RPG Design and Ludonarrative Dissonance

Analog role-playing games distinguish themselves from other games* by their inclusion of a fiction layer: imagined events the players devise, share, and engage with as an integral part of play. Designing the interface between mechanics and fiction is thus the central challenge in creating an RPG, and often acts as the point of divergence between different philosophies of design and play preferences. For example, hardcore character immersionists prefer that wherever possible, the player’s input into the fiction should be limited to the reach and will of a single character in the narrative. Mechanics that allow a player to take on a broader authorial role, editing the environment or dictating the actions of other characters, run counter to HCI play preferences.

When setting out to design an RPG, then, it helps to know what design patterns already exist for navigating the mechanics/fiction interface, and what pitfalls those design patterns sometimes hold for players’ engagement and enjoyment. This post is more of a braindump of things I’ve run into over the years than a comprehensive thesis, but I hope it will provoke some thought!

To start, the RPG chicken/egg question: which comes first, fiction or mechanics?

Mechanics first: A player makes decisions grounded in mechanical systems, and engaging with those systems helps generate fictional content. In Dungeons & Dragons 5th Edition, I expend my action for the turn and a spell slot of 3rd level, choose a target location, and roll eight six-sided dice. That then translates into a fictional event where an explosion of magical flame bursts forth, scorching enemies and setting scenery on fire!

Fiction first: A player narrates fictional action, the content of which activates mechanical systems. In Dungeon World, I describe my character darting across a stone bridge over which “Pit and the Pendulum” style blades swing. That triggers the “Defy Danger” move, and I must roll two six-sided dice adding my DEX score to see if the character makes the crossing without mishap.

Note that a game will almost inevitably feature both modes, as mechanics and fiction move one another forward in a cycle. But considering where a player’s decision-making process is likely to start, or which of the two directions of flow has greater emphasis, can help inform your game’s core priorities.

If an RPG features constant mechanical engagement, but fictional content is optional, thin, or an afterthought, that produces a problem we might call boardgame regression. The design risks sliding out of RPG territory by neglecting the fiction layer altogether, resembling a board game (albeit perhaps a flavor-rich one, something like Dead of Winter). This is the most common criticism I hear leveled against the fourth edition of Dungeons & Dragons. It was all too easy to play the game as a series of boardgame skirmishes, with little to no reference to characterization, plot, or even an imagined scene beyond the positions of figurines and their associated fluctuating numbers. To avoid this pitfall, ask yourself: is it possible for players to disengage from the fiction and still play? How does fictional content help drive mechanical decisions?

When an RPG features narrated fiction that only rarely or tenuously grounds itself in mechanical systems, I call this trouble slipperiness. Players are uncertain if their narrative contributions make any difference to the game state, or lack trust that the rules will back them up in the event of dispute or ambiguity. I run into this most frequently with games that give one player authority to secretly edit or override (“fudge”) game mechanics, the “Rule Zero” espoused in texts like Exalted. To guard against slipperiness, consider: how does my game differ from a minimalist collaborative storytelling setup, where players volunteer bits of story to be adopted by consensus? How do mechanical decisions and outcomes generate fiction? Is a situation where players break or drift the rules distinguishable from one where they play by them as written?

Friction in the mechanics/fiction interface needn’t be so pervasive as the above, however! It can occur sporadically in play, within specific rules or procedures. Even an overall functional D&D4 game sometimes hits moments where a mechanical outcome has occurred, but it’s difficult to picture what happened fictionally. “I use Arterial Slice on the skeleton! It’s now bleeding for 5 ongoing damage.” “Wait, what? Skeletons don’t have any arteries to bleed from.” In a Wicked Age features mechanics that only activate for physical conflict, so if a player narrates a character intimidating, bribing, or otherwise attempting to persuade another, the rules cannot help determine if their ploy is successful–a moment of slipperiness in an otherwise grounded game. The general term for these jolts is “ludonarrative dissonance” (hat tip to Kevin Weiser for that!), a place where game and fiction aren’t quite harmonious.

Ludonarrative dissonance can also arise within a game’s reward cycles. Mechanics might encourage an action that doesn’t make fictional sense. In Burning Wheel, given a minor expense a group of player characters would like to pay for, it is often in the group’s mechanical best interest to have the poorest character make the purchase with the help of more wealthy characters, rather than the wealthiest character dispensing with the buy alone. But coming up with a justification for that approach from the characters’ perspective tends to be tortured at best! Or an action that flows naturally from narrative and characterization could prove a terrible choice mechanically. In a recent D&D5 game, a player attempted to win over a villain driven by anger and despair, putting the spell “Beacon of Hope” on him to instill a sense of optimism and possibility. Reasonable, yes? But all that really accomplished, game-wise, was to make said villain more resistant to the heroes’ magic in the ensuing battle scene.

