercise all the agency - while pretending the players have it – by controlling the goal of the game.
It’s not enough to say the goal of the game is ‘fun’. This can be the goal of any game – tennis or chess. But the goal of play in each of these games is different. One may have fun playing, but in playing you are trying to achieve a different goal – checkmate or clock the opponent, score points. Those are the goals of the game.
It isn’t enough to say the goal in D&D is to level up. This is like saying the goal of chess is to improve your rating. Without a goal for a game of chess (checkmate, clocking the opponent) there is no way to win a game such that you improve your rating. D&D is the same – levelling up is a measure of your success at achieving goals. It is a tortology to also claim that levelling up is the goal. You can only level up through instances of play – so what is the goal of each instance of play?
A character in an RPGs need someone to author their goal. The person who authors the goal for a character is the person who exercises agency over that character.
Dungeon-crawling, with no GM agenda, and no ‘plot’ or ‘story’, can easily have player-authored goals – I’m going to leave this dungeon alive and with as much treasure as I can carry. I’m going to own a +2 sword or die trying. Ironically, or perhaps tellingly, it’s easiest without a GM – I did so nearly 40 years ago with a few friends and the random tables at the back of the AD&D DMG, and more recently with the roll and write game Four Against Darkness and boardgame League of Dungeoneers.
Jump forward a few decades from random-table D&D and you can play games such as Burning Wheel and Apocalypse World, which (using different techniques) prompt the players to author the goals for their characters and then defines the role of the GM to make achieving the player-authored goals for each character exciting and difficult.
However, I see little traditional play where the goal has been authored by a player. There’s very little where the goals are even known by the players – instead they are withheld in the name of ‘mystery’ or ‘tension’ or ‘realism’.
What often happens is the breadcrumb trail – endless instructions to go here and do that, and then there and do this, and then being attacked by this thing, and ambushed by that thing. Fetch this, kill that, uhoh there’s a skeleton army on the way. You must destroy it. Fairly soon you don’t need the overt instructions anymore – it’s implicit.
Even when this doesn’t happen the secondary presumption is that the players will choose from a menu of goals pre-scripted by the GM. These are still the GMs goals.
D&D doesn’t ask you to create characters with stuff they need to achieve. It asks you to create blank slates waiting to serve the GM. The GM gives you a hook and you are expected to bite. This is the ceding of your agency – of the infinite things your character could want to achieve and choose to pursue, you are expected to do the things and go to the places and talk to the people that the GM has written this adventure about. D&D gives primacy to the GMs world and the GMs prep. It makes the players, through their characters, subservient to both.
In games featuring player agency, the starting point is the creation of characters with their own purpose and their own goals. I may have been outcast for my sorcery, but I will become King. I will avenge my murdered wife, taken from me by an apparition which appears only during a lunar eclipse. The world is secondary. Only once we know the characters, can we understand the world we need to create to bring these exact characters to life. When the players have agency, the world serves the characters, not the other way around.
But my players have agency!
Assuming the truth of a conclusion isn’t persuasive. Agency in chess, in tennis, in poker, in bridge, in Monopoly, comes from them having inviolable rules, which are known and which allow the player to reliably achieve a known goal.
The fact that a roleplaying game is played in the imagination doesn’t change the requirements for agency. We can play chess without a board and pieces, purely in the imagination too. In the absence of inviolable rules which the players know and can rely on to achieve known goals imaginary chess falls apart exactly the same as over the board chess.
The imaginary component of RPGs does, however, provide additional cover for claims of agency in games which feature very little.
Common arguments used to argue that players have agency include:
1: Pseudo-resolution processes
When I talk about resolution which players can rely on that means that the outcome is codified into the rules. Taking a piece in chess is an outcome I can rely on. But moving your piece and asking me, the GM, to tell you what happens is completely devoid of reliability. It reliably produces the outcomes the GM chooses – and that is all. Roll a dice and the GM describes what happens gives the same player agency. None. Of course, it is also beloved of illusionist / railroading GMs.
Much of D&D functions in this way – roll a dice, add a number from your character sheet, and I tell you what happens.
