Systems, Game Systems, and Systemic Games

Systems are a big deal — they’re much more pervasive and important than games. But, for reasons I’ll get to later, I believe that games, and game design in particular, are a uniquely effective avenue for understanding what systems are and why they’re important.

A full discussion of what systems are is beyond the scope of one blog post. But I want to lay out some basics as the foundation for some later posts. This may be a bit abstract at first, but I’ll bring it back to games in a way that I believe illuminates important parts of game design.

All systems have a few things in common: a system is made of parts that interact to form a purposeful whole. There’s a lot packed in there, so let’s take that bit by bit.

Each ‘part’ in a system can be thought of an object in software terms, with its own state (defined by attributes) and behavior. Importantly, this definition is recursive: each part is itself a whole, a sub-system, with its own internal workings: a part’s attributes are themselves objects with their own states, attributes, and behaviors. This recursive nature is vital to understand and can also be a little dizzying. We often want to cling to (or design to!) a single level of interactions without any additional structure. This leads to a poor understanding of systems and how the world works, and to shallow, lackluster gameplay. In terms of thinking about systems, just remember that whenever you’re looking at a system, there are parts within it to be unpacked — and the system is itself just a part of a larger system.

The behavior of an object in a system may change its own state, and it may communicate with other objects, potentially changing their state and behavior. This communication is the interaction between parts mentioned above. While objects interact, not all objects in a system interact equally often. As is often seen in any network, there are dense interactions between some objects, and far fewer between others. This has the effect of forming hierarchies: dense connections of interactions between one set of objects creates, implicitly or explicitly, a boundary that defines a super-object at the next level up in the hierarchy.

The result of this is that two or more sub-systems (with their own internal sub-systems) interact in ways that create a holistic system and drive the whole system forward with purposeful behavior. “Forward” might be a bacterium traveling along a sugar gradient, a football team traveling down-field, a school of fish evading a predator, or a game player working toward domination of a game world. In each of those, the “system” is a complex whole made up of parts acting together: the bacterium, the football team, the school of fish, and the combination of game and player.

That last example brings up an important point: while we often talk about combat systems, skill systems, etc., in games, we have to remember that the overall system that creates the experience of playing a game involves at least two important sub-systems: the game, and the player (or players!). Each is a complex sub-system with its own internal objects, states, and behaviors; and each interacts with the other to create the overall system of “the game being played.” The game systems are sub-systems within the game that interact together and with the player as the game is played.

This brings us to the last part of the definition I gave above, that of “a purposeful whole.” Systems are often spoken of as having a purpose or teleology: fish avoiding a predator, a football team moving the ball downfield, etc. In a similar way, games are designed to create a particular kind of experience through the use of their internal sub-systems, but it is the system overall, the combination of game and player, that bring that experience to life. That experience is the purpose, the meaning, of any game.

So what is a game system? And how does this relate to what we might call systemic games?

Here I’m going to turn to the MDA framework (or alternatively, its closely parallel cousin the Function-Behavior-Structure ontological framework) as scaffolding for talking about game systems.

In these terms, the mechanics (or structure) of a game system are its objects and their attributes. These are, for example, the pieces on the chessboard or the weapons in an RPG and, just as important, their current state (location on the board, amount of durability) and their behavior (how each chess piece moves, or how a rapier, longbow, or broadsword attack).

The dynamics or behavior of a game system are the events and actions that emerge as the result of game-based objects interacting (though remember that dynamics at one level create the mechanics of the aggregated parts at the next system-level up). As one example, “dribbling” was not found in the original rules of basketball, and didn’t appear as an acknowledged part of the game until several years after its introduction. Players could pass the ball to others to move it down court, and they eventually discovered they could also pass the ball to themselves by bouncing it on the floor as they advanced, passing it back to their own hand. Dribbling is a dynamic behavior that emerges from the mechanics in the stated rules of the game. Creating similar dynamics, as by combining the roles of tank, DPS, and healer characters in many MMOs, is an important part of game systems design. If the game objects don’t create any new, significant combined effects — if each object works pretty much on its own with little value gained by interacting with others — then the game system (if it’s a system at all!) is not likely to engage players for long.

Finally, the aesthetics or function of a system refer to the system’s purpose, or in game terms, the desired experience for the players: the game may engender feelings of achievement, fast-paced action, of hypnotic flow, terror, love, companionship, loss, wonder, etc. This experience requires the interactions of multiple game sub-system and the game and player as sub-systems within an overall system of the game experience.

