3 In a Game, You Can Be Whoever You Want to Be

Play enables us to rearrange our capacities and our very identity so that they can be used in unforeseen ways.

—Stephen Nachmanovitch

In different ways, scholars have talked about games as protected spaces, separated from the concerns or pressures that occupy us in the “real,” that is, non-game world. Stephen Sniderman talks about a game’s frame. Katie Salen and Eric Zimmerman, borrowing a phrase from Johan Huizinga, talk about the magic circle. Inside the magic circle, we are all free to play, constrained only by the game’s rules and, of course, those ethical and social norms we won’t violate, even within a game. What interests us for the purposes of this discussion is the freedom of identity that players exercise within the magic circle. In a game, you can be whoever you want to be (provided it doesn’t break the game).

It is striking to us that within a game, so many players are fiercely determined, persistent problem solvers, an identity they adopt for the duration of the game, but one they often leave behind when they exit the magic circle. Why are so many children who are committed game players unable to commit to school? Are these players simply lazy, unwilling to apply themselves to any other pursuits besides games? Are they unimaginative, unable to see how traditional schooling offers equally engaging challenges? As may be obvious, these are straw questions, and we reject the idea that children are fundamentally lazy or unimaginative. Rather we see two deficits in the larger system that contribute to this disconnect between committed play identities and indifferent school identities. The first is that we rarely offer students in school problems to solve that are as engaging or relevant to them as their games, and the second is that we fail to reinforce or validate the problem-solving identities students form in their game play.

When we plead for more relevant and engaging schoolwork, we’re not simply asking that students be better entertained. We understand that to succeed students must learn to apply themselves to problems with uncertain outcomes and the promise of at best delayed gratification. But if we present algebra, or biology, or history as devoid of any meaning in students’ real lives, we are in effect telling them that school is a place where they will be rated on the performance of unnecessary tasks whose only purpose is to sort them for future rewards. Some students may be sufficiently acculturated to the “game” of school to want to perform these tasks for the promised rewards, but in the end, school performance winds up being a measure of that acculturation. Qualities like curiosity, passion, or creativity may be of little value to students trying to perform many school tasks.

So in an ideal world, school would be full of more meaningful activities. Note that relevance is not limited to those themes or topics that students are already interested in. If a problem is presented in ways that spark student curiosity, it becomes relevant. If students learn about real-world problems related to such topics, they become relevant. If topics are embedded in narratives that evoke empathy or excitement, they also become relevant. Challenges of intellectual or emotional depth have the possibility of becoming relevant to many students (though no one activity will be relevant to every student).

And of course, many good teachers and some good school systems are working hard to make academics more meaningful to their students. But even in such environments, students arrive with a range of prior experiences and cultural influences. Too few have been rewarded for their problem-solving abilities. Strong research evidence shows that students perform better when praised for their persistence rather than for their intelligence, but the prevailing culture celebrates talent, as if it were naturally endowed, God given, or somehow magic, rather than something accessible to all students. If only all students could have the experience of seeing themselves as hard working and accomplished. Or rather, if only they could see their game-playing identities merge with their academic identities.

Lure of the Labyrinth, funded by the US Department of Education, and designed by the Education Arcade working in collaboration with Maryland Public Television and Fablevision, was an attempt to give middle school students the opportunity to apply their game-playing skills to pre-algebraic math, and in the process shift their identities to nascent problem solvers, mathematicians, and scholars. Designed as an online long-form puzzle adventure game, Labyrinth is also an attempt to use games to nudge the culture of school in the direction of greater relevance, and maybe even help teachers recognize capabilities in their students that often go unrecognized in traditional classroom settings.

In the process of describing Labyrinth, we also discuss the following principles of resonant games:

• Cultivating player identity

• Scaling across time, space, and users

• Preparing for future learning

• Deepening complexity

• Guiding reflection

Settling on an Approach

Labyrinth was funded under the US DOE’s Star Schools program, and it had several ambitious, overlapping goals. The game was primarily meant to address middle school mathematics (pre-algebra), but it was secondarily focused on literacy. Its intended audience was underserved populations, and since Maryland Public Television was the recipient of the grant, it was tested in schools in urban Baltimore and in rural parts of the state. We anticipated developing Labyrinth for game platforms accessible to the target audience, and if possible, the game was to be collaborative and multiplayer.

