Jason Grinblat, Max Kreminski, Graeme Lennon, Cat Manning, Alexei Pepers, Lauren Scott
Why is it fun to learn the rules of a videogame? Not just to play the game, but to learn it; to travel from the state of unknowing to knowing, to be surprised when a revelation subverts your expectations, to play in a state of suspense?
We know the feeling of suspense from other mediums. In literature and film, it’s largely a narrative phenomenon. As the metaphor implies, we experience an ungroundedness, a sensation of charged uncertainty at what might come next – whether it’s anticipating the killer’s next victim, or wondering if a pair of star-crossed lovers might end up together, or imagining exactly how a situation poised for tragedy will fall apart. In music, suspense is evoked through the use of dissonant intervals that prompt us to pay attention for when their harmonic tensions will resolve.
Because videogames are multimedia works, they can import the structures that produce these types of suspense. Narrative games might evoke narrative suspense through good writing; a horror game might heighten a tense situation with suspenseful music. But there’s another kind of suspense unique to videogames that we aim to identify and characterize in this report. Systems suspense, and the correlated phenomenon of systems surprise, are spun out of the uncertainty inherent to engaging with a videogame whose rules are not fully known to you. Just as the author draws the reader along a path through their work from unknowing to knowing, taking care to wind the path at certain angles to offer views that impart specific emotional affects, so can the designer draw a path through their game’s systems to produce a particular tenor of experience for their players and tap into the aesthetics of suspense.[1]
In this report we attempt to draw boundaries around the phenomena of systems surprise and systems suspense, to understand what their aesthetic textures are and how they relate to each other, and to explore how we can design systems to produce their peculiar joys.
This report was motivated by the authors’ experiences of delight at playing certain games that swerved in unexpected ways. Those moments of “whoa!” and “what else might happen??” (the two affectual pillars of surprise and suspense, respectively) are positively charged for us, but they may not be for others. That is a question of aesthetic preference, and as such, we won’t be addressing it in this report. Instead, we assume surprise and suspense are desirable in certain situations, and we focus our lens on the question of how to build game systems that produce them.[2]
We also assume the phenomena we’re describing are close enough to the types of surprise and suspense referenced from other mediums to make the analogy useful. But we invite further investigation into the mechanics of surprise and suspense in those original contexts, both to understand them better and to enrich the new meanings presented here.[3]
Finally, surprise and suspense are constructed in the course of maneuvering through expectations, which can vary greatly among audiences. What is surprising or suspenseful to one player will fall flat for another, or even worse, evoke disappointment and frustration. As in all of design, the decision of how and where to invoke surprise and suspense will be in conversation with the larger set of thematics and aesthetics operative in your game; there are no universal answers. However, we will spend some time picking apart the layers of material that compose expectation as a category of thing.
In FromSoftware’s 2022 open-world action RPG Elden Ring, after you complete the tutorial dungeon, you emerge onto an overworld fantasy landscape. Shortly after interacting with the first above-ground Site of Grace (the game’s version of a checkpoint), you get a popup tutorial message about the world map.
The map itself is a bounded rectangle shaded by fog at its borders and emptiness in its center. A few discovered points of interest are marked.
A mixture of genre convention, visual cues, and explicit tutorialization set expectations for engagement with this map. It’s typical for RPGs to feature a world map, and the few points of interest laid out against a sparse background imply that more will be filled in as discovered. The size of the map appears to conform with reasonable expectations for a game of this scale, and so players are prompted to understand the borders of the map as representative of the borders of the playable world. The metaphor that coheres is as follows: the map is a visualization of the player’s and the player character’s knowledge of the playable world, the blank spaces are the gaps in that knowledge, and the bounded rectangle is the full possibility space.
But something peculiar happens when you reach the edge of the map, either by walking far enough toward it or activating a cleverly placed teleport trap at the bottom of an early dungeon. Instead of merely filling in the blank space with new discoveries, the edges of the map expand to a much larger rectangle; the game world is bigger than the game has let on.
This swerve goes beyond the simple revelation of new content. It changes the contract the game had set up with its players by “breaking” the rules, and thus it forces players to reexamine their deepest beliefs about the game. This moment of delightful, systemic surprise (“Whoa!”) is complemented by an emerging sense of suspense (“Will this happen again? What else might happen?”); the game has proved itself willing to subvert our expectations of its systems, and so we’re on edge as to whether it’ll happen again.
With assumptions set forth and a shared example at hand, let’s define the terms that are central to the rest of the report.
Surprise – a revelatory moment where a player’s understanding of the game is sharply altered
Suspense – a state in which the player is actively and excitedly aware of uncertainty in their understanding of the game
In the following sections, we occasionally draw a distinction between systems surprise/suspense and content surprise/suspense. Though the distinction is not always clean (due to the fact that content and systems are not always cleanly disentangled), the affects of surprise and suspense can exhibit factors that push them in one direction or the other.
Systems surprise/suspense conforms to the definitions and examples we’ve given up until now. In particular, these are affects produced by an evolving picture of a game’s very rules. They are characterized by multiplicativity; systems surprises tend to recontextualize and enrich past experiences with the game. Imagine, for instance, the sudden realization that previously experienced, “scripted” moments of gameplay were in fact emergent outcomes powered by systems more sophisticated than assumed, and so those outcomes you thought were preset could have in fact unfolded very differently.
Contrast this with content surprise/suspense, which may function similarly to their systems analogs (by providing revelatory moments or putting the player in a state of tense anticipation), but which aren’t executed through mental refactors of the very rules of the game. For instance, compare the Elden Ring map example above with how the map works in Nintendo’s action RPG The Legend of Zelda: Breath of the Wild. In the latter, undiscovered regions are similarly kept from the map, but they are instead outlined in blue until discovered. Even though the revelation of new areas may spark delight or prompt questions about future discoveries, the metaphoric language of the game’s map is kept consistent; no contract is set up to later be subverted, and so this class of surprise and suspense isn’t systemic in the same way.
