GDnD Wiki

GDnD Wiki

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

Game Loops

12 min read

Summary

Game loops are the cyclic interaction patterns between players and games that drive ongoing engagement. Sellers (Advanced Game Design, 2018) identifies four principal loops that exist in any game system, and provides a vocabulary for designing and analysing them. The concept of “core loop” — what the player does most of the time — is the most practitioner-relevant, but understanding all four gives designers insight into how engagement is structured at different timescales.

(Sellers, Advanced Game Design, see source-advanced-game-design)

The four principal loops

Every game system contains four principal loops, each at a different level of the hierarchy:

LoopDescriptionWho is the active agent?
Game’s internal loopThe game’s own simulation running — AI, physics, economy tickingThe game itself
Player’s mental loopThe player forming a mental model of the game, making decisions, updating understandingThe player’s cognition
Interactive loopThe reciprocating exchange: player acts → game responds → player perceives → player actsPlayer and game together
Designer’s loopThe designer observing the game+player system and adjusting the game modelThe designer

The first three loops operate during play. The designer’s loop operates during development — but it is the most powerful, as it shapes all the others.

The interactive loop (core gameplay)

The interactive loop is the essence of interactivity:

“This give-and-take between the player and the game is often referred to as the game’s core loop.” — Sellers, Ch. 4

The cycle: player forms an intent → acts within the game → game state changes → game provides feedback → player perceives and updates mental model → player forms next intent.

This cycle must:

  • Complete quickly enough that the player feels in control (see game-feel)
  • Provide feedback that is clear and informative enough to update the player’s mental model
  • Leave the player with a meaningful next choice

The designer’s loop

“You can think of the game design process as a balancing loop, where the design as created by the designer is experienced by the player, who provides feedback to the designer. This feedback demonstrates the difference between the designer’s intent and the player’s experience. The designer then changes the design to (hopefully) reduce this difference, and the loop begins again.” — Sellers, Ch. 7

The designer’s loop is a balancing loop in the systems-thinking sense (see systems-thinking): the designer observes a gap between intent and experience, adjusts the design to close the gap, and tests again. This loop cannot operate until a playable version exists — “a bunch of rules does not make a game” (Sellers). Prototyping and playtesting are the mechanism that makes the designer’s loop possible.

Core loops

Definition

“A game’s core loops are its primary interactive systems — the interactions between the game and the player that form the player’s primary focus.” — Sellers, Ch. 7

A core loop is identifiable by what the player is doing most of the time and most focused on at any given moment. It is not necessarily the fastest loop; it is the most important for player engagement at a given level of attention.

A game must have at least one core loop. Simple games may have only one; more complex games have multiple, operating at different timescales.

Inner loops and outer loops

Core loops exist at different timescales and nest within each other:

Inner loops (fast, tight):

  • Short feedback cycle — seconds to minutes
  • Action/feedback dominated — physical reactions, immediate decisions
  • Combat, platformer movement, puzzle solving, card play
  • Primarily engage the player’s reaction speed and short-term cognition

Outer loops (slow, strategic):

  • Long feedback cycle — minutes to hours to sessions
  • Cognitive/emotional/social dominated — planning, progression, narrative
  • Levelling up, base construction, deck building, story advancement
  • Primarily engage the player’s long-term cognition and emotional investment

Most engaging games combine at least one fast inner loop with at least one slower outer loop. The fast loop provides immediate satisfaction; the slow loop provides a reason to keep returning.

“Many games use similar sets of core loops, often combining a high-tension, action-oriented battle, puzzle, or similar interaction loop with a lower-tension construction, crafting, trading, or similar loop. The former tend to employ faster action/feedback and short-term cognitive interactivity, while the latter tend toward slower long-term cognitive, emotional, and social interactions.” — Sellers, Ch. 7

The biological basis Sellers notes: fast sympathetic “fight or flight” responses power the inner loop; slower parasympathetic “rest and digest” functions power the outer loop. This is why high-tension and low-tension phases feel qualitatively different and why alternating them reduces fatigue.

