because deliberation preceeds action

No more Excuses!

Mar 10

Descriptions of Girl Gamers, by male and female children

In the games industry, games often cater to the imaginary male gamer majority that wants male protagonists in everything they play. Games like Remember Me are rejected by publishers if they have female leads. The "reality of production" caused the high budget Assassin's Creed Unity to abandon their plans for female characters in co-op. A 2012 study found that "Games with a female-only protagonist … [received] only 40% of the marketing budget of male-led games. Less than that, actually." and consequently didn't sell as well, creating a vicious cycle of less marketing in the future.

Last week, I attended GDC, the annual Game Developers Conference, for the first time. It was exhausting, exhilarating, and surprising.

In the midst of panels and lectures on breaking in to the game industry, game engine pricing announcements, and overviews of successful game creation and marketing techniques, a number of speakers targeted awareness and challenging the preconceptions we have about games and the people who make and play them.

Unexpectedly, my favorite of these was Curiosity, Courage, and Camouflage: Revealing the Gaming Habits of Teen Girls by Ashly Burch and Rosalind Wiseman.

Their premise: everything we think about girls and games as an industry is wrong.

They surveyed 1400 schoolchildren in the 6th through 12th grades (11-18 year olds), 46% female, 50% male, and 4% "prefer not to answer" spread across 39% private, 58% public, and 3% charter schools.

Wiseman and Burch used a combination of statistics and anecdotes to weave a strong argument for us to make games that are designed for male and female audiences alike.

For example, while speaking to 700 girls about conflict resolution, Wiseman successfully used imagery from games such as the first person shooter series Call of Duty. The response from the audience was unexpected: the girls started cheering and high-fiving one another. Girls play more games in more genres than we expect them to:

Games girls play

Especially in the aftermath of Gamergate, many in the games industry think that the young consumer base is a "boys club," but this is a serious misconception. In fact, the recent controversies in the game world are not even relevant to young audiences, who tend not to read gaming websites except for walkthroughs. Boys don't want girls to stop playing games; in fact, they actively want more girls to play the same games they do and already think that they are:

Boys want more girls playing Games boys think girls play

As far as protagonists in games go, as children get older, boys don't care about their avatar's gender and girls want more female characters to play.

In general, children would like to see less sexualization in their games; they want characters to be treated with respect. Notably, 61% of 11 year old boys think that there is too much sexual objectification of women in games.

The main takeaway here is that we have no more excuses not to make more games with strong, non-sexualized female characters. Lots of us have been arguing this for a long time, but now, thanks to Wiseman and Burch, we have lots of actual data with which to convince our teammates, publishers, and marketing executives.

Managing Complexity: Spelunky vs League of Legends

Nov 29

I'm not afraid of complexity. As I've mentioned in the past, I love learning new things across many different domains, from mathematics to art to experience design. However, some things are easier to learn than others because of their mental models.

It is easy to learn things that behave in accordance with your expectations, and learning something new that builds on your existing understanding of the world is relatively simple. Learning something that doesn't is much harder. Someone who speaks two romance languages probably won't have as much trouble learning a new romance language as learning Chinese, for example. Someone who experiences math as formula memorization probably won't be able to apply Calculus outside of the classroom context within which they learned it.

A few days ago, I went to a Thanksgiving party, helped make and eat lots of great food, and played digital games with some of my friends. Among them were Spelunky and a 5v5 game of League of Legends (a MOBA commonly known as LOL).

Spelunky and LOL are fantastically complex. They each have many items and characters, each with different properties that substantially affect gameplay.

While becoming proficient at these games is quite challenging, the different ways in which they manage complexity makes Spelunky much easier to understand than LOL, and I think it has a much better overall design as a result. Why does Spelunky excel where LOL falls short?

  1. Consistency

    To be consistent, a system has to follow its own rules and conventions. Spelunky is very consistent. Treasure chests always contain items. Gold bars give you a predictable amount of gold. You can only carry one thing in your hands. The environment is deadly. Even though every level is procedurally generated, it will always have one exit that leads to the next level. Every game has the same progression of stages.

    LOL has problems with consistency, especially with respect to items and the item shop:

    Item Shop

    LOL has many different items that all fall into different categories. The most important ones (from what I can tell as a first-time player) are consumable (single-use), passive (always in-use), or active (need to be triggered with an associated cooldown).

    However, the item shop doesn't use that classification for items. All the items look consistent even though they aren't. Players can switch to the list of all items to see this:

    All Item Shop Items

    Since LOL is a fast-paced game of team competition, the only way to make effective item choices is to memorize the descriptions of the items to determine how they work when the game isn't running.

    Even more problematic is the "trinket" slot next to the 6 other item slots. A certain type of item, called a trinket/ward can only fill that one spot. Again, the store presents this item type consistently with the others. When I had filled 6 of my item slots, I couldn't understand why I couldn't buy another item even though I still had an empty slot. Since 6 of my slots were filled, the only visual distinction (the low-contrast eye emblem) failed to help me differentiate the function of the 7th slot from the others.

  2. Affordances

    An affordance is part of a system that broadcasts its use by its form. Door plates afford pushing while vertical door handles afford pulling. The spikes in Spelunky look particularly deadly – and they are. The only locked chest in Spelunky has a large golden lock with a matching golden key, and the purpose of the key is evident from its appearance.

    LOL has many poor affordances. Why does clicking on the store (which looks similar to the other buildings) make a modal dialog fill the screen, but clicking on the other buildings does not? Why do radar dishes on the ground explode when you get close to them when real radar dishes are used to scan objects at a distance?

  3. Feedback

    True learning comes at the price of repeated failure. When something doesn't behave as you expect, you adjust your expectations based on your experiences. When a system encourages failing fast and provides good feedback that helps you understand your mistakes, you can build a better mental model of the system more quickly.

    Spelunky is a brutal game; death means restarting from the beginning. However, it associates every failure with the feedback you need in order to prevent making the same mistake again. Since every playthrough is different but pits you against a new configuration of the same obstacles, the game encourages true learning rather than memorizing a sequence of actions.

    Theoretically, LOL has a framework for providing much more relevant feedback than Spelunky because death means respawning back at the base. In fact, LOL does use feedback to manage complexity more effectively than consistency and affordances. LOL displays a death recap modal after every death. While this modal is probably extrordinarily useful to intermediate players, it didn't help me, as a beginner, understand many of my deaths because they were so sudden and revolved around placement as much as on the abilities of the other players. I think LOL beginners would benefit from a death replay similar to Towerfall Ascension, a game that also has very rapid deaths.

    LOL Death Recap Towerfall Death Recap

Spelunky is an original title while League of Legends was inspired by a mod of Warcraft III called Defense of the Ancients. LOL is built around mental models that don't make much sense in the context of a standalone game. Consequently, it would probably be much harder to perfect LOL than to create a much better game by making a completely original MOBA (a path a number of other games have attempted). Nonetheless, I think LOL could be substantially improved with more thought given to consistency, affordances, and feedback while still meeting the needs of its legacy userbase.

The Problem with Video Tutorials

Nov 23

I try to spend most of my time learning new things. Some of the things I've been learning most recently include how to bake, how to make props, how to program in Clojurescript, and how to work with Quaternions.

Since I need to use a lot of software to create interactive media, learning to use software and programming APIs is part of my daily life.

At the moment, the major software teaching approaches on the internet fall into the following categories:

  • Q&A Style Forums
  • Text Tutorials
  • Video Tutorials

I really like forum style answers from sites like Stack Overflow because they answer specific questions and are easily searchable. However, because they are so limited in scope, they are rarely effective at providing introductions to techniques; deferring to text or video tutorials for that purpose.

I like text tutorials because they are easily glanceable. If the tutorial is well-organized, I can easily find the answer to my specific question without having to waste my time by wading through explanations of things I already know. Unfortunately, text is static and can't convey complex visual techniques. Well-chosen and annotated images can help to alleviate this problem, but they can't replace someone actually demonstrating how to do something.

Demonstration is the selling point of video tutorials. Recording a video of doing something seems much easier than writing a comprehensive explanation. Because of this apparently low barrier to entry, the quantity of video tutorials far outstrips the quantity of other software explanations on the internet. Ultimately, video tutorials have two severe flaws.

  1. They are not glanceable. Skimming a video of a software demonstration isn't very helpful because the audio is just as or more important than the visuals. Most video player interfaces are designed around watching a video from start to finish. Tutorials do not benefit from this structure.