If we are to identify systems surprise and suspense as distinct experiences a player can have within a game, a closer examination of what is occurring to the player throughout that process is required. Both are defined as states the player can be in, and while output from the game does provoke them, the experience itself unfolds within the player’s mind.
To examine this internal experience, we use the term mental model to refer to the framework the player builds that maps their understanding of how the game operates. We call it a model because it is not a catalog of the actual moments the player has experienced up to this point, but rather their understanding of the game as a process that has certain rules and constraints. Specific moments within the game feed into this larger possibility space that encompasses all the moments that the player believes could be possible given the systems of the game. As we’re concerned with suspense and surprise within the context of systems and rules, considering how mental models are formed and modified will allow us to better explain how the player’s system understanding can provoke these feelings.
The most direct method a player has of updating their mental model is also the one we have the most control over as game developers – that of experiencing the game themselves and learning from how it behaves. Whether the player is actively focused on learning the rules of the game or not, the experience of playing provides them with test cases from which they can extrapolate patterns. As a piece of hardware, a videogame does necessarily have an underpinning set of rules and processes, and so it’s possible for a player’s mental model to become an increasingly ‘correct’ one as they gain more information through play.
Explicit testing of one’s own mental model is a form of play that is often encouraged and recognized as a type of player motivation. Scenarios are designed to give players the opportunity to form hypotheses and test them – what happens if I bring this item over there, what happens if I combine this with that, what happens if I do A before I do B, what happens if… The results of such testing might be intentionally obscured by the game, or be misconstrued by the player, but by internally forming a hypothesis and taking actions to test it the player is specifically engaging with their mental model and updating it based on the results. By pushing on the edges of the possibility space, they work to find a wall where the game sets a limit on what it is capable of, which becomes a defined edge in the shape of their mental model.
There are also techniques we use to try and impact a player’s mental model in a more directed way. Tutorialization and providing explicit instructions on how the game works are ways we provide information that is meant to directly inform a player’s mental model. This lets us attempt to prevent the formation of an incorrect mental model which could lead to frustration or confusion. It is also a kind of scaffolding that defines the basic structure of a model, to aid players whose initial model might be overwhelmingly open-ended. Tutorialization often serves to make explicit the base rules that must be understood in order to productively engage with the game at all, while leaving out the rules that can be unknown for a time without negatively impacting the experience.
Finally, there are ways we can draw a player’s attention to their mental model without actually filling it in. One is to plant breadcrumbs – small pieces of content within a much larger system can make the player aware that a system exists that isn’t currently captured in their mental model, while providing them only hints at the shape of that system. Examining an item in your inventory and seeing it has tags like ‘Vegetable’ and ‘Spicy’ associated with it can cause a player to consider what purpose those tags have. If the idea of a cooking system already exists in their mental model, it can reinforce their belief that this particular game does include cooking even if they haven’t been given the opportunity to interact with it yet.
Another technique on a more systems level is to establish patterns of behaviour, like the concept of regular and rare outcomes. Multiple systems within the game can be set up so that there is a regular outcome that occurs in most cases, but also a noticeably distinct rare outcome that only occurs sometimes. An enemy sometimes spawning as a rare variant, a loot table sometimes including a rare drop, a chance for a rare event to occur on an otherwise regular day – if the player is made aware of the exceptionality of the rare outcome and they experience it enough to get a sense for the frequency, then they can form an expectation that other systems in the game likely have a possibility of rare outputs even if they haven’t yet seen them. For example, in Caves of Qud a bug resulted in an inanimate table being chosen for a village pet instead of a living creature.[4] The experience was consistent enough with the desired feeling of Caves of Qud to inspire a function that encapsulates this “most of the time to this, but rarely do that” design pattern, so that even after the bug fix, an inanimate object can be chosen as village pet 1 in 1000 times.
We have focused so far on how the player’s mental model is formed and modified while actively playing the game, but equally important is how the player’s model is formed before they even launch the game for the first time, as well as in all the moments between play sessions. We’ve touched on this difference in mental models in mentioning how moments of suspense and surprise might land for some players but not others, as well as how mechanisms like tutorials are sometimes used specifically to make up for basic rules that some players might lack coming in to the game, while others will have them already formed and can jump right in.
Any game exists within many contexts that carry with them sets of expectations. Some examples that are especially relevant to the size and shape of a player’s mental model are:
Genre Conventions – a benefit and drawback of a game associating itself with a given genre is that with a brief descriptor, the player is expected to bring in a set of expectations for what systems are likely present and how the game broadly behaves.
Budget/Scope – whether it’s broad strokes terms such as indie and AAA, or more specific markers such as budget a team size, a player’s perception regarding the generalised budget of the game impacts their expectation of scope and complexity.
Franchise History – more specifically than genre, a game might be associated with others in a shared franchise, previous titles from the same studio, or other projects from notable members of the development team. All of these can result in a player coming into the game with a mental model already shaped by those previous experiences.
Advertising/Framing – the process of selling a player on a game involves giving them information about what experiences it will provide. Appealing to a particular player fantasy, sharing stories of memorable moments within the game, and associating the game with particular demographics are all examples that impact player expectations and can be found in marketing materials as well as critical reviews or recommendations from friends.
An important detail for when we get into a consideration of surprise and suspense is that the expectations that are informing a player’s mental model are not just concrete details like advertised systems or returning characters from previous games in the series, but more abstract expectations of patterns and the kind of game this is.
