A sixth-grade boy pounds the arrow keys and space bar on his laptop keyboard, expertly maneuvering an avatar that looks like a fat baby wearing a blue bonnet through a brightly colored maze. The video game goes “Ding!” each time the avatar eats a little yellow disc or key, and with each key, a door to a different part of the maze pops open. When the avatar reaches the end of the maze, the screen flashes “You Win!”

The boy pumps his arms in the air. To the girl sitting next to him, he says, “How neat. That was easy.” The girl tries to suppress a smile: “I didn’t have time to build a second level yet,” she says.

This scene takes place at the public school Quest to Learn, which opened last year on the East Side of Manhattan. Quest is one of 139 schools created by the city Department of Education in collaboration with New Visions for Public Schools, a city-based nonprofit specializing in helping schools adopt innovative curricula. It’s also the first school in the U.S. to teach entirely through lessons, called “quests,” that are structured like games.

In the class “Sports for the Mind,” students build their own video games, using a program called Gamestar Mechanic that was designed by the school’s executive director and founder, Katie Salen. In another class, “The Way Things Work,” students help fictional characters called Troggles whose homes are falling apart, teaching them about standard measurement so they can build sturdier structures.

“We present kids with a series of problems that build them up to be able to solve the next one, just as a game would do,” explains Quest curriculum designer Arana Shapiro. “So as you level up, the quests become increasingly complex, and they require you to draw on skills that you learned in the previous level.”

“Sports for the Mind” teacher Al Doyle says his class helps students learn “systems thinking,” understanding the relationships of parts to wholes. One analogy, he says, is the subway system: “What is the goal of the subway system? To get passengers from one place to another. What are the rules? You have to buy a MetroCard. What is the hidden part of the system? The electricity that runs it. What are some other components? Invariably, someone says rats. There are rats in the system, and if you ignore them, the system will break. You may not want them in your system, but they are there, so the system has to include ways of dealing with them.”

When the students design their own games, they must incorporate the components of a system: goals, rules, and stakes. Eventually, says Doyle, students will learn the more complex aspects of a system, such as choice and balance, and build those into their games.

Doyle says he saw signs of systems thinking in his students’ first assignment: to design what was called a Birthday Game. The player had to collect points, and the game had to be light and colorful. No enemies.

When one girl showed her game on the class’s Smart Board, it was dark and gloomy. Doyle recalls that when he pointed this out, the girl rolled her eyes and said, “If you played the game, you’d know that you have to get through the street to get to the party.”

Doyle realized that the game actually had five levels. First, he had to escape the house because the avatar’s parents didn’t want it to go to the party. Then he had to find the key to get out. Once on the street, he dodged thieves and muggers. All the while, he collected points. There were several challenges before he arrived at the colorful party.

The premise of using video games to engage students in advanced thinking is drawing more advocates. “Virtual worlds have the added benefit of play,” says Jan Plass, co-director of New York University’s Games for Learning Institute, a collaboration of researchers from seven U.S. universities looking to find out what makes video games so engaging. “Video games have the power of visualizing things, of creating open-ended environments for people to explore things, of engaging and motivating learners. What you have is a strong learning approach that should be added to the educator’s toolbox.”

Both Salen and Plass like to say that schools are stuck in the 20th century, where rote learning reigns. Proponents of video gaming claim that a new set of skills—problem solving, communication, collaboration, and critical thinking—should be taught to prepare kids for the 21st century.

Of course, teachers have long beenpromoting these skills as a remedy forrote learning.

“[Video gaming] is hardly a 21st-century skill,” says Diane Ravitch, an education historian and NYU professor whose book Kid Stuff: Marketing Sex and Violence to America’s Children includes essays on video games’ effects. “What [students] need most is to know history, civics, foreign languages, economics, literature, and to engage in the arts. These are knowledge and skills not acquired in the blink of an eye. They require thinking, self-discipline, practice, concentration, study, intellectual energy—not the same skills one learns when playing with video games, which give instant gratification and reward the lucky.”

Other education experts are more positive. “There are a lot of benefits to putting lessons into a game format,” says Charles Kinzer, a professor of education at Teachers College. “You have record-keeping functions that are built in. Players know how well they’re doing compared to other players, and that’s motivating.” Games can also adapt their difficulty levels to players’ abilities, Kinzer says, staying challenging without becoming frustratingly hard. “I’m not saying that teachers don’t do this. But it’s when a teacher is dealing with 30 students, those students don’t necessarily get immediate feedback. That lag time is detrimental.”

“You’re never left for very long without knowing how you’re doing,” agrees Quest designer Shapiro. “You can’t move further if you don’t solve some problem. You’ll see kids try over and over again to move a block out of the way when they’re playing a game, but they won’t try over and over again to add a fraction together.”

Scot Osterweil, creative director of MIT’s Education Arcade, which researches games for use in the classroom, says that learning through games does happen, but that it’s more through player activity and discovery, not through instruction, much the way artists, engineers, and scientists solve problems: “They observe, probe, hypothesize and test their hypothesis, and solve problems, and they do that across a wide range of games.” But, he adds, “there is no guarantee that when a player experiences these things in a game—even if they develop those skills to a high degree in games—they will recognize how to use those skills in real life. Teachers and parents have to make those connections and show kids how these skills apply to real life. No matter how powerful games are, they can never replace teachers.”

Osterweil also questions whether video games can match the complexity of our current teaching strategies. “Right now, the quality of the narrative in video games is weak,” he says. “The all-powerful hero who slays the dragon and saves the world is a necessary part of kids’ fantasy lives when they are young. But in literature, in Hamlet, Oedipus, or Macbeth, acting forcefully or heroically doesn’t always work. Literature can get at the greater complications of life. Games are still stuck in the simple forms of narrative.”

At NYU’s Games for Learning lab, researchers design prototypes of educational video games based on their observations of students playing video games at 10 New York City schools. One invention is a posture sensor that will measure a game player’s engagement from the way he sits and moves. The purpose is essentially to engineer learning—to discover exactly where the students are most engaged and transfer that knowledge to making educational games that engross kids so fully they don’t even realize they are learning.

This reduction of student behavior to reproducible action begins to make more sense in light of the commercial interests behind gaming research. E-Line Media will sell an expanded version of Gamestar in May as well as offer a simpler version for free. A portion of the proceeds from Gamestar will go to Salen’s nonprofit Institute of Play to further its research on games-based learning.

E-Line has a captive group of beta testers at Quest. “I just stood there with a notebook, writing down the students’ ideas about how to make Gamestar more interactive,” says Doyle. “They wanted powers. They wanted to fly. They wanted to customize their avatars. They came up with a list of about 15 ideas that I have to now e-mail to the Gamestar people.”

Although Osterweil sees a future for video games in classrooms, he doesn’t see them as a panacea. “If you had a really interesting open-ended, lab-type classroom,” he says, “you’d see a lot of play, experimentation, and game playing, and you wouldn’t need formal game playing to make it happen. But as schools become ever more bolted down, with kids tied to their desks, people who care about education are looking to other venues in which that learning can happen.”