Elegance

The ratio of the depth to complexity of a game system. The "efficiency" of the design.

Elegance is a term that can elicit groans from game designers and reviewers alike when it is invoked. It is thrown around often, and with many different meanings in discussions about mechanical systems. It is often conflated with the subjective quality of a game's systems or sometimes with their simplicity. This scattershot use of the term means that some people view it as an empty piece of praise or a relatively meaningless descriptor. However, the term has a distinct and meaningful use as outlined Cameron Browne's paper from 2012, and it's a definition of which I think most people actually have a vague or intuitive idea, even if they couldn't define it when prompted.

The elegance of a system is the ratio of the complexity of that system to the depth that it produces. A system with a high depth and low complexity is said to be "elegant," though where this line lies is entirely subjective. But to really understand what this definition is saying, first we need to define both complexity and depth.

Complexity

Complexity seems the more straightforward of the two terms at first. It might seem like an objective state of the system we can quantify: a measure of the total scope an intricacy of all the mechanics in the system. However, this is only part of the equation. The dynamics that emerge from the interactions of these mechanics can create wildly complex interactions. Think about how a small number of relatively-simple cards in Magic: the Gathering can interact in ways that require experienced judges to come in a mediate at tournaments. In this way, a collection of simple mechanics can create a set of wildly complex interactions.

Now, some of this emergent complexity is good! As we'll get into with depth, these kind of complex interactions are exactly what creates deep game experiences. So how do we distinguish how much an interactions adds to the complexity of the system, versus adds to the depth of a system? The meaningful distinction here is what interactions the player needs to understand to be able to play the game, not necessarily to play the game perfectly (or even well), just to play it to completion. Looking back to our initial mention of mechanics then, we can apply this distinction there as well: only mechanics the the player needs to understand to be able to play the game add to complexity (for its relevance to elegance at least).

Finally, bringing these two elements together we can define the complexity of a system as the difficulty for a player to understand all the rules and interactions needed to play the game to completion.

Depth

The depth of the game system is, put simply, the possibility space of all of that complexity interacting in play. It is all the ways that the pieces of the system can interact to create meaningfully different outcomes, meaningfully different decisions for the players, or meaningfully different experiences.

As you might be able to guess from its use three times in a row, the phrase "meaningfully different" is important here. You can create a system that generates, on paper, an infinite number of different outcomes, but if those all feel relatively similar to the player, the number is irrelevant. My favorite illustration of this is Kate Compton's "10,000 bowls of oatmeal problem" in her article on procedural generation:

I can easily generate 10,000 bowls of plain oatmeal, with each oat being in a different position and different orientation, and mathematically speaking they will all be completely unique. But the user will likely just see a lot of oatmeal.

If a system creates 10,000 different outcomes, it's important that players can perceive those differences.

When covering complexity above, we discussed how the dynamics that a player needs to understand to be able to play the game contribute to complexity. This does not mean that all the other dynamics that make up the depth of the system shouldn't be understood by a player, merely that they don't need to be to play the game at first. One of the beautiful things about a deep system is that players are constantly learning and and discovering new aspects of the interactions, improving their abilities and being surprised by a game. When it comes to some of the most deep, elegant games in the world like Go and Chess, players can spend their entire lives exploring the depth of those games.

Elegance in Games

Now that we have a functional definition for depth, complexity, and elegance let's take a look at some practical examples. The most straightforward examples, mentioned above, are combinatorial games like Go and Chess. These games have a small, fixed set of known rules that interact in incredibly deep ways. They are incontrovertibly elegant games whose depth many players spend entire careers exploring.

Examples of the other extreme, games with very high complexity yet low depth, are harder to come by. High complexity can create a barrier to entry, and if there isn't depth on the other side, players don't take the time to overcome that barrier. So games like this don’t often excel. I think the example I've settled on for this category is Warhammer 40k (don't @ me, I'm actually an avid player). This miniatures game has hundreds of pages of necessary rules across its rulebooks and codices specific to each army. Games play out over the course of four hours where dozens of units each have dozens of rules all interacting with each other. This ends to create an experience that, while it has some pockets of depth, does not have a depth commiserate to the complexity it offers. The game instead exists on what it offers from a storytelling and creative hobbying perspective. Being a system to simulate massive, gorgeous dioramas. That still has value, but isn't the mechanical depth we're discussing and charting today.

Most games don't fall into either of these extremes, however. Most games sit somewhere in the middle with closer to what we'll call a "1:1" complexity-to-depth ratio. Where exactly one game's elegance compares to another is a question left to arguments across social media, but I'd like to use this metric to discuss one other phenomenon we see in the modern games industry: live games.

In addition to the marketing value live games bring, we can use this model of elegance to look at the value that games with "content treadmills" bring to the players. Games like League of Legends or Magic the Gathering, for example, are both fairly elegant games at their core. But even the extent of their depth would probably not be enough to sustain a large community's interest across decades. Think about if Magic only had the alpha set, or if League only had its starting champions. Players would have likely moved on years ago. Instead, by constantly releasing new content (new complexity), these games incrementally increase the depth of their game, giving new possibility space to explore. And as this chart we've been using shows, the added possibility space isn't just "X depth," but actually "X depth multiplied all previous complexity." Each addition to the game multiplicatively increases the space players can explore. These games just have to be careful about how high that complexity creeps along the way.