One example of this is the fact that a player’s mental model can include an expectation that whatever rules the game initially presents, they are going to be subverted later on. Puzzle games as a genre commonly use this pattern of setting up rules and then bending them, such that players familiar with the genre may be expecting just such a twist, even if they don’t know exactly how it will manifest. More specifically, Frog Fractions is a game best known for its unpredictable transitions between drastically different modes of gameplay. The twists were especially novel for players who didn’t know what to expect going into the experience. When Frog Fractions 2 was released, the mental model of players who were familiar with the first game were in a very different state as they were now primed to expect these kinds of maneuvers. Their mental model included the expectation of unpredictability and subversion, even before Frog Fractions 2 had set up a system to subvert (making it much harder for the sequel to evoke the same kind of surprise and suspense!)
A broader type of abstract expectation can take the form of familiar patterns. The three-act structure of narrative can be familiar to players from experiences outside games, and is also common enough in the structure of video game narratives that a mental model which includes an expectation of a second act setback before the third act climax will often prove correct. The UI conventions around party slots is established enough that if a player has experienced enough party-based games, they only need to see a character select screen with four slots to update their mental model to expect at least four playable characters. As the Elden Ring example showed, patterns of RPG tutorialization are strong enough for players’ mental models to include an assumption of what the first map they’re given represents about the edges of the world – one which is proven wrong, to their surprise.
With this definition of the mental model to work with, we can now redefine our central concepts of surprise and suspense in terms of what is happening to the mental model, instead of as a description of the state it evokes in the player. Beginning with surprise, we can understand it as:
Surprise – a sharp and sudden re-orientation of the player’s mental model into a new and noticeably different one.
A way to visualize this definition of surprise is to think of the player’s mental model as a box that the player has drawn around the possibility space of everything the game contains. As the player builds up a model of the game, they layer rules and expectations outward, expanding the box. The reality is much more complex but the core idea is that it’s a demarcation between what is possible in the game, and what isn’t. When surprise occurs, that box is suddenly tilted and sheared, dropping some of the previously included possibilities while just as rapidly gaining new ones.
Earlier, we defined systems surprise/suspense as characterized by shifts in the game’s rules that lead to recontextualizing of past experiences in a new light, versus content surprise/suspense which doesn’t have this reframing effect. With this visualization of surprise in mind, we can think of systems surprise as the shearing effect where everything inside the box is resting on expectations that may have just been disproven, and so large areas are suddenly on the outside as what the game isn’t and are replaced with the new understanding of what the game is. The impact of such a surprise can feel so large and powerful because of the chain-reaction where the adjustment of a single rule ripples out to impact all derived rules and expectations.
Content surprise on the other hand is an additive change to the mental model. For it to register as a surprise, the addition must be large enough to not just feel like you’re learning more about the game. It also needs to be an actual addition to the possibility space, which as we discussed includes not just what has been proven to exist in the game, but also all the other player expectations based on the bigger picture such as genre conventions and advertisement. Finding a piece of content that they didn’t know was in the game isn’t enough to trigger a feeling of surprise for a player, if that type of content already fit into their mental model of what kind of content they were likely going to find. Without the multiplicative effect of systems surprise, it is much more difficult for content alone to generate the same feeling of surprise. Examples like the reveal of more areas of the map in The Legend of Zelda: Breath of the Wild need to provide a large amount of content in terms of scale and novelty in order to cross the threshold of surprise.
Next we revisit suspense which we define in relation to another term – fuzziness. If we revisit our visualization of the player’s mental model as a box, then the hardness of the edges of that box represent how confident the player is in the boundary they’re drawing around the game’s capabilities, and their ability of their model to predict what is inside and outside the realm of possibility. In contrast, a fuzzy line is not strongly defined and still has room for multiple contradictory possibilities. There is not enough information to form a unified set of rules that fit all the known facts, and so that edge of the model can’t yet collapse down into a predictable ruleset. This doesn’t mean that it’s impossible to reach a state of better prediction – just that at this moment in time, the player’s understanding is insufficient and the resulting demarcation between what’s inside and outside the possibility space is undefined.
We can now define suspense with slight changes in reference to these terms::
Suspense – a state in which the player is actively and excitedly aware of a fuzzy edge within their mental model of the game.
What this change in terminology helps us hone in on is that systems suspense is such an exciting feeling because it relates to the possibility space of not just what the player knew was in the game or knew wasn’t in the game, but everything they thought was possible. If they expect a game will have several party members, then having the identity of all those party members be unknown will not provoke systems suspense, because the player’s mental model can accommodate a clear understanding of “this is the kind of game that will let me gather additional party members.” If however the player collects one of those party members and their playstyle is drastically different than what the game had offered so far, the feeling might become suspense, as the player realizes that the firmness of their mental model’s boundary was based on assumptions like “the party members will offer different flavours of the same core gameplay” that are no longer certain. It might be that this was a single exception and the others will behave like the player originally thought, it might be that the next party member is yet another shift – the suspense and fuzzy edge stems from not knowing which of those options it is.
An important element to this definition is the player’s active awareness of the fuzzy edge within their mental model. It isn’t enough for the fuzziness to just exist – and given how large a mental task it would be to confidently map out all of what’s possible in a game, it’s normal for some areas to be fuzzy. The player might incorrectly dismiss some possibilities as impossible, or come to an incorrect conclusion of how a rule behaves without having yet seen any proof to the contrary. They can go through the entire game with a confident but incorrectly drawn model, and experience no contradictions and no suspense. It is only when they’re given the chance to test the impossible case and find it’s actually possible, or encounter a situation that reveals a previously unnoticed contradiction, that they realise their understanding is insufficient. If they do come to that realization that their current model is incorrect in a way – and that the model can’t just be fixed by replacing one clear rule with another, due to the lack of a clear rule – then that is when they can enter a state of suspense.
systems suspense in this framing is easy to describe because if our mental model is largely a set of understood rules, and the system element refers to the player’s awareness of those rules directly as opposed to the content that operates using those rules, then it follows that becoming aware of uncertainty in system-understanding leads to uncertainty of the game rules making up the mental model, and therefore suspense. But how can suspense be generated on the content side of the axis?
