Inevitably, players tell each other stories about their gameplay experiences. But some games inspire more, like an electric ecosystem of fan art and narrativized play, or a collaborative worldbuilding effort that elevates to the level of shared canon, or stories that rise to such prominence they become community legends.
Player stories have long functioned as an informal metric of success for a certain type of game, and recent research in interactive narrative theory has even formalized the concept as re-tellings, shared player experiences that “can be considered an indicator for a well designed INS (interactive narrative system) and as an instrument of critique”1. Community storytelling drives interest in games, leading at its highest pitch to free marketing and organic virality. Beyond the sales benefits, to designers, player stories are often their own reward.
Why do some games provide richer soil for player stories than others? What can we do as designers to encourage and support community storytelling?
We can start with the hands-off approach that has driven a lot of design: make a good game with compelling dynamics and the stories will come. Perhaps we can do one better: make a good game and provide some social scaffolding — say, a community space like a Discord server — for players to share their experiences and tell their stories. But we wish to go deeper: through the lens of tellability2, we can slow down the ludic lifecycle and take a close look at each moment of phase change in the process of player story formation. From when a player first encounters a game (and even before then, in design and development) to when they relay a noteworthy experience to other members of their community, we can identify points of intervention for us as designers to catalyze the process.
This report aims to offer strategies and tools for designers to foster player storytelling about our games. The rest of this introduction will give some outside perspective to tellability. Then we’ll lay out a model pipeline for how design and development flow into play and then into community. This will give context to the meat of the report, the seven tools we describe for seeding player stories (with some empirical examples). We’ll situate each of them along a lifecycle of play in order to hone in on exactly when and how they motivate sharing. By understanding how moments of play become memorable and tellable (individually and collectively), we hope to contribute to a new and fruitful exploration of meaningful, personal, and enduring player stories.
We limit our discussion here to games that play in the space of storyworlds. This includes but isn’t limited to what is commonly construed as the narrative genre. Any game whose mechanics support a persistent fiction would qualify. Abstract games, like Tetris, may produce tellable moments in their own right, but our tools and examples aren’t geared for them. Additionally, our tools work best in games that embrace some amount of emergent play3, as emergence is good at producing the conditions of tellability.
These words are either new or used with particular meanings throughout this report that may deviate from standard usage.
Devtime: the design and development phase of a game’s lifecycle, as separate from the play and sharing phases
Framing: the deliberately added context in which game content and experiences are understood, shared, or remembered. A frame can be literal or visual, such as putting a golden frame around a screenshot to make it seem more historically significant, or it can be abstract, a point of view or set of assumptions that imparts meaning to an experience.
Lacunae: intentional gaps or holes; deliberately omitted information that makes a world, system, or mechanic more intriguing by its omission
Legend: a widely-repeated tale of an event or action, reaching far beyond its initial participants and witnesses
Playtime: the phase of a game’s lifecycle where players actually play the game, as separate from the dev and sharing phases
Sharetime: the phase of a game’s lifecycle where players converse and share stories, as separate from the dev and play phases
Story: a description of events, as interpreted and told by a person. Not necessarily narratively complete, coherent, or accurate.
Symbolic vocabulary: the symbols, terms, audio/visual motifs, and game mechanics that establish a game’s voice and characterize the specific texture of its storyworld
Tellability: how likely a game experience is deemed worthy of sharing with others. Influenced by factors such as unexpectedness, particularity, and genre convention.
In this report, we look at the ways communities take piles of generated events and curate and collage and embroider them to create shared stories. Of course, digital games aren’t the only source of generated events to grow stories around. We can look to other social storytelling traditions to gain inspiration and note what makes their stories tellable.
One of the most common of these traditions is sports. Any kind of sports game generates a mountain of individual data: RBIs, three-pointers, lifetime bests, records broken, games won, overtime, injuries, fouls, and red flags. By themselves, these data don’t mean anything. But sit with an aficionado, and they’ll tell you which ones contribute to a greater story, which are the dimensions of tellability specific to the systems of their particular sport. There are tellable staples in the sports genre, too. Consider the player who was traded away, came back to face a rival team, was shaken for the first half but rallied towards triumph in the final moments; or the gymnast who was injured and risked greater injury in the name of a career-defining victory for the team. These are story formats we recognize as compelling even when we lack familiarity with the specifics of their sport. Fans, too, become subjects in sports stories, despite the lack of interactive agency we associate with digital game playing. The game that was won on my birthday, the team I support despite their losses, my favorite player who was injured on the day I forgot to wear my lucky jersey: tellability lives not just in the events themselves but in our relationship to the events and the way we frame them.
