The phrase “game mechanics” sends a pleasant shiver down my spine. At the heart of every game are these mysterious whirring clicking mechanisms that deliver to the player pleasure and thrills.
We use them, we build them, but I’ve never seen a good unified definition of game mechanics that gives us a practical base upon which to build great games. Here is one. It is clobbered together from a variety of influences though many of you will recognize some central tenets from ‘A Theory of Fun’ by Raph Koster.
Game mechanics are rule based systems / simulations that facilitate and encourage a user to explore and learn the properties of their possibility space through the use of feedback mechanisms.
It is a simple definition, but it offers a good amount of insight into why games work and how we can make them better.
Central to the model is the concept of feedback loops that encourage learning. Here is a diagram that should explain the concept in a more visual format:
- Player performs an action.
- The action causes an effect within the simulated game world. The simulation contains public and private tokens and the causal rules that affect the states of the tokens. The player rarely knows all the rules and is highly unlikely to be able to instantly describe the complete possibility space described by the rules. The unknown portion of the simulation is a “black box” that the player must attempt to decipher.
- The player receives feedback.
- With new tools and information in hand, the player performs another action. Using what we’ve learned, we pursue additional pleasure.
Linking game mechanics to create a system of systems
Interconnected networks of game mechanics make up the game as a whole. You can think of the game as a set of interlinked of puzzles where solutions to one puzzle lead to clues that help on additional puzzles.
The info treats that a game provides to the user need not be used to solve the immediate black box at hand. Humans horde potentially useful information like squirrels horde nuts for the winter time. We’ll store hints in our copious long term memory in the hope that there will be another black box down the line that will yield to our improved tool chest of knowledge.
The traditional metagame that sits on top of a game’s core mechanics is a good example of how one black box feeds into another. In this situation, the game mechanics are arrange in a temporal hierarchy where rapid feedback loops (often part of the basic control scheme) provide tools that enable the mastery of longer term feedback loops. The potential patterns of linking game mechanics together are nearly endless. This is a wonderful area of future study.
Humans are infovores
Humans are wired to solve black boxes. It is a fundamental aspect of our neurological learning wetware. We get real chemical rewards when we grok a problem or gain information that we suspect will help in grokking a black box. Evolution has selected for this behavior over thousands of generations since it is the biological reward system that encourages tool use and technological adoption. Without this built in addiction to problem solving, we would lack agriculture, medicine, architecture and other fundamental survival techniques that make the human species such a remarkably successful animal.
A key aspect of our model is that games actively encourage learning. I can put a black box on the table with a hidden button. Unbeknownst to a potential user, pressing the button enough times and the black box will spew out a thousand shiny silver coins. This is not a game. This is a bizarre gizmo.
To turn it into a game, a game designer would need to do several things.
- Encourage Discovery: First, make it obvious that the button in meant to be pushed. Humans are naturally curious creatures, but as game designers, we need to explicitly direct them to take certain actions.
- Encourage Exploration: Second, the designer would put a counter on the front of the machines that lets the user know that their actions are having some impact on the system. The counter provides delightful drips of feedback and it is up to the user to interpret that feedback
- Provide Tool Mastery: Third, the designer would post a note “Payout: 1,000, coins!” Not all games need explicit winning conditions, but hinting at future utility is a highly useful technique for encourage the player to begin interacting with a particular game mechanic.
We’ve turned a gizmo into a simple game of chance. The difference between the two is that our primitive 1-armed bandit is explicitly designed to encourage player learning.
Existing games are richly laden with techniques that encourage learning. A few that come immediately to mind:
- Levels take complex systems and encourage players to explore and master one aspect of the possibility space at a time
- The use of scores, coin collecting and experience points are all simple feedback mechanisms that let the user know they are making progress towards some future state.
- The classic “See the treasure chest you can’t reach” in Zelda acts as a promise of future utility.
A system alone is not a game. A dump of information is not a game. A system that encourages learning through strong feedback mechanisms is a game.
I’ve just described the foundation of a game mechanic. Now lets dig into several of the secondary effects that immediate appear when you attempt to put this system into practice:
- Red herrings
- Human factors
Burnout: A definition
After merrily harvesting tidbits of information by plunking coins into the virtual pachinko machine, the player will eventually grok the system. The game mechanisms may still serve up information, but the tidbits are not longer as tempting. The info we receive has no resonance with problems that we are solving or problems we have solved. It activates no curious networks in the brain. We begin subconsciously filtering out the feedback from these mechanisms. Burnout is a state of completed learning where the player finally figures out that a particular action no longer yields meaningful results.
