what sort of game is an RPG? part 2: goals

When we sit down to play an RPG, what are we trying to get out of it?

Last time we set the stage with various attempts to address the broad subject of games, and I brought up The Elusive Shift on players’ earliest attempts to describe their new game.

In this article, we’ll get into the big question that’s going to recur throughout this series: do goals factor into RPGs?

1. Roleplaying games and goals
   1. Vi Huntsman—RPGs as non-Suitsian
   2. C. Thi Nguyen—two types of play
   3. is this just GNS lite? etc.
2. Vincent Baker’s arrows
3. some points in RPGspace
   1. stance theory
   2. blorb vs quantum
4. how does the GM make decisions?
   1. kriegsspiel
5. So, is it Suitsian?
   1. what about the DM?

Roleplaying games and goals

There are many, many, many different species of roleplaying game, and the question of goals varies a great deal.

Naturally, the question comes up in The Elusive Shift. Peterson discusses how, in early RPGs, the first goals were articulated as trying become an overwhelmingly powerful ‘superperson’ through levelling up:

D&D does not specify any ultimate objective of its play—unlike its close imitator Tunnels & Trolls, which would state quite bluntly in its 1975 edition that “the true object of this game is to accumulate as many experience points as possible and by this means advance your first level character into as much of a superperson as you can.” Gygax in 1976 would say much the same of a D&D campaign: “progression, rather than winning per se, is the object” (SFF 87). If the life of a D&D adventurer may be compared to a game of pinball, then the experience point total for each character is the glowing score in the back box; players sneak glances at its steady rise throughout the game. The progression system of D&D implies a character arc: characters begin as inexperienced, weak, and undifferentiated yet will over time grow in power, gain confidence, and develop a personal history, if not a legend.

By this definition, there is an obvious ‘prelusory goal’ or ‘quantifiable outcome’: the number of experience points is higher. The ‘constitutive rules’ you take on are that you can only gain experience points by overcoming the challenges of the dungeon. Just as Suits observed, it’s not that players particularly care about this number in itself, but pursuing ‘number go up’ enables the activity of dungeon-exploring.

However, before long, Peterson observes, the ‘superperson’ goal became controversial. For example, in 1977 Peter Cerrato was complaining that the focus on experience points got in the way of players identifying with their characters. Games such as Bushido attempted to attach the incrementing score to more desirable behaviours, requiring that you behave like a proper samurai in order to level up.

Some players began to view trying to become a “superperson” as childish, the domain of younger “minimax” players flooding into the hobby…

In their early encounters with role-playing games, Seligman explained, “many players, and I admit that at one time I was one of them, had no other goal than to become as powerful as possible no matter what the means.” After he attained further experience with the potential of role-playing games, new vistas opened to him. “But is this sort of ultimate search for ultimate power the only form of interesting experience one can find in an FRPG? I feel the answer is no.”

Moreover, players and designers started focusing on other goals besides accumulating points by whatever means. As early as 1975, someone cooked up a system for accumulating ‘psychosis’ and losing ‘sanity’, so character progression could be negative. Others focused on the mystery of learning more about the world, or the backstory motivating a character, and devised ways to systematise them.

Nevertheless, in various forms, ‘progression’ systems have remained a near-universal element of RPGs, even up to the present. Usually these situations give the player characters more ability, i.e. the player more ways to narrate their character succeeding. Only relatively radical story games such as Fiasco, usually oriented-towards one-shot play, tend to eschew them entirely.

How does the goal of ‘make character more powerful’ relate to Suits’s story about the goal of ‘keep the game going’, or other goals that might be articulated, like Apocalypse World’s ‘play to find out what happens’?

Vi Huntsman—RPGs as non-Suitsian

Last year, I wrote a whole series inspired by a video by Vi Huntsman criticising certain PbtA-tradition games for their misuse of incentive systems and attempts to exhaustively define the possibilities in the game, something she views as a pernicious application of behaviourism.