Narrators have to compile and edit their stories, and announcers must do so live, as events are generated in real time. Automating this storytelling process has long been a goal for digital practitioners, to create virtual play-by-play announcers for soccer games or to process game events into generated newspaper copy. How does one tell when an individual event is part of a greater story? How does one decide what story to tell in the first place? When does one discard the story they thought they were telling when a more exciting new thread emerges? These are the questions narrators, including our player-narrators, must answer.
Collaborative storytelling exists outside of digital games, too. Both diffuse group storytelling efforts, like those inside fandoms, or more centrally-hosted projects like SCP develop around (or develop their own) toolkits for storytelling. These are reusable templates for telling specific kinds of stories with known plot structures and themes that can be recombined even by an unpracticed beginner (example: the SCP beginner guide) to guide them toward that individual community’s watermark for tellability. Collaborative storytelling happens outside individual subject positions, too. We can look at the histories of cities, or movements, or objects as organic, communally-driven stories.
As we move through the lifecycle of play in the next section and think through the raw narrative material and shaping tools we give our players, it’ll be useful to remember that player stories are situated in a greater storytelling tradition.
Players’ experiences cohere into tellable stories during and after the sequence of play. But the process begins earlier, when devs establish the storytelling conditions and conventions by building the digital artifact players will interact with. Somewhere inside the bounds of this wider picture is embedded the full transformational chain to player story and — beyond — to community legend. Laying it out in sequence gives us insight into where the intervention points may be.
There are multiple models of digital games and interactive narrative that break along the lines of 1) development 2) play and 3) interpretation.4 We use a similar set of phases, refocused through the lens of player storytelling. In devtime, devs make the game, fabricating its architecture and generative systems but also developing a symbolic vocabulary that gets threaded through these systems and, when spun out into a storyworld through the play sequence, imparts its characteristic texture. In playtime, players engage with the generative systems, producing a playtrace that gets tracked by the game itself but also in the mind of the player, who is framing their experience and iteratively making meaning from it. In sharetime, players engage with a community of other players, sharing their experience in variously raw and refined forms. Some of these experiences transcend the main loop and become ubiquitous legends in the community and beyond.
These phases overlap intentionally. Games that update frequently, like live-service games, alter their generative systems as and after players engage with them. Likewise, the play/share sequence is often cyclical; players look to community resources to understand and contextualize their experiences (for example: by looking something up on a wiki or asking a question in a Discord channel), and their contact with other players’ stories can inform further play.
Repeating the central question of this report: what moves can we as designers make to catalyze the player storytelling process? Where in the lifecycle do we make these moves, and where do we see the results? The goal of the pipeline is twofold: 1) to provide a context for the tools that appear in the following sections, and 2) to function as a model for the comparison and categorization of future tools. To these ends, for each tool, we specify:
We assembled this toolkit through observation, analysis, and personal design experience. It’s our claim that these techniques yield a higher volume and richer quality of player stories in the games we’ve created and played. It’s not meant to be an exhaustive list but rather a starting point for exploration into the mechanisms of tellability and their support structures.
| Intervention phase: devtime Where the results show up: devtime, playtime, sharetime Sharing motivation: sets expectations for what stories look like “here” |
Some games are particularly successful in inviting players to share their storytelling space — often through the same techniques that make a setting or IP good for a long-running series. Part of the challenge is to help players easily imagine what sorts of stories can be told in this world, and how these stories might be different from stories in other worlds. The most fruitful spaces for player engagement often feel both accessible (we know the general type of story that might happen here) and distinctive (the game offers us a new perspective or set of questions to bring to this story).
These dev actions help players work effectively within the world of the game:
Often a few memorable items do more to guide players toward what is tellable than a large selection of less-notable elements.