Applying Elegance in Design

When a designer is iterating on their game design, elegance can be an incredibly useful tool for honing the systems of the game. Instead of looking at the complexity-to-depth of the system as a whole, a designer can take individual mechanics and make estimates to what it is individually contributing to each of these values. This is especially useful when trying to reduce bloat in a system, or identify mechanics that aren't "pulling their weight" in the amount of dynamics they are creating.

As an example, let's look at Ori and the Blind Forest, a beautiful metroidvania by Moon Studios. This game works like a standard platformer. To progress through the game, Ori:

• jumps through levels

• dodges projectile

• attacks enemies

As the game progresses, Ori unlocks new abilities that add new functionality to these limited moves. Let's say, due to a scope change during development, one of these abilities needs to be cut. As the designer, you have to choose between:

Charge Flame

Hold down the attack button to do a charged version of the attack that does additional damage and breaks special barriers.

Bash

When approaching a projectile, press 'Y' and point to a direction to cause Ori to launch off off the projectile, sending it in another direction and giving him a jump-like boost.

Looking at their complexity, Bash is undoubtedly more difficult to understand and execute. It gets a higher complexity score. So then, let's look at the depth they bring to the system by listing some of the dynamics they create with other mechanics in the game.

Charge Flame

• Extra damage

• Breaks barriers

Bash

• Extra jumps

• Change direction midair

• Avoid hitting projectiles

• Send projectile back at enemy

Of the three core systems listed above, Charge Flame interacts with one, and Bash interacts with all three! It also adds a multitude of new ways for level designers to create levels with creative platforming puzzles. So, while Bash might be more complex, their elegances looks something like this:

As the designer looking at this comparison, you would likely decide to cut Charge Flame (not to mention that cutting Bash would probably break most of your level designers' work), and it should come as no surprise that Charge Flame was cut in the sequel Ori and the Will of the Wisps in favor of mechanics that provided a bit more depth.

So use elegance as a lens through which to inspect mechanics for the value they're bringing to a system and to identify places you can build in more depth or trim complexity. And the next time someone calls a game "elegant" as a catch-all compliment for a game, ask them why. I bet you'll get some insights into the depth of the game system that they have discovered.

Recommended Reading

Elegance in Game Design. Cameron Browne (2012)

Lenticular Design. Mark Rosewater (2014)

Depth vs. Complexity. Extra Credits (2013)

Obfuscation

The necessity in a game system to make the outcome uncertain to the players

For most of my design principles so far, I've covered well-established terms with standardized definitions. For today's article, however, I wanted to discuss something that I've been unable to find a standardized term for. Through my reading, I have seen a lot of methods discussed that all share a similar overall objective of disguising the ultimate outcome of the system. It’s that objective that I want to discuss more in-depth today. So, for the sake of this article, I've coined a term to describe this common objective. If any of you reading are familiar with a term already in use for this that I was just unaware of, let me know!

Tension in game systems is created by the players be unaware of the outcome of the game. The moment they know how things are going to play out, the game loses almost all of its interest. For example, the game of tic-tac-toe is interesting to a new player when they are first introduced to it, but once they realize the game can only play out in a finite number of ways that they can be easily calculate, it loses all tension. Even a game whose outcome is already predetermined like the card game 'War' (we'll skip the discussion about whether it is or is not a game for today), retains tension due to that already-determined outcome being hidden from the people playing it.

A visualization of all the possible outcomes of a tic-tac-toe game. Each square visualizes the potential outcomes of ‘O’ choosing that space.

When designing a game system then, a designer needs to build in systems to obfuscate the potential outcomes from the player for as long as possible. There are potentially-infinite ways to create this obfuscation, but most methods fall under a number of standardized categories:

Execution

A lot of games have an even smaller number of outcomes than tic-tac-toe, but that doesn’t mean they are “solved” in the same way. A game of Mario, for example, can basically only end with two outcomes: Mario dying or reaching the far side of the level (or reaching a star or moon in a newer iteration like Mario Odyssey). Or in most sports, the score usually ends in only a couple dozen different configurations.

What keeps these outcomes uncertain and interesting then, is the actual physical execution and skill of the player(s). While a game of tic-tac-toe has a discrete number of decisions to reach its outcomes, a sport has a near-infinite number of interactions with the physics of the ball, other players, and the physical execution of strenuous activities. No one knows exactly how the game is going to end, and a lot of joy of the game comes not from which outcome was reached, but how it was reached.

Mario’s controls offer a ton of space for expression through execution. It is not about the binary outcome of “acquire star” vs “die,” but about how the player reaches those outcomes through their execution of the complex controls.

Randomness

A favorite of tactics games (and, of course, gambling), randomness is an easy and cheap way to add uncertainty to the outcome of a game system. Examples of this is the randomness of rolling dice for a skill check in Dungeons and Dragons, or drawing cards from a shuffled deck at the start of a turn of Magic the Gathering.

Randomness is often looked down on a bit in the design community. Admittedly, even I called it "cheap" in the above paragraph. But it is an important and versatile tool when applied correctly. Consideration needs to be taken for how the probabilities feel to the player and where in the game loop the randomness is sitting to create a feeling of variance for the player without stripping a feeling of control.

Hidden Information

This form of obfuscation plays most directly to the “obfuscation” name by literally obscuring information from the player. This category comprises any techniques used to hide parts of the systems from the player for some or all of the play time. A really clear example of this is in card games, where players' hand or the deck may be hidden. For example, the card game Solitaire would be very uninteresting if you knew the entire order of the deck before you started playing.