At the time, my interest was mainly in trying to understand the relationship between the printed RPG book, and the actual game that emerges at a table. But as it happens, Bernard Suits came up in that account too—by way of the theories of C. Thi Nguyen, who built on Suits in his book Games: Agency as Art.

In her analysis, Huntsman discusses how Suitsian games have a standardised goal—similar to Salen/Zimmerman’s ‘quantifiable outcome’—against which players in different sessions can compare their decisions. If you’re cheating, Huntsman/Nguyen say, you’re not expressing Suits’s lusory attitude or really playing the game at all—cheating is beyond the bounds of Suitsian analysis.

(Suits indeed says as much, that ‘cheats are not playing the game’—and likewise that a runner who ran along a racetrack for another reason, and only coincidentally happened to follow it because it was the most efficient route to somewhere they needed to be, would also not be racing. This is in contrast to the discussion of cheats and spoilsports in Huizinga, for whom cheats have more respect for the magic circle than spoilsports who disregard it entirely.)

Suits’s frame, Huntsman says, considers a set of rules admirable when the constraints ‘consistently mold our play into a certain category of shape’. She quotes Nguyen’s book saying the reliability of game experiences being what the designer can take credit for. Huntsman immediately starts questioning where the designer’s partial credit ends. She says:

If Root has an ergodic contract, the rules are a finite list of actions that are ‘the game’.

If you break the rules, as a cheat does, your experience ‘exits the game’s artistic frame’.

Skipping forward in the video, at 2:12:20 she gets into the question of whether TTRPGs fit this definition of a Suitsian game.

Playing a TTRPG will inevitably involve goals, obstacles, abilities, and an environment, but I don’t think any of them match the Suitsian model at all.

This is in part because it’s not winnable, it doesn’t have a Suitian goal. I think at some point in the transition from The Grasshopper to Huntsman’s video, the discussion of ‘open games’ got elided, but in any case…

Huntsman further argues that, while the space of moves in a game of Root the board game is finite, this is not at all true in a TTRPG: there are endless possible situations that could come up, and there is no way that even a game like Root The RPG which tries to exhaustively define every special case could address all of them.

This is an interesting objection, and one we’ll get into later.

Huntsman’s ultimate argument is that the game designers in the post-Forge tradition assert a dubious claim at auteurship in which they take undue credit for the player’s creativity and actions that are not part of the space of agency defined by the game’s actual mechanical rules, through writing down ‘agenda’ rules and ‘GM moves’ that are more like vague suggestions; and that game design involves ideological claims of shaping player behaviour through forcibly constraining their actions, which she likens to ABA. We’ll get into that whole subject much later in this series; for now, I want to focus on the ‘are RPGs Suitsian’ question.

C. Thi Nguyen—two types of play

To support her argument, Huntsman cites a paper by Nguyen arguing that make-believe play is not Suitsian, which also discusses an alternative model of make-believe play advanced by Kendall Walton in Mimesis as Make-Believe—one that follows in the footsteps of Huizinga.

Nguyen’s paper argues that ‘striving play’ as described by Suits and ‘make-believe’ as described by Huizinga are two different things, and neither can be reduced to the other. For the reduction of make-believe to striving play, he takes on the argument by Suits we discussed above; for the reduction of striving play to make believe, he suggests that Walton’s follwers focus too hard on the magic circle and erase the aspects of games which are centred in Suits’s analysis.

For Suits, he says his basic claim is wrong: in a game of ‘cops and robbers’, children are willing to ‘die’ in-character when narratively appropriate, even though this will not extend the story, so this cannot be their goal. But even if the goal was ‘create the most dramatic story’ or something along those lines, this makes the concept too expansive: you’re playing for the sake of the dramatic story, not with the dramatic story as an arbitrary goal where the real activity is the effort to overcome voluntary obstacles in the way of telling it.

A useful test, Nguyen says, is whether you can cheat at an activity. If you can, it’s a striving game. An improv troupe, for example, could ‘cheat’ with audience plants, so they’re playing a Suitsian game. But if he was imagining some scene in his head, just for the fun of it, there’s no way to cheat.