The pirate story genre has been popularized by numerous books and films. They’ve helped set reader, viewer, and player expectations for the presence of mischievous thieves, glorious adventure, and treasure hunting. These genre expectations informed the setting and gameplay of Sea of Thieves, contributing to the game’s expressive range and giving a specific context for the play experiences that come out of it. Because of the public’s familiarity with pirates, players come equipped with an internal compass for how to act in a pirate-ridden world, how to interpret their game actions as a particular kind of social performance. This guidance is both a driver for play decisions and a storytelling aid.
For a specific example, consider the several distinct factions in Sea of Thieves, called Trading Companies. From the treasure-obsessed Gold Hoarders to the mystical Order of Souls, several pirate genre staples are given representation in the storyworld through these factions. Players can support or oppose them and will interact with their representative NPCs from time to time. Mechanically, the player’s standing with these factions determines what type of missions they can take on and what sorts of rewards they get for completing them. Beyond strict gameplay, though, faction allegiances act as social stage direction; players will enthusiastically roleplay the implied values of their chosen faction.
This leads to a virtuous cycle of roleplay and storytelling; players take cues from their faction’s imagery, NPC behavior, and overall narrative tone, and then they take actions unto themselves, reinforcing these inspiring traits but also feeling out new territory in the game’s expressive range. Because of its multiplayer ecosystem, these player actions taken while representing a Trading Company impact the community’s perceptions of that faction.
For instance, the Merchant Alliance faction is represented by NPCs who are seen as stiff, proper business folk who take no nonsense. Their missions often center on transporting valuable goods in a timely fashion. By reflection, players who opt to take on this identify often play in a reserved and defensive way, avoiding danger and seeking treasure only opportunistically. Since they have invested substantial time and effort into their commodities, they have a reputation as some of the most intense and ruthless opponents when challenged for their treasure.
Another faction, the Reaper’s Bones, are explicitly aggressive and hostile. Their NPC representatives speak in villainous tones and often task players with preying on emissaries of other Trading Companies. They are the game’s PvP advocates; a pirate who raises the Reaper’s flag is perceived as someone to be cautious around or to avoid entirely.
Roleplaying as directed by these strongly characterized factions has led to enduring player stories that both reinforce expectations and subvert them. During the limited-time “Lost Sands” event, players could choose to support the Reaper’s Bones efforts to destroy a beloved outpost called Golden Sands or, alternatively, they could help defend it. One player created a mock interview featuring his character Reaper Ron, a good-natured pirate who innocently joined the Reaper’s Bones thinking their reputation was undeserved and that they were simply misunderstood. Reaper Ron humorously recounts his horror at seeing the damage accrued at Golden Sands, whereupon he realizes he was catastrophically mistaken. This story with its subversive take is the contrivance of a creative player-author, but its telling was enabled by the Sea of Thieves devs staking out a clear and compelling expressive range.
| Intervention phase: devtime Where the results show up: playtime, sharetime Sharing motivation: invites players to collaborate in interpretation and participate in narrative completion |
While defining the game’s expressive range, it can be fruitful to intentionally omit details. As game designers and worldbuilders we often intuitively strive to craft beautifully deep, rich lore that entwines with our elegantly complex systems. There is an impulse to retain full authorial control. But lacunae are powerful opportunities for players to speak up and contribute their own interpretations, hopefully supported in their efforts by a well-defined expressive range. When a game has a strong voice, lacunae can gesture at satisfying kinds of answers to the questions they pose.
In some cases, lacunae are best implemented as known but deliberately omitted facets of a storyworld. This approach allows players to feel clever and “invent” explanations that fit the clues implied by the extant design. As designers, we can retain control of the possibility space while offering players inroads to their own role in authorship. We can even introduce our canon answers to these omissions down the line, in an update, sequel, or DLC, and give players the chance to compare and contrast their interpretations with the game’s across time.
In other cases, lacunae are best implemented as genuinely uncrafted, undesigned spaces for players to fill in. Whether or not the community perceives the difference, its contributions are filling in real holes, and it’s acting as genuine co-author. This wilder, more divergent approach loosens the authorial reins, but it has the potential to tap into the unique aesthetic space of reparative play5, where players are actively invited to bring their own meaning-making processes to bear on their experiences. This kind of personalization is a catalyst for tellability.