Case study: Clash of Clans

Sellers analyses Clash of Clans as an example of multiple nested core loops:

Inner loops:

  • Collect resources → Battle → bring back coins and elixir

Outer loops:

  • Build and train → construct/upgrade buildings, train troops
  • Clan social loop → join a clan, help allies, rise in ranking

The “collect and battle” loop has the shortest cycle (minutes) and the highest tension. The “build and train” loop is longer and lower-tension — the restorative counterpart to battle. The clan loop is the longest and most social, driving long-term retention. These loops are hierarchically nested, each forming part of the next outer loop.

The monetisation model is integrated into the loops: real money buys gems that reduce construction/training time, allowing players to fast-forward the outer loop.

Core loops and game mechanics

Sellers proposes a systems-based definition of game mechanics that emerges from the loop concept:

“Each game mechanic forms an interactive loop between the player(s) and the game. This loop is identifiable and, as its own system, is generally free of any particular game context.” — Sellers, Ch. 7

A game mechanic is a reusable, identifiable loop — “draw a card,” “territory control,” “push your luck.” When a particular mechanic forms the dominant loop of many games, it constitutes a genre. Platformers: jumping as core mechanic. RPGs: combat + loot + levelling as core mechanics.

Levels of interactivity across loops

Sellers (Ch. 4) identifies six timescales of interactivity, each corresponding to a different kind of loop:

TimescaleLabelExamples
Milliseconds–secondsAction/feedbackShooting, jumping, card plays
Seconds–minutesShort-term cognitiveTactical decisions, puzzle steps
Minutes–hoursLong-term cognitiveStrategic planning, tech trees
Sessions–daysSocialGuilds, clans, multiplayer communities
Sessions–weeksEmotionalNarrative investment, character bonds
Months–lifetimeCulturalGenre literacy, speedrunning culture

Well-designed games engage players at multiple timescales simultaneously. A game that only engages the first timescale (pure action/feedback) provides short sessions and low retention; adding long-term cognitive loops creates replay. Adding social loops creates communities.

In practice

Designing a core loop:

  1. Define the player’s primary activity — what are they doing most of the time?
  2. Ensure the activity creates a feedback cycle: intent → action → result → updated intent
  3. Make the feedback informative enough to update the player’s mental model
  4. Make the result meaningful enough to create a next decision
  5. Test: can you play the loop in isolation, with no outer context, and find it engaging?

Adding outer loops:

  • What does the inner loop accumulate? (Resources, skills, progress, story)
  • What does the outer loop do with those accumulations? (Build, unlock, advance narrative)
  • Does the outer loop give the player a reason to come back to the inner loop?
  • Do the two loops feel qualitatively different in tension and timescale?

Unity (C#) — a minimal core loop structure:

// A core loop lives in the game manager, not in individual objects
// Inner loop: player acts each frame
void Update() {
    // 1. Player intent → input
    Vector2 input = new Vector2(Input.GetAxisRaw("Horizontal"), 0);
    
    // 2. Game state changes
    player.Move(input);
    
    // 3. Game provides feedback
    UpdateHUD();
    CheckForCollisions();
}
 
// Outer loop: checked on significant state changes, not every frame
void OnEnemyDefeated() {
    playerData.AddExperience(enemyXP);
    if (playerData.CheckLevelUp()) {
        ShowLevelUpScreen(); // Outer loop reward moment
    }
}

Clark’s Anatomy of Play and the Context Loop

Oscar Clark’s Games as a Service (Clark, 2014, see source-games-as-a-service) offers a complementary loop framework to Sellers’, grounding each layer in a different function of the body:

Clark’s layerBiological analogyFunction
Core mechanicBonesThe structural skeleton; what the game physically is
Success/failure statesMusclesWhat drives movement and tension; challenge and consequence
RewardsNervous systemThe feedback signals that motivate continued action
Context loopCirculationThe rhythm that sustains engagement across sessions
MetagamePsychologyThe cognitive/social layer that drives long-term commitment

Clark’s context loop maps roughly onto Sellers’ outer loop but emphasises the narrative and situational frame: the game world, its story state, and the player’s position within it. The context loop is what gives the inner loop meaning — the same puzzle mechanic feels different depending on whether it is presented as “a training exercise” or “defusing a bomb with the hero’s family at stake.”