A less obvious example of hidden information is present in almost every digital game. When a player first launches the game, almost everything inside of it is hidden to them until they start unveiling it as they play. For example, in a game like Mass Effect, the outcome of the story and all possible choices they will make will only become clear as the game unfolds, obscuring the exact outcome of their choices until after they make them.

Competing Decisions

This category could arguably be folded under "Hidden Information," but I think it has enough nuance to warrant its own entry. In a competitive game, the fact that you can't know what your opponent is thinking or planning for their next move can keep the outcome of a game uncertain. Unlike some of the other categories though, this one has some qualifiers:

First, an opponent's decisions only create uncertainty if the game system is of a suitable complexity level. If, like in the tic-tac-toe example earlier, the game is simple enough that players can conceptualize every possible outcome, it doesn't matter what decision the other player is making.

Secondly, the game needs to have some semblance of balance in the different decisions that can be made. If a player only has one clearly-optimal choice, the decision is rendered irrelevant as they are forced to follow the optimal path.

Complexity

Finally, this last form of obfuscation is often paired with others, but plays a critical role. A game system can use the sheer complexity of the different potential outcomes to obfuscate the outcome from players. The outcome of the game may be able to be theoretically calculated, but that calculation would be so staggeringly complex that no lone person could complete it. This is not to say that the game system itself should be so complex that it defies comprehension, but that the total possibility space of future decisions should branch to a complexity level that defies comprehension. The human brain is excellent at looking at massive datasets and filtering down to only the most important data to what it needs to be processing right now. A good game presents simple mechanics that offer complex and nuanced dynamics that produce an interesting dataset to give the players' brains opportunities to do this filtering.

For example, given infinite time or computing power, every possible game outcome of chess could theoretically be calculated, much like is possible with tic-tac-toe now. This would effectively "solve" the game. Then, playing would just be a matter of making choices that lead you down one of the possibility branches that lead to your victory. In fact, chess grand masters can actually "see" a good amount of this possibility tree with years of training and practice. But, even for them, the full breadth of all moves it too much to keep in mind. That's how new chess moves get "discovered." Those possible moves were always there, players has thus-far just been filtering out these moves as not not worth it, not realizing it could lead to a whole branch of possibilities until someone reframed their understanding.

Conclusion

This list of obfuscation categories is far from exhaustive, but hopefully offers broad enough categorization to discuss them while designing. In actual practice, this kind of categorization is far messier as most games use a combination of different systems in all sorts of categories. A game like Civilization VI, for example, obfuscates with the sheer complexity of its systems, with the hidden information of its fog of war, with the competing decisions of the different players, and the randomness of the map's generation.

When designing your own systems, try to layer and combine different forms of obfuscation in this way to create multiplicative considerations as your players try to figure out the outcome. But, be cautious of applying too much obfuscation. While you want players to be unsure of the outcome of the game system, you want them to have enough of a guess to plan their own tactics around. Finding that balance of “just enough” information is more of an art than a science, and can only really be reached through a lot of iteration.

Affordances

Physical properties of an object or environment that defines or elucidates its possible uses.

Some objects or environments are better suited for certain uses than for others. A steep slope with stairs built into it is better suited to a human's climbing than one with a smooth surface, a round wheel is better suited to rolling than a square one, or a door with a protruding handle is better suited to being pulled than one with a flat metal plate. These objects' physical properties make them more convenient for people to perform these actions, and their properties can even imply these actions to an observer simply through their physical form. These objects and environments are said to better "afford" these actions.

The "affordances" of an object then, are all the actions to which its physical properties make it better suited. As you can probably imagine, depending on the object, the list of its potential affordances could be very long. However when people refer to an object’s affordances, they usually aren’t referring to every single action it could afford, but to the one or two things that it most obviously affords to the people who are most likely to be using it. In this way, affordances exist in this strange place between the objective and the subjective. What actions an object affords are based on objective truths about the object: its size, shape, and other properties; but they are also based on how those truths relate to capabilities of the individual using the object. A teapot affords pouring to an fully-abled person, but not to one who can't easily grasp objects.

The subjectivity of affordances goes beyond the capabilities of a users, but also into that user's motivations and intentions for an object. Upon seeing a fist-size rock, one person might see an object that affords weighing down their papers, while another might see an object that affords being thrown through a window. In his article that first coined the term (The Theory of Affordances), James Gibson explores how humans will actually conceive of and contextualize objects based on the actions the objects afford them, instead of on the actual physical properties of objects. For example, when asked to sort a metal knife, a stone tool, and a smooth stone, people are more likely to group the two objects that afford cutting together than they are to group together the two made out of stone.

When making a design, it's important to consider the affordances of the various components of the design as well as the affordances of the environment in which it will be used. When the obvious affordances of objects correspond with their intended use in the design, the design will perform more efficiently and its use will be more naturally intuitive to the user. Users will naturally make assumptions about components of the design based on their affordances, and if those assumption line up with the design's intention, that user has just learned something about the design without ever needing to be explicitly taught.

This can be true of the components of a game's design, but also control methods. A good example is item selection in modern console games. The limited inputs of a console controller didn't easily afford selecting from linear lists of items that were common on PC games, so over time more and more console games have moved to radial menus that the joysticks of a console controller easily afford selecting from.

