Enforced Agent Behavior

From gdp3
Revision as of 12:00, 20 March 2018 by Staffan Björk (Talk | contribs)

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search

The enforcement of certain actions in order to maintain or develop an agent’s personality.

Games that have developed characters may have rules to ensure that these characters behavior is consistent to their personality, or that the personality develops in a certain direction. For those characters that are under players' control this may require limiting what actions they can make the characters perform, or taking control way from them.


The dialogue choices available in games such as the Mass Effect series or the Witcher series gives players' some degree of freedom while at the same time guaranteeing that the responses chosen are in line with the personality of the players' characters.

The Thief series uses goals to enforce certain behaviors. Players that try to complete the game will make their characters steal valuable items since this is dictated by mandatory goals. However, they are further encouraged to do similar actions through optional goals, and are encouraged to not behave in certain other ways (e.g. killing guards) by the lack of game rewards for engaging in these types of activities.

The roleplaying game Call of Cthulhu has rules for how player characters can become insane. When insanity checks are failed, if the character sees some horrific monsters or heinous acts, the character, for example, freezes or flees despite the player’s will. The suggested Mind Module[1] shows a way where gameplay can emerge from providing a computer-facilitated model of psychological state of player characters.

In Space Alert players decide which actions their characters should perform in one phase of the game, and all the actions are then performed in a later phase without a possibility for players to change them (with a partial exception of being able to correct wrongly selected actions).

Using the pattern

The initial concern with making use of Enforced Agent Behavior is to determine what behavior the Agent should be. Typically this may be to provide Thematic Consistency (including Thematically Consistent Dialogues) or maintain Predetermined Story Structures (especially concerning Character Development). However, it may also be used to make it impossible to actively act against team mates in games with Teams.

The easiest way of making sure of Enforced Agent Behavior is to only make those actions which are wanted possible, which comes down to defining a Limited Set of Actions. This set of actions may be enough to sustain the intended behavior if the only interesting part of the personalities one wishes to present is noticeable from actions uninformed by previous events. In these cases the state of the Agents may only need to expressed through use of Avatars, NPCs, or Units. However, for more complex behaviors the development of Characters to keep track of internal states is typically necessary.

There is a clear difference between achieving Enforced Character Behavior depending on if the Agents in question are under players' control or not. Algorithmic Agents, Dedicated Game Facilitators, and Game Masters are natural choices for guaranteeing Enforced Character Behavior when players are not intended to be directly involved in choosing or performing the actions, i.e. for controlling the behavior of NPCs. However, Zero-Player Games and those using No Direct Player Influence also achieve this as the behaviors are set earlier (typically through Action Programming from a Limited Set of Actions) and then enforced.

Regarding Agents controlled by players, AI Players is an easy solution but does not of course work for all players except when used in Zero-Player Games. Looking beyond AI Players, two strategies present themselves in how to ensure Enforced Character Behavior: making players perform the actions or taking control away from them so that the game system can perform the actions. Game Masters provide the most flexible solution to this, being able to apply specific solutions and switch between the two as circumstances dictate. Another option is Player Augmentations, which if done skillfully, can make players feel that they performed the actions although it was the augmentation that made it done at the right time and in the right way.

Goals that are both Predefined and Enforced force players to perform certain specific actions. However, they do so without guaranteeing the specific behavior in all the actions leading up to the ones that complete the goals. A Selectable Set of Goals provides some Freedom of Choice while still guaranteeing at least one action will occur. Supporting Goals and Optional Goals encourage behaviors but do not actually enforce them unless they are Committed Goals with related Penalties. Since they are related to player choices and Characters, the use of Player-Created Characters and Player-Planned Development are ways of achieving this, especially when combined with Diegetic Social Norms and Trait Regulated Behavior.

Cutscenes take away player control and thereby can provide Enforced Character Behavior as well as the perspectives to be used when presenting the scenes to the players. By doing so they can ensure Predetermined Story Structures but at least temporarily remove any Exaggerated Perception of Influence. Automated Responses can in contrast provide Enforced Character Behavior in specific case during gameplay.

Being constructed from rules game designers have chosen, Algorithmic Agents (and AI Players) by definition have Enforced Agent Behavior, but the pattern can also be used to give them consider specific Enforced Agent Behavior as compared to other Agents.

Diegetic Aspects

Enforced Agent Behavior is, regardless of if it is applied to Agents controlled by players or not, essential for making actions have Thematic Consistency.

Narration Aspects

As stated above, Enforced Character Behavior can guarantee Predetermined Story Structures, either be making sure events take place when they should or by making sure that certain events do not take place.


Since Enforced Agent Behavior can ensure that specific actions and events take place, they can ensure Character Defining Actions of Characters. By doing so, Character Development can be steered so that Predetermined Story Structures are maintained and told as planned. Enforced Agent Behavior can also make certain that some action do not occur, e.g. that Characters do not kill each other and thereby finish Internal Rivalry structures prematurely.

Enforced Agent Behavior can help Thematic Consistency since it can ensure that Agents behave in accordance to what is expected within a Game World. It can also help modulate Player/Character Skill Composites by giving Characters agency but this also makes the combinations into Player/System Action Composites. However, the use of Enforced Agent Behavior typically makes it difficult to have Exaggerated Perception of Influence and Freedom of Choice, especially when player control is explicitly removed through use of Cutscenes. Even when games provide some Freedom of Choice in regards to Enforced Agent Behavior, it can conflict with, rather than modulate, Roleplaying if the character design is not transferred consistently to the gameplay. Player may also actively recent or oppose Enforced Agent Behavior if they created the Characters through Player-Designed Characters.

For games with Characters, Enforced Agent Behavior can also interfere with the possibility of having Player-Created Characters and Player-Planned Development. The exception to this is when it is a self-inflicted effect due to choices players have done regarding their Characters and their intended development.


Can Instantiate

Predetermined Story Structures, Thematic Consistency, Thematic Consistent Dialogues, Teams

with Characters

Character Defining Actions

with Player/Character Skill Composites

Player/System Action Composites

Can Modulate

Agents, AI Players, Algorithmic Agents, Character Development, Game Worlds, Internal Rivalry, Player/Character Skill Composites, Roleplaying

Can Be Instantiated By

AI Players, Algorithmic Agents, Action Programming, Automated Responses, Characters, Committed Goals, Cutscenes, Dedicated Game Facilitators, Diegetic Social Norms, Game Masters, Limited Set of Actions, Enforced Goals, No Direct Player Influence, Player Augmentations, Player-Created Characters, Player-Planned Development, Predefined Goals, Selectable Set of Goals, Trait Regulated Behavior, Zero-Player Games

Can Be Modulated By


Possible Closure Effects


Potentially Conflicting With

Exaggerated Perception of Influence, Freedom of Choice, Player-Designed Characters, Player-Planned Development, Roleplaying


An reworked pattern based upon Enforced Character Behavior, first introduced in Lankoski 2010[2].


  1. M. P. Eladhari (2009). Characterising Action Potential in Virtual Game Worlds applied with the Mind Module. Ph.D. thesis, Teesside University, UK.
  2. Lankoski (2010). Character-Driven Game Design - A Design Approach and Its Foundations in Character Engagement. D.A. thesis at Aalto University. Publication Series of the School of Art and Design A 101.