Systems like Burning Wheel and Apocalypse World have codified outcomes. But – and this is critically important – those codified outcomes work in tandem with player-set goals. The elements of agency don’t work in isolation.
So, a common argument is to try and focus in on a moment of resolution – such as a die roll and resulting narration - in a game with agency and a game without. Frequently they look quite similar. GMs from games dictated entirely by them say ‘Aha! it’s all the same’. What they elide is that when the GM has already set the goal – your characters are all tasked with killing A or solving B or discovering C – the resolution process is immaterial. If your goal has been set for you, play is all about doing what you’re told, not what you want.
A game can feature a player-authored goal such as ‘get rich’. But if that game doesn’t feature codified mechanics for acquiring wealth (Monopoly, Acquire, 1830 all feature such mechanics) there’s no agency. It’s the combination of goals and mechanics that are vital. Try playing Monopoly where another player (let’s call them the GM) has complete control of the game’s money supply, asset availability and asset prices – and come back and tell me how much agency you have.
2: ‘Real life’
People keen to maintain the illusion of agency in their roleplaying games often try to equate gameplay agency with 'real life' agency.
Real life has no person who chooses the weather. It has no person who chose to create mountain ranges or oceans. No person who decides every building in your local town, or everyone who lives within a 10-mile radius of you, or how close the nearest policeman is right now. These things are not determined by individual agency - even though in some cases they may be the accretion of dozens, or thousands, or billions of human decisions.
For these things to be included in a game someone has to decide them. Making stuff up is not analogous to the real-world’s deterministic factors. Nothing about who makes stuff up makes any of it more realistic. Denial of your agency to make stuff up is nothing to do with 'realism' - it's simply denial of your agency to make stuff up. The primacy of GM world-building and GM prep is given cover by claims of ‘realism’ and ‘consistency’. Both realism and consistency are entirely possible (and, in my experience, overwhelmingly likely) in games featuring player agency.
3: ‘Character’ Agency
This goes back to the chess board with the mutable rules and hidden goals. Just because the chess pieces move doesn’t mean you have agency. I can let you move pieces and still retain all the agency by controlling the property of the pieces and the goals of the game. Claims that characters doing things in the fiction are, in themselves, evidence of agency are smoke and mirrors.
4: Perfect Information
Some people take the definition 'inviolable rules which the players know and can rely on to achieve known goals' and insert their own strawman of 'perfect information' or 'perfect situational awareness'.
Nothing in the definition precludes games from featuring hidden information. Poker, for example, is predicated on exploiting hidden information for advantage. So is Bridge. Both are considered games with player agency under the definition. In both cases, the rules are known and agreed and abided by. The rules of Poker can be relied on by players to achieve the goal of winning hands - either by bidding more than others are willing to or by having the best hand at showdown. Likewise, you win at Bridge by making bids about how many tricks you can win in a hand and then scoring points for doing so (or losing points if you can't).
Indeed, hidden cards are essential in poker and bridge – without that they are solvable mathematical exercises. But in both cases the rules are not hidden or negatable, the reliability of outcomes cannot be undermined by any participants and the goals are clear for each player.
Aargh! This guy is attacking my playstyle!
You might think that. I’m not. I’ve played and run many years of low-to-zero player agency RPGs and had plenty of good times doing so. Observations over 40 years suggest agency isn’t a particularly valued currency in RPGs – the vast majority of players are perfectly willing to cede it in return for subjective experiences such as mystery or immersion.
I’ve also played many high player agency RPGs and had plenty of fun doing so. They have an entirely different feeling of creativity and collaboration which generate their own mystery and their own immersion.
However, there is a very clear difference between the two – and if I want one, I’m going to bounce off the other really hard.
The problem isn’t the playstyle – the presence or lack of agency. However - the claims about it, the illusion that the players’ actions matter in games where any or all of the goals of play, permissible moves, processes of play or results are held by the GM, are untenable.
If player agency matters to you, keep your eyes open for inviolable rules which you know and can rely on to achieve your goals. Watch for GMs with opaque processes, resolution which doesn’t give you any actual say in your character’s outcomes, GMs that expect you to agree to them setting your character’s goals for you.
Last edited:
