From Serious Games to Serious Gaming (Part Three): Backflow

This is part Three in a multipart series showcasing the serious games work being done by the MIT Comparative Media Studies Program. Today, we focus onBackflow, one of the games developed this summer as part of our newly launched Singapore-MIT GAMBIT Lab. Under the guidance of Eric Klopfer, Judy Perry, and Marleigh Norton, Backflow was developed by Zulfiki bin Mohamed Salleh, Neal Grigsby, Chen Renhao, Nguyen Hoai Anh, Wang Xun, Fabian Teo, Brendan Callahan, Guo Yuan, and Hoo "Fezz" Shuyi from the Singapore-MIT GAMBIT Game Lab. To download Backflow, visit the GAMBIT homepage. Backflow

By Neal Grigsby, Philip Tan, and Teo Chor Guan

The GAMBIT Summer Program

The Singapore-MIT GAMBIT Game Lab was established in 2006 between MIT's program in Comparative Media Studies and Singapore's Media Development Authority as a five year project to sponsor new research about video games, to development new and innovative games, and to train students from Singapore's tertiary education institutions, its universities and polytechnics, in preparation for entering the games industry. The GAMBIT name refers to the many axes along which the project aims to generate research: Gamers, Aesthetics, Mechanics, Business, Innovation, and Technology.

The project adheres to the ethos of "applied humanism," which is the conceptual core of the Comparative Media Studies program. All CMS research projects share the goal of putting theoretical research into practical application, of testing theoretical precepts in contexts outside of the academy, for example in the realm of childhood education (the project for New Media Literacies), the media industry (the Convergence Culture Consortium), and journalism (the Center for Future Civic Media). For GAMBIT, this philosophy demands not only the generation of writing about games, but also the development of games themselves in support of research goals.

In 2007, GAMBIT research managers prepared for their first summer of game development. They selected over 30 students from Singapore based on the strength of their academic records, portfolios, and their demonstrated passion for video games. Successful applicants were flown to MIT to participate in the equivalent of a professional internship, the GAMBIT summer program. Over a short 9 week period, they worked closely with MIT students and faculty to develop 6 new games, each designed to answer a specific research question.

Most of the research questions and initial game concepts came out of a semester-long process of investigation, preliminary research, collaboration between academic departments, and, finally, the selection of brief written proposals. They were chosen based on mutual interest from faculty at MIT and Singaporean institutions, and the viability of projects for such a short development cycle. For each game, the faculty member in charge of the research would meet with the development teams at the beginning to outline the project requirements and help in brainstorming game ideas, and then reconnect systematically throughout the cycle to ensure ongoing compliance and provide help when needed.

In addition to the specific research goals of individual teams, there were also some shared goals. Games should be innovative but should not ignore the lessons of thirty years of gaming history, that is, they should retain what is interesting and successful about current games. The teams should strive to meet the academic research goals, but also aspire to a professional level of polish. Most importantly, the development of the games should contribute to the education of the team members, to expand their intellectual and professional horizons. To this effect, the first week of the 9 week program was spent in orientation, with faculty and staff lecturing students on game design, usability, animation, and related topics. In addition, local game industry professionals provided insider perspectives on design challenges and careers. This week-long orientation served as a tool-kit to help prepare the students for the intense 8 weeks of software development that followed.

With such a short development cycle, and a demanding variety of projects to complete, the organization and management of teams was considered of highest importance. Teams were kept small in the context of software development, with just 7 members each, but taken together exceeded the threshold for effective centralized control. Therefore, they needed to be relatively self-sufficient and able to respond to challenges with speed and independence. Each team consisted of students from a variety of backgrounds and specialties: two programmers, two artists, a test lead, a game designer, and a project manager. In addition, a wildcard 2-person team of musicians provided audio services to all of the game teams.

Programmers were, of course, responsible for writing the software itself, and artists for the generation of art assets such as character designs, backgrounds, menu screens, etc. The day-to-day tasks of the test lead would change often throughout the cycle, but always with the intention of placing him or her in a position to represent the interests of the end user. In the early design phase, the tester would assist in design tasks, but while the researcher and designer might come up with innovative, "blue sky" possibilities for their games, the tester was to help keep their feet on the ground, to remind them of their responsibilities to the player. Later, the tester would be responsible to systematically evaluate each new iteration of the game.

The designer was responsible for translating the educational goals of the project into a game concept and play mechanics. However, GAMBIT designers were encouraged to put their egos aside and serve more as facilitators of the design, as design leaders rather than authoritative authors. Each team member needed to feel a sense of ownership over the design if they were to successfully complete the project requirements. After all, the project would be their full time job over the next two months. The team could not afford to have even one of its members feeling disengaged, or believing that their ideas had not been taken seriously by the rest. In addition, while it was thought that the Singaporean students would be very technically adept, GAMBIT hoped to engage the students on a much more creative level than had previously been demanded of them. Therefore the concept for the game should come from the team itself and should be selected by consensus, with the designer stepping in as tie breaker when necessary.