In Monkeyball, researchers were astounded to find the the biggest jolt of pleasured occurred when you fell off a cliff and died. People loved it! If you look at falling off the cliff as a huge learning experience, this makes perfect sense. However, when they replayed the animation, people hated it. Same stimulus, radically different response. The animation of falling off cliff lost its ability to teach the second time around. Ultimately, users are subconsciously constantly asking the question “Is this activity worth my time? Does it gain me anything useful?”
There are multiple paths that learning can take and not all are ones that game designers desire. We would like to imagine that groking a system results in complete and utter mastery of that system. In reality, ‘grokking’ means that that the user has stabilized on a mental model of the system they no longer feel like improving further. This model can be simple or complex, depending on the inclinations of the user.
- A complex model of Black Jack might take into account probabilities of cards appearing based off what has already been played.
- A simple model of Black Jack might conclude that cards appear pretty much randomly. There is more depth for the user to explore, but if they are a casual player, saying it is random is ‘good enough’ to judge the game.
A big frustration to game designers is that many users settle on a very simplistic model of how a particular game mechanic works. Players will claim that a game is unfair or too difficult and immediately toss it in a rubbish bin because the designer misjudged their reaction to a game mechanic.
Some mechanisms have highly predictable burnout rates. Most players immediately figure out that watching a cutscene again isn’t going to provide much additional information. Other mechanisms demonstrate a large variation in burnout rates depending on the person who is playing the game and their personal preferences and disposition towards addiction. Some players try a slot machine once and then never again. Others will ruin their lives in pursuit of the next reward, never grokking the simple truth that such machines exist to take money, not give.
The factors that influence burnout are numerous.
- Personal history.
- Practical importance of imagined future rewards that stem from mastery.
- The ability for the mechanism to signal that there is additional depth of mastery possible.
The first two factors are not possible to derive by simply exercising your superior intellect. A deep understanding of your target audience’s psychology is most helpful here. The second two factors are very much under the designer’s control and can be refined through heavy prototyping and player observation.
Milking: The transition from learning to tool use
The flip side of burnout is grinding. If burnout is when a player discards a game mechanism because it is no longer useful, milking is when a player continues to exercise a game mechanic long after they’ve reached the state of mastery because the game mechanics continues to provide value.
When a player has learned one system, they will often keep interacting with it. On first blush, this seems mildly demented. The activity no longer provides burst of juicy learning. It is a bit like jawing on a piece of gum that long ago lost its flavor.
However, remember that games are networks of linked game mechanics. Player will continue to interact with a mastered game system in order to create a useful game state for exploring another black box. Mastery gives the player predictable pragmatic tools that helps them advance in other aspects of the game. The learning and mastery that occurs in other portions of the game provide the necessary reward that goads the player into revisiting old game mechanics.
You can extend the time that a player spends with a set of a game mechanics by ensuring that a mastered system still provides utility to the player. Designs techniques that build tools result in more gameplay for less development work.
Red Herrings: Black boxes external the game
The network of blackboxes that the player considers valid can extend far beyond the systems in the game itself. Often, the player will collect strange bits of info that have no real impact on the game mechanics that the game designer built into the game. These pieces rattle around in our heads like a collection of oddball keys for a set of locks that we may never find.
Game designers can tease the player with hints to systems that do not exist in order to suggest depth to their games. A sly arched eyebrow in a cutscene triggers as massive cascade of meaning alerts. Our brains love people and faces and relationships and the breeding opportunities and politics! Surely, that eyebrow is important? The player greedily stores the memory away.
What impact will the collected information have on their gameplay? None. What impact will it have on their lives? Very little. This virtual person in a cut scene is no one they will ever meet. But our brains were not evolved to deal with such things. As apes, the tale of an arched eyebrow by a potential mate from our little tribe always meant something very, very important. So our brain rewards us with a little jolt of pleasure for noticing such an “obviously” beneficial tidbit.
The designer managed to suggest a system and get some of the benefits of that system without actually building it. It is not going too far to suggest that paintings, sculpture, movies and television all thrive on this simple quirk of our brain’s learning systems.
