| Field | Value |
|---|
| Module | level |
| Version | 1.8.2 |
| Effort | high |
| Tools | — |
Keywords: design, environment, level design, map design
- Designing battle arenas, dungeon rooms, or open-world zones
- Placing obstacles, cover, or spawn points for AI units
- Planning encounter pacing, difficulty waves, or boss arenas
- Designing player movement flow (linear corridor, hub-and-spoke, open world)
- Adding environmental hazards, traps, or interactive elements
- Sizing arenas for unit counts (how big should this arena be?)
- Player must understand the space within 3 seconds of entering
- Landmark (weenie) visible from spawn — directs movement without text
- Distinct zones by floor/wall texture, lighting color, or height variation
- Every path leads somewhere meaningful (reward, shortcut, information)
- Dead ends only if they contain loot or environmental storytelling
- Chokepoints are designed — never accidental narrowing
- Off-path content gives optional rewards (not mandatory progress)
- Curiosity is rewarded: 80% exploration → bonus loot, 100% → achievement
- Secrets require skill or observation, never luck
| Shape | Best For | Weakness | Key Rule |
|---|
| Circular | Boss fights, duel arenas | Predictable rotation | Add center obstacle to break circling |
| Corridor | Ambush, chokepoint tension | Limited tactics | Widen mid-fight: destructible walls, branching |
| Multi-room | Dungeon crawl, stealth | Pacing breaks between rooms | Connect rooms with visible reward glimpse |
| Open field | Cavalry/ranged battles | Melee struggles for cover | Scatter 4–6 cover clusters asymmetrically |
- Ranged behind melee: ranged units start 8–12m behind front line; melee absorbs aggro
- Flankers on sides: 1–2 fast units spawn 45° off-axis from player entry — forces rotation
- Never spawn behind player: staging zone extends 3s travel from entry point (enemy-free)
- Elite placement: elite/mini-boss at arena center-back — player must push through adds
- Density rule: max 1 ranged per 2 melee at standard difficulty; invert for hard encounters
- Risk-reward distance: high-value loot placed at maximum travel from safe spawn
- Tier system: common → safe path; rare → off-path detour; legendary → boss/hazard-gated
- Visual breadcrumb: player can see reward before committing — respects decision-making
- Chest trap rule: any chest in open sight without obstacles is suspicious — teaches awareness
- Wave entry points: 2–4 entry directions; rotate per wave to prevent corner camping
- Reinforcement lanes: separate lane from main combat area; arrives from flanks on timer
- Staging buffer: 2–3s before spawned enemies reach player — prevents spawn-kill
- Visual telegraph: spawn FX (portal glow, dust cloud) 1s before unit appears
| Encounter # | Relative Difficulty | Mechanic Introduced |
|---|
| 1 | Tutorial (50%) | Basic attack pattern |
| 2 | Easy (70%) | Ranged + melee mix |
| 3 | Medium (90%) | Environmental hazard |
| 4 | Hard (110%) | Elite + adds combo |
| 5 | Rest/reward | No combat or trivial |
| 6 | Boss (150%) | All mechanics combined |
Stat scaling per tier: HP ×1.15/tier, damage ×1.10/tier, speed ×1.05/tier.
Within a single level that introduces a new mechanic (new enemy type, hazard, ability, traversal rule), structure the space as three acts:
| Act | Spatial Pattern | Player Experience |
|---|
| Teaching | Safe arena entry, mechanic appears in isolation, clear sightlines, low/no enemy pressure | Discovers the rule by interacting; particle/environmental hints replace text where possible |
| Main Body | Mechanic mixed with previously-mastered systems, light pressure (1-2 enemies, soft timer) | Applies the rule under manageable distraction |
| Finale | Harder variant of the mechanic, time/resource pressure, optional show-off route for collectables | Stress-tests mastery; an alternate path lets stuck players progress while experts express skill |
This is the level-scoped expression of the difficulty saw (per-mechanic cycles, not a monotonic curve). Across an arc of N levels, each new mechanic gets one full Teaching→Main→Finale cycle before the next mechanic is introduced. → See puzzle-game-design/references/difficulty-saw-framework.md.
