Today I want to talk about Gamestate Volatility, and specifically how it relates to obstacles and enemies and the idea of agency. How much can obstacles, enemies, and non-player-characters, in the ways they act upon the world, interact with each other and react to the player, change the overall gamestate. How much “agency” do they have over their situation.
Now, when I use the word “agency,” we have to acknowledge that enemies (and other non-player actors) in games don’t have any true agency per-se. Their actions are either completely deterministic or have elements of randomness. So, at the most basic level, when we ask how much agency an enemy or NPC in a game has, what we really mean is what is the range of potential impact they can have on the gamestate, and how much they can react to the game state.
So a low agency enemy would be an enemy that sticks to a pattern, or reacts in very simplistic, predictable ways that don’t have a lot of potential for emergence. A high-agency enemy would be one that acts based on numerous sources of environmental feedback, and can effect the environment in multiple ways, which could then in turn lead different feedback, and different reactions, and so on.
For an example of very low-agency enemies, let’s take a look at the Megaman games. Enemies in those games tend to be very tightly deterministic, and don’t have a very wide scope of influence. The Mets are a perfect example. They are unable to move. Their only abilities are to pop out of their hiding place and shoot in a predefined pattern when the player comes within a specified distance.
The narrow scope of these enemies allows the designers to create very tightly controlled encounters. The levels can be built around making these obstacles pose a threat. In fact, they must, since obstacles placed in a position where they don’t pose a threat have no way of adapting to their situation or reversing their fortune. They will sit their ineffectually.
Building levels around low-agency enemies is how you create action-puzzles. The designer places the obstacles in such a way that the player has to come up with a plan as to how to get past them. All the elements are predictable enough that you can plan based on their reactions, and placed in such a way that just running at them and attacking won’t be effective. And because of the limited agency of the obstacles, one the player moves past the obstacle’s threatened area, there’s not really anything it can do about it. And if the player stands in a location which isn’t threatened, they can stay there indefinitely without fear of danger.
On the other side of the spectrum would be a strategy game, where your opponent ostensibly has the same level of agency as you, the player. They can build settlements, research technologies, attack your cities, and even win the game.
Let’s look at an in-between case. In Bioshock, enemies are able to roam around the level. In addition to that, they are able to start fights with the Big Daddy, or attempt to destroy security devices which you’ve hacked. The elements of the game are not only reacting to the player, but also interacting with each other. They are not as predictable as the enemies in Megaman, but they don’t have the same level of agency as the player. They can’t leave their area, and they certainly couldn’t beat the game instead of you.
However, their increased scope, both in terms of movement and methods of interaction with other elements of the game, means that the game-state can change independently of player action. A safe location could become dangerous due to enemies wandering in, and situations could arise that the player could take advantage of (for example, enemies fighting the Big Daddy could make it easier to take the Big Daddy down).
Now, an interesting case is the stealth game. Many stealth games have enemies who have very low agency some of the time, and higher agency other times. Specifically, if the enemy has not been alerted to the player’s presence in any way, it will walk on a predefined patrol route. However, they have a whole host of actions available to them when some sign of the player appears. They can investigate, they can call in friends, they can pull an alarm, they can investigate hiding places, they can move freely, so on and so forth.
Many stealth games have been reenforcing the barriers between these states, a prime example being Mark of the Ninja. They’ll have hiding places where it’s impossible for the player to be spotted by an un-alert guard, meaning that the game-state is completely stable as long as the player doesn’t act. They’ll give the player easy ways to see the enemies around them and notice their patrol patterns so they can find the safe route through. However there is a large amount of potential volatility if the player does something to disrupt the balance, such as getting spotted.
Then on the other hand, there will be a whole host of combat, escape and distraction techniques available to the player which help them to survive the volatile alert phase of the game. The thing is, these techniques are essentially useless if the player is playing “optimally.” Why would the player opt to make the gamestate more volatile when the non-volatile state can not only be “solved,” but solved in the way that the game tells you is best (ie. “ghosting” the level)? Even something as simple as throwing a pebble to distract the guard breaks the carefully designed-to-be-beaten pattern, makes the guard slightly more aware, and thus raises the volatility of the gamestate. These tools are ostensibly there to give the player options when they make a mistake, but for many players trying to get the best ranking on the level, the best tool at their disposal in this regard is to let the enemies kill them so that they’ll restart at the last checkpoint, which sets the volatility back to its base state, and allows the player to attempt to solve the puzzle the “correct” way this time. The potential volatility of the system remains hidden.
How much volatility you give your system will depend on the type of experience you’re trying to give the player. Do you want a series of carefully crafted challenges with puzzle-like solutions? do you want constantly shifting situations that the player has to react and adapt to, and which are, in turn, reacting and adapting to the player? Or do you want something in between? The important thing to think about is how not just the player, but the game itself, can drive emergence. How the elements of your game, especially the elements that act against or independently of the player, contribute to the overall volatility.