GDnD Wiki

GDnD Wiki

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

Source Summary: Breaking Conventions with The Legend of Zelda: Breath of the Wild (GDC 2017)

8 min read

Source metadata

  • Type: Conference talk (GDC 2017 panel, about 88 minutes), delivered in Japanese with consecutive English interpretation.
  • Title: Change and Constant: Breaking Conventions with The Legend of Zelda: Breath of the Wild
  • Speakers (Nintendo): Hidemaro Fujibayashi (game director), Takuhiro Dohta (technical director), Satoru Takizawa (art director).
  • Date and place: 1 March 2017, Game Developers Conference, San Francisco, two days before the game’s release.
  • Where to watch: official GDC upload on YouTube (https://www.youtube.com/watch?v=QyMsF31NdNc, uploaded 10 March 2017); GDC Vault session 1024562.
  • Cross-checked against: Daniel New, “Breaking series conventions with Breath of the Wild” and “Breath of the Wild, science, and clever little lies”, Thumbsticks, 2 March 2017; Tori Vaz, “Zelda Team discusses Breath of the Wild’s development and 2D Zelda prototype at GDC”, Zelda Universe, 9 March 2017. These agree with the transcript on the three chemistry rules, the elements/materials split, the use of Havok and the 2D prototype.

Key takeaways

The talk has three parts, one per speaker, framed around the question of which Zelda conventions to change and which to keep.

Fujibayashi: from a passive game to an active one (00:02 to 00:22).

  • The goal was “a game where the user can truly experience freedom in an expansive play field”, deciding for themselves “where they want to go and what they want to do”. He traced this back to the original NES Zelda, where every scroll of the screen offered a new discovery.
  • He named the conventions that blocked this: events that need a set path, walls you cannot get round because of level design, difficulty that assumes a predetermined order of experiences, and dungeon puzzle solutions that are all on the internet. These made earlier games passive: the player plays “within the confines of a pre-prepared mechanism”.
  • First experiment: make every wall climbable, so boundaries become optional paths, then add gliding so that “once you climb, you can fly to wherever you want”. He called gliding “the crucial ingredient” for freedom, and the simple multiplication of player action and field led him to the idea of multiplicative gameplay.
  • Earlier Zelda puzzles used natural phenomena and simple science, but each object was “made specifically for those puzzles”. That “development by addition” polishes well but costs a lot of resources and would not scale to the new world.
  • His hypothesis: “make the game so that objects react to the player’s action, and the object[s] themselves also influence each other.”
  • He tested it with a 2D, NES-style prototype (blue-clothed Link, a leaf, a log). It “does not include any puzzles”: only a river, some trees, and one rule: “There’s a situation and a goal. Can you reach it?” Multiplying varied actions, items, landscape and reactive objects under that rule produced “an active game … where countless different events occur for which the user can freely create solutions.”

Dohta: physics plus chemistry (00:23 to 00:57).