  2. They are easy to structure poorly. While making a well-structured video tutorial is actually much harder than creating a well-structured text tutorial; it just seems easier to get started with a video tutorial: you're just talking to a camera, how hard could that be?

I think the best tutorials and explanations consist of a hybrid of text, video, and interactivity. Using this approach, teachers can create well-structured, glanceable representations that allow learners to experiment in the context of the lesson.

Bret Victor has created some early experiments in this direction:

Direct Manipulation Tool Construction

Nov 19

I've been busy exchanging skills with my classmates at USC's graduate Interactive Media and Games Division. I taught Unity programming and learned prop making and baking:

Skitter Mask

Journey Cake

Both of these arts revolve around direct manipulation; you bend foam and fondant, cut components into the right shape, and use tools to shape your work.

It's important to note that the end result rarely matches the original designs, and that this is a Very Good Thing. Working on physical objects involves many implicit constraints that shape the end result and inspire solutions that would otherwise never have been considered.

This is the power of direct manipulation: constantly getting feedback as you are working and changing your work in response to that feedback. Interfaces (and more generally, systems) without constant feedback are inherently limited to an initial design, which is a problem because, as you work, you often realize that the original design is flawed. For a commonly accepted example of this in software and engineering, consider the derided Waterfall development technique.

Working with physical materials, I constantly felt the urge to improvise. If I didn't have a tool I thought I would need, I would take a minute or two to create an ad-hoc tool for the specific task I was working on at the time. How can I bend this part into the shape that I want? I can just use a bowl. I'm missing a rolling pin small enough to flatten this gum paste, so I'll use this thin wooden dowel. How can I draw a large perfect circle without a compass? I'll tape two pencils to a wooden board.

I think the ease with which you can construct useful one-off tools through direct manipulation is sorely lacking in the digital space. Yes, some interfaces are scriptable through simple programming languages like Lua and Python, but the barrier to entry is much higher than taping miscellany together. Moreover, since it is so hard to make a tool to begin with, once you have decided to make it, you are much more inclined to waste the time you could have been using to make the thing you were working on by making the tool 'reusable' even if you'll only actually need it once.

The closest analogue I can think of to direct manipulation tool construction in the software space is a recorded macro that can automate repetitive actions. Unfortunately, they are a poor substitute for the flexibility of making your own tools.

While our software involves so little user feedback and so much indirect manipulation of hidden state (see Bret Victor's Learnable Programming)), I don't think we can do much better than macros and scripting languages. If you can directly manipulate something, you can have many more opportunities to easily make your own tools. Play Euclid: The Game to get an inkling of what that might be like.

Show me the Data: Learning Math through Exploration

Nov 9

Scalars. Vectors. Matrices. Quaternions. Most digital games are full of geometric mathematical manipulations and concepts. All games that represent movement visually through screens or projectors involve some form of conceptual mapping from abstract mathematical spaces to concrete geometric spaces we can see. Pac-man projects a toroidal space onto a 2d cartesian plane. Most 3d games involve transformations in a 3d cartesian space that are ultimately projected on a flat screen. Some, like the upcoming 4d puzzle game Miegakure, involve even more complex spaces.

Although most of us have at least some experience playing these games on our phones, computers, and consoles from a young age, we tend not to think about the mathematical manipulations directly. Instead, we think about how our actions as players affect the 2d image that we can see.

Traditional math classes teach geometry through proofs and formulas, involving students in memorization and symbolic manipulation. Some of the students can see beyond the drudgery in these tasks to the elegance of the actual mathematics. Those are the students who will become mathematicians, engineers, and computer graphics experts. The rest become bored and go on to avoid "math" in their lives as much as possible.

I think seeing and changing the hidden transformations in some games, perhaps through visual debug modes, could help us form more accurate mental models of geometry while contextualizing its importance in the media with which we interact on a frequent basis.

Games that Invite Change

Nov 2

The element that substantially differentiates games from other media is the degree to which they enable and encourage audience participation. Games respond when the audience acts. We might say that paintings, film, and novels, also respond, in that each audience member brings a context and set of interpretations to these media forms. However, games are much more dynamic. The best games encourage players to experiment with the mechanics in as many different ways as possible. Jonathan Blow's game design philosophy revolves around guiding players to discover the deep emergent consequences of a simple and elegant ruleset.