Early in the process, we held a series of meetings with Maryland math teachers to assess their needs, and from these meetings we drew an additional set of requirements. The teachers made it clear that they were unfamiliar with games in the classroom, were skeptical about their effectiveness, and because of curricular pressures were not sure how much time they could allot for game play during the school day. In addition, they were anxious about the introduction of any new technology into their classrooms because they could not afford the time to break in anything new and buggy, and because they were uncomfortable introducing any technology that they themselves had not mastered. And of course, they doubted they would ever master games. Despite all these cautions, the teachers acknowledged that they had trouble teaching some core concepts in the curriculum, such as proportionality, bases, and fractions, and that they would welcome any new tools to help students master those concepts.

One element that makes design, including game design, an interesting discipline is that designers often find themselves trying to design solutions that reconcile seemingly irreconcilable conflicts. In this case, serious needs weren’t being met by traditional means of instruction, but multiple obstacles militated against using games in the classroom. (We want to be clear here that these obstacles are systemic, not simply the fault of teachers.)

Fortunately, in the system’s resistance to games, we also saw an opportunity. In our prior work, we had observed that when children were imaginatively engaged in a game, playing of their own volition, they enthusiastically solved the kinds of complex problems that they rarely encountered in the classroom. In a subsequent chapter, we talk about the role of imaginative engagement in sparking kids’ deeper learning, but suffice it to say we believed that children would do better work while playing games without the heavy hand of school resting on their shoulders. So the challenge we took on was to design a game that could largely be played outside school, one that bore little resemblance to traditional curriculum materials, but that could nevertheless have connections to the classroom and be of value to the teacher.

In thinking about an out-of-school game, the first question we had to resolve was what game platform? In 2006, when this project began, plenty of studies suggested that the digital devices most common in underserved homes were game consoles, such as the PlayStation or Xbox, or the smaller handheld equivalents, such as the Nintendo DS or Sony PSP. We knew from the outset that these would not be an option. Developing games for these systems required purchasing a costly license from the device manufacturers, and even if we could have afforded the multiple licenses required to cover all the options, the manufacturers were not prone to sell license for “educational” games, which they feared would weaken the image and cache of their devices (a sad commentary on the status of educational games).

So if dedicated game consoles were out, we were left with personal computers—smartphones did not exist. And yet PCs were not sufficiently prevalent in the homes of our intended players. We knew, however, that most students had access to computers in computer labs, afterschool programs, and libraries. Most of these computers would be Internet enabled. If we created a web-served game it would have several affordances:

• We would not have to deal with the barriers found in many schools to downloading and installing new applications.

• Students with log-in identities could continue the game on any machine they had access to.

• There were possibilities for multiplayer games.

• We could aggregate student data in ways that might be helpful to teachers.

This was at a time when Flash had come into its own as a development environment. Flash plugins worked with every browser, and most institutions had already installed them. Flash had the computing and graphic capabilities we needed to create an attractive and immersive game. Although no solution would bridge the last mile into every student’s home, this approach would reach the largest possible number of students, and with home Internet penetration increasing, the game’s reach would only continue to improve.

Labyrinth was designed as a long-form game, requiring twenty hours of game play to complete. We hoped that its involved plot about monsters and kidnapped pets would excite players’ imaginations and empathy, and in the process become relevant to them. Because it was set in a factory, we thought we could develop puzzles in which numerical and mathematical functions would be embedded within the workings of the factory, making concrete what is usually abstract in math instruction.

The Story

When players first log in to Labyrinth, they are immediately asked to pick a pet from a set of ten choices (figure 3.1). Along with the typical cute dog or cat, they can pick more unlikely choices, such as a skunk, a porcupine, or a pig (named Rover). In this, the very first game action, they begin to assert their own identities through the pets they pick, whether cute, fierce, or just a little weird.