In these terms then, a game system has well-defined objects with their own (typically recursively defined) attributes, state, and behaviors. Those objects have to interact in ways sufficient to create new aggregate behaviors (the dynamics) that support the creation of a particular game experience. The object interactions aren’t random; they’re carefully constructed to create dependencies between the objects in ways that evoke meaningful dynamics in their combined use. If an object isn’t used in combination with others; or if a mechanic doesn’t support some larger-scope game behavior; or if the in-game behaviors don’t create a cohesive experience for the player — then your system is failing. It lacks purpose (aka teleology, function, aesthetic), and you need to work on the pieces that make it up and how they interact.

It’s worth noting that it can be incredibly difficult to keep this all in your head (or even in your design docs!) at the same time — being able to go from “we have these mechanics to create these dynamics in support of this player experience” taxes the abilities of almost any designer. In my experience, different designers gravitate toward one part of that — usually mechanics or aesthetics — at the expense of others. Some designers are all about the “nouns and verbs” of a design, but don’t really have a solid, intuitive feeling for what kind of game experience the nouns and verbs will create. Others know exactly the feeling, the experience, they want to devise, but are fuzzy on exactly what that means in terms of the specific underlying mechanics that will get them there.

The necessity of being a game designer who can span this range was summed up nicely by Paul Stephanouk (@gamegeek) who said that a systems designer should be able “to turn a game into a spreadsheet and a spreadsheet into a game.”

Okay, given all that, what about systemic games? These are games that depend on their internal sub-systems to generate repeatable, engaging gameplay for the player(s). The most common of these currently are those often called “roguelikes” due to the fact that they procedurally re-generate the game world each time, providing at least some amount of novelty and engagement.

Compare that with other games (often high-end AAA console games) that depend on carefully crafted, often beautiful, and generally expensive set piece levels, characters, cut-scenes, and/or scnearios to create gameplay. These present the same gameplay the to the player each time, with at most minor variations in play. There may be a “combat system” for example, but this often isn’t really much of a system: there’s one sniper rifle that beats the others, one shotgun to use close, a machine gun that does lots of damage, etc. In many role-playing games, players quickly gravitate toward the one best “character build” for making the most effective tank, DPS, healer, etc. And in many strategy games there are at most two or three well-honed ways of playing to win, and all others are doomed to failure. In games designed like this there are game objects but few interactions that support different and effective strategies for playing the game. There are choices for the player, but they quickly separate into “effective” and “ineffective,” with the latter leaving the player frustrated (why have a choice in the game if it only leads to sub-optimal gameplay?). Or there might be endless combinations of objects (as in Borderlands’ or Diablo’s weapons), but these only sometimes create interesting (game-relevant) dynamics, as the interactions between random weapons-parts are too broad and not systemically deep.

Another example that I think sometimes trips up game designers are designer-systems that don’t result in game-systems. For example, World of Warcraft has a complicated and effective quest-creation system — if you’re one of the designers. If you’re a player, the result is much like the set pieces described above: the same quests always contain to the same obstacles and rewards, and lead to the same follow-on quests. As a player these are great for awhile, until you realize that at level 90 you’re still doing essentially the same thing you did at level 1, but the rats you have to fight to bring back ten of their tails are just bigger and meaner. There is a system in there for the designers that helps them create a complicated (but not complex) network of quests. For the players, there are no quest dynamics, as the interactions are all frozen at design time. There are pre-defined quest lines, but no interactive quest system.

Systemic games, by contrast, rely on interactions between game objects, from the setting to available weapons or tools. These are generated new each time the game is played based on underlying algorithms and systemic interactions. This necessarily means the game designers can’t depend on expensive, static set pieces, and so must lean more heavily on the dynamics created by the interactions between game objects and their mechanics.  It also means the designers have to create their actual game system — the system that, via its sub-systems, creates the game-to-be-played each time — in such a way that it will provide both systemic variability and maintain a satisfying desired aesthetic each time.

As difficult as this is, when it works, it works extremely well, creating highly replayable games with great longevity. That leads us to the concept of game depth — but that will have to wait for another (hopefully shorter!) post.

[This post originally appeared on Mike Sellers' blog.]