While all games encourage interaction, some actually invite players to explore the consequences of modifying their systems, to add or remove rules and mechanics and to explore the implications of those changes. These games are inherently moddable.

On the other hand, some games are so compactly designed that changing them effectively would be very difficult. As I discussed here, the card game Set discourages modding because any alterations would oppose the simplicity and elegance of the core system design.

Recently, I've been working on a creative problem solving card game called Evil Lairs for Dummies. We designed it to be moddable, to have systems that beg the players to change them to see what will happen. However, we didn't frame the game in a way that invited this type of change; the players felt that they didn't have permission to mod the game.

In order for a game to truly invite change, I think it needs to be explicitly encouraged by the designers and the game systems.

Beyond Defaults: A Philosophy of Learning through Representation

Oct 26

I have a personal philosophy that is all about learning: I can't be happy unless I'm doing it.

This notion is aligned with Raph Coster's arguments in A Theory of Fun for Game Design; he cites Tic-Tac-Toe as an example of a game that is only fun for children until they learn to recognize the patterns in the game. Under trivial optimal play, every match will end in a stalemate, and the game becomes boring.

As media designers, we frequently struggle to create "engaging" content. One simple way to do this is to ensure that everything we create teaches something. Think of all the books we enjoy even more deeply when we reread them. We say things like "there are so many amazing things I missed when I read it the first time."

This is the essence of "deep", "strategic" play in great games: it's well-accepted that each time we play them, we find new strategies and new meaning in the game mechanics. However, games aren't just about mechanics, they are also about representation through visuals, audio, and other means. It should be obvious that representation and mechanics together influence what we (consciously or not) learn from games.

In a shooting game, it matters whether the entities we shoot are vicious demons (Doom, 1993) or unarmed civilians (Hatred, planned 2015).

When designing a game, there is a big tendency to rely on defaults, on representations that match unconscious cultural or social stereotypes. The problem with that approach is that the resulting games have a tendency to have white cisgender male characters, to disempower and objectify female characters, and to portray black characters as agents of violence.

In The Power of Play: The Portrayal and Performance of Race in Video Games, Everett and Watkins examine some of negative stereotypes reinforced by racial representation in games.

If we can use representations to teach and games are fun and engaging when we can learn from them, won't games be all the more engaging if the representations in them introduce us to new contexts?

Programming vs Games in Discourse

Oct 12

The general discussion on the merits of games seems very similar to the general discussion regarding those of programming.

Programming and Games as a Distraction from "What is Important"

"When will you stop playing games and get back to work?"

We often view games as a distraction, as a 'fun' respite from the cold realities of life and work. This is the mindset that led one of my colleagues to say "I just don't have time for games."

Nowadays, our technology-saturated culture is so focused on programming as a work-related pursuit that we can find it difficult to draw a parallel with programming, but I think "Stop sitting in front of a computer screen and go outside!" is still a common sentiment.

Games/Programming as Treatment

In meanwhile, at, Bret Victor bemoans the corruption of Seymour Papert's thoughts on programming and pedagogy:

Papert's goal was to give children a path to powerful mathematical ideas, and he saw programming as a better carrier of the ideas than pencil-and-paper.

Others misinterpret him as making the dubious claim that simply learning to program will make you think better. Papert is astonished that anyone would think he was saying that.

Much of the work on gamification of education takes a similar approach, treating games as a "more fun" or less teacher-intensive method to transmit knowledge through rote memorization and award structures supported by Skinner box mechanics.

Games/Programming and Encoded Ideas

Bret Victor:

Papert explains that programming can serve as a medium in which powerful ideas can be brought within reach. But the focus, of course, must be on the powerful ideas, not the programming itself.

In The Rhetoric of Video Games, Ian Bogost argues that games represent a powerful way of representing ideas that other media cannot as readily communicate.

Ultimately, the connection between games and programming in discourse isn't truly surprising, because both involve encoding and exploring procedural behavior. In the strictest sense, all games are programs, but that fact is easy to forget because we usually think of programs as computer software while games more readily call to mind social or physical interactions. Nonetheless, I think recognizing something in one domain can help us reconsider our views on the other.

Unstructured Plots?