Boyfriend Dungeon was a viral phenomenon in the fall of 2017 due to its easily shareable, humorous premise of taking swords on romantic dates. In particular, the visual moment of transformed sword-to-shirtless-man was shared not just across social media and games websites but also major news organizations and morning talk shows. A successful Kickstarter the following year gathered a community of thousands into the developer’s Discord server, who immediately began sharing questions and theories about the details of how exactly a sword turns into a weapon. Here are some examples of their theorization:
The devs internally decided on a few answers to some of these questions while creating the game’s content, but they didn’t communicate these answers to the fans. Even after the game’s release in 2021, some details remain intentionally unexplained or were given limited or conflicting answers; different characters follow slightly different paradigms of transformation and make comments that can be read as explanations but may just be jokes.
Normally the dev studio, Kitfox Games, practices deliberate transparency and a willingness to discuss its process with fans. They place strong emphasis on grassroots community communication, so the use of lacunae here was unusual. The reasons for this deviation in Kitfox’s approach to Boyfriend Dungeon lore specifically was threefold: to make room for future content development, to preserve tone (as a primarily comedic game, too much exposition would drag it down), and to give players room to formulate their own interpretations.
Though it’s hard to isolate factors, this strategy seems to have been successful. Boyfriend Dungeon has spawned 70+ fan works of fan fiction on the fanfic site Archive of Our Own (AO3)6 — despite never being linked or directly referred to by the devs — and years of chatter on Discord, despite being a single-player game that’s not particularly replayable.
For comparison, another humorous dating game, Hooked on You: A Dead By Daylight Dating Sim, appears to have sold approximately 2-3x copies as Boyfriend Dungeon (judging by the imprecise but useful metric of volume of Steam reviews at time of this report7) perhaps due to being attached to the large IP Dead by Daylight. Both games have similar numbers of Steam forum posts, and despite Boyfriend Dungeon being released over a year earlier, both have extremely similar frequency and recency of posts. It’s difficult to compare the presence of the two directly on social media due to Hooked on You appearing largely under the parent Dead by Daylight IP, but Hooked on You has just 18 fan works on AO3, and Boyfriend Dungeon has 33% more followers on Twitch (4.3k8 vs 3.3k9 at the time of this report).
| Intervention phase: devtime Where the results show up: playtime Sharing motivation: gives recognition to unique play experiences in a shared language other players can understand |
Often, players want to feel seen. In particular, they want to feel recognized as unique individuals with unique behaviors. Acknowledgements by a game can reinforce a player’s perspective on their play experience and support or kickstart a self-told narrative that might become more broadly tellable. With acknowledgements, the designer hopes to break through the mechanical predictability of the gameplay loop and let the player know that the game is a collaborator (or, potentially, a challenger) to their ongoing perspective.
Acknowledgements can take many forms. They can appear in character — say, as affirming dialog by an NPC — or out of character, in the interface or metagame. They can be tonally varied, from celebratory rewards for jobs well done, to neutral historical annals, to punishing reminders of failure. They can be big or small. Consider how the player’s photoreel is displayed during the credits of Campo Santo’s Firewatch. It’s a small thing, appearing after the action in the game’s coda, but it sends a powerful message that the game saw what you saw, and it cares.
“I prefer to think in terms of ‘yes, and’ / ‘yes, but’ reactions to player action: they get something they expected, and also something that they didn’t expect. The first part is their reward and reinforcement for taking action; the unexpected thing is the avenue by which I’m able to drive the story forward, whether that’s introducing a new conflict or changing up the stakes.”
-Emily Short
The most widely used tool for acknowledgement in games is the achievement, an out-of-character, metagame callout of a particular player action. Due to their ubiquity across nearly all genres and platforms, achievements are used by designers as a standardized tool to recognize progression through and mastery of a game. Progression achievements may induce player stories because they appeal to desires for completion and competition, but those stories are likely to be flat, since achievements of this kind are expected and impersonal, two anti-traits for tellability.
However, achievements can also plot outlier points in the game’s possibility space. By doing so, they can reward unusual behavior and at the same time help define the game’s expressive range. These kinds of achievements are themselves tellable moments — what it means to do something noteworthy in this storyworld — and their presence can help validate a player’s perspective on their play experience, perhaps eventually getting stitched into a broader narrative.