In Prey (2017), on PC you could select weapons using number keys which afford quickly pressing one of 10 (shown left), or on console you could use the thumbstick to select from a radial menu (shown right).

Beyond the consideration of physical input, it might be tempting to dismiss affordance as not being relevant to games that exist entirely in the digital space, since virtual worlds aren't constrained by the same rules as our physical reality. But, in some ways, in this space affordances become even more important. Users bring their preconceptions and knowledge from the physical world into your digital one. By using the language of physical affordances that are already universally understood by your users, you can more easily demonstrate the uses and restrictions of your digital world without ever having to teach our users anything.

A great example of this is in the early days of touchscreen phone UI design. Apple especially known for their skeuomorphic UI designs with the original iPhone. For each piece of UI on the iPhone, the designers modeled it after a physical object who's affordances a user would already be familiar with, then made sure that their digital facsimile had similar functions in the program - a physical dial affords spinning to set the correct date, a physical button with beveled height affords pushing in, a physical paper page affords turning it with a swipe of your finger. Users already knew how to use the UI because they knew how to use the objects off of which it was modeled.

Similar to these UI practices, when designing a game, try to model different actions, resources, and environments off of real-world equivalents that afford the same functions you would like their in-game equivalents to perform in the game systems. This will assist in the player learning the game and create a more immersive, realistic-feeling experience for them in the end.

Affordances in UX Design

As you probably noted above, digital interfaces have moved away from the skeuomorphism of the early iPhone UI over the years. New interfaces have cleaner, flatter designs that take advantage of all the things digital spaces can do that their physical counterparts cannot. As this evolution occurred, UX (user experience) designers needed to spend a lot of time evolving the understanding of how users related to and understand their digital environments. They found a lot of use for the term "affordances," and in this subdomain, it has taken on a much more specific definition than the broad one I've offered above.

A UX designer mapping out wireframes for an app (though I have never seen someone do it physically on a wall like this in practice

In this new digital world, there aren't physical properties for users to intuit affordances from, so the term is instead used to refer to several different methods of signaling the function of the interface to a user. It has become a very useful term to this domain, but I feel that its use has really started to be conflated with Signifiers (a different term which I’ll cover in a separate article one day). This is muddying what makes affordances unique, but as the term is used in this way commonly now, I wanted to make sure to discuss it here.

There are six different types of affordances in UX design that people discuss. I'm not sure who originally coined these six terms, but the earliest reference to them I can find is in this article by Paula Borowska.

Explicit Affordances

Explicit affordances are any visual hints given off by the physical appearance of the design or any text that appears in in. Examples of this would be a button with bevels to appear like a physical button, or an entry field labeled with "Address" so users know what to write.

Pattern Affordances

Pattern affordances are those actions that are used so commonly by other digital UIs that they have established a convention. Settings menus being represented by a gear, or swiping right to bring up settings on Android are both examples of this.

Hidden Affordances

Hidden affordances are those that aren't readily apparent until the user actually starts interacting with the interface. Drop down menus or right-click options on PC are both good examples of this. Nesting functions into hidden affordances can simplify the initial appearance of the design, but designers should be cautious, as hidden affordances can often be missed by users.

False Affordances

False affordances are created when a component of the UI signals an affordance to the user for a functionality that it doesn't have. These are almost always errors, and are rarely purposefully built into a design. Examples might include a button that does nothing, or a gear icon that leads to a menu other than settings.

Metaphorical Affordances

Metaphorical affordances signal the use of a digital object by drawing a comparison to something the user already understands. The skeuomorphism from the original iPhone UI discussed above leaned heavily into this, as does a lot of the concept of the "desktop" on PC operating systems - folders, cut, copy, paste - all these original functions were named based on physical objects or functions people already understood.

Negative affordances

Negative affordances aren't actually an affordance at all, but a signifier of a current lack of affordances. Greyed out buttons are the quintessential example of this. This signals that there is sometimes an affordance there, but there currently isn't.

Recommended Reading

The Theory of Affordances. James Gibson (1977)

The Design of Everyday Things. Don Norman (1990)

6 Types of Digital Affordances. Paula Borowska (2015)

Bartle's Taxonomy

If you playtest your game with twelve different people, you’ll get twelve different opinions. This kind of qualitative feedback can be invaluable, but sometimes you need to step back from the tangled web of personal opinion and get a broader view of how your game will be interacting with a player community overall. To do this, most designers (and other professionals analyzing large communities) use a categorization system to break the community down into different categories based on their motivations.

Bartle’s Taxonomy is the categorization system of game players that gets taught the most in modern design circles. This system originated from discussion on an internet bulletin board in the early nineties between the top players of a popular MUD in the UK. MUDs (multi-user dungeons) are text-based, multiplayer environments where players can explore, go on adventures, and interact with each other. Think of them as the early ancestors of MMORPGs. Richard Bartle was the creator of MUD1, one of the very first MUDs in the world, and was an administrator for the MUD where these taxonomic discussions originated.

A screenshot from MUD1, the first “multi-user dungeon” created by Richard Bartle and Roy Trubshaw in 1978.

As an admin for this MUD, Bartle took it upon himself to collect the findings of this discussion and publish them in an article. This article poses that a given player’s motivation for coming to the MUD can be charted onto a graph with two axes. The first axis shows how the player prefers to interface with the elements of the game - from interacting with the elements of the game back-and-forth, to acting upon those elements by imposing their will on the game. The second axis shows which elements of the game world the player prefers to engage with on a spectrum - from the other players of the game, to the game world which includes the rules, systems, and content of the game.