Our way of explaining how content can lead to the same state of suspense is to once again consider how expectations play into the mental model. In our example above of gaining new party members, if the additive change of a new playable character doesn’t “make sense” within the expectations the player used to build their mental model, then that edge can become fuzzy as the player asks “well if this exists, then what else could exist?” If they never expected this corner of their mental model of the game to expand, then the addition of content can lead to an uncertainty regarding just how far this new content will go – not just in terms of scope, but in terms of how it will relate to existing rules and content. The state of suspense can be experienced until the player feels like they’ve mapped the solid boundaries around this additive change. Crucially this all revolves around how confident the player is in their mental model, not the actual correctness of their model. The exception might actually be a one-off exception, but if the player thinks it might be a larger pattern and isn’t sure, then they can still feel suspense.
Now that we’ve defined the types of modification to the mental model that we call surprise and suspense, we can think about why they’re often but not always related.
When we began discussing the phenomena of systems suspense, our focus was just on that idea of suspense – but we found that often, our examples of games that gave us memorable feelings of suspense weren’t just notable due to the suspense, but due to the surprise that led into that suspenseful state. The state of suspense requires an awareness of the fuzzy edges of the mental model, and the sudden drastic shift of a surprise is a very effective way to draw attention to how much certainty you as a player have in the rules underlying your mental model that was just shaken up so dramatically.
While we do normally think of surprise as the inciting action to lead into a state of suspense, it’s also possible for a surprise to provide a resolution to suspense. The fuzzy aspect of the mental model during the state of suspense means that multiple resolutions to that aspect of the model feel possible to the player. If they are suddenly given a definitive answer to how that aspect of the model works, and that answer is unexpected enough to have been outside even the fuzzy set of possibilities that were up for consideration, the result will be suspense collapsing down into a solid but surprising edge of the mental model. At best, this will feel like a cherry on the cake of one last unexpected twist in the arc of learning the rules of the game; at worst, the distance of the truth from the player’s expectations will make the resolution feel arbitrary, depriving the player of the satisfaction of having at least one of their guesses be proven correct.
The connection between surprise and suspense is what we find the most interesting to explore, but it’s worth also recognizing the effectiveness of surprise even without suspense. A feeling of systems surprise can be experienced when you have one very clear understanding of the type of game you’re playing, and suddenly it twists into an equally clear but very different shape. Uncertainty isn’t required to be delighted at the new possibilities afforded to you by what the game has suddenly revealed.
We’ve already called out the fact that suspense relies on the player’s awareness of the fuzziness of their model, and so the state of suspense isn’t just triggered by a previously solid boundary becoming fuzzy, but also the realization that a fuzzy boundary exists that was there the whole time. This kind of dawning realization is especially notable for how much it allows a player to look back on their past experiences in a new light. They might wonder how long the uncertainty was hiding, and mentally replay past scenarios to see what clues they missed. To fully correct their misunderstanding they must sort through all the information they’ve gathered via previous interactions, and consider whether or not their new awareness of uncertainty casts doubt on the conclusions they’d made at the time. This recontextualization leads to suspense of this type feeling especially impactful, and gets even more impactful the more the player has experienced the game and built up a mental model to be disrupted.
By running through these examples, what we hope to make clear is that while suspense and surprise are often found together, they are two distinct concepts. They both rely on the player’s mental model, which we define as a set of rules and expectations that encompass the entirety of what the player believes the game is capable of. Suspense is then a state the player can be in, where they’re made aware that part of their mental model is fuzzy – meaning that they can see multiple incompatible possibilities of how the game could behave in some areas, but are currently unable to decide which are actually possible and which are not. Finally, surprise is just one way of calling attention to this kind of fuzziness with a sudden re-orienting of the mental model. Planning for a moment of systemic surprise as the player discovers a hidden rule or is nudged to explore an edge case is a great way to prompt them to more closely investigate their own mental model and discover those fuzzy edges – but it’s just one method of many, and we now aim to give some more detailed examples of how you can integrate ideas of systems surprise and suspense into your own projects.
So, we’ve gone over the concepts we’re exploring, and you’re thinking one of two things:
Or
Why? and How?
So, let’s get into it! We’ll go over each way in which systems suspense can be utilized in YOUR design, with as many concrete examples as possible. Each method will be followed by summary takeaways of “why” you’d want to do it and precisely “how” to do it.
Systems suspense can be used as a tool for increasing the player’s overall interest in your game – picking it up in the first place, sticking with it, and completing it.
This inspiration of intrigue can begin as early as the trailer for your game: hinting at deeper systems while not revealing them fully. This begins the makings of the player’s mental model of your game, and gives them something to latch on to as a reason to want to try it out. In this trailer for Crowsworn, as an example, much of the trailer depicts a fairly standard side-scrolling hack-and-slash game. At 0:51, though, the trailer shows hints of a progression mechanic that can’t be defined as simply as “what you see on the screen is what you get”, like the combat or aesthetics. It is a piece of the gameplay that will deepen their experience as a player in ways that they can’t yet foresee. This grabs players who otherwise feel like they’ve seen this type of game before, and piques their interest about what these new systems bring to the table.