Once they’ve picked their pets, players are launched into the story, which is told in comic book form. It is a dark and stormy night, and their avatars are taking their chosen pets out for a walk. Because of the weather, each avatar is wearing a hooded yellow rain slicker. The gender and ethnicity is left entirely to the player’s imagination.

Players leave their pets tied up outside the pet store as they briefly step in to buy more pet food. While they are inside, a monstrous-looking character lures their pets away with a tasty treat, and when they come out, their pets are gone (figure 3.2).

The players’ avatars follow the footprints, which lead to a sewer pipe and through that to a vast and mysterious underground space, in which they promptly get lost. Fortunately, a young woman appears and offers to help. Though she is blue skinned and winged, they have no choice but to trust her. She explains that the pet has been kidnapped by monsters, and to get it back, the player will have to infiltrate the underground factory that is the monsters’ lair. She invites players to don monster costumes to be safe. Again, players are given a chance to assert their identities through the costumes they pick. There are forty-eight possible costume combinations, and the identity they each choose will be theirs throughout the remainder of the game (figure 3.3).

Most of the characters in the story emerge from worldwide mythologies. The young woman helping players is eventually revealed to be Iris, the daughter of Hermes, but we don’t limit ourselves to the Greeks. Monsters from Mayan, Chinese, and Indian mythologies also make appearances. In creating a world that may activate young players’ imaginations, we hope to increase their investment in the game’s outcome.

Following Iris’s instructions, players enter the factory and present themselves to the HR (Horrible Resources) Department. From there they are assigned a series of jobs, each one taking place in a different location, and each representing a different mathematical puzzle to crack. There are twenty-seven puzzles in all. While on the job, they discover that many pets have been kidnapped, and that the monsters have a plan for world domination. In the course of the game, players free hundreds of pets, including their own, and foil the monsters’ fiendish plot.

The storytelling is done with a fair measure of humor. Although players will be called on to thwart the monsters’ machinations, we try to avoid the sodden adolescent fantasy that too many video games traffic in, that of the emotionally distant superhuman hero who saves the world while toiling in isolation.

Games aimed earnestly at kids tend to fall into one of two traps in the way they address their audience: they are either flattering or patronizing. Flattering games tend to portray the player as cool or superhuman, taking a quick and easy route to wish fulfillment. Unfortunately, adult attempts to portray “cool” usually ring hollow to kids. And while we all may occasionally fantasize about being all powerful, as we mature, we come to value problem solving over storybook omnipotence. Indeed, in the best games, we experience our potency through the hard problems we master, not through any superpowers the games may give our fictional characters.

Patronizing games take the route of assuming too little of players. In these game worlds, everything is cute and free of dark narrative threads. Games designed for educational settings are particularly guilty of this approach, the designers apparently afraid of injecting anything that any adult could find objectionable, and equally afraid of challenging kids with genuinely thorny problems. It’s always sunny in these game worlds, and frankly a little boring.

In Labyrinth the world can be somewhat dark, and players are represented as just regular kids prone to errors and setbacks, who nevertheless eventually succeed by virtue of pluck and determination, and with the help and support of others. These story elements mirror our hope for the players: that they will succeed in school and in life through perseverance and collaboration. To quote the experience of Lynette Barr, a teacher in Australia who, inspired by Labyrinth, built a whole curriculum around using games:

The students in 5/6B began the year looking for the “right” answer, or the answer they thought I might be looking for, during any given task. It has taken some time, but this mindset has changed throughout the year as students learned to tackle open-ended questions, solve problems without explicit instruction, take risks in their learning and accept that their teacher will not provide them with all the answers but will instead learn alongside them.

During this trial the student/teacher dynamic has changed as students became familiar with my teacher-as-facilitator approach. Labyrinth has assisted with this transition because it is truly student driven, and my role as a facilitator meant I could provide varying levels of support, modeling and explicit teaching to individuals, small groups or the whole grade in response to the needs of the students.

My teaching practice has continued to transform and evolve with the GBL [games-based learning] approach I am taking with my grade. I am witnessing an increase in the depth of student learning taking place in maths and, subsequently, in other areas of learning. Dealing with student misbehaviour is not at the forefront of my practice when using LOTL [Lure of the Labyrinth] with my students because they are truly engaged in their learning and classroom behaviour is excellent. I take a back seat when students play LOTL and watch them drive their own learning experiences. This trial has reinforced my belief in my own philosophy for teaching and learning—powerful learning occurs when I am a facilitator for student learning, and a learner alongside them. —“Case Study of Lure of the Labyrinth,” Playing to Learn, March 16, 2012, https://lynettebarr.wordpress.com/2012/03/16/lotlcasestudy/.

Puzzles

After describing a typical Labyrinth puzzle, we discuss the principles that animate our puzzle design. In our example, “The Employee Lounge,” HR sends players to the break room, where they meet a large green gelatinous blob of a monster. Totally lacking in appendages, it asks them to do it a favor and get it something out of a vending machine. That’s all the information they have when they arrive at the game screen shown in figure 3.4.

Typically, players start exploring by clicking on the food items or the numbers below them, but the game is unresponsive. Eventually players discover that they can pick up the colored disks/blinking eyes in the lower right of the screen. Some players try to place the disks on the food items but get no response. Eventually they figure out that disks will fall into place with a pleasing “pop” sound when set in the circular spaces on top of the machine. Once three disks have been placed, one in each spot, they fall into the machine with a satisfying “ka-ching,” and a food item drops out of its position and lands below. After that, symbols and numbers appear in the upper right black space. The picture shown in figure 3.4represents the typical results of a player’s first move. In this instance, the player placed green, pink, and blue disks in that order, left to right, and the item that dropped out was in the slot marked 13. (Because this is a monsters’ lounge, the food in the vending machine is humorously disgusting. The delivered cupcake has an eyeball in the frosting.)

Typically, many players’ next move is to try the same three colors in a different sequence. The result is no additional food dropping out, while the symbols in the upper right again mirror the pattern of disks placed in slots, and again equal the same number as before.

Players may do this several more times, but most eventually have the insight that maybe the disks are coins, and maybe each one has a different value, adding up to the cost of the food item dispensed (and the number displayed upper right). Once they’ve made that realization, they hit on the strategy of depositing three coins of the same color. Inevitably, this results in a value that is a multiple of 3, confirming their hypothesis. Once they have constructed this model of the system, they are on the road to solving the puzzle.

The complicating factor is that after three tries, one of the food items starts flashing. If players fail on their next try to submit coins equaling the price of the flashing food, the puzzle shuts down. Only players who figured out the system after one try have gained enough information to succeed. But if the puzzle shuts down, players are free to replay it immediately. The next time they play the puzzle, however, the coins will all have different values. In other words, the puzzle is not about figuring out what the blue coin is worth, but about figuring out a strategy for solving for all values.

When players have displayed mastery by solving this puzzle three times, they graduate to another version of the puzzle, one in which they have the same limited number of tries but can no longer put three of the same color in the slots. The puzzle is still solvable, but it requires players to explore more deeply what they can deduce from different patterns of numbers.

As may be apparent, this is a game about algebra, about “solving for x,” though we never use the word variable or call attention to the concept. Like all the puzzles in Labyrinth, it invites players to explore a system that initially seems opaque, but that reveals itself through probing and experimentation. In the process, the player begins to construct a model of the system.

This puzzle is an illustration of what we mean by relevance. We give the players an interesting problem, one that feels analogous to the real world even as it is embellished with the fantastical. These kinds of problems are not foreign to players from their game-play experiences, even if they bear little resemblance to the way algebra is taught in school.

This example is also an illustration of what we mean by depth. Every puzzle in Labyrinth is procedurally generated—every time it’s played, there is a different, randomly generated solution. To advance, the player must solve it three times. The first time players solve a puzzle, they are often just beginning to grasp the underlying model. Solving it a second and third time gives them the opportunity to solidify their mental model and take pleasure in their growing problem-solving abilities.

Having successfully solved a puzzle three times, they should be ready for us to further complicate the challenge and deepen the complexity of the model. Every Labyrinth puzzle has three levels of difficulty, and as they work their way through them, we hope that players are developing ever more sophisticated models of the embedded concepts. Our goal is not simply to have players demonstrate mastery of a limited number of mathematical procedures, but rather to build robust cognitive scaffolds, and to become increasingly confident in their own reasoning processes.

“The Employee Lounge” is one of nine puzzles, and each puzzle has three distinct levels of difficulty, for a total of twenty-seven puzzles to be mastered. Topics of other puzzles include proportionality, fractions, modal arithmetic, alternate bases, vectors, and the relationship between the area and the perimeter of a rectangle.

In addition, puzzles all take place inside “rooms” in the factory, and to play a puzzle, players must first find the room. This involves deciphering maps, which themselves represent mathematical challenges related to number systems, Cartesian coordinates, and fractals (figures 3.5 and 3.6). As the puzzles increase in difficulty, so the maps grow increasingly obscure and more challenging to interpret.

Game Structure

As deep and satisfying as we hope the Labyrinth puzzles are, giving players an assemblage of interesting puzzles isn’t sufficient. The design of the game’s overall structure is also critical. The structure determines the multiple paths players have to achieve their goals as well as the nature of the incentives they receive for that achievement. Incentives include points earned, feedback about levels of mastery, and the players’ advancement through the game’s story. Labyrinth is structured to encourage both freedom of exploration and persistence in the face of difficult challenges.

Even in unsuccessful attempts to solve puzzles in Labyrinth, the player earns points (called “credits”) that will accumulate toward winning the overall game. In other words, some credits are awarded simply for effort, though more are awarded for success. The purpose of these partial rewards is not to make the player feel good, but rather to honor the fact that the player persisted, that problems can be solved only through repeated effort, and that strategies learned from unsuccessful attempts usually lead to eventual success. Repeated failed attempts may cost the player time, but they don’t otherwise impede the game. Indeed, players usually need to collect several failures just to acquire the information they need to solve each puzzle. The game makes it easy to replay a puzzle immediately after a failed attempt to smooth the way for the kind of iterative problem solving often demanded of us in real life.

In fact, a player could conceivably reach a conclusion of the game by accumulating credits entirely through unsuccessful play (though this would take significantly longer than the twenty hours of game play we expect). This is not to say that we don’t provide incentives for genuine success. Players who play and beat every puzzle will get to the end of the game much quicker, with tangible markers in the game that they’ve played well. But no one is penalized for mistakes (except for lost time), so every player who eventually solves a puzzle ends up with an equally high score. The game puts a premium on problem solving, not on speed.

This structure also enables players who feel stymied by one puzzle to put it aside and pursue success in other puzzles. Players are free to follow their interest and curiosity into whatever part of the game excites them most at the moment. Real learning is driven by interest, and interest in every player will be different and idiosyncratic. Today I may be intrigued by one puzzle, tomorrow by another, and the game allows me to follow whichever path is calling to me now.

Through the course of game play, players are accumulating credits. In turn they can use these credits to buy items that they use to free pets trapped in the factory. The accumulation of credits is also what triggers new chapters in the game’s narrative as well as new revelations in the mystery at the center of that narrative. Over time the game world gets more expansive as new rooms are opened to the player, and it also gets deeper as the complexities of the story are revealed.

Creating such a broad and deep world was always a significant design goal. We wanted the game world to feel “large,” so players’ imaginations would be stimulated, and their investment in the outcome would be strong enough to motivate them even when the going got tough. Small games can of course be artful and engrossing, but when done well, epic games invite players to make a deeper commitment; when one’s friends and classmates are simultaneously playing them, they acquire even greater relevance.

Pedagogy

As is evident from “The Employee Lounge” puzzle, the game is not “curricular.” It doesn’t cover all the material one would find in a textbook chapter, nor does it attempt to instruct students on traditional procedures or algorithms used when “solving for x.” In fact, we wouldn’t expect most students who play this game to automatically see the connection between the game and their formal math instruction. But we do think that by playing the game, they’re doing important work. In response to the mysterious phenomena they encounter in the game, they initially probe (clicking around the screen) and observe (what actions work, what feedback do I receive?). What starts out as trial-and-error exploration gradually evolves into more organized testing of hypotheses, and along the way, they learn practices like controlling for variables in their testing. They are learning how to model systems in order to make sense of them. And they are building mental models of certain specific mathematical concepts.