When charted on this graph, players can be placed into four different categories based on which quadrant of the graph they fall in. Bartle associated each of these with a playing card suit:

Killers enjoy imposing their will on the other players of the game. As the name suggests, in a lot of more combat-focused games, this manifests itself as killing other players in PvP combat. Killers enjoy demonstrating mastery over other players and testing their skills not against a system, but against other human beings.

Achievers strive to accomplish demonstrable goals and milestones in the game’s systems. They might socialize, explore, or kill, but it is all a means to the end of accomplishing whatever goal has been set out for them. Achievers gather points and resources, attempt to overcome the game’s greatest challenges, and strive to “beat” the game (or at least some aspect of it).

Socializers are interested in other people. The game is merely a set piece in which they can interact and share meaningful experiences with other human beings. Socializers prefer collaborative activities, coordinating with other players, trading and bartering, and systems that allow them to better form relationships with others.

Explorers seek to uncover and understand as much about the game as possible. This can manifest in the very literal exploration of the game’s space, but it also might be an exploration of the game’s systems and their inner workings. Explorers seek to find hidden secrets and build knowledge of the game and its systems and quirks.

---

Because these categories rose out of the multiplayer environment of a MUD, Bartle also took time analyzing how fluctuations in the number of a given player type in a community could affect the other player types. As a game’s community increases in the number of killers, for example, it can drive away socializers, who aren’t generally interested in the adversarial dynamic. Bartle charted this out in an ASCII diagram you can find in the original article. It’s a bit hard to parse. I have gone ahead and recreated a more simplified version here. The descriptions for what each arrow type means can be found below the diagram:

These interactions have been simplified from Bartle’s original article.

White arrow show a proportional relationship from one category to another. As it increases or decreases, so does the other.

Black arrows show an inversely-proportional relationship. As the source category increases, the other decreases and vice-versa.

*The grey arrow shows a slightly a-typical case in Bartle’s diagram. When achievers are at an equilibrium, it can increase the number of socializers. But radically increase or decrease the number of achievers, and the number of socializers will fall.

This diagram of interactions doesn’t get shared nearly as much as the chart of the four categories themselves. I think there are a couple reasons for this. First, this diagram can be a bit hard to parse. But second, as you might have done yourself while looking it over, it’s pretty easy to start poking holes in these various interactions. In this diagram, it becomes a lot more obvious that this taxonomy and the thought that went into it are primarily centered around one single type of game (MUDs), and their specific communities. There are a lot of well-established dynamics observed in modern player communities that aren’t really represented here. Even Bartle himself has warned against the application of this taxonomy to other game types, as he’s concerned it’s likely incomplete.

So, if this taxonomy is incomplete, why has it been so popular across game design circles? I definitely think it’s simplicity makes it easier to grok (heck, it’s why I chose it as the first categorization technique to cover on this blog). This simplicity also means that it’s a good foundation for other designers to build upon with the specifics of their game and their community. Categorization like this can be an invaluable tool for analyzing how changes to a game might impact a player community, and tons of people have built more in-depth and specific models that hopefully I can cover in future articles:

• Researchers such as Dan Dixon have built on and responded to the foundation of Bartle and others to create a more in-depth discussion.

• My favorite model of player motivation comes from the market research company Quantic Foundry. They map motivations across 12 factors in 6 categories. Their site also has a slick survey that you can take for your own personalized “Gamer Motivation Profile.”

• Lots of designers have created their own taxonomies specific to their games. Most famously Mark Rosewater, the lead designer on Magic the Gathering, has created a number of psychographics that categorize the different players of Magic.

So, in the end, I don’t think I recommend using Bartle’s Taxonomy unaltered to analyze your own game. However, I do recommend reading into some of these more advanced categorizations and spending some time thinking about how they map to your game and its community (or potential community). Also, I would be curious to hear from all of you on what motivates you in games. Jump over to Quantic Foundry’s motivation profile survey and let me know here or on Twitter what kind of player you are!

Recommended Reading

HEARTS, CLUBS, DIAMONDS, SPADES: PLAYERS WHO SUIT MUDS. Bartle, Richard (1996)

Designing Virtual Worlds, Bartle, Richard (2004)

MDA Framework

In these “game design principles” articles, my goal is to introduce fundamental concepts to game design and establish a theoretical framework to use for future analysis and discussion of games. With that goal in mind, I would be remiss if I didn’t discuss one of the first big pieces of academic work on game design: MDA Framework.

This framework was originally taught as part of a workshop at the Game Developer’s Conference, and was later published in a paper in conjunction with researchers from Northwestern. It introduces a fundamental way of deconstructing game systems to give a shared vocabulary and methodology for people of all disciplines working with games. The framework proposes that games can be understood by dividing their components into three distinct categories: Mechanics, Dynamics, and Aesthetics.

Mechanics are the basic building blocks of a game. The smallest elemental pieces that a game system can be broken into—a single rule, a particular resource, an action available to the player etc. Like the gears that make up clockwork or the atoms that make up a molecule, these are the individual pieces that combine to create a greater whole. In a card game, mechanics are things like suits, numbering, shuffling, or trick-taking. In a first-person shooter mechanics are things like WASD movement, weapons, ammunition, or spawn points.

