I’m a new member of the lab here, and that means that I’ve got a lot of learning to do. I need to learn about the different projects in the lab, learn about the various systems involved in those projects, and even about programming languages used in those systems. But I also need to learn about the theory that drives those systems, and more broadly, the theory that motivates the work in the lab. So for the past few days, I’ve been reading articles–and even a short book–about the theory of fun in games.
Why is this important to the lab? Here at EIS, we build AI systems that work with games. Many of them are focused on building a particular type of experience, or enabling developers to do something new. And since we’re building game experiences and tools to assist in their creation, we need to be able to talk and think seriously about what makes games good or bad. If you build an AI system that enables some new mode of interaction in a game, it’s often not enough to say that it’s new; you usually want to make a claim that it’s good, and that involves either user testing, or some theory that you can point to to claim that it’s probably good. Additionally, before you even build the system, you want to be able to design it according to principles that will help ensure that the finished product is good. And so we study the theory of fun in games in order to better understand the medium with which we work, and the ways in which we can expand and contribute to that medium.
These are my thoughts about fun in games, distilled from the reading that I’ve been doing. First, fun isn’t something that games are. Fun is something that people have (often while playing games). To put it technically, fun is a property of an experience (which is what occurs during a particular instance of gameplay). When we say that a game is fun, we’re making a generalization about the kinds of experiences it helps us have. The reasons that a game helps us have fun experiences are not always directly under the control of the game creator, so it’s important to recognize this. As game creators, we need to think about what aspects of the game experience we can control, and what aspects we cannot control. When these uncontrolled aspects get in the way of fun, can we adjust the controlled aspects to compensate for this? Can we even detect the uncontrolled aspects? A good example of this is player behavior in multiplayer games. The game creator doesn’t have direct control over player behavior, and player behavior is important for whether or not the game is fun. But the game creator can exercise limited influence over player behavior through psychological cues, and can give players control over their collective behavior to try to ensure a positive experience. Things like the ability to mute other players or vote to kick players in first person shooters are an example of this. The game creator can’t prevent players from behaving in a way that annoys other players, but ey can give the players tools to help mitigate this behavior. So the first step towards an understanding of fun in games is to recognize that the fun isn’t in the game, it’s in the experience of the game, and because of that, game designers should pay attention to all aspects of that experience, including ones that they cannot control directly.
Given this understanding of fun, one might ask: “What makes game experiences fun?” But I think that there’s another question that should be answered (or at least considered) first: “What does it mean to be fun?” To be more precise: is fun a monolithic concept, and we classify experiences as simply fun or not, or is it a group of related concepts, meaning that things can be fun for different reasons or in different ways? My answer to this question is that to some degree, both descriptions are valid. If you ask someone whether something was fun or not, they can often give a simple yes or no answer in the broad sense, and even though a more detailed analysis is possible, this simple, broad, yes-or-no sense of fun is still valid. In fact, I think that sometimes it’s not useful to break down why an experience was fun, because the reasons can be extremely complex and specific. On the other side of things, there are multiple ways in which an experience can be fun. Things can be fun because they’re challenging, or because they’re beautiful, or for many other reasons. There are a few of these reasons that feature prominently in most games, and those reasons are what the literature focuses on. But I believe that it’s important to also acknowledge the role of other, more minor varieties of fun, because each way in which a game can be fun has the potential to become the main variety of fun in a game that targets it. The “observational immersionist” style of games mentioned in the previous post is a good example of this: in many games, part of the fun lies in exploration and appreciation of beauty. But when this kind of fun becomes the focus of the game, the result is a game that feels very different than many mainstream games, which often focus heavily on challenge as a source of fun.
So using this idea of many varieties of fun, we can ask what the sources of fun are. How do we create the different types of fun? This distinction between a kind of fun and a cause of fun is often subtle, and in some cases perhaps unnecessary, but I think that it’s important to keep the two separate. For a simple example, we can identify challenge as a particular kind of fun. A more thorough analysis would perhaps be more specific, but we can say that in general, people have fun through experiencing and overcoming challenges. This is a kind of fun. How do you create it? You create challenges and let the player experience them. So challenging situations create “challenge fun”. Seems pretty obvious, right? But it’s important to note that there are some constraints on the challenging situation: in very broad terms, it can’t be too simple or too difficult (Of course, this means different things for different people, which is a reason for thinking of fun as a property of an experience rather than a property of a game). More subtly, there are any number of situations that could count as challenges, and produce challenge fun. Maybe the game itself is quite easy, but a community of players has taken to playing it without looking at the screen. In this case, the experience may have the property of “challenge fun”, even though the challenge is not a part of the game: it’s a part of the particular game-experience. So the separation of “types of fun” from their causes is useful in order to think clearly about the goal: the creation of fun experiences (of course, other goals are possible, but that’s perhaps a subject for another blog post).
I’ve rambled on a bit about fun and games, and I haven’t really addressed the meat of the problem in detail, but I think I’ve given an overview of one way to think about and look for fun in games. The hard work, which is actually classifying the different types of fun and their causes, is the subject for more than a blog post; in fact, it’s the subject of several books and research papers that I’ve been reading through. But the important points are that fun arises from experiences, not directly from games, and that there are many types and causes of fun, some of which have not been explored much in the context of games.
For anyone curious about what I’ve been reading, here’s the list of what I’ve read to get an introduction to this area:
- “Why We Play Games: Four Keys to More Emotion Without Story” by Nicole Lazzaro.
- “GameFlow: A Model for Evalucating Player Enjoyment in Games” by Penelope Sweetser and Peta Wyeth.
- “An Experiment in Automatic Game Design” by Julian Togelius and Jürgen Schmidhuber.
- A Theory of Fun for Game Design by Raph Koster.
One other thing that I’ve not yet read but am interested in is Mihaly Csikszentmihalyi’s Flow: The Psychology of Optimal Experience. It’s not targeted at games, and in fact looks at fun from a psychological perspective, but it’s cited by most of what I’ve read so far, and is the product of some very thorough research.
About the author: Peter is a 2nd year PhD student interested in most of what goes on in the lab. He's done some work with StarCraft and level generation, and is working on joint generation of levels with stories right now.