There’s one last rules/fiction pattern I’d like to call attention to, as it’s one I’ve struggled with in recent memory. I’ll call it the justification veto. In a justification veto setup, a player has access to certain mechanical resources–skills, character traits, or what have you–that need to be brought into the fiction in a meaningful way for them to grant bonuses. A classic example is Aspects in FATE, freeform descriptions of a character’s tropes that if I can explain how they help me with the task at hand, grant the ability to spend a Fate point for a reroll or dice result boost. That “if I can explain how” is the rub, though. If I’m a couple points away from succeeding on a roll, the rules urge me to find a way to bring one of my Aspects into the scene. The success or failure of that effort, however, rides on my ability to narrate that Aspect in a convincing manner for the context. If the other players (particularly the “GM” whose word on such matters is final) feel it’s too much of a stretch, the use is vetoed: neither the proposed narration nor the bonus take effect.

Justification vetoes are a very natural pattern to draw upon, helping ensure that mechanical bonuses are grounded in coherent fiction and vice versa. I’ve used them myself, in my game Blazing Rose! But the experience of pausing game flow for a “Mother may I” petition can frustrate players, especially those with different levels of skill in navigating mechanical systems vs. weaving persuasive narrative-grounded arguments. (I would not be at all surprised if neuroscience revealed these skill sets operate in disparate regions of the brain.) If that’s an experience you’d like to avoid in your design, put this pattern in a “use with caution” column.

A few games work around the justification veto’s drawbacks in clever fashion. In Chuubo’s Marvelous Wish-Granting Engine, applicability is not a binary “yes, you may” / “no, you may not”; a player may use any skill on their character sheet for literally any purpose. (An example in the book describes using a Cooking skill for the intended action “I blow up the Earth with my mind.”) Rather, the GM’s assessment of how much the action stretches the skill diminishes its effectiveness, making it more costly to get favorable results with. That still encourages matching mechanical elements to appropriate fiction, without the inherent frustration of shutting down a player’s contribution outright!

What pitfalls have you encountered in the interface between rules and fiction? What design patterns or play behaviors help avoid them? What other insights have you gleaned from the matter of clouds and boxes?


* “What’s a game, then?” Well, hypothetical wiseass, I don’t have an essentialist definition for you that would reliably include all games and exclude all non-games. As a usually-useful approximation, though: a game is a rules-structured, temporally bounded activity sustained by one or more behavioral reward cycles.

Blood Isn’t Good to the Last Drop

I respect feminist games criticism because it shows us how much better we could do with our designs. Many of the tropes dissected aren’t behind the times only in social awareness, but in the state of the art in gameplay. Where you find a lazy portrayal of gender, race, or violence, lazy design choices often follow, and we can improve both by amending or eliminating our use of trite patterns.

This excerpt from Phil Owen’s WTF Is Wrong With Video Games? and this tongue-in-cheek list of in-game activities that would count as war crimes in real life primed me to think about these topics in my current gaming. (Both articles have their problems, but I can appreciate and recommend them for getting those thoughts rolling.) In particular, Cracked‘s critique of “giving no quarter” resonates with gameplay irritations I’ve run into in both video games and tabletop RPGs. Why is it still so often necessary to kill every opponent on a map before concluding a mission?

Feeling the itch for some turn-based squad tactics (perhaps in anticipation of XCOM 2?), I’ve recently restarted playing the WWII skirmish game Silent Storm, originally published in the early 2000s. It does one thing well with respect to the No Quarter trope: mission objectives almost always require that you obtain information, not kill everyone. Your goal on a given map is to procure documents, film reels, prototype technologies, etc., or to subdue and capture personnel with crucial intelligence. I find that quite refreshing! Real-world military objectives–at least for forces we see as admirable or heroic–rarely focus on annihilation, and it’s great to see that in a game.

Unfortunately, the rest of the game’s design undercuts that commendable concept. Level layouts, enemy AI, and the fact that you can’t leave the mission zone with visible enemies even if you’ve accomplished all objectives, mean that most of the time you must wipe out all opposition to advance anyway. The intelligence targets you must capture don’t surrender; you have to fill them with lead to “knock them unconscious” and carry their limp bodies away. (There are a few nonlethal weapons in the game, but their game statistics are terrible, heavily disincentivizing their use.) Enemy units sometimes flee, but they can’t actually leave the level, so they reach the black expanse of nothingness at the map’s edge, then turn around to start shooting again.