Oct 4

At a panel at GDC, Tom Abernathy (Riot Games narrative lead) and Richard Rouse III (designer at Microsoft Game Studios) asserted that plot is "highly overrated" in games, that players are less likely to remember the plots for their favorite games than their favorite movies, and that games should stop pursuing a three act narrative structure in favor of unstructured stories that focus on characters. See Narrative designers discuss benefits of ditching the three act structure.

Based on my own experiences, the longer and more complex a storyline is, the less likely I am to remember it, especially if it is interrupted by sequences that provide no discernible story progression.

"Game narrative is almost always extremely spaced out," Abernathy said. "There might be ongoing narratives with individual characters, but plot twists that come every 90 minutes or so happen right before you are about to get back into the gameplay and you're less likely to remember it."

If all our favorite story games provide 'story' in the form of cutscenes that interrupt gameplay every 90 minutes or so, then yes, it is quite clear that we will have trouble remembering the 'plot' in those games.

However, I can think of many examples in other media that suffer from the same problem: interrupting main plot progression with convoluted side stories, long-winded descriptions (certain fantasy books are particularly at fault here), long and flashy action sequences (most modern action movies), etc…

Perhaps the reason players can remember the stories in their favorite movies more readily than in their favorite games is that the games simply employ comparatively poor storytelling techniques.

The best told stories, regardless of medium, rely on disordering incidents, on frequent character development, on understandable motivations, and on interaction between characters. The stories in games like Portal and Ghost Trick are particularly well executed in this regard, and I can remember their plots quite well as a result.

The improv-heavy role playing game Fiasco makes character interaction part of its core mechanic; it consists of scenes in which players assume the roles of characters attempting to achieve specific goals.

In cases where we want to tell stories with our media, I think what we really want is more structure, rather than less.

The Game Webserial

Sep 28

Some of the most enduring online communities gather around webserials (serialized online fiction like Worm), webcomics (generally serialized online comics like Questionable Content), and web series (serialized online video series like The Guild).

As we traverse the spectrum from static text to the moving image, we see the complexity of authoring these types of experiences increase dramatically, although there isn't as much variation in production times as one might expect. While some webserials update several times a week and some webcomics take several months to create, we can also find well-produced video content updated on a weekly basis.

The interactive medium is conspiciously absent from this list. While some fantastic episodic games exist, they are generally released as downloadable titles with months or years between releases. The idea of a single creator or small group regularly releasing games is not unheard of, I know of no serialized online games.

Will "Game Webserials" ever become popular? What form will the episodic 30 second - 30 minute online game take? Will it be a classic point-and-click adventure with a linear storyline? Will it be a procedural world with new content produced each week? Perhaps something we've never seen before?

Are the tools we have to create these experiences approachable and advanced enough to make this kind of creation feasible for creators as diverse as those who create more traditional webserials?

A "Magic Threshold"

Sep 21

In There is No Magic Circle (2009), Consalvo argues that games cannot be viewed from a purely structural perspective, that Huizinga's "magic circle" is absurd given its implication that the ordinary rules of everyday life no longer apply within its confines. Instead, we must recognize that the ordinary rules very well do apply "in addition to, in competition with, other rules and in relation to multiple contexts, across varying cultures, and into different groups, legal situations, and homes."

Nonetheless, I think the concept of a magic circle as a boundary between the mindset for being in a game and outside a game is still very valuable. Context is everything when it comes to games and play. For example, when we speak of "cat and mouse" games, we mean that the cat is turning what would otherwise be a hunt into a form of play. From the mouse' perspective, there is no game, simply a struggle for everyday survival. In the 1997 film The Game, the protagonist participates in an experience that he percieves as a game others are playing with him rather than a game that he is playing.

As UX designers, I think we need to ensure that the experiences we create respect this distinction by making passage through the magic threshold and back very clear to our users and players.

On Narrative

Sep 14

In 2002, Henry Jenkins published the article Game Design as Narrative Architecture. In it, he addresses many of my misgivings about the views Jesper Juul expresses in his A Clash Between Games and Narrative and Games Telling Stories?

In the following decade, we have witnessed the emergence of many innovative experiences that have advanced the state of the art in many of the categories Jenkins proposes.

Transmedia Storytelling

Experience franchises spreading to diverse media is nothing new. Look no further than the superhero characters from comics appearing in books, movies, television, web-series, and games. More interesting is the phenomenon of interactive experiences that transcend media boundaries. A recent example is Google's Ingress, a worldwide multiplayer alternate reality game that enables player participation through smartphone-world interaction, online code breaking, and storytelling/partaking through online videos and social media.