How exactly you subdivide a system up into discrete “mechanics” isn’t something that can be specifically defined, and really depends on the needs of how you are currently analyzing that system. Based on how granular an analysis needs to be, the designer of that first-person shooter might be looking at the exact hit box, damage numbers, traced path, and fire rate of an individual bullet as separate mechanics; or they might just take the entire weapon as a single mechanic. It depend on their current needs.

If mechanics are the atoms, dynamics are the molecules. They are the systems that emerge at “run-time” from the mechanics of the game interacting with each other and with the players’ decisions. These are the heart of the game. They are the interactions that create the players’ experiences—feeling tension, excitement, friendship, betrayal.

Dynamics are simultaneously the most important elements of a game design, but also the most ethereal. To players, they are their tactics and strategy, or how the elements of the game are behaving. To the designers, dynamics are intangible behaviors that come about as the mechanics begin interacting. It takes a lot of practice as a designer to begin to reliably predict what dynamics will emerge from a system, and even the most seasoned designers are sometimes surprised at what they find.

Finally, aesthetics are the actual experience of the game itself. These are the emotions, stories, or fantasies that the dynamics are invoking inside the brains of the players. Hunicke, Leblanc, and Zubek break down a taxonomy of eight different aesthetics, but admit that this list isn’t exhaustive:

“Sensation – Game as sense-pleasure
Fantasy – Game as make-believe
Narrative – Game as drama
Challenge – Game as obstacle course
Fellowship – Game as social framework
Discovery – Game as uncharted territory
Expression – Game as self-discovery
Submission – Game as pastime”

— Hunicke, LeBlanc, Zubek, 2004

Each game doesn’t focus equally on all these different aesthetic goals. Instead, a game usually focuses on 3-4 of these as its central focus (with larger, more sprawling games sometimes encompassing more). In the end, everything else inside of a game is in service of these aesthetic experiences.

---

All three of these are fairly abstract concepts, so let’s look at a quick example of them in action to get some context. In a video game, you might have several mechanics tied to your character:

• A numerical value that counts up and down.

• Every time your character is hit by something, that value is reduced by 5.

• When that value reaches zero, you lose the game.

These mechanics interact in dynamics that create a desire to avoid getting hit, and potentially some tension in the player about their performance. Aesthetically, I’m sure you’ve figured out what this would be in most games: a player’s health bar. But in other games it might represent something totally different.

So, what is the goal of breaking down a game into these different categories? Designing games can be an incredibly complicated pursuit, especially when you start looking at some of the larger video games created by teams of hundreds of people. Every little decision being made by all of these people every day changes mechanics in the game. These little tweaks alters how those mechanics interact with each other, which can have a rippling impact on the dynamics and by extension the aesthetics of the game. Quickly, what seems like inconsequential changes can add up and dramatically change a game. This is a complex problem for a designer to wrap their head around.

MDA provides a high-level theoretical framework to start abstracting these complex problems and help deconstructing a game and understanding how ideas flow from the “bottom” to the “top” of its systems. Put more plainly, it is creating a shared vocabulary and structure with which to understand and discuss complicated ideas.

For example, let’s look really quickly back at our health bar example and what happens if you add in a single additional mechanic:

• This value recovers 2 every second the player is not in combat.

With this change, the dynamics are altered to cause a lot less long-term tension in the player, and focus on individual engagements, and the aesthetics change to representing the character’s moment-to-moment stamina instead of their overall closeness to death. Conversely however, if instead the mechanic reads:

• This value is only restored to full upon reaching one of a few rare checkpoints.

Suddenly, the dynamics of the game will be tense, and every hit will have long-term consequences. The aesthetics of the character’s health feel a lot more final and permanent, and the game feels a lot more grim.

---

One other powerful element of MDA framework is that it considers how the elements of a game relate to both the designers and players of the game. As we touched on above, designers view the game from the “bottom-up.” They create mechanics, which give rise to dynamics, producing the aesthetics of play. Meanwhile, players consume the game “top-down.” They first experience the aesthetics of the game, then beginning to understand the dynamics, and finally identifying the component mechanics of the system.

Yeah, I know I’m using “bottom” and “top” but this diagram is side-to-side. It just looks way better this way!

These different perspectives are important to keep in mind, and figuring out how ideas translate “up” and “down” through the system is critical to understanding its impact. In the end, that’s really what a framework like MDA is for. It is a lens through which a game system can be viewed to give people of all different disciplines perspective when trying to solve problems, and to give them a shared vocabulary and understanding to more coherently discuss and deconstruct games.

Recommended Reading

MDA: A Formal Approach to Game Design and Game Research. Hunicke, Robin; LeBlanc, Marc; Zubek, Robert (2004)

Engagement Loops

Welcome back, dear readers, to the second installment of the Magic Circle blog. I’m going to be posting about once a week, probably on Mondays. If you’re interested in being updated each time a new article is available, consider signing up for my mailing list at the bottom of this article. I will only be sending out an e-mail with each new article, and promise not to spam you with anything else. Also, on the last article I got a lot of good feedback via messages. So, this time around, I’ve decided to enable comments on the article. If you have any thoughts after reading, I really encourage you to post a comment and I’ll try to respond to each of you.

For this second article, I wanted to stick to defining another basic game design principles to lay a strong foundation of vocabulary that we can use when discussing more complex subjects down the line. You don’t get more basic than today’s topic—the engagement loop.