@Crowsworn Trailer – Kickstarter Anniversary 2
Once your player has actually begun playing the game, systems suspense can begin to build intrigue and anticipation at a more measured and deliberate pace. Use tutorials and initial quests and missions to foreshadow larger systems, creating psychological “hooks” to draw in players’ imagination and attention; this can kickstart the formation of mental models for what the game’s form may actually take. Control does a great job of this in their first few minutes of play. The attached video shows the first time you ever see something you can pick up in the game. The experience functions as a built-in UI tutorial and also implicitly sets up intrigue: it sends you into the bottom level of the collectibles menu, forcing you to back out of the menu after seeing the pickup, thus showing you all of the other potential pickup categories in the game, all of which are unknown at this time. The categories include things like “Board Countermeasures” and “Clearance Level” and other deliciously intriguing phrases that pique the player’s interest and push them onward to figure out what all of those things could be. This judicious building-out of your game’s world and the rules that govern it, and hinting at rules that will be important but are heretofore undefined, inspires in players the determination to solve those mysteries and confirm or disprove the assumptions they build in their mental models.
@https://www.youtube.com/watch?v=e2jyznEfH70&ab_channel=LaurenScott
Apply systems suspense to engage players and hook them into the parts of your game that are intriguing and create the desire for more.
Why?
How?
Resolving the suspense built throughout your game is a great way to reward players with a sense of having gained mastery over systems they didn’t previously quite understand. Create challenges and blockers in your systems that have an unknown resolution, to begin building suspense. Then, make it clear that they must be solved via the player’s actions and continued progress through the game. Metroidvanias have this method down pat: games like Metroid intentionally introduce doors locked behind arcane symbols, enemies that are unbeatable in the player’s current state, even areas that seem like they COULD be reachable but are JUST out of reach, without a discernible reason why. [5]
Create suspense around wondering when and how these things can be achieved. Then, award the player through gameplay new tools, abilities, and movement tech that make those challenges possible to beat. The closing of this loop, the resolution of these points of suspense, creates a feeling of contentment and mastery in a player that can overall increase their satisfaction with your game.
Why?
How?
Similarly, resolving suspense or providing a fun surprise can act as a reward for players experimenting and exploring within your game’s world. This is a direct application of “Building Mental Models through Play” described earlier in the paper.
Ever punch an item in a game early on and become disappointed that it didn’t do anything? This is an example of an extremely short suspense loop:
Here, the player’s short period of suspense is resolved in an unsatisfying way. This probably means they’ll do less exploring and experimentation, even later in parts of the game where you’d like them to. By changing the outcome of #6 here – having the box react to a punch, having it bounce around or destroy and maybe drop loot – the player’s suspense is resolved in a way that either confirms their hypothesis or, even better, extends their understanding of the world while not completely defining exactly what that fuzzy area is. For example, #6 could look like:
Now, the player has a) had their hope that the box does something resolved in a satisfying way, and b) gained even more suspense about what exactly boxes do in this game, because the box just did something totally unexpected, granting them an item or stat point that they don’t even know what to do with yet. This leaves them feeling both satisfied and raring to go find out how this all plays into the rest of the game!
Why?
How?
Your game is undoubtedly a collection of systems that you care about, and that you’d like the player to engage with. Some systems, though, are a bit more buried or take a bit more effort to engage with. Systems suspense can be used to guide players through the usage of other systems. By having them engage with a relatively simple system, and putting up a block to that system that can only be answered by the product of a different, more complicated system, they will more naturally be incentivized to go and try out that second system, to get the answers they want for the one they’d initially set out on.
The flower cross-breeding system in Animal Crossing: New Leaf is a very fun part of the game. But, it is quite hidden! Players can plant flowers from seeds they buy, which seems like a simple system – water them, and they spawn new flowers nearby. However, planting certain colors of flowers next to each other creates hybrid colors that are not purchaseable. How do you steer players toward engaging with a feature that you know they will love, but that is easy to miss?
You, as the designer, know that furniture is a core feature that the player is sure to engage with during their normal course through the game: they need it to build out their home in ways that they like and that the Happy Home Academy scores well; animal buddies may also ask for specific pieces. Furniture is also in general a much more visible feature, requiring specific items detailed in the UI. You can introduce furniture crafting recipes that require flower colors that the player can’t buy. The systems suspense begins when the player sees this new color requirement. They may begin by thinking, “Okay, a new flower color! I guess I have to wait and see if that color comes up at Nook’s shop.” Over time, though, as the days go by and they don’t see the color ever come up in stock… the suspense begins to build. How do I get this flower color? This curiosity, paired with the practical need for the input into the furniture system, may cause the player to begin to experiment with the colors of flowers that they DO have. Drawing on their kindergarten-level color theory: “The game says I need a purple… Well, I have blue, and red… what if…?” The suspense piques here, and the player makes the leap into the flower cross-breeding system of their own volition. Their experimentation is rewarded with the purple flower that they’ve sought! A satisfying systems suspense loop closed – and with a path that you specifically curated, leading from one of your systems to the other.
Put content that you want highlighted at the end of a systems suspense resolution point for added oomph, and also added surety that the player will get to the end of that loop and see the content that you (and they!) care about.
Why?
How?
By their very nature, progression systems often include some amount of systems suspense. A skill tree with blanked-out nodes immediately inspires in a player questions about what those nodes might be, about how they might unlock the power of their character.
To bump up the innate systems suspense in your progression systems, you might consider employing some additional techniques to enhance this suspense.
Adding breadcrumbs in the world that point to their integration into a potential progression system can be an excellent tool for generating additional suspense for that progression system. In Final Fantasy X, the Sphere Grid is known to the player very early on: it is a skill graph that players use spheres earned from battle to unlock nodes that relate to stat boosts and new abilities.