Nguyen further argues that the spoilsport, following Huizinga’s distinction, is the test for the other kind of play. The cheat breaks Suitsian striving play, the spoilsport breaks make-believe play. Nguyen argues that you can’t be a spoil-sport with regard to chess by interrupting the game to ask about a rules clarification, because there is no illusion to break.

I don’t think I buy this particular argument. You could spoilsport chess very easily by moving the chess pieces around in invalid ways without trying to hide it, or tipping over the table, or even just making a loud noise nearby while the players are trying to play the game. Or, less aggressively, you could intrude some outside context that dissolves the magic circle, e.g. by ringing the bell that marks the end of chess club. If the players are really enjoying their game, could they plausibly accuse the person trying to close up the classroom a ‘spoilsport’?

Also, in exceptional cases, there are also some pretty wild examples in the history of chess that trouble this account of chess lacking a dimension of make-believe. When Soviet grandmaster Anatoly Karpov played defector Viktor Korchnoi at the World Chess Championship of 1978, the whole match was mixed up in elaborate geopolitical theatre. Karpov brought a hypnotherapist, Vladimir Zukhar, along to the match. Convinced that Zukhar was trying to hypnotise him, Korchnoi brought in two American yogis to psychically defend him; when this prompted Zukhar to leave, he felt vindicated. (He later accused the Soviet team of passing a secret message using yoghurt.)

The bizarre antics are considerably more remembered than the actual games of chess, and in a sense you could think of it as a roleplaying game between the two players and the spy agencies backing them. Couldn’t you act as a spoilsport here by coming in and saying hey, psychics aren’t real? Shit, maybe I just reinvented Umineko…

Moving on, Nguyen says many games are hybrid games, notably including RPGs:

When a group of players are playing a fantasy role playing game, like Dungeons and Dragons, it often seems that they care both about the pleasures of overcoming obstacles and the pleasure of imaginative absorption. We can see that these are distinct activities, because we can see that it is possible to break play in two entirely distinct ways. A player that quietly erases her character sheet and changes the amount of gold and experience points she’s accumulated is a cheat, but hasn’t shattered the imaginative illusion. On the other hand, a player that constantly calls for rules clarifications and accuses other players of not playing by the rules is not a cheat, but she is a spoilsport - she is shattering the shared illusion. In fact, there is a special term for such a person among game-players - they are a “rules lawyer”, and they are the opposite of a cheat.

It seems Nguyen’s account of a spoilsport is less broad than mine, and primarily means a player who interferes with the magic circle by violating the unwritten rules like “don’t ask for too many rules clarifications” considered by Sniderman. I prefer Salen and Zimmerman’s account, which would categorise this as ‘unsportsmanlike play’.

Nguyen argues that these two elements can complement each other. Make-believe is what turns a rubber sphere into a football and a patch of grass into a pitch, and thus creates a context in which a striving game like football can be played. But it’s not necessary to play within a magic circle, and you could also strive against real, physical obstacles, like throwing yourself into the wilderness to try to make your way out. (I’m not entirely sure I agree with this. While the physical hazards may be entirely real, the magic circle still defines a particular location as a ‘goal’, something which has no meaning outside the game.)

I think, while I might quibble with some of the details here, Nguyen’s frame still makes sense overall: there are at least two types of play that games can evoke, and Suits does not sufficiently account for the ‘make-believe’ better addressed by people like Huizinga. But while this conceptual division helps, it still leaves open a lot of space for how these two (or more!) types of play relate might relate to each other in any given game.

is this just GNS lite? etc.

It’s easy to identify Nguyen’s ‘striving’ and ‘make-believe’ with the traditional roleplayer dichotomy of ‘rollplaying’ and ‘roleplaying’, but honestly I think it’s a bit more nuanced than that.

If you’re versed in the history of trying to theorise RPGs, you might also make an identification with a taxonomy like the Forge ‘GNS’ or its precursor in the 90s ‘threefold model’, or perhaps if you’re more oldschool about it and read The Elusive Shift, you might go for Glenn Blacow’s four forms of roleplaying, wargaming, ego-tripping/minmax, and story-telling. (I’ll brush over exactly how Blacow defined these.)

