Summary
Systemic depth and elegance are two related properties that distinguish great games from merely functional ones. Sellers (Advanced Game Design, 2018) treats them as measurable design targets — not aesthetic preferences — that emerge from how well the designer has structured the system’s hierarchy.
Depth is the number of meaningful levels of organisation a player can explore. Elegance is depth achieved with minimal rules, exceptions, and cognitive overhead. Together they produce the experience Nolan Bushnell identified: easy to learn, difficult to master.
(Sellers, Advanced Game Design, see source-advanced-game-design)
Systemic depth
A system has depth when its parts are themselves systems — when there are multiple levels of organisation the player can zoom into and out of, each revealing new complexity:
“A system can be said to have depth when its parts exist at multiple levels of organization — when they are themselves subsystems composed of lower-level parts interacting together.” — Sellers, Ch. 2
The analogy Sellers uses: quarks → protons → atoms → molecules → water drops. At each level, a different set of properties and behaviours becomes visible. The depth is real and explorable.
In games, depth works analogously. In Go, placing a stone is a simple action — but that action exists within the context of a local group, which exists within a regional territory, which exists within the whole board. A novice sees only the single stone. An expert sees all four levels simultaneously and can reason at each.
Why depth is valuable:
- It lets players build a mental model over time, rewarded at each new level discovered — “like opening a present to find another present inside” (Sellers)
- It creates enormous variability in gameplay from systemic rather than content-based design — the designer has built a space rather than a path
- It supports mastery: there is always more to learn, so the game stays engaging far longer than its rules complexity would suggest
Elegance
Elegance is the quality that makes depth accessible. An elegant game has hierarchical depth, but the player is not overwhelmed by it — each level of the hierarchy is discoverable, comprehensible, and feels consonant with levels already learned.
Sellers identifies five properties of elegant games:
| Property | Description |
|---|---|
| Metastable familiarity | The game changes each play but retains an overarching, consistent feel — the player can keep returning without the experience feeling alien |
| Simple rules, deep hierarchy | High-level systems are defined with few rules, but those rules cascade into rich, complex behaviour as the player descends the hierarchy |
| Self-similarity across levels | Each lower-level system reflects the structure of the higher-level one. Players feel they “almost already know” what they see for the first time — scaffolding from the familiar to the new |
| Few exceptions | Rules exceptions and special cases ruin the mental symmetry of hierarchical systems — they demand memory rather than comprehension |
| Metacognitive appreciation | Once a player has thoroughly internalised the hierarchy, they can reflect on its structure and appreciate its design — satisfying even outside play |
“Elegance of this degree is rarely attained. It requires a masterful comprehension of the game systems by the designer, who must apprehend them all at once, as if they were laid out, while at the same time seeing them in linear form as the players experience them.” — Sellers, Ch. 2
Bushnell’s Law
Bushnell’s Law (attributed to Atari founder Nolan Bushnell) states that a great game should be:
“Easy to learn, difficult to master.”
In systems terms, this requires:
- Low cognitive floor — the top-level rules are simple; the player can form a valid starting mental model quickly without exceptions or special cases
- High skill ceiling — the hierarchical depth means mastery takes a long time; even advanced players continue to discover new dimensions of the system
Go is the canonical example:
- Rules state: place stones on intersections; capture surrounded groups; most territory wins
- A competent player can be taught in minutes
- Professional players study for decades and still find new strategies
Chess and Tetris have similar properties. Contrast with games that have apparent complexity — many rules, many exceptions — but shallow depth: they are hard to learn and easy to exhaust.
Bushnell’s Law and player retention: The “difficult to master” half is a retention mechanism. Once the player has seen all the game’s content and all the systemic interactions and all the viable strategies, there is nothing left to discover, and the game loses its pull. Bushnell’s Law is effectively a requirement for long-term engagement.
Depth vs. complexity
A common misconception: more rules = more depth. Sellers explicitly rejects this.
| Property | With depth | With mere complexity |
|---|---|---|
| Learning curve | Gradual; each new concept scaffolds from what the player already knows | Steep; many rules must be learned before play is possible |
| Skill ceiling | Very high — mastery is never total | Often low — once rules are memorised, there is no deeper level |
| Exception handling | Few or none | Many special cases; constant reference to rules |
| Emergent behaviour | Rich; behaviour arises from the interaction of few rules | Limited; behaviour is mostly prescribed by the rule set |
More rules add complexity (more to memorise) without necessarily adding depth (more levels to explore). Systemic depth requires well-structured hierarchical loops, not a longer rulebook.
In practice
When reviewing a design for depth:
- Can the player zoom into the system and find a meaningful sub-system to engage with?
- Can the player zoom out and see how that sub-system functions as a part in a higher-level system?
- How many meaningful levels of zoom exist?
When reviewing for elegance:
- How many exceptions or special cases exist at each level?
- Do lower-level rules resemble the higher-level rules, or are they disconnected?
- Can a player new to a given level recognise familiar structure, even if the details are new?
Second-order design and depth: Sellers connects depth directly to second-order-design — a game with hierarchical depth creates a vast state-space (many possible paths through the game) rather than a narrow scripted path. The player’s choices at different levels of the hierarchy interact to produce unique play experiences each session.
Open questions
- Bushnell’s Law describes an ideal, but many successful games have a high floor and a low ceiling (casual mobile games). Is Bushnell’s Law a universal principle, or a principle for a particular audience?
- Self-similarity across levels is identified as an elegance property, but it is also a risk: if the sub-systems are too similar, they may feel repetitive. Where is the line?
- Go and Chess approach this ideal through mathematical abstraction. Can narrative, character-driven games achieve comparable depth and elegance through systemic means?
Related
- systems-thinking — Depth and elegance emerge from well-structured systems
- second-order-design — Depth creates a state-space for the player to explore
- game-loops — Deep games have nested loops at multiple timescales
- fun-as-learning — Koster’s chunking/grokking model explains why depth sustains fun: there is always a new level of pattern to learn
- flow — High skill ceiling maintains the flow channel far longer than shallow games
- game-balance — Depth without balance produces dominant strategies that collapse the decision space
- mda-framework — Depth operates at the mechanics level; elegance manifests at the dynamics and aesthetics levels
- inherent-vs-emergent-complexity — Burgun’s analytic tool: depth is high emergent complexity from low inherent complexity
- burgun-taxonomy — Burgun’s broader framework; minimalism and elegance are central to his design philosophy
- source-advanced-game-design
- source-game-design-theory