With one of the more elaborate dialogue systems ever created, Andrew Stern and Michael Matteas’s Façade offers a responsive, nuanced encounter between the player and two non-player characters, Trip and Grace. Despite being a small, non-commercial experiment, it’s been downloaded more than 5 million times10, and its shared player stories exploded from 2004 to 2007, even appearing in The Atlantic Monthly. The game’s differentiator was its ability to acknowledge seemingly any player input, with voice-acted character responses changing based on context and mood. This varied, textured responsiveness makes players feel recognized in their choices of how to interact with the Trip and Grace.
As described by the Atlantic Monthly reporter Jonathan Rauch: “They try to draw me into their simmering argument, nudging me to take sides. I can say anything I like; there are no rules. I can be sullen and unresponsive (that got me kicked out of their apartment), or I can talk nonsense, but in most of my visits I try to behave like an improv actor, picking up on their lines and shooting back cues of my own — agreeing with one, criticizing the other, flirting with either or both. No two plays are identical.” In the same article, Will Wright marvels at how Trip and Grace can be only slightly annoyed at the player, a social subtlety rare for videogames.
The relatability of the scenario and its characters, combined with the first-person perspective, full voice-acting, and open-ended input, result in deep and amplified feelings of recognition. When videogame Youtuber Jacksepticeye played the game (his video has 4 million views at the time of writing), he was immediately delighted at how in the first few seconds of play the game acknowledged his awkward kiss on the cheek with a zoom-in (an invasion of personal space delicately ignored by most first-person games), exclaiming “I love this!”. Titled “WORST GUEST EVER”, this video details the narrator’s continuous attempts to be rude and insulting, an almost combative demand for acknowledgement.
Whether the player engages with the characters in good faith as a conversational partner or tries to cause mischief through misbehavior, Façade offers ample acknowledgement. This has resulted in countless, varied, and personal player stories.
| Intervention phase: devtime Where the results show up: playtime, sharetime Sharing motivation: gives players micro-authoring tools to put emphasis on pieces of their play experience they find compelling |
One way to offer the player expressive space and encourage a storyteller’s mindset is to ask open-ended, subjective questions that don’t affect outcomes within the game, but that invite the player to draw story elements out of their own experiences in play.
Fallen London does this by giving each player “mantlepiece” and “scrapbook” spaces in their profile, where they can display objects from their inventory that they consider particularly important to their character. Those objects could come from any story or event achieved in years of play, giving players a broad range of possibilities when expressing who they are. Because the game is a long-running live project, players may keep the same mantlepiece items for a long time, or change them out when they play a fresh event that significantly changes their ideas about their own roleplaying.
Framing invitations like this can occur embedded in gameplay as well as in journal or profile elements. Elsewhere in Fallen London, the game asks choices that are again entirely about story formation, such as “why are you joining this quest?” or “how does your character now feel about those past events?” Frequently the game also provides a “none of the above” option, for role-players who would like to imagine a response not anticipated by the developers.
The question “what do you want on your mantlepiece” is, of course, very open-ended and allows players to make up their own minds about what kinds of stories even interest them. But we can also gain by asking much more specific questions that relay something about this game world. Character formation questions asked in tabletop roleplaying games are a good source of inspiration for questions like this.
It’s often especially effective to invite players to consider character motivations. Why is the protagonist doing this? What past event or encounter set them on their current path? These are narrative decisions that add character and richness to the story but are often orthogonal to the elements actually narrated by the game.
| Intervention phase: devtime Where the results show up: playtime Sharing motivation: provides players with a reduced set of storyful moments to compile their own stories from |
You can ease the authorial burden on your players by providing curated summaries of their experiences, which can then be more easily picked through and assembled into their own stories. To use a mining analogy, you’ve taken them on a tour through the central mine with a well-defined expressive range, you’ve handed them a pickaxe and shovel with your framing support tools, now with curation, you’re giving them a map to the rich veins of ore. Curation is meant to help the player understand what happened in the first place, to make something otherwise-abstract concrete and otherwise-complex straightforward.
A simple example is a high-score table or progression chart provided at the end of a game session. The Civilization series has experimented with different flavors of this endgame summary, to varying levels of success. During play, your player might not have noticed that from the years 200 to 300 AD their empire was on the brink of collapse, but the graph comparing them to their neighboring countries can provide this insight and motivate a re-telling.
More elaborately, the Legends mode of Dwarf Fortress is a great example of curating a huge volume of content into a meaning-rich subset. While generating hundreds or thousands of years of history for several civilizations, the game’s world generation engine will print callouts to notable events, such as the defeat of powerful creatures, the creation of artifacts, or the founding of cities.