With such a theory in mind, you might argue that ‘striving play’ is another word for ‘gamism’ or ‘wargaming’, and ‘make-believe’ is glomming together the pleasures of, say, narrativism and simulationism. However, in agreement with Vincent Baker’s more recent stance, I think you could come up with endless theoretical motivations:

But you and I could put our heads together and come up with interesting game dynamics all day long, limited only by our inventiveness as creators. Narrativism is exactly one of them. We’re supposed to divide the rest between gamism and simulationism? Why?

Vincent later links a blogpost by Eero Tuovinen which divides the GNS ‘simulationism’ into five further more or less incompatible ‘creative agendas’. While Tuovinen still believes in GNS (with a bit of work to repair Ron’s faulty description of ‘simulationism’) and attempts to justify the idea that the simulationist modes he describes are more compatible with each other than the other families, Vincent seems to be suggesting that the lines drawn by GNS are fairly arbitrary.

Right now I’m focused on the theory of Bernard Suits, and trying to find its limits. I’m not going for a complete taxonomy of all types of enjoyment to be found in a roleplaying game, except to consider how far these different approaches can be placed into the schema of arbitrary goal + voluntary restrictions to enable the activity.

Vincent Baker’s arrows

The last part of Nguyen’s paper (after a brief and rather uncritical summary of GNS theory) addresses a debate over whether representational game objects such as the gun in Duck Hunt are ‘fictional’ or ‘actual’. To the player, is the object in their hands a representation of a fictional gun that they are firing within the fantasy, or an actual light gun?

He argues that objects in a game can function as ‘actual’ if they are ‘recalcitrant’, ‘refuse to bend instantly to our will’ and thus enable striving play. But in general they’re both:

A World of Warcraft player who treats an in-game dragon as an obstacle to get over is focussed on the in-game properties of that dragon - its health, its range of motion, how much damage it can inflict. For the striving player, that dragon is actual. A World of Warcraft player who treats an in-game dragon as a prop can use it to imagine herself into the world of Warcraft - to imagine herself there, holding a sword, overcoming fear, to best the dragon. For this player, the video dragon is a prop for an act of fictive imagination. And, of course, many players do both, and so interact both with actual and fictional objects.

This last part of Nguyen calls to mind some old articles by Vincent Baker, written in 2009, on the way RPG mechanics connect the nebulous world of ‘the fiction’ to external mechanical representations like dice and character sheets. On the left, we have the game’s fiction, primarily the shared fiction that has been stated but also the private fictions in the head of each player. On the right, we have dice, character sheets, miniatures, and whatever other objects (jenga towers, for example) are drafted into the magic circle to hold some kind of game state.

Baker discusses rules as being like arrows which go from one to the other. For example, a rule that says ‘when you have zero hitpoints, your character dies’ goes from the tokens to the fiction; a rule that says ‘whenever you tell a lie, roll the Deception skill. If you succeed, they believe you’ goes from fiction to character sheet (read the Deception stat) to dice (roll the skill) and ultimately back to the fiction.

(If you were programming a computer game with an ECS architecture, you might characterise this as the components a system needs to query and mutate.)

This is on the old part of Vincent’s site, when he was still more engaged with the Forge, and his thinking might have changed since then. (This post from 2020 addresses a similar subject, more granularly.) But I find this model very helpful for understanding how RPG mechanics work.

The things that make RPGs especially interesting, and distinct from e.g. board games, are Vincent’s ‘fiction->tokens’ arrows, mechanics which change the state of the game mechanics based on the fiction. This means, returning to Nguyen’s account, that in RPGs the fictional things are in a sense ‘actual’: what the Forgeites call ‘fictional positioning’ is a large part of how the game evolves, and something you can strive against!

By contrast, in a board game, this isn’t true. There may be a fiction represented by the game, but only the state of the tokens is relevant to the game mechanics. I might say that my pawn is terrified by the bishop lurking nearby, but the game will play out no differently for that terror.