Before locking encounter difficulty, gather real-player data:
| Method | What to Track | Threshold / Signal |
|---|
| Death-location heatmap | (level, position, attempt#) of every failure | Cluster on one tile/spawn = local spike; cluster at entry = readability problem |
| Encounter completion drop-off | % reached and % completed per encounter | >40% consecutive drop = panic signal; <5% drop = encounter adds no friction |
| Footage / silent observation | First failed attempts of unfamiliar players | If 4 of 5 testers try the same wrong move first, the level’s visual language is misdirecting |
| Structured friend testing | Scripted, specific questions (“How long on encounter 4?”, “Which arena made you put the controller down?”) | Outliers across 5+ testers are signal; single anecdote is not |
→ See game-balance-tools/references/validation-methods.md for paired simulation methods.
| Design Task | Key Rule | Reference |
|---|
| Arena sizing | size = max_units × 2-3m (melee) / 5m (ranged) | references/arena-layout.md |
| Obstacle coverage | 15-25% of floor area for tactical depth | references/arena-layout.md |
| Encounter density | 1 encounter per 30-60s travel | references/encounter-pacing.md |
| Difficulty pattern | Easy → medium → easy → hard → rest → boss | references/encounter-pacing.md |
| Hazard telegraphing | 0.5-1s wind-up warning before damage | references/environmental-hazards.md |
| # | Anti-Pattern | Problem | Fix |
|---|
| 1 | Invisible walls | Player frustration, breaks immersion | Use natural barriers (cliff, water, locked door with visible key) |
| 2 | Dead ends with nothing | Feels like wasted space | Add loot, lore item, or shortcut unlock to every dead end |
| 3 | Unfair spawns | Players spawned in combat range | Staging zone: 2-3s of safety before enemy can reach spawn |
| 4 | Symmetric arenas | Predictable, no exploration value | Asymmetric with mirrored spawn distances (fair but visually varied) |
| 5 | Random hazard timing | Feels unfair | All rhythm hazards use predictable, learnable cycles |
| 6 | Over-coverage obstacles | Units clip, pathfinding breaks | Cap obstacles at 25%; leave clear corridors for Navigation mesh |
| 7 | Reward high performers with harder-only content | Only experts ever reach the asset → wasted budget | Place expert content as optional bonus path off the main route, not gated behind elite-only access |
| 8 | Punish poor play with harder future encounters | Death spiral: failing player gets harder spawns, churns | Silent rubber-band: detect 3× consecutive failure → invisible 10-15% reduction; never announce |
| 9 | Unskippable cutscenes in high-retry areas | Each retry replays narrative → rage-quit signal | Skip-after-first-view; auto-skip after death in same encounter |
game-feel-juice — environmental feedback (rumble on hazard hit, particle on trap trigger)
game-ux-design — spatial UI (minimap, objective markers, zone transition prompts)
- engine navigation/scene skills (auto-activated via registry) — arena geometry and pathfinding tools
- engine navigation skills (auto-activated via registry) — pathfinding constraints for obstacle placement
game-encounter-design → (use game-balance-tools for encounter difficulty tuning)
puzzle-game-design — room-based level design for puzzle dungeons
| File | Coverage |
|---|
references/arena-layout.md | Shapes, obstacle %, spawn zones, cover, chokepoints, sizing formula, height variation |
references/spatial-flow.md | Weenies, breadcrumbs, hub-and-spoke, linear corridor, open world zones, verticality |
references/encounter-pacing.md | Encounter density, difficulty wave, boss arena, reward placement, roguelike structure |
references/environmental-hazards.md | Hazard types, telegraphing, fair placement, interactive hazards, rhythm patterns, scaling |
- Spatial flow ≠ encounter pacing — a beautifully laid-out arena can still feel monotonous if encounters all peak at the same intensity.
- Spawn zones must be reachable AND defensible — testing this with paper-prototype spawn lists before geometry saves art rework.
- Gating mechanics that bottleneck player progression vs gating mechanics that punish exploration are different design choices — name which one you’re doing.