The process of modeling described above actually occurs in any game that’s sufficiently challenging, and yet we know that dedicated gamers don’t automatically adopt the identity of scientists or engineers even if they are mastering analogous problems through their game play. We also know that playing interesting games won’t make everyone a good student. We believe that only through continuing to work with the models they’ve constructed will students solidify the understanding they’ve been approximating through their game play. Fostering this continued engagement with the concepts learned in games remains a critical role for educators in the process.

So imagine a teacher walking into class one day and projecting “The Employee Lounge” puzzle on the whiteboard. Because she can see logs of online activity, she happens to know that most of the class has already mastered this puzzle. She knows that a student who hasn’t participated much in class discussion was actually quick to solve it. So she invites that student up to the front of the room to describe her strategies for solving the puzzle. Other students also volunteer their strategies, and a discussion follows about both the particular puzzle and problem-solving strategies in general. The design of Labyrinth makes it easy for teachers to access discrete puzzles within the game to facilitate this kind of classroom discussion whenever it aligns with the class’s given curriculum.

In traditional math classrooms, the introduction of each new topic—like variables—is an opportunity for more students to get left behind, to finally decide that they’re “not good at math,” as most students will conclude by the time they graduate high school. And once they get left behind, they rarely catch up again. In the story above, before the topic of variables is introduced, students have actually mastered the process of solving for x, and they’ve done it in a relevant context. So when the teacher introduces variables, she’s not throwing a new obstacle in her students’ path, but rather she’s building on their hard-won expertise. The classroom discussion becomes an opportunity to solidify the understanding students may have begun to acquire through game play but have not yet connected to the formal procedures of mathematics.

In the above example, the teacher is also learning. From looking at the player logs, she’s discovered that a quiet student is actually engaged and capable. Through the logs and the ensuing classroom discussion, she learns more about the competencies her students bring to the work, and about their ways of thinking about problems. It is an opportunity for her to give positive reinforcement to her students for their persistence and for mathematical reasoning they’ve displayed. It’s also an opportunity to provide further attention to any areas of weakness that may have been revealed.

Of course, this entire discussion turns on a hypothesis of how learning concepts can be initiated through game play and fortified through classroom experiences. Initially, we intuited these ideas through our own experiences of learning and game playing, but they have been supported in the research literature and characterized as preparation for future learning(Schwartz et al., 2005). Building on earlier work by Harry S. Broudy (1977), preparation for future learning focuses on “interpretive knowledge” and the ways in which prior learning experiences, when they are experientially rich, can give learners the interpretive tools to make sense of new problems. Even when learners have forgotten all the factual details, or content, of a prior learning experience, if they’ve had opportunities to build robust mental models, they are better prepared to learn new concepts when they encounter them.

In one study that supports this approach (Schwartz & Martin, 2004), researchers divided learners into two groups. One group was given reading material about a specific concept, and the other group was given a hands-on experience that enabled them to explore the concept much the way our games enable students to explore concepts. Each group was further subdivided in two, with half of each group subsequently hearing a lecture on the same concept, and the other group getting no further instruction. All groups were then assessed on their understanding of the concept. The group that had only the hands-on experience fared the poorest on the assessment. The two groups that started with reading did about equally well regardless of whether they received further instruction. But the group that followed the hands-on experience with formal instruction did significantly better on the assessment than any of the other groups. The earlier exploratory experience had better prepared the students to truly learn the concept in a more formal context.

We would never claim that playing a game like Labyrinth is all that’s required for a student to learn a concept. Rather we see game play as the kind of engaging and relevant experience that leaves students prepared to make fuller sense of the concepts they’ve begun to master when they encounter them again in school, or in their daily lives.