For Vi Huntsman, this is one reason why RPGs are not (her narrower definition of) Suitsian. The actions available to the player are anything that can be narrated, not limited to a list of actions defined by the game. Suits doesn’t have this restriction, so perhaps for him an RPG could fit his definition, just like ‘cops and robbers’.

But if there is ‘striving’ in RPGs, then it’s worth asking how it actually works, if not within a self-contained universe of rules… like, what might players be striving to achieve, what are the restrictions they introduce, and how much is that striving the focus of the game?

some points in RPGspace

To address this question, we need to understand how RPGs work. But this means we need to have a suitably broad understanding, not just focus on one playstyle of one family of games. So let’s briefly look at a few of the theories that RPG players have come up with over the years…

stance theory

One of the points of contention between the ‘story games’ school and the ‘old school’ is how the player should relate to their character.

The traditional stance, which I’ll gloss as ‘character stance’ is that the player identifies with their character and try to make them succeed as far as possible. The GM or referee’s task is to give them a suitably challenging environment to pursue their character’s goal as skillfully as they can.

An alternative stance, most often identified with the 2010s ‘story game’ movement, is that the player approaches their character as a writer would, and aim to create dramatically interesting events with them. At a limited end, they might choose a course of action that is mechanically suboptimal but narratively an interesting character moment. Pushing it further, in games with more distributed narrative authority, they might deliberately make trouble for their character.

Many elements of the ‘author’ stance are not all that new. Peterson cites numerous examples in the 70s of proponents of the oldschool mode, such as Kevin Slimak or Lewis Pulsipher, railing against roleplayers who wanted to treat the game like a novel rather than a game of skill.

For Slimak, considering D&D as a role-playing game encouraged players to make poor decisions when, in their estimation, the characterization system required it: “I’ve just seen too many people doing stupid things BECAUSE of the character they’ve rolled and are role-playing.”

DMs like Slimak, who were interested in throwing characters into a highly lethal dungeon where the players must figure out how to survive, found it frustrating to be confronted with knowingly suboptimal play. At another table, though, this might be a great example of co-creation of a story by the players. Players also came up with various intermediate approaches to resolve the tension. For example, the ‘intelligence’ stat could be interpreted narrowly as ‘magic skill’, so it would not oblige a player with low ‘intelligence’ to have their character make bad decisions.

In practice, in a lot of games, players slip freely between the different stances from moment to moment, making trouble for their character at one turn, and working hard to solve it at another. I might at one turn put my character in a compromising position where another character could overhear their secret for the drama of it, but when a fight breaks out, furiously rack my brain for the best strategy to win it. Some game rule structures may be more compatible with one or another approach, but many are flexible (or if you prefer, undefined) enough to work with both.

blorb vs quantum

These are jargon terms for preparation-focused, mechanistic games vs. highly improvisational games.

The term ‘Blorb’ was coined by Sandra Snan as a kind of manifesto; it’s sometimes also called Clockwerk though Sandra dislikes this. I’m not sure the exact timeline, but it can be contrasted with pieces like No Myth that originated with the Forge, originally coined by someone called Le Joueur.

At issue is the matter of what is decided in advance of the game, vs. improvised on the fly. Blorb is a limit case of defining as much as possible in advance of the game, or by procedures like dice rolls, to enforce as much as possible the sense that there is an underlying ‘reality’ that the characters are exploring.

‘Quantum’ approaches such as ‘No Myth’ establish less in advance, and improvise the story stochastically based on only what has been stated out loud. It is as if everything else in the world is in a state of quantum superposition of many possibilities, only ‘collapsing’ once it is observed and someone makes a decision.

A paradigmatic case: a character walks into a room and searches for dangers. A GM in a ‘Blorb’ game would look up their pre-prepared map of the dungeon and observe that there is a trap, and ask the player to make a ‘find traps’ roll. A ‘no myth’ GM might ask the character to make a ‘find traps’ roll, and if they get a bad outcome, retroactively decide that there is a trap (which might be defined as a ‘hard move’ in a PbtA game). Or on a good outcome, they might decide that the player found a trap but skillfully avoided it.