Spatial Narratives

Gone Home is a perfect example of what Jenkins calls a "Memory Palace," a place that players can explore at will, unravelling stories left in narrative fragments throughout a space. While Gone Home communicates one primary narrative, that of the protagonist's sister, it also contains several tangentially-related story threads involving the other members of the family.


Most modern games, particularly big budget titles, embed many "memorable moments" into their games. Some of these are developed to the extent that they are known as "side-quests," with mostly self-contained, sometimes optional narrative arcs that may enhance the primary storyline or serve as exercises in world-building.

Punch Drunk's Sleep No More is particularly notable in this category. It encourages audiences to explore a mass of related but distinct narrative threads at their own pace as they follow actors and explore a physical three story building with maze-like passages.

Emergent Narratives

Unplanned narratives often emerge from games with large player-bases. The space game Eve Online is a prime example; it provides a sandbox-like simulation environment with enough constraints that players self-organize to take on roles in many different stories involving themes such as faction conflict, diplomacy, corporate espionage, and piracy.

It seems to me, however, that very few experiences explore smaller-scale emergent storytelling, especially in the category of single-player computer games with an emphasis on player choice. Some of the best games in this category (such as the fantastic Spelunky) involve procedurally generated worlds and enemies and explore few emotions unrelated to conquest.

Beyond Definitions

Sep 7

What constitutes a game? What constitutes play?

Philosophers, academics, and game creators have been debating the answers to questions of this ilk for a very long time.

In Rules of Play, Katie Salen and Eric Zimmerman describe a game as

a system in which players engage in an artificial conflict, defined by rules, that results in a quantifiable outcome.

In The Definition of Play, The Classification of Games, Roger Caillois describes play via the properties Free (voluntary), Separate (with predefined limits of space and time), Uncertain (the results are not guaranteed beforehand), Unproductive (creating nothing of value during the course of the game and ending in a situation identical to that prior to the beginning of the game), Governed by Rules, and Make-believe.

Jesper Juul's definition of "game" (from The Game, the Player, the World: Looking for a Heart of Gameness) is

a rule-based formal system with a variable and quantifiable outcome, where different outcomes are assigned different values, the player exerts effort in order to influence the outcome, the player feels attached to the outcome, and the consequences of the activity are optional and negotiable.

There are so many conflicting definitions that Molleindustria created a whole site devoted to game definitions.

The great variety suggests that it is futile to determine one true meaning for these terms; that they are simply ambiguous and context-dependent.

However, rather than ignoring them, I think that designers can think of them as artificial constraints to surmount.

For example, does an experience really have to have a "variable and quantifiable outcome" to be considered a game? Why not try to create an emotionally moving interactive experience that involves an emotional journey but doesn't vary in its outcome?

Does play truly have to be "unproductive"? What about games that involve creating lasting artifacts and social change?

Trying to envision experiences that challenge specific definitions seems like a fantastic way to develop games that have yet to be experienced.

Set, System Analysis

Sep 1

I'm studying how systems work at USC's Interactive Media and Games Division, and I decided to explore how the design of Set cards is tied to the rules for the game. I initially wrote this with no resources other than a deck of Set cards and a Python interpreter, but I have since amended some of my initial impressions after reading this.

We can use playing cards in myriad ways, including building card castles, throwing cards at targets, using cards to propel projectiles, floating cards down streams ala Poohsticks, laying cards side-by side like dominos, and bending cards to create three dimensional shapes.

Set is a simple card game with a deck of 81 standard-size playing cards that are uniquely suited for exactly one purpose that doesn't involve the destruction of the cards: for playing the game of Set.

One side of each Set card bears the logo of the game on a matte surface, while the other is devoted to one or more large symbols displayed on a glossy white canvas.

The symbols on any given card can differ in the following ways:

  • Quantity (each card has one, two, or three symbols)
  • Color (a given symbol can be green, red, or purple, and all symbols on a card share the same color)
  • Shading (symbols can be solid-colored, striped in parallel with the longest side of the card, or consist of empty - outlines, and all symbols on a given card share the same shading)
  • Shape (each symbol on a card is either tilde-shaped, diamond-shaped, or capsule-shaped, and each symbol on a given card has the same shape)

Note that each of these four properties has exactly three variants, which allows for the 34=81 unique cards in the deck.

