Failure by Design: Encouraging learning through failure with games.

People generally hate to fail. Failure makes you feel like you’re not up to a certain standard, and is often seen as an endpoint. In videogames, however, failure is expected and at times even celebrated. Since the creation of Space War, videogames have been challenging players to revise their strategies and revisit their conceptual models in order to progress. Game scholar Jesper Juul theorizes that failure is actually directly related to enjoyment of a game. If a game is too easy, or resolves in a deterministic way, players feel cheated. This is not to say that players must fail to enjoy a game, but that games should allow for failure because it gives success meaning and increases the fun. This goes a long way toward explaining the popularity of ridiculously hard games like the previously discussed Super Hexagon, Dark Souls and even QWOP.

So why do players generally persevere when playing a game, but are more prone to giving up in other areas, like school? What might we learn from games if we want to foster this kind of perseverance in other contexts? And what happens if we don’t allow for failure?

Good games encourage players to take risks, explore, and try out new possibilities. In everyday life, the stakes are often too high to just try things out. For instance, one would never dream of building a rocketship for a manned space mission without any prior understanding of astrophysics. However, in games like the Kerbal Space program, players are allowed to do just that. Sure, most of the early attempts explode on the launch pad, but this failure doesn’t discourage players. On the contrary, it invites them to revise their mental model of the task at hand and try again in a process known as “recursive play.” Some players even play with the intention of creating the biggest explosion and learn a great deal about the system as a result. This revision process isn’t unique to games, but in a world where “failure is not an option” games are one of the few places that actively encourage it.

Recently, researcher Kevin Dunbar studied the scientific process by observing four of the world’s leading molecular biology labs. During this time, Dunbar found that over 50 percent of the experiments conducted resulted in an unexpected outcome requiring the research lab to revisit their process. These revisions resulted in the refinement of methods, better control of variables, a deeper understanding of the problem and innovation. It makes sense that new knowledge would spawn from behaviors and models that we don’t completely understand. Yet, we rarely read about these null results and failures because they are rarely published. Instead of being discouraged that our approach was incorrect, we should follow in the steps of historic figures such as Albert Einstein and adopt the maxim, “Failure is success in progress.” Science is built on a foundation of failure and, as Karl Popper said, we must subject our understanding of natural phenomena to rigorous tests of falsification in order to strengthen our understanding.

A typical Kerbal Space Program launch.

There is debate as to why players tolerate failure in games, but avoid it in other contexts like school. The most common of these is embodied in the saying “it’s just a game” which implies that the outcome doesn’t matter. Granted, this is a logical assumption to make, but I believe it’s too simplistic. Anyone who says failure in a game doesn’t matter should watch (or participate in) a heated Mario Kart race, a tense poker tournament or think back to the last time their favorite sports team met with defeat. It’s not that the outcome to a game doesn’t matter, rather that the impact of the outcome can be negotiated or attributed to the system, or circumstance. Games allow us to declare “the game is dumb” if we fail too often (like in the image above) and lets us retain a bit of social capital when we don’t perform well. This attribution of failure to something other than ourselves also creates a safe space to practice. While we’re learning, we are able to diminish the role our performance has on the outcome until we improve and can take credit for the success. This ability to negotiate the impact of failure is something we lack in other domains. We can’t just say schools are dumb even if their assessments are sometimes ludicrous.

Games also foster learning by the way they handle, and allow for, failure. Well-ordered problems serve to keep the player challenged while discouraging player attrition. Game designers take great care to order their problems so that they’re difficult, but not impossible for the player to overcome, allowing players to utilize knowledge learned in other contexts. A good example of this can be found in the first level of Super Mario Brothers. The entire first level was specifically designed to introduce players to all the actions they would need for the entire game. During the first few minutes, players are introduced to slow moving enemies, “?” blocks, pits and pipes. As players improve, more complex actions are introduced such as hitting a block and getting a power up mushroom, or using warp pipes to discover hidden areas. If a player dies at any point, the level is restarted and the player re-visits the level drawing on the experience of their previously failed attempt to move forward.

A New York Post headline rationalizing a USA failure in the world cup.

Problems that lie outside of a player’s level of proficiency also work as gatekeepers by making sure a player has mastered a specific skill or concept before moving on. If a player is allowed to progress past the point of proficiency, he or she may find themselves in a situation where they can no longer make progress, at which point they may become frustrated and quit. The similarities to Mihaly Csikszentmihalyi’s concept of “flow,” Seymour Papert’s idea of hard fun, or Lev Vygotsky’s zone of proximal development are not coincidental. Good games challenge players with tasks that are just outside of their level of expertise (and are therefore still interesting). In order to do so, the game has to have some sort of idea of the player’s skill. In place of a standard assessment, beating a boss at the end of a level gives some sort of evidence that the player is proficient with the new skills, and allows the designer to introduce more complex challenges later in the game. In a game with well-ordered problems, failure doesn’t stop players from playing. Rather, it allows them to become experts.

Failure is especially discouraged in the majority of schools. This isn’t too outlandish considering that standardized testing is a de facto measure of progress and often has very severe consequences for failure. These consequences include, but are not limited to labeling a student as deficient (and thereby reducing the number of opportunities available to them such as AP classes) and reducing the amount of funding a school receives if enough students fail. This in turn has driven many teachers to “teach to the test,” and many students to cheat in an attempt to avoid the consequences of failure. The fact that students avoid failure, which if treated differently could serve as an avenue to discovery, is quite troubling.

Failures played back during Super Meat-Boy

Even if there’s a chance for failure, there is something engrossing about working on an interesting difficult task. If we believe, as Popper did, that the creation of knowledge is an iterative process, we must allow for misconceptions and revisions. Games that challenge the player seem to give them some form of dignity. They do not assume the player won’t live up to the challenges — they expect them to rise to the occasion of understanding and executing complex commands or solving fiendishly difficult puzzles. The game Super Meat Boy does a wonderful job of showing players just how far they’ve progressed. Just like Super Mario Brothers, Super Meat Boy was designed to teach the player different aspects of the game as a series of well-ordered problems. However, Super Meat Boy is not as forgiving as Super Mario and the player often dies quickly, resulting in a faster feedback loop. Upon success, Super Meat Boy plays back all of your previous failures. Rather than this being an embarrassing experience, it’s actually quite rewarding to see how far along you’ve come. The screen becomes flooded with Meat Boys who are mercilessly crushed and pulverized until there is only one left — the last Super Meat Boy that you controlled and completed the level with. It’s the Super Meat Boy that overcame all the obstacles. You.

I talk games, code, and education on twitter. For educational game consulting please visit my website VideogameResearcher.com

This post was originally published on the Wisconsin Institute For Discovery website.