The downside is that such red herrings burnout quickly. Our brains becomes quite good at recognizing false, useless information. Almost no one watches a cut scene more than once. What would be the point?
My personal bias is to use red herring game mechanics sparingly. As game designers, we have deeper skills at our disposal. We can tailor potent electronic cascades of feedback loops that spin out a complex duet between computer and the player. Such system are highly effective at causing visceral pleasure and encouraging deep long term learning. As game designers, we conduct a majestic symphony of explicit learning and entrancing interactivity, something no static media will ever manage.
Sometimes though, it is worthwhile to suggest great mysteries with broad brush strokes. Setting, character design and plot can be crucial hooks that help make a game meaningful to players before they even press a single button.
Human factors: Emphasizing the humanity of games
Some folks read about models and immediately see them as reductionist mechanisms that strip the humanity out of the soul out of creating artistic games. The game mechanics I’ve described in this article attempts to avoid this trap. They explicitly include social, narrative and emotional elements in addition to purely analytical problems. All aspects of the human experience, that have an impact on our ability to process and learn from stimuli, fall within the domain of potential game play.
This definition of game design is much broader than the current range of games available on the market. Though it works quite well with hit points, button mashing and high scores, the breadth of the definition is intended to encourage exploration of a much wider range of human learning. Some open questions that I find immediately suggested by the model include:
- What are the feedback mechanisms that impact learning about relationships, love, hate or spirituality?
- How do we build games around such topics that leverage these feedback mechanisms?
Existing games give us the foundation of practical knowledge that lets us make the same thing in a reliable fashion. A good theoretical framework helps game designers create future titles that are inclusive of a wider range of human experience.
The goal of any model of game design worth its salt is that it both explains existing behavior and predicts future behavior of medium. In my experience so far, this model seem rather robust at explaining almost any existing game on the market ranging from board games to slot machines to social games. There is certainly room for improvement, but it is a good enough for my main goal.
I want a practical model that lets the good folks in this grand industry describe game designs in more exacting terms. The model should give insight into why their prototypes suck. It should allow them to discuss potential issues and solutions with shorthand language that cuts to the meat of the matter. A good predictive model allows for more intelligent design decisions with less waste and unnecessary rework.
So some of aspects of the model that I find useful:
- It treats game mechanics as well defined, comprehensive atomic units. These units can be discussed individually and they can also be linked together in interesting ways.
- Explicit identification of user value. Fun is not a nigh spiritual activity that spontaneously bursts forth from the ether. It has a testable neurological basis.
- There exist clearly inputs and outputs that easily identified. You can easily tell when a specific game mechanic has all component elements such as actions, rules, tokens and feedback systems. Through observation, you can identify the player’s reaction to each mechanism and then adjust its impact.
All and all, the hope is that this model of game mechanics is a good foundation for future discussion. It is one that I’ll be leaning on heavily as I continue to meander through this lovely little series of essays on game design.
The pleasure of killing monkeys
See research lesson #1. I don’t agree with their conclusion about what causes the reported result, but I find the data fascinating.
A theory of fun for game design: Raph Koster
Many of the basic concepts in this essay build upon the ideas in this book. I find it helps my thinking to rework what I’ve read in essay form. Call it a form of active listening if you must. (
A slightly different definition of feedback loops that comes from control theory.
Games are designer foods for infovores
Other loose ends
This essay became too long and started budding little essays. Some have been planted in new documents that may one day emerge in full blossom. The rest are here for your reading pleasure.
Is a book a game? With this big emphasis on learning, there is bound to be a wiseass who asks “Is a text book a game? It too encourages learning.” The problem here is that there are few strong feedback mechanisms evident. The user reads the book and without a doubt they get a burst of pleasure from ingesting the info. However, the act of turning the pages, and interpreting language are skills mastered through other activities ages early. At best, reading the book is an example of milking, where a player uses a mastered technique to advance the grokking of some larger blackbox.
The primary role of content. In this model of game mechanics, content in the game is meaningful only through it’s association with a feedback mechanism. Plot points become reward and hints, Damage becomes a punishment that clues that player into the fact they shouldn’t be doing something. There is no such thing as an inherently pretty picture that exists ‘just because.’ The image is pretty because it activates the brain’s learning systems which in turn feed back into actions.
In order to answer the question “what content does my game need?” you need to first answer the question “What feedback should my game mechanics provide to the user based on their actions?”