Evolving Game Design from Alchemy to Science
by By Alexander "DmGoober" Jhin

Sales Pitch

Do you ever wish you had a magical formula that allowed you to create the next Quake or Everquest? Better yet, how about if you had a scientific formula that allowed you not only to create the next hit, but also understand, fundamentally, why it was successful? This article presents some ideas how to approach game design from a scientific tilt that will help designers to truly understand why their games work.

Introduction

Currently, game design more closely resembles alchemy than chemistry. Like alchemists, game designers often use a "guess and check" approach when designing games. The designers take a "best guess" at preliminary game play, game art, and story, which are then "checked" by publishers, other designers, beta tests, and play tests. Based on feedback, the game is then tweaked until it seems "right" or time and money run out.

As in alchemy, this guess and check approach makes it difficult for the designer to nail down, at a fundamental level, what caused her game to succeed or fail. She knows what tweaks made the test audience happy, but she is left guessing as to why they worked.

The chemist, on the other hand, approaches her problem by first understanding proven principles and theories, then using these principles to achieve her goal. By using principles, the chemist knows that if her work fails she can research and understand where she violated chemical principles or find a way to modify the principles to explain her failure. Whether she succeeds or fails, she has an understanding why things turned out as they did.

For example, if an alchemist wanted to make gold out of base metals, she would try melting lead and silver together, then, if that approach seemed right, she would continue. A chemist, on the other hand, would understand that gold differs from other metals because of the differing number of protons and electrons. The chemist would then focus on trying to add extra protons and electrons into base metals. The chemist's approach is obviously favorable, as the chemist actually has some basis for her work. The alchemist is merely guessing.

So how does the designer act more like a chemist? Unfortunately, unlike chemists, designers do not have the resources to perform original research in order to discover and isolate underlying game design principles. So the designer must rely on appropriate, existing principles from other fields of study and apply them to game design. Many applicable fields of study are hundreds of years old and can offer tried and true principles to game design. By borrowing principles, game designers can build a concrete foundation of principles to work from and can allow designers to begin to create their own concrete principles specific to game design.

The Process of Liberal Borrowing

Using principles from other fields is a simple process of:

Stating the goal correctly.
Brainstorming fields of study that may offer insight into achieving the goal.
Choosing appropriate principles from those fields.
Checking principles against existing games and genres.
Implementing principles in game design.

This process creates principles that are useful both in game creation and game analysis.

Example 1: Making A "Realistic" Game

For example, assume that our goal is to "create a game that displays a realistic 3D environment." To state the goal correctly, it is important to rephrase the goal so that the goal reflects the interactive nature of games; games cannot really do anything without the gamer. Thus, every goal in game design should include the word "gamer" or "player." The goal could be rephrased as "to create a game that the gamer perceives as visually realistic and three dimensional on a flat computer monitor." This goal is bulkier but more accurately states what we are trying to do; we are trying to fool a human into perceiving three dimensions.

Brainstorming yields two obvious fields of study that can help us achieve our goal. The first field is art, specifically drawing and painting. For millennia, artists have studied how to achieve 3D on a 2D canvas. The second option is to examine principles from psychology, specifically visual perception principles.

Using psychological principles to achieve our goal is the better option as psychology can be considered a "lower-level science" than art. That is, psychology can explain why many artists' principles work. Psychology often yields principles that are useful to game design, as psychology deals with the perceptions and behaviors of humans. The goal of game design is, after all, to get the player to perceive and behave within a game environment.

The next step is to choose specific principles from psychology that are applicable. We open a high school introductory psychology textbook to the chapter on visual perception. The book states that there are seven depth cues:

Linear Perspective - As parallel lines recede, they appear to come together.
Aerial Perspective - Objects that are farther away appear hazier because as distance increases, so does the amount of interceding dust and other particles.
Texture Gradients - Objects that are farther away appear rougher and appear to have more detailed textures.
Interposition - Objects that obscure other objects appear closer.
Relative Size - Objects that are farther away look smaller.
Light and Shadow - Bright objects appear closer than dark objects. Light and from irregular surfaces also give depth cues.
Stereoscopic Vision - Human's left and right eyes receive slightly different images that allow the brain to perceive depth.

Ok, so our 3D game should attempt to have all or some of these features.

The last step is to check these principles against existing games. As it turns out, all of these principles have been used in 3D games at some time or another. In game design these principles have been implemented by use of viewing frustrum, fog, textures, z-buffer, viewing frustrum, Gouraud shading, and stereoscopic glasses respectively. While we have not discovered any "new" principles in 3D graphics, the power of this technique is two-fold: First, we know this list is "correct" as it has been "proven" by psychologists. Second, we really did not have to work hard at all - we just looked in a high school textbook, and voila, we had a features list to achieve our goal!

Actually implementing these features from scratch would require a lot of math, creativity, and coding. However, in implementing these features we can rest assured that we are not wasting our time - we are working with "proven" 3D depth cues.

Example 2: Making an Addictive Game

Now let us assume that we have a more exciting goal of, "Create a game that gets the player to play for as long as possible." Again, notice the goal includes the gamer. A quick brainstorm of fields of study that involve "getting someone to do something for as long as possible" yields: coercion or torture, marketing, and psychology of motivation or addiction. Psychology is the obvious choice, as psychology can explain why marketing and coercion work.