Some way into the game, the player begins earning previously-unseen spheres, like Teleport Spheres and Master Spheres. Finding a new sphere immediately expands the player’s mental model for what the sphere grid is and how they can move through it, and that added range of possibilities adds overall depth to the system and excitement for continued engagement with it.
In reverse, the Sphere Grid itself inspires its own sense of suspense, showing nodes on its surface called Locked Nodes, with different levels. This communicates to the player that they WILL eventually earn keys for these sorts of nodes, and leaves the player in a state of suspense wondering when they’ll find them, and what sorts of goodies they’ll get once they do.
Why?
How?
You may have a goal in your game to evoke a specific emotion or response – fear, loss, elation. Narrative is not the only tool that can achieve this – in fact, game mechanics are often uniquely successful at inspiring a real, visceral reaction in a player, as the reaction is tied to game components that the player cares deeply about, like their progression, their powers, their currency, or something else they’ve grown to care about.
Systems surprise is often used to evoke a jarring response, or a sense of loss or fear. Because systems surprise is all about subverting the expectations that you’ve been building in your game, and throwing a curve ball at a player, it maps very naturally onto these emotions.
For example, say you want to evoke a feeling of deep loss in the player. Sure, you could have one of the player’s favorite characters in the game die. And on the face of things, that’s just what Final Fantasy 7 did with the iconic death of Aerith. But, looking closer, what really made this death real for the player is the systems surprise baked into the moment. Up until that moment, it was a given that, if a player levels up a character throughout a game, committing resources and EXP and time into that character’s powers and growth, that investment is honored by always allowing that character to be part of the party. The player brought all of their genre expectations with them throughout the game, leveling up Aerith because she was obviously a strong magic user. They envision bringing her all the way through the game, to the final fight, to stand alongside Cloud in the final battle. The moment that Aerith is killed – and the subsequent moments, as the player realizes she’s not being resurrected, is no longer in the party, her items and resources permanently gone – pairs the narrative loss with the systemic loss of all of those gameplay goodies, in a way that the player was absolutely not expecting. The success of this method is apparent in the long lasting fame of that moment, and the memory of that loss that players feel even today, when they think of it.
To replicate a similar sense of loss, tie resources or other gameplay components that a player cares about to narrative instances of loss. How do you know what a player cares about? The sentiment that a player has for a given gameplay component is often directly related to the time investment spent growing or maintaining that component; or, the amount of important features that that component is tied into.
Horror games often employ systems surprise to evoke fear in the player. Horror games commonly employ shocks and scares that are unexpected, and put the player in a state of perceived vulnerability. You could just throw some super scary-looking character models at the player – and sure, maybe that is scary in itself. But you can achieve a very scary effect using the content that you already have, just deploying it in a strategic way.
Similar to the loss example, players are building a mental model for the game as they play. This model is a safe place for them – a set of rules that they think they can rely on, to not take damage or get jump-scared or die. A safe, comfy box. How do you cause some true fear? By throwing the safety of that box into question, of course! Resident Evil 3 does this in a very literal way. [6] Throughout the game – indeed, throughout the franchise, up to that point – safe rooms are known by the player to be unbreachable by enemies.
At a certain point, the biggest baddie in the game, Nemesis, is chasing the player. The player ducks into a haven, which their mental model tells them will shield them from this threat. The moment that Nemesis pushes its huge bulk through the safe room door, piercing that veil of safety, it invalidates a section of the contract the player had with the game that they thought was unbreachable. The floor of their mental model drops out from under them, and their mind, in that moment, is falling. Thus an extremely vivid sense of fear is achieved, in this moment of systems surprise.
Now, disrupting the safety of the safe mental model box is a powerful tactic, but there’s always the danger of messing with your game’s rules so much that the player doesn’t have a solid idea of what the game is anymore. You want to maintain the integrity of the box, even as you do things to challenge it.
Systems surprise can be very effectively employed to produce “negative” emotions; systems surprise and suspense can be used to evoke happy emotions, too! Many of our examples throughout the paper show how you can use systems suspense to set up and produce rewarding experiences for the player. Systems surprise can take known and expected conventions in games and flip them on their heads to produce feelings of delight, as Inscryption does multiple times throughout the game.
Why?
How?
So, a player begins playing your game. They start out intrigued, but over the course of a couple of hours, feel like they “get” the game and don’t need to play anymore. What happened? Well, among other things, the player felt like the mental model they created for your game was “complete”. That is, there were no more secrets to uncover, no more mechanics to master – no more reasons to play. Systems suspense can help stave off this state until you want it to happen, which is ideally at the end of the game, not somewhere at the beginning or in the middle!
Dave the Diver has a rather open example of this, which is similar to what a level select screen might serve to do in another game: One of its features, the fish farm, contains areas to breed fish from different “depths,” or areas of the ocean. Ones that haven’t been found yet are shown but netted off. This shows the player that there are some areas of the game the system hasn’t touched yet, and that they can expect the game to continue going until at least this system is resolved.
And, even if a player does get all the way to the end of the game feeling satisfied, you can continue to apply systems suspense to lengthen the natural tail of your game. In Starfield, the player effectively ends their campaign when they jump into the Unity at the end of a run. But, the devs built in a narrative and systemic conceit that can actually inspire players to continue playing: They’ve just re-started in a new world! One that is very similar, but with key systemic and narrative differences. The actual success of this ploy in Starfield is debatable, but it is an example of an attempt to use the intrigue of “what might be different in this world? Where might my mental model be incorrect or different here?” to inspire play even past the obvious end point of the game.
Why?
How?
