Issue 27 07.15.2011

Issue 27
Stop Putting the Hero’s Journey in a Skinner Box

What happens when you decompose a game? When you take a game and remove the goals, remove the structured levels, remove the explicit challenges, remove everything except the components that make it up?

Someone who makes games can tell you that get kind of a glorified toy box. Sometimes it’s fun to play in, sometimes it’s not.

It’s interesting how players and playtesters react when you do this. Without explicit or implicit goals, a unifying structure, or anything to unlock, most people think something like this: Hey, it’s maybe kind of neat that this guy can walk around, and bump into stuff, and you can put this thing on top of that thing, but this game is not ultimately fun enough to spend time or money on.

A lot of games have a sandbox mode or a level editor. These routinely get less attention from most players than the “actual game”.

Meanwhile, if you add goals and structure to your game, players will decide that it is worth playing.

Why is this?

Well, there are nicer and nastier reasons, and they depend on how the player is motivated.

On the one hand, we have players motivated by the challenge of the thing. Their goal with the game is to improve their skills, maybe show off to friends, maybe compete with other players. It’s the same stuff that motivates athletes. Players might learn valuable life skills: resource management, teamwork, timing jumps.

On the other hand, we have players motivated by the promise of later content. Their goal with the game is to see everything that the game has to offer, and games offer all kinds of stuff. Art, environments, characters, a story, music, the breadth and depth of the game mechanics themselves — save for the last, the kinds of things we find and value in other media.

If you’re motivated by challenge, goals and structure are a very good thing. A game can act like a coach, telling you what and how you need to improve. It can withhold complicated challenges until you’ve mastered the fundamentals. It can reward you for your especially hard work, providing extrinsic reward when the intrinsic reward isn’t strong enough. This kind of approach can be tremendously helpful when you’re making an educational game, especially one that teaches and measures a skill.

If you’re motivated by content, goals and structure are poisonous. You’re not playing the game because you enjoy the mechanics; you’re playing the game because you want to see the next thing. That is, of course, a simplification — most content-oriented games have at least somewhat compelling mechanics. But if they’re subsidiary to the content, you’ll become more frustrated and resentful with repetitive or difficult gameplay than if the mechanics were the point. Since the reward is extrinsic, you’re less willing to apply yourself to difficult tasks.

Moreover, the structure of intermittent, inconsistent reward (in the form of content) for your actions (in the form of gameplay) effectively turns the game into a Skinner box. Like casinos!

To create a game with entertaining content, you have to accept this tradeoff: if you want your players to work to earn the content, you can’t challenge them too much, or they’ll get frustrated along the way.

So players expect that a game will contain goals and structure, and there’s a temptation for developers to conform to this expectation.

But while this is appealing to players — for better or for worse — and therefore dominant in games, these conventions aren’t really necessitated by the medium. If you’ve been playing traditionally structured games for a long time, it’s hard to think of games without explicit goals or reward structures. But doing so can be instructive in making better games, and in extending the benefits of games into other domains.

Moreover, by omitting goals and reward structures, you can avoid the detrimental effects they have on many games:

An Interlude Leading Up To More Detrimental Effects that Goals and Reward Structures Have on Many Games

A common way of coming up with a full game is actually to start by making a toy — a gameplay prototype that doesn’t have goals or structure, whose sole purpose is to find a nugget of fun that can be expanded into a full game.

World of Goo and Crayon Physics Deluxe, for example, started as simple one-week prototypes. Nintendo has a trick that I will call an empty room test: before a protagonist is allowed to explore the game world at large — that is, before the developers are allowed to create the game world to begin with —, the protagonist has to be fun to play with while just running and messing around in an empty room. The idea is that if the core interaction of the game — building a tower of goo, drawing physical objects with a crayon, and navigating the environment, respectively — is fun, that fun will spread throughout the entire game. It’s a simple strategy, and it’s yielded a lot of success for a lot of developers. And it’s very easy to tell, as a player, whether it’s a process that was applied during development: do you ever find yourself messing around instead of making progress?

To add a goal and a challenge to a toy, in order to make it into a game, you need to decide what you want your player to get good at, and you need to find some way of measuring their success.

In Super Mario Bros., for example, the challenge is to master Mario’s running and jumping behavior and quirks, so that you’re able to navigate a 2D environment, touching the things you want to touch (coins, mushrooms), and avoiding the ones you don’t (fire, mushrooms).

The Mario development team measured this by designing a sequence of levels — if you get past this level, it’s because you were able to jump over this, and under this, and avoid this turtle.

In Tetris, the challenge is to get really fast at assembling a 2D puzzle. Inventor Alexey Pajitnov measured this by recording your progress one horizontal line at a time.

Kyle Gabler designed World of Goo to reward players for using as few goo balls as possible in building structures as fast as possible. The game doesn’t let you win levels unless a certain number of goo balls are left after you’ve finished constructing, and you get a little badge for winning levels using even fewer goo balls.

Often, as a game maker, you’ll come up with a toy that you can’t easily turn into a game for one reason or another.

Maybe you can’t because it’s hard to measure player success.

For example, in a game based on dancing, drawing, or any other kind of creative activity, the computer can’t really be programmed to assess aesthetic value. So those games either aren’t graded at all, or their grading is based on your ability to recreate something that already exists — in Dance Dance Revolution, you have to follow a pre-existing dance, and in Picross, you have to follow the numbers to uncover an existing picture.

Maybe you can’t because it’s hard to determine what constitutes player success.

For example, teddy bears and rattles and books and travel vacations and model airplanes are awesome. (We can agree on at least a few of those, right?) But if you try to grade them, you end up just marginalizing the breadth of value of the thing. If I get really inspired by this book, whereas it taught you a lot on a topic you didn’t know much about, who won? If I love building model planes and you love flying them, can’t that be enough?

You don’t need to be good at these things in order to enjoy them. But a lot of games do require that you become good at them in order to see everything they offer.

Why don’t most games allow players to simply look at everything in the game? Why don’t they let players fly through the game world, watch all cutscenes at the very beginning, play with the weapons that you’re only supposed to unlock at the very end?

Because limiting players in this way — making them work to attain cool things — makes them appreciate their unlockables more once they get to them.

But it also wastes players’ time and effort. You know how games have a reputation for that? Well, that’s why.

Most games today have a linear, do-this-to-get-this structure. And it’s not always a bad thing! It’s useful for telling players a story. It’s useful for training players to gradually get better at in-game skills. It’s useful for avoiding overwhelming players with too many options for what to do at any given time.


If you make a large game nowadays, most of your players won’t see most of your game, because they don’t have the time or interest to invest in beating everything. After all, some games take an entire workweek to complete! Even a short, brilliant game like Braid has only a 30% completion rate. Which is a crying shame, given that it has a terrific ending.

Blocking players off from content is probably the most frustrating thing games do. If a game offers a challenge with no reward, and a player is unable to beat it, the player only blames themselves, and decides, perhaps, that meeting the challenge isn’t worth their time and effort. But if the game prevents the player from seeing the rest of the game because he or she hasn’t done this really hard or boring thing yet, the they will become aggravated that they can’t actually see the things they played the game to see.

The technology we have available to us lets us make amazing things. The psychological frameworks we’ve become accustomed to let us keep players artificially engaged with what we make. But this comes at the expense of them actually enjoying it as much, and often it comes at the expense of them playing it at all.

These psychological frameworks shouldn’t go away for good, because they’re very useful in some circumstances. But in the circumstances where they’re detrimental, it’s worth thinking about alternative ways to achieve their benefits, and make our games, our industry, and our players a little bit healthier.

By Kelsey Higham