We open an introductory psychology textbook to the chapter on motivation and addiction. Two principles seem appropriate:

Operant Conditioning. In a nutshell, operant conditioning is the psychological principle that states that a person is motivated to do or not do an action based on whether they have been rewarded or punished for that action in the past. Operant conditioning principles also explain how to schedule rewards in order to maximize motivation to perform the action. These principles have been discussed in the context of games in other Gamasutra articles, and I refer you to them.
Maslow's Hierarchy of Needs. Maslow's theory dictates that people are first motivated to satisfy basic needs. When these are satisfied, people try to achieve higher and higher needs. The needs, according to Maslow and beginning with the most basic are: Physiological Needs such as hunger, Safety, Belonging and Love, Esteem and Approval, Self-Actualization.

We can use these two principles that have been "proven" by scientific methods, to both create addictive games and analyze the addictiveness of existing games. There are other psychological principles of motivation: chemical addiction, sexual motivation, and the Opponent Process theory may be applicable to games. However, I ignore them for brevity's sake. The Opponent Process Theory seems particularly applicable, as it deals with both success and failure and should be thought about further.

In order to briefly validate that these two principles apply to games, we need only look as far as Everquest and the Sims. Both unconsciously (or possibly consciously) utilize these principles to great effect. Everquest is well known for its addictive properties and its ability to get players to play for very long periods of time. I encourage the reader to attempt to list all of the uses of operant conditioning in EQ. A few examples are the use of variable schedule reinforcement in striking a blow, killing a monster, getting good treasure, finding good monsters, and skill crafts just to name a few. Maslow's Hierarchy also applies to EQ. Newbie players struggle for safety then as they become stronger, they begin to form friendships and look for belonging and people to party with. Finally, as the players begin to reach top levels, they begin to look for esteem and approval, often in the form of distinctive looking armor, weapons, and distinguished powers and levels. Thus players are allowed to advance up the hierarchy and are motivated to reach the top.

The Sims use of Maslow's Motivations is blatantly obvious. New characters struggle for physiological needs such as food, sleep, and bodily functions. As the players become richer, they can focus more on safety, then belonging and love, then esteem and approval; even if the love, esteem, and approval are only from other Sims. The use of operant conditioning is less obvious, but still present. Rewards take the form of home improvement, neat object animations, and friends. These rewards also follow a conditioning schedule, making game play addictive.

Implementing operant conditioning in a game in order to make it addictive is pretty easy. Simply ensure that your game doles out rewards following a variable interval/variable ratio schedule (look at Everquest, Diablo, or a slot machine in Vegas to see some successful examples of operant conditioning.) Implementing Maslow's Hierarchy in a game is a slightly more difficult, less tangible goal. Allowing players to express their "affection" to other players, allowing players to show team allegiance, and allowing players to track their own success (and ensuring that players eventually succeed) are all good ways of allowing players to feel that they are advancing up Maslow's Hierarchy, thus encouraging them to play further.

These two examples demonstrate the relative simplicity of generating principles that are both useful in a game design and game analysis.

Counter-Arguments

Unfortunately, many do not agree with this approach. First, some may argue, successful games have been designed without understanding underlying principles, thus why use them at all? Certainly, many game designers, such as Will Wright, seem to have an innate ability to design games without the use of established principles (at least I assume that Will Wright did use psychology textbooks in designing The Sims.) However, arguing against using principles simply because some people have succeeded without them is similar to arguing a return to alchemy, simply because some alchemists had some success (albeit in this case, Will Wright is one damned talented alchemist.) Chemists have almost completely replaced alchemists, despite alchemists' limited success, because the chemists' approach is simply more efficient and allows a better understanding of underlying issues. This is not to say that this approach to game design will ever replace the Will Wrights of the world. This approach will simply help the rest of us.

The second argument is that game design is an art not a science, therefore, established principles should not apply. Game design is an art - however, even arts have established principles. Painters use the principles of perspective to portray three dimensions, while most scriptwriters use general principles of conflict and climax in order to keep their plots moving. Plus, the principles borrowed from other fields will never rob game design of its creativity. If, for example, we tried to create a game that only used operant conditioning sans creativity, all we would have was a dispenser with a lever that popped out food pellets using a variable reinforcement schedule. Nobody would be willing to "play" this "game." No, great amounts of creativity must be applied to these principles in order to make them fun and enjoyable. However, these principles give a basis of understanding with which to work from.

The final argument is that game designers don't have time to muck around with psychological principles. First, the two examples that I used above took about fifteen minutes each to research - certainly designers have fifteen minutes to spare. Second, by understanding fundamental principles, designers can avoid some guesswork and therefore save time in the future. Finally, if game designers are willing to share the principles that they have discovered, then other game designers will not have to "muck around" with psychological principles. Rather, they will simply be able to read about them in the context of game design.

Conclusion

The process of borrowing principles from other fields is as simple as stating an objective, brainstorming applicable fields, choosing appropriate principles from those fields, checking those principles against existing games, and implementing.

Would you rather be an alchemist or a chemist?

Alexander Jhin was almost a psychology major at Yale University. Instead, he became a computer science major. He has worked on the Asheron's Call and Zone.com teams and is currently the head of a small, independent, international team. He will be working for Microsoft's DirectX in the coming year. He can be reached at alexander.jhin@yale.edu and he welcomes any comments, especially negative ones.

Discuss this article in the forums

Date this article was posted to GameDev.net: 3/20/2003
(Note that this date does not necessarily correspond to the date the article was written)

See Also:
Game Design