Crafting Satisfying Game Narrative

Telling stories that keep players engaged

I’ve been doing a little re-reading and research on narrative¹ in games and a lot of thinking about how it shapes not only our writing but all aspects of a game’s design². Specifically, I’ve been exploring ways in which developers might create a narrative in which satisfaction can be experienced at any scale — from a single encounter to a longer quest or even an entire game. I know that “satisfaction” is one of those normative terms that should usually be rejected principle³, but for the purposes of this post I’m treating it as being synonymous with the build-up and release of narrative tension.

Satisfaction and Structure

If you’ll forgive a quick trip to the remedial writer’s workshop, I want to talk about dramatic structure in order to establish that satisfaction requires a complete cycle of setup, conflict and resolution.

We’re probably all familiar with Freytag’s pyramid⁴. It is such an easily understandable tool for analyzing the dramatic structure of classic five-act plays that it is often introduced in primary school classrooms. With a little effort, we can map a more modern three-act (setup, conflict, resolution) structure to this pyramid. The exact details of the mapping don’t matter, but the catalyst that kicks off rising tension usually occurs late in the first act and the climax occurs within the third act.

Mapping 3-act structure to Freytag’s Pyramid

One important thing to keep in mind when mapping this to the pyramid is that the visual doesn’t represent a constant progression of time. If we adjust for this, we can re-draw our diagram overlaid on top of a tension graph. Our sense of dramatic tension is tied to our narrative’s conflict⁵, so the graph generally follows the rising and falling action (with the exception of the inciting incident acting to drive up tension in the first act.)

Three-act structure and tension graph

We can see from this graph that if satisfaction comes from the build-up and release of narrative tension, we require all three acts to create satisfaction. The key, I think, is to realize that it is three acts all the way down: from the macro to the micro, it is all setup, conflict, and resolution.

To illustrate this point, Your Plot Is a Fractal references YA fantasy novels⁶. Harry Potter’s struggle against Lord Voldemort spans all seven books in the series and follows this dramatic structure. Meanwhile, individual books feature enemies like Quirinus Quirrell who can be defeated in a similarly structured 300 pages. This “fractal” structure continues down into each chapter and paragraph and provides satisfaction at every level.

Maybe it works better to think of this as the Narrative Sierpinski Triangle

If we were to zoom in on our resulting tension graph, we would see that the smooth curve isn’t smooth at all — it is made up of tiny cycles of build-up and release. Zooming in further, we would see that each of those cycles consists of even smaller cycles. Intuitively, we might think that these small tension buildups would be unable to elicit a sense of satisfaction — tiny bumps on the side of a mountain are hardly noticeable. But if we continue with that analogy, it is the author who acts to adjust our horizon and imbue each tiny rise and fall of the terrain with a sense of dramatic scale.

The resulting “fractal” tension graph

Returning to the example of Harry Potter, J. K. Rowling has complete and total control⁷ over the narrative as we experience it — she can exercise the authority of authorship to make a school Quiddich match seem relevant even against the backdrop of a long-term epic battle against He-Who-Must-Not-Be-Named. This allows her to curate a satisfying narrative experience whether we read a single chapter before going to bed or binge-read all 4,224 pages of the series in a single sitting.

Game narrative is multicursal and participatory. Unlike J. K. Rowling, we cede a significant degree of authorship to the player. This collaborative authorship disrupts our ability to create a clean fractal structure. In our mountain analogy, the author’s intent to focus on the smaller bumps can be rendered ridiculous by a player who is focusing on the summit. We’ve all experienced the resulting unsatisfying narrative in games: How many hours has “the chosen one” in an RPG spent killing rats and helping farmers find their sheep?

Our goal is to create cycles of tension so that narrative satisfaction can be experienced at any scale and the connection between dramatic tension and rising conflict inherent to traditional western narrative requires us to create a fractal structure based on setup, conflict, and resolution in order to achieve that goal. Unfortunately, the nature of collaborative authorship makes ensuring this structure impossible.

With A Japanese Twist

Let’s switch gears for a moment to talk about Kishōtenketsu — a four-part⁸ structure found in classic Japanese (and Chinese and Korean) narratives. The word kishōtenketsu comes from each of its four parts: Ki (introduction — establishes the characters and location). Shō (development — advances the plot towards ten without making significant changes). Ten (twist — usually in the form of a literary non-sequitur). And ketsu (conclusion — denouement; connecting the twist back to the plot). Kishōtenketsu is elegantly illustrated by the poet Rai Sanyō:

Daughters of Itoya, in the Honmachi of Osaka.
The elder daughter is sixteen, and the younger one is fourteen.
Throughout history, daimyo killed the enemy with bows and arrows.
The daughters of Itoya kill with their eyes.

In this example, the opening line introduces the daughters and the location. The second line follows logically from the first and develops the narrative by providing more details about the daughters. Then comes the third line: Daimyo? War? It is a twist that we could not have seen coming. Finally, the fourth line connects everything back together and contextualizes the twist by reconciling it with the first two lines.

If there is something eerily familiar with all of this, and you don’t happen to be an avid reader of Edo period poetry or 4-koma manga, you might also recognize it from someplace else: childhood nights sitting around a campfire. It turns out that the structure commonly used for ghost stories and urban legends tends to mirror kishōtenketsu. Introducing (and developing) innocuous elements and following them with a twist that recontextualizes them and makes them scary just so happens to be one of the natural ways to relate tales of horror. (edit: It turns out that tofugu.com has a great post exploring this connection)

An interesting thing about ghost stories is that there often isn’t any (narrative) conflict at all — they are just creepy. Similarly, an aspect of kishōtenketsu is that it decouples the need for conflict from the structure itself. The tension that conflict might otherwise introduce is instead imparted by the presence of the twist and the juxtaposition of the twist against the rest of the narrative. Of course, it isn’t as if the kishōtenketsu can’t contain conflict (figuratively or literally), just that it isn’t required by the structure for the whole thing to be satisfying.

