State-action Visualization Systems

State-action transition type visualizations are more concerned with how players make decisions given the game’s state. This is often useful because it allows designers to understand players’ strategies. Designers can also use such analysis to understand how to tune difficulty or to determine how to best balance or pace the game content and skills. Furthermore, such analysis can also shed light on dominant strategies found by players or other unexpected solutions to game obstacles.

Representing states and actions

The game state is the virtual context, which would typically include the factors inside the game that directly or indirectly relate to the actions that players make. The definition of game state uses the metaphor of state machines, where a player performs an action, and the game calculates a response. These responses trigger an update to the game state (the state describes the global state of the game), and the player is then confronted with new choices to perform a new action. We can envision playing a game as a cyclical process of players performing actions, followed by updates to the game state.

Visualizing decision-making process

To visualize and understand players’ decision-making processes through the game state, one must define players’ actions and the game state, as well as identify how the game state changes as the player make decisions within the game. Such information is not usually recorded within the log file but can be computed from the log data. For example, by knowing the players’ position, we can deduce the spatial context, such as how many team members are around them, how many enemies are around them, and whether they are hidden by trees or other spatial elements in the level.