Collaboration and Reflection

We’ve already discussed the critical role that classroom experiences can play in solidifying a student’s understanding of mathematical concepts first encountered in games. As important as we believe the classroom is in this process, we also believe that learning can be fortified through collaboration with other students, and through self-reflection. As we designed Labyrinth we kept thinking about how gamers often engage in fairly sophisticated conversations about their strategies in game chat channels, in game FAQs, and in fan forums (Steinkuehler & Duncan, 2008). For example, students who show no interest in writing in school will sometimes write elaborate, well-constructed walk-throughs of games for an audience of their fellow players, because this audience has meaning for them. The culture among gamers is one of reflection and collaboration, even if gamers might not label it as such.

We therefore wanted a way for Labyrinth players to engage in similar behaviors. The most direct way of fostering this would have been to make Labyrinth a multiplayer game with a built-in live chat system, like World of Warcraft’s. For several reasons, however, this wasn’t possible. Multiplayer games are much more expensive to build and maintain, which would have strained the limited resources of the grant. Of equal significance, creating an open online environment to be used by middle schoolers would have had serious issues. Parents and teachers are understandably concerned about who children of that age might encounter in unmoderated spaces, and laws limit what can be offered to children younger than age thirteen. We firmly believed that some form of collaborative play was desirable, but for such a game to be adopted widely, it would have to be a walled garden, where students encounter only students they know personally. And since there are opportunities for mischief in even such bounded environments, it would have to be a space that teachers could monitor with relative ease.

We had a second reason for wanting some form of collaborative play, beyond its benefits for learning math. A subsidiary goal of Labyrinth was to have a positive impact on student literacy. The choice of comics as a storytelling mode was part of this strategy. But we also believed that if we could inspire students to read and write about their mathematical reasoning, it would influence their reading and writing as well as their reasoning.

As much as we desired collaborative play, we knew about additional constraints related to the lives of the students we were most interested in reaching. Remember that we were designing for use after school, and for a platform (the Internet-enabled computer) that was not in every home. We presumed that our target audience didn’t necessarily have adequate control over their own schedules to guarantee that they could play online at the same time, so the game would have to work when played asynchronously.

Our solution was to enable teachers to put the students in their classes on teams of up to six players, and we suggested they set up a friendly competition between teams to get the highest score. Although all students would be playing the game on their own, they would have incentives to help their teammates succeed as well. To facilitate this, we built into the game a messaging system, but one in which students could see only the messages of their teammates. We imagined students would use this system to ask each other for help or to share successful strategies. Since all the puzzles were procedurally generated, we knew students couldn’t give each other “right” answers (every time they played, the answers would be different). But students could write to each other about their strategies, or try to explain to each other the underlying model for a given puzzle. Research consistently shows that when students have opportunities to teach each other, both the “teacher” and the “learner” benefit. Our hope was that the messaging system would become a major venue for collaborative learning.

Labyrinth is still widely used nationwide, but because of changes in federal funding in the midst of the project (the entire Star Schools program that supported Labyrinth was defunded), we’ve never had opportunities to adequately study the game in use. As such, we don’t know how much players use the messaging system, though a small-scale study undertaken several years after the game’s release suggests that the messaging is not used as much as we’d hoped. This may be because students within a class have plenty of opportunities to talk without depending on the messaging system. It may be because the students don’t perceive that they’ll get answers in real time, so they don’t bother to use it. We certainly see the behavior we were hoping for in the live chat channels of many multiplayer games, so perhaps it would have worked better if students could have had live chat with the community of all players, but of course that wasn’t feasible for reasons of safety and security already mentioned. Nevertheless, teachers report to us that the team competition motivates their students to collaborate, and they do see some interesting math talk from time to time in the chat channels.

Even when we are working to make robust, fully useable games like Labyrinth, we also see our work as design research, and we specifically saw the inclusion of chat as one of the more experimental features of the game. Our experiences with collaborative play in Labyrinth informed our subsequent work on Radix and Vanished (both described later in this volume), and we continue to explore it in much of our ongoing work.

Integrating Game Play and Narrative

In designing Labyrinth, we built on what we had learned from the development of an earlier game, Zoombinis, designed in 1996 by Chris Hancock and Scot Osterweil at the Cambridge research and development firm TERC, and developed by the respected software publisher Brøderbund. In particular, we pursued a similar structure for Labyrinth, one that makes it easy for players to persist with a puzzle through repeated play, but that also allows players to explore an epic space and apply their efforts wherever their interests or impulses lead them. As in Zoombinis, we created puzzles that require players to start with opaque challenges and gradually construct new models in order to eventually prevail.