Consequently, it would be impossible to play a standard match-2 game with a Set deck; by virtue of having three variants for each property, the game materials are optimized for 3-card combinations. The logo of the game reinforces this notion; it contains a visualization of stacks of three cards with different shading, calling to mind the card properties. While there are 511920 unique 3-card combinations possible in a deck of Set cards, not all of these combinations can be meaningful. If every 3-card combination with at least one shared property variant were considered a "match," then around 36% of the possible combinations would qualify, which would eliminate much of the depth present in the Set card game. Instead, a match, known as a "Set," occurs when each property varies either entirely or not at all. People playing the game compete to find and collect as many such Sets as possible.

The following three cards form a Set because they all have the same quantity variant but differ in every other property:

  • 3-Green-Empty-Tilde
  • 3-Purple-Solid-Capsule
  • 3-Red-Shaded-Diamond

These cards do not form a Set because two of the cards have the solid shading variant:

  • 3-Green-Empty-Tilde
  • 3-Purple-Solid-Capsule
  • 3-Red-Solid-Diamond

Only around 1% of the possible 3-card combinations are Sets. However, the probability of discovering a 3-card Set in a collection of n cards increases quite quickly with n.

These are the probabilities of a Set in a given collection of n cards drawn from the deck at random, at the start of the game (determined via numerical simulation in Python, using 100,000 samples for each n-card collection):


Card games typically involve cards held in the player's hands, but holding more than seven cards is ungainly, and there is only a 39% chance that seven random cards will contain a Set. Furthermore, the size and position of the symbols on the cards does not afford holding several of them in a hand; doing so obscures the property information which is integral to forming Sets. The cards are clearly designed to be placed face up on a level surface.

From the diagram above, we can see that we get increasingly diminishing returns in the probability that an initial collection of n cards will contain a Set for n>10. In fact, for n=20 and greater, the probability is 100%. When n=12, the quantity used in the game of Set, the probability is initially 97% (although it quickly drops to about 93% as the game progresses and Sets are removed and replaced with new cards from the deck). Furthermore, the average amount of Sets present in a collection increases rapidly with n. For the sake of challenging players to find Sets, n should never be too high, so n=12 is a logical choice, although it does sometimes require the addition of new cards from the deck when no Sets are present (the rules to Set call for the addition of three new cards in this situation, which ensures that the amount of face-up cards on the table is always divisible by the amount of cards in a Set). Elegantly, 12 is 4x3, the number of properties multiplied by the number of variants for each property.

Although there are several Set variants, I would argue that the cards for Set are perfectly designed to match the primary game for which they were created, and that many of these alterations oppose the simplicity of the core game and/or the design of the cards. For example, the variants frequently call for holding some cards in one's hand, removing cards, or adding betting tokens. Nonetheless, I applaud the ingenuity of the player base.


Aug 30

Design is all about constraints. Designers must be able to work within limitations, and the designs they create guide user interactions.

Recently, I've been thinking about how the constraints in some systems influence the way we engage with other systems. This idea is tied to the concept of consistency in interaction design. As Norman explains in The Design of Everyday Things, the conceptual model for how we expect something to behave influences how we interact with it, and we take our knowledge of existing systems with us when we interact with others. Left-clicking in computer apps is a simple example: we expect to left-click to select virtual objects in drawing applications. When we encounter an application that breaks this convention, like Blender, we experience what Cooper (The Inmates are Running the Asylum) calls cognitive friction.

After witnessing the deplorable behavior perpetuated by a subset of the people who identify themselves as "Gamers," I spent some time reflecting on the reasons behind their outrage at whom they bafflingly brand "Social Justice Warriors," as if that could possibly be a pejorative.

I'm certain that part of the problem is simply that old standby, fear of change. However, I think it might also be affected by the the proliferation of binary experiences, both in games and across social media platforms. For many users, the core interaction with the latter is content consumption augmented with an expression of "thumbs up" or "thumbs down." In many games, players take on the role of an unassailable hero or anti-hero character who is always justified in whatever he (because characters in games are rarely female) does.

I wonder if the constraints inherent in these experiences have trained people to to view the world in binary, to view criticism as condemnation rather than constructive critique.