A curated list of important events seeds the storytelling soil and gestures at which artifacts, characters, and places may be worth further investigation. Future encounters with these highlighted game entities are likely to be more notable to the player.
Consider combining this technique with support for player framing; letting players tag characters, places, and events as notable — in the same language the game itself does its curation — is like giving your players a writer’s notepad and a shorthand for notetaking.
| Intervention phase: devtime Where the results show up: playtime Sharing motivation: invites players to compare and contrast counterpart play experiences and their narrative implications. Invites players to collaboratively map out a possibility space. |
As discussed, underdetermining a game’s storyspace through the use of lacunae can encourage player participation in authorship. Procedural generation techniques can work in a similar way. Procedural systems produce their own kind of lacunae, but not directly sculpted out of the storyworld in the same way a narrative designer might do explicitly. Instead they’re formed in the gaps implied by traversals through an open possibility space. Procedurally generated content takes strict canonicity out of a playthrough, instead requiring players to use their own interpretative tools to make sense of their experiences and narrate what’s happening. When executed well, these techniques can produce moments that are singular, non-repeatable, and unexpected: all traits ripe for tellability.
In particular, procedural generation lends itself to collaborative storytelling, both because interpretation is discursive (players will want to discuss readings of experiences that have no canonical answers) and because traversals through a generator’s possibility space can be compared and contrasted for a richer picture of the whole. A playthrough is given contextual meaning in relation to others; players situate their experiences inside a larger story volume11. This provides an additional impetus to share. Here’s where establishing a strong voice for the game and defining an expressive range (per tool #1) pays off; you will have trained players on what stories look like “here”, and together they will write the tales of your game world.
Roguelike/RPG Caves of Qud uses a hybrid approach of authored and procedural content to generate its NPCs. One character, Sheba Hagadias, has hand-authored traits, is always found in the same location, but has her species procedurally-generated each game. Her backstory is always the same: she fled a hostile family life to become a librarian in a knowledge-poor world. As fantastical social and biodiversity is part of the Caves of Qud expressive range, this variance is fodder for a broad spectrum of interpretation and reinterpretation of Sheba and her role in the storyworld of Qud. Players speculate on what Sheba’s harsh family conditions might have been when she’s a sapient baboon vs. a pig farmer vs. a sentient laser turret. By providing a familiar baseline and dynamic material for remixing that baseline, players are motivated to collaboratively explore the range of possibilities and what, all together, they say about the storyworld.
Indeed, Sheba gets a lot of fan fiction treatment in the Caves of Qud discord. Whether it’s extended stories or small snippets (it’s common to post “rare Shebas”), Sheba as a character has staying power in the communal imagination, and that is driven by her underlying, thoughtful procedural generation.
| Intervention phase: sharetime Where the results show up: sharetime Sharing motivation: gives players a place to share stories and moderation to feel comfortable & empowered to do so |
Not all interventions happen at devtime. As gestured to in the introduction, providing social spaces for your community can help engender a storytelling culture. Virtual community spaces change formats over time (web forums, blog comment sections, Discord servers), but their core remains the same: places where individuals come together to discuss your game and form community. From that, stories will follow.
At their most straightforward, community spaces can function as sites where player stories live. Additionally, they can put players in contact with each other while playing, allowing their conversations to enrich the play experiences and their interpretations. Consider designing your spaces with both of these modalities in mind. For example, the Caves of Qud Discord has one set of chat channels for talking about live play, another set dedicated to posting screenshots of memorable moments (that may become fodder for more complete stories), and another channel for posting longer form fan art and fan fiction. These delineations help players navigate your spaces and catalyze the storytelling process through community structure. They can even dovetail with other strategies by providing another avenue for framing support and helping to define and reinforce the game’s expressive range.
In order for these spaces to thrive as community campfires players tell stories around, however, you cannot simply build them and walk away. Managing community through active moderation protects player safety, at a minimum, and gives room to marginalized voices that may otherwise remain quiet in an anything-goes environment.