Zoombinis was a very successful game in the marketplace, and we think in Labyrinth we created a game that was wholly original, but that built on what was learned from the creation of the earlier product. Nevertheless, one way in which we believe Labyrinth didn’t fully match the success of Zoombinis was in the full integration of game play and narrative, and the reason for that is instructive about the design and development process.

Zoombinis is a game about combinatorics and the mathematics of data. The game actually emerged out of earlier work by Hancock and Osterweil on ways of representing data for young children. A classroom product they designed called Tabletop Jr. included little blue creatures whose traits (hair, eyes, nose, feet) were independently variable, like data records in a database. When Brøderbund asked Hancock and Osterweil to create a game based on these characters, the narrative that emerged was based on how the Zoombinis’ characteristics interacted with the world they encountered on an epic journey. As such, the whole world of Zoombinis is a world full of structures related to data and data processing. The game narrative and the game play are fully integrated.

Creating Labyrinth, we didn’t have the luxury of starting from clearly instantiated mathematical concepts, the way the Zoombinis already embodied ideas about data. Before we had opportunities to fully explore the mathematical concepts we would base the game on, we needed to develop a preliminary design document that captured the scope of the game and the underlying narrative so that the software development could be put out to competitive bid among several potential production partners. This forced us to commit to a narrative gambit before we had fully defined our mathematical ideas. We think that the narrative we developed is appealing to our target audience, and it justifies the inclusion of multiple puzzles, each embodying a different mathematical concept, but the larger story world did not fully embody those concepts.

After we had committed to a development partner and an overall approach to the game, we were able to further integrate mathematics into the larger game world through the maps the player uses to find each puzzle room, but the game remains a collection of good mathematical puzzles linked by a narrative and game world that is itself not fundamentally mathematical.

In an ideal design process, we might have had more time to develop the core ideas of the game before committing to a development plan. With Labyrinth’s focus on pre-algebra, we might particularly have looked to create a world that reflected issues related to ratios, proportions, and fractions. For example, in retrospect we can imagine a game narrative about a character who, like Alice in Wonderland, has to navigate a world in which items frequently grow or shrink, and more of the puzzles and the larger game world might have involved the proportional reasoning necessary to navigate those changes.

We bring this up not to lament the outcome of Labyrinth, a game we view as very successful, but rather to highlight the importance of a design and development process that, as much as possible, enables the design team to fully explore the ideas that will undergird the game. We strive always to create games that help illuminate what is authentically playful and engaging about their subject, and the more that design can occur before the high cost of software development kicks in, the more likely we are to fully realize our goals.

Conclusions

We’ve discussed how we designed Labyrinth to work in the particular academic environment we were given, one in which teachers were at best cautious about using games, and one in which students would not all have access to the same computers or game devices. We’ve also talked about how the underlying pedagogy of Labyrinth promoted collaboration, reflection, and most importantly, preparation for future learning (although it should be emphasized that the “preparation” is itself also a form of learning).

It would be easy to conclude from this discussion that if we design a good game, and one that fits into the ecosystem of school, then somehow learning will automatically occur. That couldn’t be further from our intent. No matter how well we work to ease the way for students to learn through well-designed, engaging games, our core belief is that students learn best when they choose to learn, when they’re motivated and inspired to learn. Which brings us back to the question of identity.

We want students playing Labyrinth to master new math concepts and exercise literacy skills, but we particularly want them to experience themselves as capable problem solvers, learners, and collaborators. When they are confronted with a challenge that seems difficult or opaque, we want them to remember the value of probing, observing, hypothesizing, and leveraging wrong answers as necessary steps on the road to right answers. We want kids not to just see themselves as the superheroes and dragon slayers of childhood fantasies, but to take on an identity as members of a team of heroic problem solvers, the kind of identity they will all need to become lifelong learners and actively engaged citizens.