The Engagement Loop

The engagement loop (also known as a core loop) is a simple cycle, present in every game, that a player goes through as they play. Visualized, it looks something like this:

Players constantly cycle through this loop as they play the game, and each cycle starts with that player’s current mental model of the game. This mental model includes everything they know about the game and its systems—their understanding of the story, of the rules, of their final goal, of other players’ strategies etc. A player’s model is (almost) always incomplete. Building and refining this understanding of the game system is an eternal process that players undergo over time as they play the game more and more. This mental model informs the player as to what they want to accomplish in the game. At any given moment of play, a player has formed a goal for themselves for what to do next based on their model.

Acting on whatever their current goal is, a player performs an action in the game by conducting some kind of input into the game system. This could be as simple as pressing the ‘X’ button on their controller, moving a game piece on a board, kicking the ball down the field—whatever kind of action the rules of the game and affordances of their interface allows.

The player’s input is then processed by the rules of the game to create some sort of feedback for the player about what effect their action has had. My description here might make the game system sound like an active player in the process, or like a computer programmed to respond. Sometimes it is. However, these “rules” can also be the laws of physics in sports or written rules enforced by the players in a board game. Whatever the rules are in a given case, they will produce some sort of result from the action, and the player will observe feedback based on the results. (Or, in some cases, no feedback at all. As that is a form of feedback itself. Some actions may not have any impact, or at least no evident impact, on a system.)

After receiving feedback, the player adjusts their mental model based on any new information they may have absorbed, and starts the cycle over again. Rinse. Repeat.

Scope

As you’ve likely observed, this loop is pretty abstract. As such, it can be used to analyze all sorts of different game systems, and analyze them at all sorts of different scopes. For example, in a game of Hearthstone, a player goes through this loop card-to-card, turn-to-turn, game-to-game, and even season-to-season. One “game” loop contains many “turn” loops, each of which also contains many “card” loops, creating a nested picture of a player’s flow through the entire game:

Using the Loop

Alright, we’ve got this tool through which we can observe a player’s cyclical interaction with our game, but how do we use it? There are some… less-than-virtuous ways we can use this to leverage player psychology for profit, but we’ll get to those in the next section. For now, let’s look at less morally-complicated example—using it as a tool to analyze how randomness is affecting our players’ experiences. For this, let’s continue exploring our example of Hearthstone.

Randomness shows up all across Hearthstone, and has received a wide range of reactions from players. Early on in the life of the game, a number of cards received a lot of hate from players for being too random—cards like Unstable Portal, Lightning Storm, and Piloted Shredder.

These cards cause random effects after being played—dealing a random amount of damage, selecting a random card, summoning a random creature, etc. So, looking at this in the context of the card-to-card engagement loop of Hearthstone, the randomness occurs right after the player’s input of choosing what card they’d like to play:

This creates an effect a bit like gambling. There is an exciting buildup of uncertainty, and then either a payoff or disappointment based on the results. Whatever happens past this point, it’s totally out of the player’s hands. This can create exciting stories when things go their way, but can also create a frustrating feeling of helplessness when they don’t.

Contrast this with another form of randomness in Hearthstone—drawing a card at the start of each turn.

When we look at our engagement loop, this lands before the player makes any kind of new decisions based on their revised mental model. Even though which card a player draws can have a significant impact on the outcome of the game, it is often far less exciting, and less frustrating, than our previous examples of randomness. Because the uncertainty comes before the player’s decision, the player feels in control. They are choosing how to solve the puzzle that has been created for them by the randomness instead of just watching something unfold before them.

This is a lesson that I think Blizzard really took to heart in later Hearthstone expansions. They began designing cards with this relationship to randomness in mind. Those of you who play Hearthstone, if you have examples of modern cards that change this relationship to randomness in interesting ways, let me know in the comments below!

Using the Loop… For Evil!

As the name suggests, the engagement loop is not only a useful tool to model a player’s interaction with a game system, but is also a pattern that is naturally engaging to the human brain. Googling “engagement loop” brings up many other synonymous or similar terms such as “core loop” (mostly synonymous), “game loop” (this is for the programming side of things), and “compulsion loop.” This last one is the one I’d like to talk about here, as it is often conflated with these other terms, and has a bit of a sinister background.

Engagement loops are often used to model how a game system is psychologically rewarding its players, something we will be diving into deeper in a future article. This isn’t inherently evil, but it quickly becomes sinister when it starts to be used to psychologically manipulate people into spending their money. This is what compulsion loops are used for.

Based on compulsion loop example from this article: https://gameanalytics.com/blog/the-compulsion-loop-explained.html

Primarily a tool of the “freemium” game market, compulsion loops were created to track and optimize a player’s Pavlovian response in an effort to bring that player back again and again to pay more and more money. These predatory strategies prey on trusting customers and addictive personalities to maximize profits.

I’m not going to dive into detail on compulsion loops in this article. I just bring them up to encourage you to be cautious when talking about engagement loops, core loops, compulsion loops, or even game loops. Since these terms are often used synonymously, just ensure there is no confusion about which one you’re referring to. Engagement loops are a great tool for more precisely discussing a game’s design; we shouldn’t let their misuse by some get in the way of their greater use by all.