We see this trope time and again, and invariably it makes for a worse game. I love the XCOM series, but especially in the earliest versions, hunting down the last alien on the board to complete a map was an exercise in tedium. It comes up in tabletop play, too. A common complaint against the fourth edition of Dungeons & Dragons was that battles took too long to complete, and often at least part of the problem came down to playing things out until every monster in the encounter was dead. We’ve got No Quarter burnt into our heads by long exposure, but it’s a bad pattern.

We have the technology to do these things differently. The XCOM games already have morale algorithms, where enemies panic in the face of impending defeat, dropping their weapons and fleeing–but for some reason, the games haven’t taken the logical next step: have said enemies surrender, removing the necessity of blasting them to end the level. (Yes, you can knock them out with nonlethal weaponry in XCOM, which is a nice touch, but it’s still a waste of time and verisimilitude that you’ve got to hunt down and shoot routed enemies at all.) I appreciate the design patterns in Dungeon World, in many ways a superior set of tabletop play tech than D&D, whose principles of fictional flow and “bring every monster to life” lead naturally to combatants fleeing, laying down arms, or otherwise changing the nature of the conflict before they’re all dead.

These are more humane, progressive, feminist, etc. approaches to violence and victory than the tired No Quarter trope, and they make for better games too!

The Mistakes of Hyperoptimization

Adulath II:

This is a follow-up to my System Matters post–a caveat to it, if you will. There are places where the line of thinking in hardcore System Matters, much as I believe in such a thing, can lead you astray, and it’s useful to be mindful of the pitfalls.

The situation I’ll discuss in this post is familiar to many who’ve been involved in role-playing games for any length of time, and I’ll call it “hyperoptimization.” (You might have heard terms like “powergaming” or “munchkin” in reference to these situations. I’m avoiding those terms since they’re derogatory; “optimization” tends to be acceptable to all concerned, with the “hyper” prefix demonstrating that we’re talking about too much of a good thing.) A game presents various options, typically in the skills and equipment available to characters, to customize a player’s approach to the game. Some options, alone or in combination, are more effective at overcoming in-game challenges than others. The better a player understands the game, the better they can find and take advantage of synergies between options, a process called “character optimization” or simply “optimization.” In itself, this poses no problem and is often part and parcel of the game’s fun, but in some cases it becomes hyperoptimization: taking such thorough advantage of game mechanics and options that in-game challenges become trivial, and other players’ and characters’ contributions start to feel inconsequential by comparison.

Such lopsided play can lead to frustration at the table. Outclassed players feel left behind, and players in GM roles responsible for providing adversity and challenge find it difficult to do so in a way that works for both the hyperoptimized characters and the rest of the group. I’m going to leave that much as self-evident; what I’m concerned with is what happens next. If the frustrated players/GM approach a hyperoptimizing player with their frustration, and ask him/her to rein in the optimization, one common response bears all the resonance of a System Matters argument: “The game encourages this kind of play, I’m just doing what I’m supposed to. In fact, folks who don’t optimize could be said to be playing poorly!”

I reject this argument. It’s frequently mistaken, because games can mislead you as to what optimal play looks like, both in degree and in kind. Moreover, it’s a non sequitur rebuttal to what is at heart a social-contract complaint, which sits in a broader, overriding context to that of the game and the play it encourages.

1. Some games mislead you. At times, whether due to poor design or design that’s simply too subtle for its own good, a game looks like it purposefully encourages a particular behavior, but in fact it doesn’t work the way it appears on the surface. The first edition of Exalted was like this: it presented players with a shopping list of magic powers used by characters and their opponents, structured something like the cards in a collectible card game. It even used card-game terminology like “Combos.” This encouraged players toward tactical, winning-combination-seeking play… but the game disintegrates under such an approach. The rules are made for fast-and-loose telling of epic stories, lacking the consistency or rigor needed for tournament-caliber competition. 3:16: Carnage Amongst the Stars is an example of the “too subtle” category: it has lots of optimization-encouraging features, such as a score tally (kills) and Xbox LIVE-like achievements (medals). But the game becomes deadly boring if played to that angle. Instead it comes into its own when players dive into the narrative elements of character flashbacks, exploring how the experience of war shapes the characters. In these situations, hyperoptimization isn’t good play at all. It only pushes the game faster into territory it’s ill-equipped to cover.

