GDnD Wiki

GDnD Wiki

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

Burgun's Taxonomy of Interactive Systems

7 min read

Burgun’s Taxonomy of Interactive Systems

Summary

Keith Burgun proposes a strict, prescriptive taxonomy in Game Design Theory (2012) that defines four types of interactive systems arranged as nested subsets. His central argument is that the word game is overloaded — covering radically different types of software — and that this conceptual confusion has damaged design quality across the industry. By distinguishing games precisely from puzzles, contests, and toys, Burgun argues designers can make more principled decisions about what they are actually building (Burgun, Game Design Theory, see source-game-design-theory).

Key ideas

The nested hierarchy

All four types are interactive systems — “possibility spaces defined by explicit rules.” Each successive category adds a new defining property:

Interactive Systems (Toys)
└── + Problem/Goal → Puzzles
    └── + Competition → Contests
        └── + Ambiguous Decisions → Games

Toys / Basic interactive systems

  • Have rules but no defined goal or win condition.
  • Examples: Minecraft (before player adds goals), SimCity, Garry’s Mod, Lego.
  • Often called games culturally, but Burgun argues there is no contest and therefore no game. Players who impose their own win conditions on Minecraft are acting as game designers, not players.

Puzzles

  • Add a problem with a solution.
  • A puzzle is solved or not solved — there is no winning or losing against another party.
  • After a puzzle is solved it “stops having value to a person, much like a riddle whose answer is already known.”
  • Examples: Portal (individual chambers), dungeon puzzles in Zelda, most point-and-click adventure games.
  • Caveat: single-player video games with no random elements degenerate into puzzles regardless of intent. Super Mario Bros. levels are effectively puzzles on repeat play.

Contests

  • Add competition: a solution that cannot be shared — once one agent achieves it, others lose.
  • Winning and losing distinguish contests from puzzles.
  • Decisions may exist but are not meaningfully ambiguous — they reduce to execution, dexterity, luck, or memorisation.
  • Examples: weight-lifting contests, hot-dog eating contests, Guitar Hero, most pure racing games, Dance Dance Revolution, Candy Land (luck contest).

Games

  • Add ambiguous decisions: choices whose consequences ripple through the whole system and whose optimal move cannot be determined with certainty, even in retrospect.
  • “Even if you win, you cannot say for sure that the decisions you made were the correct ones.”
  • Decisions are endogenously meaningful — they have effects inside the system that propagate outward. Contrast with thematic decisions (which only affect the narrative layer, not the system itself).
  • Examples: Tetris, Team Fortress 2, chess, Go, NetHack, football, poker.

Ambiguous decisions — the defining property

Burgun treats the ambiguous decision as the most important and fragile property of games:

  • A decision is only meaningful if multiple options are genuinely in play. If one option is clearly optimal, there is no real decision.
  • The ambiguity must be endogenous — produced by the system’s rules and state, not by narrative or thematic framing.
  • This is what distinguishes games from contests: a push-up contest may involve micro-decisions (breathing technique, pacing), but these are not endogenously meaningful because they do not ripple through the game-state in unpredictable ways.

Games are fragile

Because the ambiguous-decision property is “surprisingly elusive,” games are inherently unstable:

  • They can degrade downward into puzzles when they become solvable (tic-tac-toe for adults; Tetris when the 7-Bag and hold-box remove uncertainty).
  • They can degrade into contests when execution requirements dominate over strategic choice (many RTS games, fighting games with complex input sequences).
  • They can become too random — if outcome is dominated by chance, player agency is severed and the game becomes a contest of luck.

Burgun describes this as games needing to “dance upon the threshold of the known and the unknown. They must live at the border between what we can understand, and what we cannot.”

The thematic / mechanical distinction

Burgun insists that theme (narrative, art, music, character) is a layer placed atop mechanisms, not part of the game itself. The game is the ruleset. This matters because:

  • Thematic decisions (choosing a dialogue option, watching a character die) may feel meaningful but have no endogenous mechanical effect. These are false choices.
  • Designers who prioritise theme over mechanism build systems that are, mechanically, contests or puzzles dressed in game clothing.
  • A player experiencing emotional meaning in Final Fantasy from a character death is responding to the theme, not the game system — the same image on a poster could produce the same reaction.

In practice

Burgun’s taxonomy is a design-time diagnostic tool, not a consumer label. When designing, ask:

  1. Is this a toy? Have I defined win conditions and competition? If not, I am building a toy or simulator, not a game.
  2. Is this a puzzle? If there is no competition and the system has a discoverable optimal solution, it is a puzzle. Puzzles have their own value — but if I intend a game, I need randomness or a human opponent to prevent solvability.
  3. Is this a contest? If the primary challenge is execution, memorisation, or luck rather than genuinely ambiguous strategic choice, I have built a contest. Contests are not worse than games — but they are different and should be designed differently.
  4. Am I protecting the ambiguous decision? Features that remove uncertainty (quicksave, hold-boxes, auto-aim, rubberbanding) can degrade a game into a puzzle or toy.

In Unity/C# terms: when designing a single-player game, ensure there is a source of genuine uncertainty that the player cannot memorise away — typically via procedural-generation or randomised parameters. Without this, the system will become a puzzle after sufficient repetition.

Evidence

Burgun opens his book with the case of Tetris (Burgun, Game Design Theory, Introduction, see source-game-design-theory): the original game was a genuine game because of risk-management decisions under uncertainty (the random piece generator, the choice to play safe or push for a tetris). Modern Tetris — with 7-Bag generation, hold boxes, and multiple next-piece previews — has eliminated the uncertainty. Modern Tetris is “more of an execution and reaction contest — almost akin to a rhythm challenge like Dance Dance Revolution.”

On Minecraft: “When they [players] do this [add win conditions], they make a game out of Minecraft in the same way that one makes a game out of a flight simulator or Legos. In these cases, a player has actually taken on the role of game designer.” (Ch. 1)

On the value of games: “The primary and direct value that games have for us is that they teach us how to learn. They provide an environment for us to focus on increasing a specific skill or set of skills.” (Ch. 1)

Implications

  • Designers who conflate games, puzzles, and toys will apply the wrong design criteria. A puzzle does not need balance or competition; a game does.
  • Genre labels (platformer, RPG, FPS) often obscure what type of system a game actually is. Spelunky is a game; Super Mario Bros. is closer to a puzzle (no randomness, memorisable). Both are called platformers.
  • The formalist position implies that story and game are structurally in tension: stories are linear sequences, games are branching webs. See narrative-design for alternative views.
  • Burgun’s taxonomy is more restrictive than Schell’s or Adams’/Poole’s definitions — compare game-definition.

Open questions

  • Burgun admits subjectivity: a game that is solved for one player may still be a genuine game for another. Does this undermine the taxonomy as a design tool, or does it simply mean the taxonomy applies to the system and not to individual player experience?
  • Competitive RTS and FPS games generate genuine ambiguous decisions under real-time pressure. Does Burgun’s critique of execution elements apply equally to all genres, or mainly to single-player contexts?
  • His taxonomy has been criticised for being too restrictive — excluding World of Warcraft, cooperative games, and experience-focused works. See mda-framework and game-definition for broader approaches.

Graph View

Backlinks