Recommended Reading

Gamification at Work: Designing Engaging Business Software. Mythily Kumar, Janaki; Herger, Mario (2013)

Magic Pixie Wonder Dust. Deterding, Sebastian (2014)

A Taxonomy of Randomness in Hearthstone. Gallant, Matthew (2016)

The Magic Circle

Welcome to The Magic Circle — a blog about games and game design. My name is Andrew Fischer. I’m a college dropout who has somehow tricked people into paying me to make games for the last decade. Over the years I’ve done a lot of reading and writing about game design, mostly just for my own edification. I put together this blog as a place to start collecting that writing, and to act as a devlog for any of my (non-NDAed) projects. Hopefully someone out there finds this information interesting or helpful, and I can trick a few more people into thinking I know what I’m talking about.

As game design has become more discussed and studied over the years, a layer of jargon has started to stratify over the language used to talk about games. It’s messy. One person’s “negative feedback loop” is another person’s “rubber banding” or another’s “catchup mechanic.” In an attempt to wrestle this haphazard collection of jargon into something you might be able to call a “disciplinary vocabulary,” a lot of smart people have spent a lot of time discussing, studying, and standardizing the use of these terms. So I’m going to freeload on their hard work. The first kind of article I’m going to be playing around with are these Game Design Principle posts. They will discuss one principle used in game design theory, give examples of it in action, and point to some of those sources I freeloaded off of in case you want to read more. For this inaugural post, I’m going to talk about the eponymous principle of this blog — the “magic circle.”

The magic circle is a state of mind. It is a buy-in to the conceit of a game. It is a social contract between players mutually interested in having a shared experience. It is the bubble in which the normal rules and realities of the world are replaced by those of the game to everyone playing. Inside of the magic circle, the fastest way around a track in track-and-field is to be the greatest athlete, to train and practice and become faster than your competitors. Outside of the magic circle, the fastest way is to just cut through the middle. Inside of the magic circle, a chess knight is a playing piece, bound to its starting square until it is activated to execute a movement in specific pattern in accordance with a grand strategy. Outside of the magic circle, it is a wooden horse head.

This concept, the concept of a contained circle in which new rules and realities apply, is incredibly important not just to chess or to sports, but to all games. It is the concept that creates the meaning of games. It is the reason people can become so invested in the advancement of their World of Warcraft characters, or the reason that friends are willing to viciously betray each other in Diplomacy. Inside the circle, it’s real.

Diplomacy - Avalon Hill

Depending on who you ask, the magic circle as a term was either coined last decade, or a century ago. Its original use came from the book Homo Ludens, by Johan Huizinga in 1938. This book was an exploration of the elements of play in culture and the importance of play to humanity. It is an important touchstone in the history of game studies. It didn’t present the term formally, but made a couple light references to it:

“The arena, the card-table, the magic circle, the temple, the stage, the screen, the tennis court, the court of justice, etc., are all in form and function play-grounds, i.e. forbidden spots, isolated, hedged round, hallowed, within which special rules obtain. All are temporary worlds within ordinary world, dedicated to the performance of an act apart.”

— Huizinga, 1938

The general use of the term in the wider games industry really didn’t begin until it was discussed in Rules of Play by Katie Salen and Eric Zimmerman in 2003. In that book, they really re-invented the concept. They cobbling together ideas from Huizinga and other authors, and reframed the concept in terms of semiotics and design. The definition they put forth is the one discussed earlier in this article, and the one used by the games industry as a whole.

---

Unsurprisingly, even as simple a concept as the magic circle hasn’t been without controversy. A lot of designers and critics have come away from Rules of Play with a fairly black-and-white view of how the magic circle constrains and controls the play of a game. They see the concept as an approach to game design that encourages the creation of game systems as walled gardens — consequence-free zones completely shielded from cultural context or external ethical influences. In their eyes the magic circle is a dangerous philosophy that seeks to cordon off games from the realities of the wider world, limiting what games can be. This critical view of the concept became widespread enough that even Eric Zimmerman addressed it in a Gamasutra article years after the release of the book.

As is probably abundantly clear from my framing of the critical argument, I don’t particularly agree with that view of the magic circle. As I opened the article: I see the magic circle as a contract between the player and the game — a contract to accept new rules and realities in a limited space, for a limited duration, to experience meaning in interactions they would be otherwise-unable to experience. This contract is necessary to give game rules meaning, and doesn’t exclude the wider contexts outside of the circle, but adds to them. Games are, and should be, informed by world around them. The magical reality inside the world of the game has meaning because it relates back to your life and the world around you.

Examples of this are everywhere. A player’s personal experience with real-world racism will inform how they react to the ostracization of an alien race in their favorite sci-fi RPG. In the reverse, someone with little experience with real-world racism might gain new perspective from experiencing the treatment of the aliens in this RPG. Or, in a board game of deception and betrayal, personal histories between players may affect how players interpret each others’ actions, and insights gleaned from these fraught interactions may inform future real-world relationships. These games don’t gain this relevance in spite of the magic circle, but because of it. This suspension of reality inside the circle is what allows in-game interactions to be meaningful enough to be relatable to the real world.

In the end, regardless of how you decide to interpret it, the concept of “the magic circle” isn’t a principle on how to design games, merely a principle on how to understand players’ relationship to games. It is a lens through which to view your players’ experiences and understand the meanings you’re crafting for them.

Recommended Reading

Homo Ludens: a Study of the Play-Element in Culture. Huizinga, Johan (1938)

Rules of Play: Game Design Fundamentals. Salen, Katie; Zimmerman, Eric (2003)

Jerked Around by the Magic Circle - Clearing the Air Ten Years Later. Zimmerman, Eric (2012)