Dwarf Fortress has already been mentioned, as it’s famous for its player legends. Its design makes use of several of the techniques discussed in this report, but often overlooked is the effort the devs put into community management. Even as a tiny team of two and occasionally literal calendar years between updates, Bay 12 Games still maintained a forum and moderated it consistently, banning toxic users, gathering and responding to feedback regularly, and posting public newsletters with progress updates. This provided an ecosystem for players to meet during sharetime and a culture that invited discussion and collaboration.
With Caves of Qud, after popular Youtuber SsethTzeentach posted a review that flooded the Discord server with bad actors, the devs and community moderators locked the server and instituted a stricter vetting and onboarding process. This approach has tradeoffs, but it did lead to the cultivation of a safe, welcoming space with many community members who might not otherwise feel comfortable sharing their stories.
The best player stories arise organically, but like all forms of constructed emergence, the conditions they arise out of matter. As designers, we cannot tell these stories ourselves — they aren’t ours to tell — but we can plant the seeds and help translate compelling play experiences (which we’re already designing for) into tellable, memorable, moving stories.
In this report we’ve introduced a pipeline for contextualizing how, why, and when players tell stories, and we’ve enumerated seven techniques that our analysis and experience have shown to be useful in catalyzing the storytelling process.
Our pipeline is a start toward understanding the transformations that yield story out of play, but it’s incomplete. We’d like to see a deeper, magnified investigation into the processes that occur when players come into contact with games, drawing insight from fields like media studies, psychology, sociology, and narratology.
This report came out of early discussions around how devs might collaborate with their communities to tell stories, potentially folding player contributions back into the games proper. In the course of that investigation we uncovered how much there was to say about player storytelling in the first place and how much collective knowledge we had on the subject. So we pivoted to the contents of this report, but the question of how to close the loop and bring player stories into the fold remains a compelling (if exploratory) one. Our preliminary findings are included in the appendix of this report; we hope someone picks up the investigation.
Some of the enduring community legends — like the World of Warcraft “Corrupted Blood incident” linked to in the introduction — arise out of accidents, bugs, exploits, and game dynamics gone awry. Clearly, designing for these types of stories is a whole different animal, if it’s possible at all. We’d like to see an examination into the possibilities of tactfully handling these kinds of incidents when they arise. Perhaps there are no interventions to be made at devtime — but what can be done at sharetime to acknowledge these incidents, playfully, and elevate the stories they produce, without putting our fingers on the scale and ruining the organic authenticity that drives interest in them in the first place?
This report has examined tools for planting the seeds of player stories. For certain kinds of community-engaged games, we can imagine harvesting those player stories, too, and incorporating some subset of them back into the games themselves. There is precedent for this, but it’s a much more exploratory topic. What are the possibilities for harvesting player stories? How can devs do it sustainably while leaving open avenues of remediation for when things go wrong? A full investigation into harvesting and how it fits into the lifecycle of story cultivation is still to be done, but what follows are some observations and an entry point into the problem space.
There are two common routes to harvesting player input:
Both approaches have significant challenges and pitfalls. Development teams may rightly be cautious about engaging with player-written content in case of legal entanglements about the ownership of intellectual property. Many game writing teams, including the Failbetter team that writes Fallen London, intentionally avoid reading player fan fiction in order to guarantee that they are never even accidentally lifting an idea from a player.
Along the same lines, many development teams have offered Kickstarter rewards that would allow players to contribute names or ideas to play, only to find themselves needing to carefully edit the results to remove trolling, off-genre, or otherwise problematic contributions.
It’s possible to solicit player input for future development in a more controlled way: for instance, for many years Fallen London incorporated yearly elections in which the player base would choose the mayor for the next year, out of a slate of three characters chosen by the developers. After the election was concluded, the development team would write the new mayor into that year’s stories. The elections were often successful in attracting a lot of player engagement, and game forums filled with players campaigning for their preferred candidates and even posting fan-made political slogans and cartoons.
This method requires significant development time and effort, and there’s usually a necessary delay between the point where the players get to make their choices and the point where the consequences are seen. In Fallen London, election results would often result in new story content a month or two later — a timeline that often felt demanding to developers but frustratingly slow to players.
Finally, elections in Fallen London suffer from some of the same communicative limits as elections in the real world: in asking the player base to communicate with one small piece of information (“who do you think should win this election?”), we weren’t able to note or respond to the often-differing reasons behind their votes. Did the player base vote for the cat mayoress because they are fans of cats? Because they expected she was promising a type of content they were excited about? For some other reason?