Either way, from the player perspective, the outcome might look the same: they walk into the room, make a ‘find trap’ roll, and the GM informs them that their charcter discovered/fell into a pit trap. However, on a strategic level, it interacts in a curious way with the character/author stances.

Towards the ‘Blorb’ end of the scale, rolling ‘find traps’ can only be a good thing for your character: if you fail, you are no worse than you started, and if you succeed, you might get important information. However, from an author stance, failing a ‘find traps’ roll when there is nothing to see could be seen as boring. You spend valuable game time on nothing happening.

Conversely, in a ‘quantum’ game, rolling ‘find traps’ is inherently risky in the ‘character’ stance, because on a failed roll, the GM might introduce something new and dangerous to hurt your character. But from a ‘writer’ stance, it makes perfect sense: both ways the roll turns out, something interesting is going to happen.

So you have a tradeoff: one mechanic/stance combo favours interesting things happening at every turn. The other works hard to foster a sense of an ‘actual’ world to push against—which may also aid a sense of fictional ‘immersion’.

As before, in practice, most games do not fit into either extreme perfectly. Even in a traditional, high-prep game, a GM will improvise and insert something plausible when the players inevitably interact with something beyond the material they’ve prepared. (Snan’s manifesto allows for this, as a step towards the ideal of full Blorb.) Apocalypse World, though it encourages a fairly ‘quantum’ approach overall, also instructs its MC to think offscreen and spend time between sessions writing out threat sheets to model various game world entities.

how does the GM make decisions?

Besides ‘what’s there’, ‘what happens’ is also a question that can be adjudicated more or less formally. Suppose that the fiction has established that NPC A is going to tell a lie to NPC B. The DM could simply decide, based on dramatic considerations, that B isn’t buying it. But they might prefer to roll dice based on A’s ‘deception’ stat… or, if the game doesn’t have a ‘deception’ stat, perhaps they just decide it should be a 50/50 chance and flip a coin.

However, even if a GM prefers to defer such a decision to dice, they would likely only follow the procedure when it’s dramatically interesting, and not try to simulate every single offscreen interaction of NPCs.

kriegsspiel

This question in fact long predates RPGs. Kriegsspiel refers to a family of wargames that originated in the 19th century as a military tool. The players, in the role of officers, give orders to their army with limited information. An umpire decides what happens in between turns and updates them on the carnage that ensued.

How should the umpire decide this? Early iterations of Kriegsspiel had elaborate, complicated formal systems for resolving the various things that could happen in a war. By the 1870s, a couple of German officers felt this was overly constraining, and proprosed that the umpire should have essentially absolute discretion to judge the outcomes however they want, while keeping the basic game structure of ‘giving orders, limited information’. This variant became known as Free Kriegsspeil, and it became popular even outside of the German military, still played to this day.

So in Free Kriegsspiel and most (but not all) roleplaying games, the umpire/GM can declare anything they please, but would they? A Kriegsspiel umpire would probably not declare that a UFO appears over the battlefield and abducts the player’s cavalry unit, and if they did, the players would not be happy unless they had agreed in advance that they were playing a scifi scenario.

So this is where we get into the subject of play culture and unwritten rules, which will be the subject of the next article in this series. These are important in every game, but they’re especially important in RPGs. That’s a big enough discussion to be its own article, so I’ll defer that for now.

So, is it Suitsian?

In Vi Huntsman’s narrower definition of a ‘Suitsian game’ as a closed system of formal rules with finite actions to take, Kriegsspiels and RPGs automatically wouldn’t qualify, because the actions available in the game are so open-ended. But Bernard Suits himself has a pretty expansive model of games that includes activities like ‘mountain climbing’, so he would probably be very on board with it.