  • An action game’s smallest parts are collision, movement and object state. Collision plus movement is what games call physics.
  • Games need “game physics”, not textbook physics: “false physics or liar’s physics”, used to raise controllability and responsiveness, to meet the game design, to save processing and sometimes for a stronger sense of realism. Real-world laws still matter because they build “trust between the player and the game”; the clever lies sit on top of that trust.
  • Freedom meant near-infinite physics situations (you can roll the starting-area rock all the way to the final boss), so the team adopted Havok for its stability, robustness and track record, and spent their effort on the fun parts.
  • Examples of physics “lies”: Stasis (the frozen object is actively working to stay still; “the law of conservation of energy … except really that’s a lie”); Magnesis (“It looks like magnetism, but it isn’t”); climbing (a dynamic constraint between Link and the surface, so he can climb moving objects); tree cutting (the tree swaps to a log object as it falls, and logs float using a custom, made-up buoyancy).
  • Key observation: “If you unify your game world with a set of consistent physics rules, the player will come up with ideas for all sorts of things they want to try” (drop the bridge plate on an enemy, ride a floating log, climb a giant enemy).
  • The chemistry engine. “One way of describing a physics engine is as a rule-based movement calculator. In comparison, we decided the chemistry engine would be a rule-based state calculator.”
    • Elements are things without a constant solid form: fire, water, ice, and also electricity and wind (“in a textbook, the answer is of course no, but in the chemistry engine, you bet they are”). Apart from rigid-body dynamics, every natural phenomenon is treated as chemistry.
    • Materials are solid things: trees, rocks, “even the player character and weapons”.
    • Rule 1: elements can change the state of materials (the leaves on a tree burn away).
    • Rule 2: elements can change each other’s state (fire goes out when it meets water).
    • Rule 3: materials cannot influence each other’s states.
    • “Given the right situation, elements can also generate energy and impact the physical world”, which is where chemistry hands over to physics.
    • The aim was not to simulate chemistry but “a state calculator for all these objects that are connected to one another”. The team’s in-house name for multiplicative play was “chemical reaction play”, a deliberate double meaning.
  • The answer to Fujibayashi’s question was “to build connections between everything in the world”, designed “to have an overall consistency and to feel instinctive”. The 2D prototype was actually 3D underneath, running early versions of both engines.
  • Situations and goals fill the world and each has an intended solution, but players should enjoy “thinking up and trying solutions on their own”; roundabout solutions and unexpected shortcuts count as “the correct way to play”. The convention broken was to make “a world where fun occurs through multiplication, as opposed to one in which fun elements are simply added together”.
  • Side topic: the Switch port was easier than expected; no Joy-Con-specific features were added so both versions share one game design.

Takizawa: art that makes it easy to lie (00:58 to 01:25).

  • Zelda’s systems change every game, so the art team looks for an art style “that makes it easy to lie”: one that meets playability (readable, responsive, a pleasant game cycle) without losing a believable reality.
  • HD tests of older games picked out The Wind Waker HD style. Its weakness for this game: its “lies … were too big”. To let the art “intuitively suggest possible physics and chemistry game play based on the player’s own experiences in the real world”, the team needed more realism and information density. The slogan became “refreshing and full-flavoured”.
  • Fast “poof” abstractions (a felled log becoming bundled firewood, an animal becoming meat, automatic cooking) are acceptable because the art style keeps them feeling fun rather than absurd.

Notable claims

“One way of describing a physics engine is as a rule-based movement calculator. In comparison, we decided the chemistry engine would be a rule-based state calculator.” (Dohta, about 00:39)

“Rule 1. Elements can change the state of materials. … Rule 2. Elements are able to change each other’s states. … Rule 3. Materials cannot influence each other’s states.” (Dohta, about 00:40 to 00:41)

“It’s not that we wanted to create an engine that replicates chemistry. Instead, we wanted to create a state calculator for all these objects that are connected to one another.” (Dohta, about 00:43)

“Make the game so that objects react to the player’s action, and the object[s] themselves also influence each other.” (Fujibayashi, about 00:17)

“In this simple setup, there was only one rule. There’s a situation and a goal. Can you reach it?” (Fujibayashi, about 00:21)

“If you unify your game world with a set of consistent physics rules, the player will come up with ideas for all sorts of things they want to try.” (Dohta, about 00:35)

All quotations are the English interpreter’s words as transcribed, not the speakers’ own English.

Relevance

Open questions raised

  • The talk does not explain how designers kept the multiplied possibility space readable and fair (for example, how shrines were tested against unintended solutions). Later sources would be needed.
  • The talk names the rules but not the data model (how states, durations and propagation are stored), so any implementation detail beyond the three rules is inference.
  • How far does the approach depend on Havok-level physics? The 2D prototype suggests the chemistry idea works at very small scale, but the talk does not test that directly.

breath-of-the-wild | emergent-design | game-atoms | inherent-vs-emergent-complexity | systemic-depth-elegance | second-order-design | prototyping | player-agency | open-world | game-feel

Graph View

Backlinks