And just as in the real world, Fallen London elections could become contentious, leading to arguments between players. It was this aspect of the design especially that led the Failbetter team to cancel and replace the event.
Meanwhile, many of the most obvious mechanics for building player storytelling back into a game require that the effects of that player input be quite simple and limited: a battle that can end with one of two sides in charge, for instance.
In many cases, this means that the development team has to invest in building both branches of a branching storyline, but only see one of those branches put into play: this can be a frustrating waste of resources for the development team, and frustrating for players as well when they learn that there’s an entire alternate timeline that they can’t access. Meanwhile, some portion of the player base may be significantly frustrated if they wished for the opposite outcome and didn’t get it.
A common alternative is to build a world that can cycle between states, with a battle that can be fought in either direction (for instance). This also has its frustrations, however: the story of the battle may feel meaningless and static to players, since it can always be replayed; the game loses the sense of that battle as a decisive communal event.
Finally, perhaps the biggest issue with this approach is that it makes the least interesting possible use of player input. Community riffs on stories are often the most fun when they take the story in surprising directions, identify lore connections that the developers hadn’t considered, or even break some aspect of gameplay in a narratively satisfying way. Offering the player base a collective choice between two or three balanced outcomes does little to leverage the collective imagination.
Even when the development team doesn’t commit to an ongoing project of engagement with the community (player-facing LiveOps and curation), the team can still choose to notice something interesting or impressive that emerged from player behavior, and then memorialize that event in-game.
An example of this type of memorial is the Destiny Loot Cave12. When Destiny launched in 2014, players quickly found an exploitable corner of a map where they could easily farm arbitrary numbers of enemies. Bungie patched the exploit several weeks later, noting that while they were glad players enjoyed the discovery of a social play space, they hadn’t intended the game to be played this way.
While similar other exploits were discovered (and quickly patched), the legend of the original Loot Cave — and the emergent player-driven etiquette around it — became community lore. When Bungie released new content to celebrate the studio’s 30th anniversary, a time ripe for memorializing cherished player moments, they took the opportunity to commemorate the Loot Cave incident. The opening of the new dungeon is set in the location of the original exploit, letting both veteran players and newcomers experience it anew. The devs wove an in-game story around the location that explained how the glitch might have come to happen within Destiny‘s storyworld. The resulting experience was both a satisfying acknowledgement to a community legend and a new means to continue telling it.
In 2021, Fallen London retired the election event and replaced it with a new, one-off summer festival event designed for collaborative (rather than competitive) play.
In some ways, this design was less responsive to player choice than the election, in the sense that there was only one overarching storyline, and developers made no commitments to write followup content based on the outcome of the event. Instead, players worked together to achieve targets and unlock new in-game content, in whatever order they collectively chose. The developers had somewhat expected that players would collaborate to work on one unlock area at a time; in practice, the players worked simultaneously on all areas, resulting in a slower initial pace followed by a burst of discovery.
The next year, the Fallen London team looked to make player-driven pacing an intentional part of the design. 2022’s summer festival featured a mystery lore hunt in which clues to a mystery were scattered throughout the game. Some clues were easily accessible and could be found by early game players with basic play; others were hidden in endgame areas or required significant investigation to reach. Each clue had a (different) number of times it needed to be discovered in order to count towards the game’s overall progression, so that easy clues needed to be found by hundreds of players, and the most difficult clue only by a single player.
The game also included signposting to help all players keep up with the discovery of lore, even when they couldn’t reach it. A central area of the game updated when each clue was partially unlocked (giving a hint about where it could be found, so that stumped players could catch up) and again when it was fully unlocked (exposing the new lore).
This mini ARG-like feature encouraged story development in real time (as players interpreted the clues found so far and speculated about where they should look next). Meanwhile, the design of the more difficult clues created “hero” experiences for specific players who were first to reach and unlock difficult elements.
Another way around the challenge is to create live events with sufficient flexibility that the developers can interact more like tabletop game-masters, within specific parameters. What those interactions look like may vary widely depending on the type of live game and the scope of support available to the devs, but this approach benefits from both the potential for surprising outcomes and the safety of dev remediation.
In addition to our references, we’ve compiled the rest of the links included in this report, categorized and in order of their inclusion.