Achieving Plot Within Plot Within Plot

When we apply the structure of kishōtenketsu to game narrative — replacing our structural dependency on conflict with an emphasis on exposition and contrast — the issues that we have with collaborative authorship disappear. We are then able to reliably create the fractal cycles of tension necessary for narrative satisfaction to be experienced at any scale.

I believe that this change allows our recursive structure to better survive because of asymmetries⁹ that are inherent to collaborative authorship of games. Specifically, that the player’s improvisation against the initial author is limited to the protagonist’s actions and implicit motives. While the perception of (and response to) conflict falls within the player’s purview, the introduction of a literary non-sequitur remains squarely with the initial author.

In a 2012 interview with Gamasutra, Nintendo’s Koichi Hayashida told Christian Nutt that the kishōtenketsu structure informs the level design of the Mario series — specifically The Super Mario 3D Land and World games. In these games, each level goes through the whole structure as it relates to that level’s specific mechanic. Over on Youtube, Mark Brown has a wonderful breakdown of how this works in his Super Mario 3D World’s 4 Step Level Design video.

That this narrative structure applies to mechanical aspects of gameplay shouldn’t come as a surprise. In fact, that story and gameplay are inseperable should follow from the phenomenological assumption that gameplay is merely performative narrative¹⁰. This means that not only can kishōtenketsu be applied to a game’s design, it can be applied recursively to a game’s design in a way that contributes to narrative satisfaction at any scale.

I want to conclude with a few questions. What happens when this thinking is applied to a game’s tutorialization? Or when it is applied to the design of an AI coordinator? What does this mean for the prospect of procedural narrative, and can it change how we think about introducing meaning to a sandbox environment? What about when it applies to player progression—are we able to find new ways to create synergy between aspects of our game that are too often distinct? And what about when we view gameplay as a sub-narrative contextualized within the player’s broader identity narrative¹¹ — can creating recursive tension contribute to the creation of meaning in that player-centric sense?

These are exciting things to think about and this scale independence begins to open up broader opportunities for looking at games holistically.

I hope you enjoyed reading this post. My intent writing it is for it to serve as a tool for my own mental organization, but just because it is more “for me” than it is “for you”, doesn’t mean that I don’t appreciate any feedback or comments. Thanks! -c

Notes

  1. Maybe some day I’ll wade into the ludology vs narratology mire, but this isn’t that day. I think that anyone writing or thinking about narrative in games should probably (re-)read the decades of academic discourse on the subject by Juul, Costikyan, Jenkins, Atkins, Frasca, et al. Similarly, we should all be brushing up on our classic narratology. Narrative studies is full of interesting ideas, and while post-structuralism is best enjoyed alone, stuff like Russian Formalism makes for great conversation over a beer. I guess where I’m trying to go with this, is that we should be able to have an informed conversation about narrative without getting too hung up on our fabula and our syuzhet.
  2. …and then someone brings up Tetris. Sure, “not all games”, but certainly more than just what we consider to be story-driven games or interactive fiction. When we view gameplay as a form of performative narrative, we find that it is exceptionally inclusive.
  3. The most common example of a word that we hear game designers should reject is ‘fun’. Was the game fun? How can we make it more fun? I’ll concede that precise language that communicates precise thinking is important and that ‘fun’ is one of those overly broad and normative words that risks meaning nothing.There are cases where we should take Raph Koster’s advice and call it ‘learning’ (the industry’s worst oversimplification on my part?) and there are times where we should really be talking about ‘flow’ or taking a page from Bogost and thinking in terms of ‘procedural rhetoric’. But you know what? Sometimes I just want to play/make something fun in the broadest, most normative sense of the word.
  4. Gustav Freytag’s ‘Die Technik des Dramas’ was written in 1863. It is a great introduction to narrative analysis because of it’s simplicity, but it doesn’t exactly represent a very modern approach to dramatic analysis (or an approach to analyzing modern drama.) Still, it acts as a good common reference from which we can depart.
  5. The connection isn’t conflict per se, but rather the role that conflict has in the narrative structure. It is through mounting conflict that we build anticipation, and it is the action-packed resolution to that conflict that releases it. It is this sequence, rather than the conflict itself, that makes tension possible.
  6. I really like using mainstream YA literature for these purposes. It manages to be culturally relevant enough that most people will have some familiarity with the source material, and it is usually constructed in a straightforward enough manner to clearly illustrate the point.
  7. In this regard, I’m ignoring the emergence of meaning through the imposition of a priori cultural schemata on a text.
  8. I wanted to diagram this, but the reference I found seemed too literal — a line of gently rising tension followed by a corkscrew twist. Let me know if you come across any useful visualizations.
  9. There are other asymmetries that are interesting to consider in this collaborative authorship, and these certainly contribute to the effect. For example, the collaboration is asynchronous in that the developer’s intent is fixed prior to the player’s participation, and the perceived degree of relationship exhibits and asymmetry (1:1 cardinality from the player’s perspective, yet 1:M from the developer’s) that shapes the relationship.
  10. Gameplay is performative narrative. This line of thinking arises from linguistic ideas around performative utterances and embraces much of the thinking explored as part of the performative turn in the humanities. In this regard, it might best be viewed as a post-ludological interpretation of gameplay.
  11. By grounding our concept of game narrative in the broader narrative identity theory, we can view the playing of a game as the act of collaboratively authoring a sub-narrative in which meaning emerges as a result of the interaction between the player’s own identity narrative and the developer’s preconceived narrative intent. What is exciting about this thinking is that it formalizes many ideas in present in player-centric game design.