Systems surprise, in particular, can be a great tool for innovating on standard game mechanics and conventions.
As mentioned in “Mental Models and Expectations,” players come to games with some set of expectations defined by external influences like genre, previous games played, etc (unless they’re a player totally new to games!). You can wield these expectations as powerful design tools, and use them in a way that flips them on their heads to create something not seen before.
Players come to the game Inscryption with the expectation that a game traditionally sticks with the visual style and setting that it starts out with. Inscryption uses this expectation to set up a moment of surprise when the game switches from a first-person, dark, 3D perspective and setting, into an old-school pixel graphic bright RPG-like setting and top-down perspective. The surprise is a very satisfying moment for the player, and it also sets up the suspense that even more unexpected moments like this may arise in the future. Although this sort of extreme switch has been used in some games in the past (like Frog Fractions, mentioned earlier), the exact design and implementation, and the player reactions that they invoke, are what makes this stand out as a true innovation.
Why?
How?
We often think of pacing as a narrative thing. With systems suspense, you’ll be able to affect pacing through game design, by designing how your systems are introduced, teased, and resolved.
Say you have an alchemy feature in your game. You could present it to the player all at once: they meet an alchemist, and that alchemist tells them all about how alchemy works, complete with some starting reagents and a whole recipe book. But, depending on how complex your alchemy system is, that might be a lot to digest at once. The player may bounce off the system, not want to dig in to it because they’re frustrated or intimidated by the wall of information!
What’s a more organic way of building out this system in the player’s mind? And not just build it out, but create an innate interest in figuring out what it is?
A scenario with systems suspense built in: The player has started the game, and is exploring the first area. As they do, they come across a “Viscous Liquid,” with a tag marked “Alchemy Ingredient” (or, if you want to heighten the intrigue even more, just tagged “Reagent”). The player immediately updates their mental model with a fuzzy node labeled, “Alchemy?”. This allows the player to begin building the idea of the alchemy system at a measured and comfortable pace, and even inspires some natural intrigue about what the system will entail. As the player continues through the game, pepper the world with a basic recipe, a merchant who sells reagents, and even finished alchemical products that the player can use (e.g., have a merchant sell a Firebomb Potion – “Can I make that eventually?!”, the player thinks). Finally, once the player has actually reached the alchemy workbench, where the feature is revealed to them in full, they already have quite a bit of the feature clear in their mind.
This is a lot more than they had in the first scenario, when the alchemy mentor hit them with all these rules at once! In this second scenario, they already have an outline in their mental model, which they can fill in and expand once they discover the feature in full. This creates a much more comfortable pacing feel for the player.
Why?
How?
When we as designers receive feedback from players, it’s not always obvious how to address it. Systems suspense can be a tool in your toolbelt to pull out when you see specific types of player feedback.
Players getting bored here? Maybe move the suspense payoff earlier. Players not engaging with this mechanic? Introduce some breadcrumbs or ease up the pacing with techniques described above.
You can also look at feedback with an eye toward systems suspense, and where it is falling short: Where are players pushing on the game, trying to interrogate their mental models? Where is their interest drawn, where are they creating hypotheses in their minds that they’re really invested in testing? Is that experimentation and curiosity being rewarded? Is enough resistance given before they are given their answer? You could potentially lengthen the period of time that they’re in suspense, to make the payoff feel that much more rewarding.
Use early playtesting feedback to find out where players tend to push on edges of the model of your game, and how quickly those edges resolve from being fuzzy to solid. If they’re resolving, too fast you can reduce their ability to test, by throwing gameplay wrinkles in their path.
Breadcrumb towards areas of suspense to draw player’s attention to where their model is fuzzy, so they’re likely to notice it by the time you resolve it. If players are saying they’re confused more than feeling suspense, breadcrumbs might help a point of uncertainty feel like a fuzzy edge that has distinct but unresolved rules, instead of a totally confusing unknown.
Why?
How?
You, as the designer of your game, know exactly where the edges of the game are. But, the player doesn’t! Players come into any game with a bigger imagination of its scope, and everything that’s possible – which is beautiful! Suspense is a way for players to keep enjoying that sense of vast possibility for longer.
The Elden Ring example from earlier in the paper is an example of this. The reveal of the extended map creates the possibility in the player’s mind that the game’s scope is boundless. In addition, Elden Ring employs many teleports, and portions of the game that require backtracking to previously-explored areas, so that the player truly has no idea where the “end” lies.
The box that is the player’s mental model is fuzzy until you as the designer define and solidify it; you can create the illusion of a huge game in a player’s mind by manipulating that model model.
Why?
How?
A game has a natural pacing and rhythm that a player eventually gets used to, and comes to expect. There may be times when you’d like to disrupt this sense of rhythm, to jar the player out of their comfortable rut or as a supplement to an intense narrative moment. Systems surprise can be employed to aid in this.
dotAge is a pretty standard worker placement-based town-builder strategy game for a time: decide on tasks for the day, assign workers to those tasks, end the day, and watch the workers perform those tasks, seeing how those tasks play out and earning resources. Factor the day’s results into your plans for the next day. Rinse and repeat. This type of game naturally builds a very regular rhythm – one that a player might get bored with after a time.
Eventually, dotAge introduces a mechanic that causes random events to occur unless the player does specific tasks, [7] tasks that may run counter to the plans that they had for building out their village. This systems surprise forces the player out of their comfortable rhythm, and forces them to engage with parts of the game that they might not have before. It also sets up the systems suspense for how these events will resolve – what more events are possible in the future, what the repercussions are for failing them, and what possible rewards come out of completing the events successfully.
All of this extra engagement serves to pull the player further into the game, hooking their interest and getting their minds embedded in your systems.
Why?
