"Moving Learning Games Forward: Obstacles, Opportunities, and Openness," by Eric Klopfer, Scot Osterweil, and Katie Salen of the Education Arcade, an MIT research division that explores games that promote learning through play, identifies a number of principles that must be followed when creating a successful educational game:

1. Choose wisely. Games are not perfect for everything. For example, there are many attempts at games about topics like photosynthesis, but most of what results is not a game at all, but a more typical rote classroom activity. "Topics should not be forced–games should be one medium among many for learning in and out of the classroom," explains the report.

2. Think small (sometimes). Some people get lost in 3-D worlds, and massively multiplayer games are expensive and complicated to make. "This isn’t to say there shouldn’t be big fancy learning games, but rather that not all games need to be large and complicated," says the report. The scale and complexity should be chosen so that they appropriately match the learning goals and context.

3. Educational games don’t always equal entertainment games. Many developers are scared off from developing educational games because they are convinced that the economics don’t work, says the report. Commercially successful video games cost millions of dollars to create. Developers should keep in mind that all educational games don’t have to meet the same mass-commercial standards. One way they could cut down on cost is by reducing aesthetics.

4. Put learning and game play first. Learning should not come first, and neither should entertainment come first; instead, they should both be considered simultaneously.

5. Find the game in the content. The report says that in any academic discipline, there are elements that are fundamentally game-like. For a historian, this might involve puzzling over the motives or needs of different historical actors. For scientists, it might mean constructing models of phenomena based on incomplete evidence. "While every academic discipline may involve a certain amount of drudgery or memorization, the role of games should not be to ‘sweeten’ those activities, for once a game begins enforcing work, it ceases being a game and ceases accessing those creative impulses so fundamental to play."

6. Break the mold for where educational games are played. Think about playing them outside of class and then discussing the game play in school.

7. Harness the "soft skill" learning from games, but connect it with content. Game players often have many 21st-century skills, such as the ability to problem-solve or collaborate. Games should help develop these skills for education by allowing access to shared content from other players, including specialized options for different players and experts, including competitive elements, allowing room for peer assessment, and including validating elements.

8. Don’t ignore, nor be limited by, teacher training and readiness. Games should be designed so that students can play without much help from the teacher. "What the game should provide the teacher with are materials that help relate the students’ game experience to existing curricula whether through discussion or other hands-on or paper and pencil activities," states the report. Also, games should be designed so that teachers can access specific game experiences easily, without necessarily having to work their way through the game the way students do.

9. Play everywhere and anywhere. This means taking advantage of mobile gaming platforms.

10. Reduce, reuse, and recycle. Assets of a game should be reusable. Creating characters, objects, and environments requires significant artistic investment. The report suggests that a set of artists could be commissioned to design an open-source library of Flash-based characters and character behaviors. Openly sharing and recombining these assets can lower production costs while "placing the emphasis on other unique aspects of the game to create an identity."

11. Define the learning goals. The report states that the question, ‘Do games help kids learn,’ is the wrong one. Just like asking if ‘books are good,’ each game is unique and there will be great variance among them. Also, games don’t stand alone in the learning process; the context and supplement materials also must be analyzed, and training must be provided. In addition, assessments must be defined, meaning that learning goals must be established with the game.

12. Forge partnerships. Use the resources provided by academia (universities), commercial companies, nonprofit organizations, foundations, and government agencies to help explore new designs, prototype ideas, and work experimentally with potential users.

13. Don’t ignore, or be too constrained by, academic/state standards.

14. Think of not just what but who, where, when, and why. It isn’t just what games people play that matters, by where, when, and with whom.