Finally, the project manager would keep the team on track to finishing their project. A flavor of agile project management called "Scrum" offered a reasonable fit with the demands of the project. For the purposes of Scrum, those working on the project are divided into three categories: the product owner, the Scrummaster, and the team. The team consists of everyone who works closely on the project, those fully committed to doing the work that will turn ideas into reality. The product owner represents the long-term view of the project, often advocating for the software end-user or the requirements of the larger organization. The Scrummaster makes sure that participants adhere to the rules of Scrum, which calls for short daily check-in meetings (the daily scrum), the management of development cycles called "sprints," and other details of the process. The role of Scrummaster tracks fairly closely with the role of "project manager" in other paradigms, and indeed GAMBIT project managers became Scrummasters.

In the GAMBIT summer program, the researchers were assigned to the role of product owner. He or she would work with the team to set and prioritize the goals of the project, but would not need to dictate exactly how those priorities were met. The work would be divided into four two-week sprints. At the beginning of each sprint, the product owner would collaborate with the team to prioritize a list of product features called the "product backlog." The team would select the amount of product backlog they believed they could achieve, which would become the "sprint backlog," and define for themselves the best way to realize these features. At the end of each sprint, the team would be called upon to demonstrate new functionality to the product owner in a "sprint review" meeting. Each cycle also offered an opportunity to discuss and refine the rules of Scrum itself to better serve the personalities of the project and the team. Running the teams this way demands that team members have to produce not just work from individuals (code, concept art, music, design documents) but also an integrated, testable and ever-improving game every other week. By adhering to this program of "iterative development," teams would have a better chance of actually producing a finished product after eight weeks of development.

The use of Scrum project management was both an adaptation to the stringent requirements of the project and an experiment in its own right. As many of the professional visitors who lectured GAMBIT students would attest, most game companies operate with a development cycle that perpetually concludes with "crunch time," or a period of company-wide overtime. One of the goals of Scrum is to eliminate crunch time by making sure management decisions are not based on unrealistic expectations. The team estimates the development time for each feature and management is forced to accept the team's estimates or make an explicit decision to scale back the project. Even in companies that have adopted the Scrum management system, however, crunch time is often unavoidable due to the demands of the commercial marketplace. As a fully funded educational research project, GAMBIT enjoyed relative independence from these demands, even as it aspired to a professional product. The managers of the project discouraged overtime as much as possible to engender a more healthy workplace environment and create a model of "sustainable development."

Backflow: The Game

Existing mobile participatory simulations such as Palmagotchi use the peer-to-peer connective capabilities of Palm and Windows Mobile handheld computers to embed a group of players inside a simulation. While each individual device is inexpensive, purchasing enough devices for an entire classroom can be a prohibitive expense for many schools. To address this, researchers from MIT Teacher Education Program worked with the Singapore-MIT GAMBIT Game Lab to develop a game for a platform more popular among teachers and students: the mobile phone. Mobile phones are a challenging game development platform in comparison to dedicated gaming consoles or the PC, with relatively tiny screens, low system memory, and low-powered microprocessors. However, every mobile phone is a communications device, incorporating networking technologies that are well suited for participatory simulations.

To that end, a team of students from the Singapore-MIT GAMBIT Game Lab were given a number of new mobile phones from a variety of manufacturers, and set to the task of developing a game that would run reliably across at least two of the devices. The Singapore-MIT GAMBIT Game Lab was established in 2006 between the Program in Comparative Media Studies at MIT and the Media Development Authority of Singapore as a five-year project to sponsor new research about video games, to develop new and innovative games, and to train students from Singapore's tertiary education institutions in preparation for entering the game industry. In 2007, over 30 Singaporean students were flown to Cambridge, Massachusetts, to participate in the equivalent of a professional internship. Over a short 9-week period, they worked closely with MIT students and faculty to develop 6 new games, each designed to tackle a specific research challenge head-on.

Working with the MIT Teacher Education Program, the team chose to address environmental issues in their simulation. The original concept had the player directing the flow of sewage through a series of pipes using switches, with the option of shunting the waste to his or her neighbor, helping to clear the game screen but potentially inviting retaliation. The game would simulate a system of environmental exchange and the interdependencies of environmental actors. The basic mechanic also offered the possibility to support a fun, casual-style single-player game themed around recycling, in which the player's frantic button mashing would direct recyclables in the waste stream to the correct recycling bins.

