GDnD Wiki

GDnD Wiki

This site is for Ulster University students and staff. Enter the class access password to continue.

Player Agency

7 min read

Summary

Player agency is the player’s capacity to make meaningful choices that affect the game world. It is the interactive core of games — the feedback loop between player intention and game response. Steven Poole argues that player control is not merely a feature of videogames but their essential defining quality: the moment interactivity is removed, the experience degrades into something less effective than other media. Jesse Schell frames this as the loop of interaction at the heart of all gameplay.

(Poole, Trigger Happy; Schell, The Art of Game Design, see source-trigger-happy, source-art-of-game-design)

Key ideas

  • Control as essence: Poole’s central claim is that what defines a videogame — what separates it from film, television, or theatre — is that you are doing, not watching. Without the ability to act, a game’s core value is destroyed.

  • Quality of agency vs. quantity of inputs: More available inputs do not necessarily mean greater agency. Poole’s “combo paradox” demonstrates that large libraries of preset moves can actually reduce player agency by funnelling players into predetermined sequences, rather than allowing genuinely open tactical decision-making.

  • Tactical depth from simple systems: Poole uses Robotron: 2084 as his canonical example. Its control scheme — one joystick to move, one to fire — sounds trivially simple. Yet the combination produces vast tactical possibilities: feints, misdirection, prioritisation of targets, spatial management. Compare to Tekken’s hundreds of button-sequence combos, which prescribe outcomes rather than enabling them.

  • The interactivity threshold: The pod-racing sequence in Star Wars Episode I illustrates what happens when interactive promise is removed. The cinematic version, no matter how polished, is “far inferior in terms of high-speed thrills” to the playable game version — because you cannot act, you can only observe.

  • Agency and immersion: Player agency is also the mechanism of immersion. Players become invested in a game world partly because their actions matter within it — decisions have consequences, and those consequences feel owned.

Evidence

“You’re not watching, you’re doing. And when videogames are at their best, what you’re doing is something vastly more creatively challenging than watching a docusoap or a quiz show. Your reasoning, reflexes and imagination are tested to exhilarating limits.” — Poole (Ch. 10, source-trigger-happy)

“Robotron gives you two joysticks: one to move, one to fire. Simple. But with those tools, there is a huge tactical potential of feints, misdirections and apocalyptic vengeance.” — Poole (Ch. 2, source-trigger-happy)

“The pod-racing sequence was nothing more than an extended advert for the actual videogame that was based on it. You couldn’t play the movie, so it was far inferior in terms of high-speed thrills.” — Poole (Ch. 4, source-trigger-happy)

Schell frames the same idea as the loop of interaction: the player perceives the game state → forms a goal → decides on an action → performs it → the game responds → repeat. The speed and quality of this loop directly determine how responsive and satisfying the game feels. (Schell, The Art of Game Design, see source-art-of-game-design)

The combo complexity paradox

Poole’s critique of fighting-game complexity is worth examining in detail because it counters the intuition that “more options = more freedom”:

SystemInput complexityTactical depthAgency type
Robotron (two joysticks)Very lowVery highOpen — player constructs outcomes
Tekken (combo sequences)Very highLower per-actionPrescriptive — player executes preset outcomes

The issue is not complexity per se but whether the input system prescribes or enables. Combo systems typically map specific inputs to specific fixed outcomes; the player’s task is execution and recall rather than open tactical construction. A simpler system that gives players genuine decision-making latitude produces richer agency.

This critique predates modern fighting game design evolution — see Open questions below.

In practice

For game designers, Poole’s argument implies:

  • Test with agency removed: If your game’s most compelling sequences work just as well as a video to watch, that is a warning sign. What is the game delivering that a film cannot?
  • Prefer enabling systems over prescriptive ones: Design mechanics that create possibility spaces rather than scripted move lists. Ask: does this mechanic let players invent responses, or does it require them to recall correct ones?
  • Interactivity should justify itself: Every barrier to player action (cutscenes, forced walks, unskippable dialogue) should be weighed against what it costs in agency. Use them when necessary, not by default.

Types of control (CRE133 taxonomy)

The CRE133 lectures introduce a taxonomy of three control types that shape how agency is experienced:

Control typeDefinitionEffect on agencyExamples
DirectReal-time interaction with immediate, responsive feedbackEnables kinesthetic skill development; agency felt in the bodySuper Mario Bros., first-person shooters
IndirectSetting parameters or decisions executed by AI/systems; delayed outcomeAgency felt through long-term consequences rather than moment-to-momentCivilization, SimCity
ScriptedPredefined sequences triggered by player input; limited interactivityCan enhance drama; reduces agency if overusedQuick-time events, dialogue trees

The type of control fundamentally shapes where and when agency is experienced. Indirect control can deliver as much agency as direct control — but the experience is temporally and cognitively different. Mismatching control type with player expectations (e.g. embedding many scripted sequences in a direct-control game) undermines the agency contract.

Types of agency (CRE133 taxonomy)

A more granular taxonomy of agency by domain — which dimension of the game world the player can meaningfully influence:

Narrative agency — influencing the story and outcome.

  • Branching storylines, moral choices, dialogue with real consequences
  • Enhances emotional engagement; drives replayability through unexplored paths
  • Examples: Mass Effect (dialogue choices), Undertale (actions determine ending)

Mechanical agency — influencing the game’s systems and mechanics.

  • Different approaches to objectives; character and equipment customisation
  • Makes the world feel responsive to the player’s identity and strategy
  • Examples: Zelda: Breath of the Wild (open approaches), Skyrim (build customisation)

Social agency — influencing other characters or players.

  • Forming relationships with NPCs; alliances, trades, rivalries in multiplayer
  • Adds emotional depth through social consequences
  • Examples: The Sims (NPC relationships), World of Warcraft (alliance formation)

Economic agency — influencing the game’s resource systems.

  • Resource allocation decisions; market participation
  • Adds strategic depth; creates sense of real consequence in the game world
  • Examples: Stardew Valley (trading decisions), EVE Online (player-driven economy)

Many games provide one or two types of agency while neglecting others. A game with deep mechanical agency but no narrative agency can feel purposeless; a game with narrative agency but no mechanical agency can feel like the player is being driven rather than driving.

For the full treatment of how agency connects to presence, immersion, and flow, see presence-and-immersion. (CRE133 Lectures, see source-cre133-lectures)

Open questions

  • Poole’s argument is challenged by the walking simulator genre (Dear Esther, Firewatch, What Remains of Edith Finch). These deliberately minimise or remove traditional agency. Are they games? Does their limited interactivity make them “inferior” to film, as Poole’s framework would imply — or do they demonstrate that his control-centric definition is too narrow?
  • Modern fighting games (Street Fighter 6, Guilty Gear Strive) have evolved input systems significantly since Poole wrote in 2000 — including optional simplified controls. Does his critique still apply?
  • Agency exists on a spectrum. Is there a useful taxonomy of agency types (narrative agency, tactical agency, expressive agency) that would let designers reason about which kinds of control matter in which contexts?

Graph View

Backlinks