2. Sometimes “optimal” does not equal “maximal.” Some games unambiguously and intentionally encourage players to plan and fine-tune their characters for tactical success, but discourage going too far with the practice. Dungeons & Dragons Fourth Edition is a great example. There’s no question that it’s a very game-oriented RPG, and that the design deliberately pushes players to seek synergies between powers, feats, and class features. However, if you go to the utter limit available in the system, building characters who cannot suffer harm, deal enough damage to kill solo opponents in one round, and/or deny opponents any capacity to act, the game breaks down. You may get the surface-level rewards of loot and XP, but you lose out on the deeper rewards of engagement in play. You lose out on making tactical decisions (because your optimal moves become pre-scripted and independent of circumstance), employing teamwork (because allies are unnecessary), adapting to adverse circumstances (because they simply don’t affect you), etc.

In essence, the player who falls prey to issue #2 has mistaken the means by which a game pursues its goals for being the goals themselves. At heart, most of these games seek to provide fun/enjoyment/entertainment for a group of people, and character effectiveness is only one feature in service of that goal. Hyperoptimization undermines that goal, even if the game legitimately encourages you toward it. It’s like playing a first-person shooter video game on “god mode”: yes, you get the rewards of blood & gore, victory cutscenes, and so forth, but in so doing you’ve disengaged from most of the game’s features. Sometimes, attaining optimal satisfaction in gameplay requires that if you solve the game, you adopt some handicap so it becomes challenging again.

Now, note that I’m not saying this phenomenon indicates good design. Discouraging a behavior by causing it to opt you and your fellow players out of the game itself is pretty messed up. And in the specific example of 4e, the balance point can be hard to find, especially in the upper levels. I mean this only as a counterargument to the assertion that “suboptimal” play = poor play. Some players realize, to their benefit, that maximization can be its own losing strategy in the overall pursuit of fun–even when the game makes it a tempting option.

3. The social contract trumps all. In the same circles of RPG theory that most champion the “system matters” concept, there is a structure called simply the “Big Model” that maps out how different features of a game environment interact. At the broadest level of that Model, encompassing everything underneath it and adding more besides, is the “social contract.” This “contract” contains all the rules and expectations of behavior, often unspoken, for the group of people playing the game. These are things like “Jeff brings the pizza on Tuesdays” or “if we find we don’t like this game, we’ll discuss what’s going on and maybe ditch it.” As you can tell from those examples, the social contract by necessity exists outside the scope of the game itself, and its rules can in many cases supersede those of the game in question.

When a complaint comes up about optimization, it’s rarely a thing that can be addressed below the level of the social contract. The complaint arises because the optimization behavior, on one level or another, is harshing on somebody’s fun. The appeal is not “you’re playing the game wrong,” it’s “we’re here to help each other have a good time, and right now that’s not working out.” So while a response of “I’m only playing the way the game encourages me to” is absolutely legitimate in the sense that it demonstrates any fun-harshing was unintentional, it is not sufficient as an argument that the behavior should be allowed to continue. You have to settle it in terms of players’ attitudes toward the game and the expectations within the social setting. When it comes to social contract, it is totally legit to respond to the hyperoptimizer’s rebuttal with “Yes, the game encourages this. Fuck what the game says. I’m saying that this gets in the way of my fun, as a player and as a person, and I’m asking you, as a player and as a person, to cut back.”

We don’t have to range too far afield to get a good allegory for this. Suppose a young woman is teaching her little sister how to play basketball. The elder sister is an accomplished athlete, practiced and fit; the younger is a novice, still learning the basic techniques of the game. The younger sister asks the older, “Hey, could you go easy on me? I can’t learn how to shoot or even dribble if you snatch the ball away as soon as I get it.” The older sister agrees. If she then goes back to a curbstomp of a keepaway match, Little Sis’s objection won’t be that Big Sis broke the rules of basketball, and it certainly won’t be that Big Sis is bad at the game. It’s that she broke the sisterly agreement they had.

Social contract issues at an RPG table are seldom so clear-cut as that, of course. In order for a productive conversation to take place on the matter, both the optimizer and the player objecting to the optimization need to articulate some principle like “players should be willing to compromise in-game effectiveness to make a better play experience for others,” then discuss whether such a thing is reasonable. And it may well be that the difference is irreconcilable, that the level of compromise that would make for a fun game for the objector would utterly wreck what the optimizer enjoys in play. But let us put aside this notion that “optimization is good play, it’s non-optimization that’s playing wrong” is any kind of conclusive response to the problem, for the reasons here articulated!

A Land of Wealth and Peril

Adulath Caracai II:

Dungeons & Dragons Fourth Edition is a pretty good Gamist design. It’s got high-level strategy and moment-to-moment tactics, and does well with them. But its carefully balanced encounters and guaranteed rewards undermine the creative agenda in places. There’s not much risk-reward calculus, and there’s surely not much competition.

Time for that to change.

Continue reading