Nonlinear Story Structure in Games

To interact is to have choices, but so much of our storytelling in games prevents choice. How do we tell the story of Joel, zombie survivor, without robbing players of their agency? Yet if they can choose to do anything at any time, how can we expect them to experience someone else’s story and not just create a new one of your own? Should we just let all games be story-generators like Minecraft? These questions have inspired several years of research on my part, and in this article I’m going to be discussing a narrative structure I have observed that helps to get closer to the answer.

In this article I’m going to take a look at conventional wisdom accepted in other media, and adapt that knowledge to games. I’m going to use examples from a few different games, and I’m going to invent a quick story to use as an example of my observed structure.

Part 1. What Is Choice? 

Choice is basically this diagram. You start off at A, and you can then move to B, C, or D. This is the foundation of Choose Your Own Adventure stories, and most interactive narrative. You can use this structure with any number of choices, so long as it’s greater than one. The result of this choice can then become the start of a new choice, on and on into forever. This is called Branching Narrative.

One of the common structures we see in interactive fiction is the ‘diamond’. It’s when two branches split and then come back together. This is usually a structure that we see when a choice has no real consequence. So you start at A and then you choose B, C, or D, but no matter which choice you’ve made you always end up back at E.

Generally, this structure is frustrating for players. It’s the illusion of choice, but it’s something that’s far easier to manage in a production setting, so it’s kind of where we end up much of the time, simply due to ease of production.

Part 2. The Three Act Structure

With those structures in mind, I want to go off on a tangent and discuss the idea of the 3 Act Structure. I’ll offer a link here for anyone unfamiliar. Each act in the 3 act structure serves a specific purpose. Act 1 sets everything up, Act 2 develops everything, Act 3 resolves everything. Within each of those acts are a number of events. Those events will convey what happens in the story. Some stories will present their events in reverse, while others may change the layout of the acts. Some stories simply don’t adhere to the 3 Act Structure and instead follow the Shakespearean default of the 5 Act Structure, which you can read about here. For the purposes of this article, all approaches are equally valid, so long as each act contains a sequence of information-rich events.

Here is a visual example of the 3 Act Structure. Pretty standard stuff, aside from some of my labelling, which I’ll get in to later. The takeaway here is that the story has a beginning, middle, and end and within each of those are a series of events. Now, keep all that in mind as I go off on one more tangent. Bear with me.

Part 3. Campfires In The Dark

A few years ago I heard this concept at a GDC talk (which I would link, but I can’t find it. If you know it, please link me) which referred to storytelling in games as campfires in the night. The concept is that your game is a dark forest. Everything is black and murky and you can’t tell where a player will be at any given time, but you’ve specifically lit a number of campfires to help you see. Because you know where those campfires are, you can interact with them and reasonably expect people to see whatever you do there, as they are attracted to the light. In other words, while you can’t know where a player will be at a given time, you can create cities out in the wastes, quest giving NPCs, environmental story moments, or any number of other elements. Knowing that these elements exist, and will attract the attention of the player, you can use them as framework to build your story. If you would like, I was able to find an article that summarizes the concept, which you can read here.

Part 4. Tying It Together

An excellent example of this campfire narrative concept is Legend of Zelda: Breath of the Wild. You start off on a plateau, but once you complete this beginning sequence, you’re given free reign of the world. Out in the world you find a series of memories located in various points of interest (‘campfires’), each of which gives you a new piece of the overarching story. Once you’ve visited all the ‘campfires’, you can visit the castle and begin an ending sequence that wraps up the rest of the game. In Breath of the Wild, you start off in a single, limited space before being given a series of choices. Once you’ve made those choices, you are funneled back into a single, limited space.

Hey wait, that’s that choice chart from the beginning of this sprawling diatribe of an article.

In Breath of the Wild, however, those middle choices aren’t multiple choice answers on a test, you can pick any or all of them. In fact, the game is setup with an expectation that you will do exactly that. By allowing you to choose all of these options, you no longer feel like your autonomy has been robbed from you when you are forced to choose E at the end. Instead, the “B, C, and D” choice becomes a free-form section where you can do whatever you like. This contrasts nicely with A and E, allowing those two sections to convey the linear story information, while B, C, and D give you non-linear pieces of the puzzle. Each of these sections now becomes its own unique part of the overall structure. A introduces the world, B,C, and D expand on that narrative, and E closes out the narrative. Wait a second, that’s that OTHER chart from earlier in the article, isn’t it?

A now becomes Act 1, then you pass through the first Gate to B, C, and D in Act 2. Then on to E, which is Act 3.

Part 5. Pizza Break

The introductory plateau of Breath of the Wild serves as the game’s first act, teaching you the mechanics and introducing the world and characters. Act 2 lets go of the reigns, allowing you to freely explore as you see fit. At any time you can choose to go to the Castle, which funnels you into a linear space that allows the game to control the intensity of your experience, and therefore ramp up into a climax.

All of this still basically follows the 3 Act Structure chart I’ve given above, but the addition of choice into Act 2 changes things. Given that all the individual ‘Events’ listed in the chart don’t really need to be in any kind of order, why not find a new way to represent this information?

I give you the ever-elegant narrative pizza diagram. You start off in a discretely linear area, open up into a world full of options, and then funnel back into a discretely linear area to wrap everything up. It works quite well!

But what happens if you need more than one Castle and more than one Plateau? What if your story isn’t just one big open space book-ended by linear spaces? Think about your average open world, usually there’s a back and forth in and out of linear spaces.