Other than the title of the game, Backflow, everything else in the original design had to change in order to accommodate testing feedback and hardware realities. A game in which players flushed waste to each other in real time required the mobile phones to be constantly connected to the Internet and maintain synchronization between all players. Network latency limitations and subscription costs made such a system very difficult to realize. Furthermore, game testing suggested that a multiplayer game based on "tragedy of the commons" would not be very fun to play.

The basic single-player mechanic of sorting garbage by flipping switches on pipes remained, but the multiplayer aspect was scaled back to work asynchronously, with interactions between players recast as an exchange of resources in a stock market-like system. Instead of dumping garbage on each other willy-nilly, players would negotiate to share waste capacity. The game designer hoped to use this waste market to simulate the process of "cap and trade" emissions credit trading, a strategy that has been used successfully in the real world to limit greenhouse gas emissions in a free market.

The final game is best described as a hybrid of several genres: a casual puzzle game, a city simulation, and a resource trading and management game. The player begins by registering a new account and creating a new city. New cities start at 45,000 residents, a value that will change with the success or failure of the player's ability to properly recycle. A maze of pipes extends from the city at the top of the screen to several recycling bins and a sewer near the bottom. The player uses the keys on the mobile phone number pad to direct items to the right place: glass to the glass recycling, organic waste to the sewer, and so on. If the player sends a recyclable item to the correct bin, the game rewards some raw materials of that type. These materials can be used to build efficiency upgrades for the player's system. But if the player makes a mistake and sends waste to the wrong bin, the pollution level for the city rises.

At the end of a round, the game calculates the city's pollution level, and adds or subtracts residents accordingly (based on the assumption that clean cities are more attractive living spaces than polluted ones). Also between rounds, the player may decide to buy system upgrades or trade resources. Players soon realize that they can easily build up a scarcity of one type of material and a surplus of another, making trading necessary for advancement. Urban growth increases the complexity of the pipe system and the speed of the waste stream. "Winning" the game means finding a balance of population and waste processing ability that a player can manage.

Despite the challenges of a new and constrained platform, the students successfully created an online mobile phone game in 9 weeks. Much of the success of Backflow is a testament to the team's adaptability: they faced the limitations of the technology and the feedback from real players and adjusted the game design and development plan to make a functional, playable, and engaging game faithful to the spirit of the MIT Teacher Education Program.

Philip Tan is the executive director for the Cambridge operations of the Singapore-MIT GAMBIT Game Lab, a multi-year game innovation initiative hosted at the Massachusetts Institute of Technology. He is concurrently a project manager for the Media Development Authority (MDA) of Singapore and a member of the steering committee of the Singapore chapter of the International Game Developers Association (IGDA).

Prior to his current position, he worked closely with Singapore game developers to launch industry-wide initiatives and administer content development grants as an assistant manager in the Animation & Games Industry Development section of MDA. He has produced and designed PC online games at The Education Arcade, a research group at the Massachusetts Institute of Technology that studied and created educational games. He complements a Master's degree in Comparative Media Studies with work in Boston's School of Museum of Fine Arts, the MIT Media Lab, WMBR 88.1FM and the MIT Assassins' Guild, the latter awarding him the title of "Master Assassin" for his live-action roleplaying game designs. He also founded a live DJ crew at MIT.

Teo Chor Guan has more than 14 years of experience in systems engineering for computer graphics and games. Her wide range of experience spans from 3-D graphics research at the Institute of Systems Science in Singapore to building air traffic control systems for MacDonald Dettwiler & Associates in Canada. She has also worked as a software developer at Electronic Arts (EA) Canada for over six years. She holds a Bachelor of Engineering (Electrical and

Electronics) from the National University of Singapore and a Masters in Computer Science from Simon Fraser University, Vancouver, Canada. In Singapore, she had worked in the Games Development Group at the School of Design at Nanyang Polytechnic. She was also the Software Engineering Manager in Lucasfilm Animation Singapore before joining Media Development Authority (MDA) as the Program Director for GAMBIT.

Before coming to MIT, Neal Grigsby worked at LookSmart.com in San Francisco for seven years, wearing a variety of hats including editor, ontologist, and content producer. He joined the Comparative Media Studies program in 2005, where he wrote his master's thesis on narratives of adolescence, including a look at representation of youth in video games. He got his feet wet in game design at the annual Storytelling and Games in the Digital Age Workshop, where he led the winning team in 2007. While a student at CMS he worked on the Project for New Media Literacies, producing educational video and curricula about media production for an audience of teens and young adults. Neal earned his bachelor's degree in film studies from UC Berkeley, which was also where he met his wife, artist Rebecca Bird Grigsby.