Clark’s rhythm of play concept extends the outer loop into the live-service domain: the pacing of notifications, energy refills, and daily return incentives forms a behavioural rhythm that operates on a clock cycle (daily, weekly, seasonal) well beyond the individual session (Clark, Games as a Service, see source-games-as-a-service). This connects to the engagement-led-design concept of the Flash Gordon Cliffhanger — the outer loop must end in a tension state that pulls the player back.

See also player-lifecycle for how loop design relates to the stages of player engagement over weeks and months, and f2p-monetisation-design for how monetisation integrates into the outer loop without undermining it.

Lovell’s Core Loop: Base Layer, Retention Layer and the Gearbox

Nicholas Lovell’s Pyramid of Game Design (Lovell, 2018, see source-pyramid-of-game-design) introduces a third loop vocabulary that partially overlaps with Sellers’ and Clark’s but is oriented specifically around service-game design.

Lovell defines the Core Loop as the cycle that takes the player from the Base Layer into the Retention Layer and back again. He explicitly differentiates his definition from the common industry shorthand where “core loop” simply means the Base Layer activity:

“My definition is that the Core Loop takes players from the Base Layer into the Retention Layer and back again.” (Lovell, Ch. 4)

Lovell’s termApproximate mappingNotes
Base LayerSellers’ inner loopMoment-to-moment gameplay; fun, repeatable action
Retention LayerSellers’ outer loopProgress systems, unlocks, social; drives return over sessions
Core LoopThe connection between inner and outerNot the inner loop itself, but the gearbox between them
Superfan LayerClark’s metagameElder-game content for the most committed players

Lovell identifies three archetypes of Core Loop, which are useful for genre analysis:

  • Linear Core Loop — complete unique missions to progress a narrative; most common in product games; expensive at scale
  • Strategic Core Loop — bidirectional connections between Base and Retention allow choices in each layer to affect the other (e.g. XCOM: Enemy Unknown, Total War)
  • Replayable Core Loop — repeat the Base Layer to earn resources that fuel Retention Layer progress; the dominant model in service games (e.g. Temple Run, Hearthstone, CSR Racing)

The mechanism that makes the Core Loop legible to players is the Gearbox — specifically the Pre-Event Screen (showing the player what the Base Layer will offer and what they bring from the Retention Layer) and the Post-Event Screen (showing what was earned and how the Retention Layer has changed). Without a designed Gearbox, the Core Loop remains invisible to players and loses its motivational power.

This vocabulary is complementary to Sellers’ and Clark’s. Where Sellers’ inner/outer loop distinction is useful for analysing engagement at different timescales, Lovell’s Base/Retention/Core Loop framing is useful for designing the service-game structure and for communicating priorities within a development team.

See pyramid-of-game-design for a full account of Lovell’s framework.

Open questions

  • Sellers’ loop taxonomy is descriptive. Is there a normative version — are certain loop configurations always better than others for specific design goals?
  • Outer loops are essential for long-term retention, but in student projects (short time horizon, small teams), are they worth the investment, or should student games focus on one excellent inner loop?
  • Live-service games often have very deep outer loop structures (seasonal progression, daily/weekly quests, battle passes). At what point do outer loops become extrinsic pressure rather than intrinsic engagement?
  • How does Clark’s context loop relate to Sellers’ outer loop — are they the same concept named differently, or does the context loop emphasise something Sellers’ outer loop misses?
  • How do Lovell’s three Core Loop archetypes (linear, strategic, replayable) map to Sellers’ model? The replayable Core Loop fits cleanly; the strategic loop suggests a bidirectional relationship between inner and outer loops that Sellers’ model implies but does not make explicit.

Graph View

Backlinks