So here’s our graduated pizza chart. Each Larger circle represents an act, within which are various events (aka “campfires”). Each larger circle could have as many “campfires” as you like, and can be split into sub circles as you see fit (much as Act 2 is usually split in the 3 Act Structure chart). The campfires represent optional information. You can sit at these campfires, soak them in for a while, but you’re free to go at any time. Gates are non-optional. Gates are the critical information your story needs to make sense. Gates are also the critical events of your story, the events that change the world. Each act of a story asks a narrative question, and while campfires help you answer those questions, the gates change what’s being asked. Once you pass through a gate you can’t go back, just as you can no longer quest for an answer you already know.

Part 6. Let’s Write A Story

Here we’ll write up a quick concept and fill out this generic chart with specifics.

Let’s say we have a kingdom in peril because it’s just lost its prince to the evil dragon who lives nearby. A brave knight has to go off to rescue the fop of a prince. The story starts with the brave knight as she deals with family troubles. We setup the conflict between her and her parents, we show how desperately she wants to go off on an adventure and prove herself. Then she discovers that OH NO! The prince has been stolen away! This is her chance, she must go! As she’s on the road she discovers an army preparing to go to war against the dragon to save the prince, but they can’t figure out how to take down the beast! If only someone could find a way! So our hero goes off and talks to a bunch of people who each know a little about slaying dragons, and she hears rumors of a magic ring, but it’s held in the treasure trove of the very dragon we’re trying to slay! An army would never make it out alive if they approached, the only way to get the ring is for one brave soul to sneak in under cover of darkness! Our hero approaches the dragon horde, but she discovers something more than the ring. The prince is being held in perfect comfort, and the dragon is gentle beast who is trying to stabilize the kingdom by keeping the fop of a prince from rising to the throne! He’s not such a bad beasty after all! But oh no! The army is still massing at the base of the mountain, preparing to strike. The attack launches in just three days! There must be some way to prevent this, some way to defend against them without killing her own people in the process! Oh, what a quandary! In the end our hero must figure what sort of knight she wants to be and find a way out of this mess! She decides to talk to the army, but that will never work as it is now, so she first sneaks into camp and kidnaps the general of the army. She uses the chaos as her moment, gathering those who will listen to her story about a peaceful dragon who wants nothing more than happiness for the kingdom. The army fractures, half of their blood is up and they want a fight, the other half knows the prince is an idiot and is willing to take any excuse to find a new leader. Without leadership, the army dissolves into factions and splinters against itself. Too busy to coordinate an assault against a massive dragon, the army retreats for now, long enough to begin diplomatic relations between the king and the dragon properly. The day is saved.

So let’s take a look at that mapped to our Breath of the Wild pizza graph

The intro establishes our hero. We see a cutscene, or whatever, to establish who we are and what we’re doing. Then it’s off and into the city for a little free range gameplay.

During this section we want to establish rumors of a dragon nearby, that the prince is missing, how the kingdom is responding to all of this, and how all of this is affecting our family life. All of this information sets the tone, but none of it is actually essential. What IS essential knowledge, however, is the first dragon sighting and a confirmation that the prince has been stolen. So while the campfire events help sell the story, the first Gate sets the stakes. No going back from there.

Act 2 is all about the enemy. In the first half, the dragon is the enemy. All your optional quests are about how to slay a giant dragon and earn glory for your kingdom, or how to make it on your own in the wilds. At some point you’ll move towards the dragon’s lair, and it’s time to hear about the ring that kills dragons. You’ll try to claim the ring and end up claiming the knowledge that things are not what they seem. This new knowledge changes the very foundation of your being, and changes the answer to the “who is the enemy” question. You’re against your own army!

Now your quests become about seeing the world for what it really is. You learn about dragons, you find out that man is the true monster after all, etc. etc. etc. Also in Act 2 I put a quest about the state of the kingdom, because it’d probably be nice to get a general status update from back home at some point during all this, but that quest can be done regardless of what you’re doing in the rest of Act 2. Just a little bonus info if you want it, ya know? But at some point your exploration and your questioning must come to an end and you must hear the proclamation: The army attacks in 3 days. OH NO!

Act 3 is then about who we want to be as a hero. Do we want to fight our own people? How can we protect the dragon? Does our army have a weakness we can exploit? There has to be some way to talk to them. Then you get the brilliant idea to kidnap the general and use the confusion to speak to the army at large, without a unified front they’ll be weaker. And so you move from Act 3 into the resolution, which wraps up the game as a whole. Roll credits!

Obviously, this story is super loose and awful, but hopefully you can get the idea of how this structure works in a more linear game.

Part 7. ...And In Conclusion

This structure is loosely present in nearly every interactive game I’ve studied. I didn’t first see this in an open world adventure, like the one I describe in part 6, but rather in a roguelike. Games like Dead Cells and Spelunky will progress from a single space where you shop or upgrade skills, and then out into an open world section. That open world section will be home to new information about the story/world, but you’ll always then move into another linear section. This pattern of expand and contract is prevalent all throughout roguelikes, and while it won’t often be acts of a story, the larger circles on my chart map directly to the levels in these games. Building a story for these games would be relatively simple.

Even linear games can be charted to this graph, they just chart as a series of gates and sub gates one after another. The less linear a game, the more campfires you find floating in the act circles until you get to the Breath of the Wild example.

This whole thing just functions as an added dimension to the 3 Act Structure, making it very easy to adapt to any production. I’ve personally used this on a number of prototypes that have never actually seen the light of day (getting a team to ship something with no money is hard, yo), but my career is still in development, so hopefully you can apply it to your own work. I’d love to hear about it if you do.

And for now, that’s all I’ve got. Hopefully you found something useful here. If you disagree with any of my assertions, or think I’m a fool for not going far enough with some aspect of this whole thing, let me know in the comments.

Thanks for reading!