How?
In addition to all the “Do!”s mentioned above, here are a couple of “Don’t!”s to look out for as you try out integrating some systems suspense into your design.
Resolving Suspense at the Wrong Time – You’ve done the delicate work of setting up some juicy systems suspense for your player. The last thing you want to do is resolve that suspense too early, or too late! This will leave the player either satisfied but with lots of game left to play, or left hanging at the end of your game feeling unsatisfied and like they didn’t get the answers the game was setting them up for.
An example of suspense resolving too early – Starfield has an “outpost” system in which players must set up outposts on planets to mine resources and extend their foothold in the galaxy. As soon as a player sees that this system exists, their mind runs wild with what it COULD be: “Do I get to build my own home base, a la The Lodge?” “Will I be able to do some farming?” “How will this expand my earning potential in the game?”. As play progresses, and the feature unfolds, its limitations become clear pretty early on: placement of mining structures is limited, the process to ship resources between planets is limited, the process to sell or make use of those resources is excessively complicated. The blank marble slate of the system is slowly chipped away by its limitations and rigidity, until after some hours of play the player’s mind is finally left with the shape of what it actually is. The problem is, in order to truly “complete” the system, the player still has many hours to put in, and they’re stuck interacting with a feature they already fully understand and are now doing just by rote.
Success in systems suspense lies in the timing of that shape coming into focus: If it becomes clear at THE moment that you’ve reached the ultimate mastery of the system, and have completed it, that’s where the satisfaction comes from. “Oh, I see… this piece goes here, and…. voila! I did it! Nice! That was cool.” But if that picture becomes clear far before you’re done interacting with the system (as is the case with outposts), you’re left with just the drudgery of acting within a known system, doing the busywork to get to the end goal.
This is definitely challenging to get exactly right, and something that likely requires rounds of playtesting and iteration to pin down exactly.
Weighing a Game Down with Systems it Doesn’t Need – We advocate for systems suspense as a powerful tool in this paper. However, there are some games that don’t need extra systems to succeed – and could even be adversely affected by an attempt to shoehorn extra systems in!
A case can be made that Unpacking is an example of this. Unpacking is fairly light in systems, relying on its core mechanic and visual storytelling to lead the players through its experience. As its designer, you may think, “Hmm, there isn’t enough here… We could maybe employ some systems suspense to make the game even stickier, or deeper.” You could give the player stars after completing a level, which feed into a shop or repository of notes that you don’t reveal until later in the game. You could have some items in the levels that the player can pick up and place into an inventory, their use unknown until some time later. But, Unpacking clearly succeeded and stands on its own without needing any of that!
Use systems suspense judiciously, not as a hose to spray across any part of the game, but as a paintbrush to fine-tune any spots that are obviously problems in terms of engagement or depth.
Ideally, this section of practical examples has armed you with several concrete ways to integrate systems suspense into your game’s design, and has given you concrete reasons for why it’s beneficial to do so. Identify parts of your design that could benefit from one of these methods, and try to introduce systems suspense using the steps described!
Game systems are experiential structures, and those structures have shape. By sculpting our systems we can invoke the notes of experience we recognize from other mediums, like surprise and suspense. In doing so, we discover new facets of those affects and the peculiar joys they bring, in the same way that movement through a videogame imparts a certain tenor of experience that is never quite replicable elsewhere.
In this paper, we unpacked a family of feelings we felt while playing games that swerved our expectations of their very rules. We decoded those swerves and built a framework for how they operate in terms of mental models, in which a complicated layer cake of personal play experience, genre, and medium literacy compose into an interpretation engine for understanding games as we play them. Those models gave us the language to talk about the subtle and substantial subversions that go into making players feel surprise and suspense. We went on to iterate through a number of example-backed techniques for shaping systems to produce these affects, and we argued for why designers might be compelled to do so in the first place.
The point of this project was to provide one view into the buzzing delight at the state of unknowing that a particular class of games has tended to play in. With one framework for understanding how the grain of a system and its friction of play yields surprise and suspense, our hope is to make those affects more directly accessible to designers who wish to invoke them.
1. This can be figured as a type of narrative suspense if we take the perspective that a playthrough of a game is an emergent narrative. See Max Kreminski and Michael Mateas’s “A Coauthorship-Centric History of Interactive Emergent Narrative” for context on emergent narrative in general, and in particular, Marie-Laure Ryan’s “expansive” definition that’s relevant here.
2. This is not to say we won’t address the fail cases; because surprise and suspense are adjacent to disappointment and frustration, we examine the distinctions between the two pairs of feelings. For another lens on this tension, see Daniel Cook’s “Building Tight Game Systems of Cause and Effect”.
3. For a deep dive on narrative suspense, consider Eric Rabkin’s Narrative Suspense.
4. The first instance of the “do this but rarely do that” pattern being surfaced publicly was via a tweet by developer Jason Grinblat.
5. For more on these sorts of locked-door patterns and how they can be used to draw players forward during tutorialization, see Max Kreminski’s “Locked doors, headaches, and intellectual need”.
6. The 2023 Polaris paper “Practical Tools for Empowering the Relationship Between Theme and Mechanics” first presented this example in the context of a game’s design evoking a sense of dissonance with itself.
7. dotAge systems suspense: Video footage of the above example, showing the introduction of the random event system that throws players’ plans out of the window and sets up suspense for what that means for their gameplay.
This paper was produced by a workgroup of senior game designers at the 2023 Polaris conference. The non-profit Polaris Game Design conference seeks to gather the industry’s top designers in order to first, discuss our toughest problems and second, share this deep wellspring for the benefit of game teams everywhere.
In 2023, this work was made possible in part by our sponsor Xsolla and others.