Probably he’d frame it something like this: although the RPG as a whole might not have a clear, single ‘prelusory goal’, episodes within the RPG generate a lot of different goals. A prelusory goal in an old-school game is ‘referee declares your character succeeds’ (at killing the monster, finding the treasure, etc.). The constitutive rules define the ways you could persuade the referee to give you a favourable ruling, for example by describing in detail the procedure your character uses to search a room, or how they set up an ambush. The lusory attitude means you use these kinds of means instead of, say, bribing the referee or threatening them with a knife.

Meanwhile, the combat subsystem of D&D 3-5e is something a lot closer to Huntsman’s narrower interpretation. By and large, in these editions, players will focus on the abilities listed on their character sheets, which guarantee a certain probability of a certain game effects taking place. The DM’s actions are also more tightly constrained here: they are expected to use monster stats similar to those provided by the game books, and follow the combat rules for the enemy NPCs.

Mind you, a good D&D combat scene is not purely a Suitsian exercise of striving and optimisation within voluntary constraints. I am also considering narrative questions: what my character wants, opportunities for a cool stunt, what would express her vibe, and so on. This is an expression of the ‘play’-as-free movement that Salens and Zimmerman discussed.

I believe for most tables these Suitsian elements are a major part of D&D. Within the fiction, my character would be in a tight spot, and I would want a certain outcome on her behalf (most often, just that she stays alive!). Just as I would in a board game, I would carefully consider the abilities and resources she has, and try to come up with a strategy to win the combat.

This ‘manoeuvring to get advantage’ aspect is much less present in other RPGs, particularly tables which build their activity on story games like Fiasco or Microscope. In these games, the considerations on my mind are mainly narrative ones—by the stance theory, I am in ‘author stance’. When my turn to frame a scene comes around, I want to add an interesting element to the story. Although Fiasco has various mechanics to do with using dice for scene framing and deciding the fates of each character, they exist mainly to provide pacing and structure to the story, not to give you obstacles to strategise against. In D&D, you can try to ‘win’ a fight or ‘solve’ a puzzle, but there is rarely anything to win or solve in Fiasco.

what about the DM?

The players in the game may at least sometimes be trying to achieve goals in a Suitsian way, but what about the DM? They are, after all, also a player of the game.

In terms of in-game victories, the DM has so much power that there is never really a challenge. They could, as the old meme goes, at any point declare that ‘rocks fall, everyone dies’, so they are generally not simply trying to kill the player characters. They might, however, try to kill the player characters within the terms of a specific encounter, respecting the limits of a monster’s abilities. But, beforehand, they would have tuned the monster to make sure it would provide a suitable challenge for the players, or if the monster is meant to be too hard to beat, they would make sure to indicate its power so the players could decide not to engage.

Instead, as mentioned above, the DM is a performer. They might be trying to have the players come away thinking ‘that was a great game’, and enjoy the unexpected turns of the story being devised. Or, they might be trying to rigorously enforce the terms of a challenge, even if the players end up disappointed that they couldn’t beat it. But is this a game goal, pursued within voluntary constraints? I don’t think so. As Nguyen argued about the ‘cops and robbers’ players, the emergent story or course of the game is the focus here, it’s not an arbitrary ‘prelusory goal’.

To compare to chess: the prelusory goal of chess is to arrange the board into one of the states that count as checkmate. Over the course of the game, you will generate a sequence of moves that is be of great interest to chess players—it’s the activity that voluntarily submitting to the rules of chess and working towards checkmate is supposed to enable. But the prelusory goal that structures chess is not to write out a series of moves in chess notation.

trait SuitsianGame {
	type State;

	fn prelusory_goal(state: State) -> bool;

	fn valid_move(before: State, after: State) -> bool;
}

The same goes in story games which allow the other players greater narrative authority, or eliminate the GM role altogether.

Even if the ‘prelusory goal’ might be ambiguous or nonexistent for RPGs, the ‘constitutive rules’—voluntary restrictions on what is and isn’t a valid move—are still a very relevant part of it. But before diving into that subject, let’s take a look at the ‘unwritten rules’ of RPGs, to get a more complete picture of how they work…