--- name: game-scoping description: MLP scoping and tracer bullet planning for games. Use when scoping a game idea, prioritizing game features, defining vertical slices, creating a game MLP, or breaking down a game concept into buildable phases. Covers core loop identification, MoSCoW prioritization, scope management, feature cutting. --- # Game Scoping & MLP Planning Scope game ideas into buildable plans. Get to playable fast, iterate from there. ## Quick Reference ### MLP vs MVP | | MVP (Software) | MLP (Games) | |---|---|---| | **Goal** | Validate market fit | Validate fun | | **Success metric** | Users complete task | Players choose to keep playing | | **Core question** | "Does it work?" | "Is it fun?" | | **Minimum bar** | Functional | Enjoyable | | **What to cut** | Non-essential features | Anything not serving the core loop | **Key insight:** An MVP that works but isn't fun is a failed game. An MLP that's fun but limited is a successful foundation. ### Core Loop Identification The core loop is what the player does **every 30-60 seconds**. Everything else is secondary. **One-sentence test:** "The player [VERB]s [OBJECT] to [REWARD], which lets them [REINVEST]." | Genre | Example Core Loop | |-------|-------------------| | Platformer | Jump through obstacles → reach checkpoint → unlock new area | | Roguelike | Enter room → fight enemies → collect upgrades → go deeper | | Puzzle | Observe pattern → manipulate pieces → solve → get harder puzzle | | Tower Defense | Analyze wave → place towers → survive → earn currency → upgrade | | Farming Sim | Plant crops → tend/wait → harvest → sell → buy better seeds | | Card Game | Draw cards → play combinations → resolve effects → draw again | ### Core Loop Validation Checklist - [ ] Can be stated in one sentence - [ ] Involves player agency (not watching) - [ ] Has a clear reinvestment path (reward feeds back into action) - [ ] Is inherently repeatable without getting stale in first 10 minutes - [ ] Can be prototyped with programmer art - [ ] Is fun with zero content (no levels, no story, no progression) ## MLP Scoping Process ### The 3-Feature Rule Your MLP gets **three features** beyond the core loop. Choose wisely. **Process:** 1. Define the core loop (non-negotiable, not counted as a feature) 2. List every feature you want 3. For each feature, ask: "Does this make the core loop more fun?" 4. Rank by impact on core loop fun 5. Take the top 3 **Why 3?** More than 3 features for an MLP means you're building content, not validating fun. If your core loop needs more than 3 supporting features to be fun, your core loop is too weak. ### Feature Impact Assessment | Impact Level | Description | Example (Roguelike) | |---|---|---| | **Core** | Loop doesn't work without it | Combat system | | **Amplifier** | Makes core loop significantly more fun | Upgrade selection between rooms | | **Enhancer** | Adds variety/polish | Different enemy types | | **Content** | More of the same | Additional levels | | **Cosmetic** | Visual/audio polish | Particle effects | MLP includes: Core + up to 3 Amplifiers. Everything else waits. ## Tracer Bullet Methodology ### 5-Layer Vertical Slice A tracer bullet proves your game works end-to-end through all layers: ``` ┌─────────────────────────────────────┐ │ Layer 1: INPUT │ │ Player presses button / clicks │ ├─────────────────────────────────────┤ │ Layer 2: CORE MECHANIC │ │ Game responds (character jumps, │ │ card plays, piece moves) │ ├─────────────────────────────────────┤ │ Layer 3: GAME STATE │ │ World updates (health, position, │ │ score, inventory) │ ├─────────────────────────────────────┤ │ Layer 4: FEEDBACK │ │ Player sees/hears result (screen │ │ shake, sound, animation) │ ├─────────────────────────────────────┤ │ Layer 5: OUTPUT / CONSEQUENCE │ │ Game state creates new situation │ │ (new challenge, reward, game over) │ └─────────────────────────────────────┘ ``` **Rule:** Your first build should touch all 5 layers for ONE action. Don't build Layer 1 for all actions before starting Layer 2. ### Tracer Bullet Examples | Genre | Tracer Bullet (First Build Target) | |-------|-----------------------------------| | Platformer | Player runs and jumps on one platform, falls and respawns | | Roguelike | Player enters one room, fights one enemy, picks up one item | | Puzzle | Player manipulates one puzzle element, solves one puzzle | | Tower Defense | Player places one tower, one enemy walks the path | | Card Game | Player draws a hand, plays one card, sees its effect | ### What a Tracer Bullet is NOT - Not a polished demo (programmer art is fine) - Not feature-complete (one path through the game) - Not balanced (numbers can be wrong) - Not content-rich (one level, one enemy type, one item) ## MoSCoW Prioritization for Games Adapted from standard MoSCoW with game-specific criteria. ### Must Have (MLP-blocking) - Core loop mechanics - One complete path from start to "one more round" - Minimum feedback (player knows what happened) - Win/loss or progress/regress state ### Should Have (First update after MLP) - 2-3 content variations (enemy types, level layouts) - Basic progression (something carries over between sessions) - Audio feedback for core actions - Pause/resume ### Could Have (Polish phase) - Settings menu - Multiple difficulty levels - Achievements/unlocks - Detailed tutorial ### Won't Have (Not this project phase) - Multiplayer (unless it IS the core loop) - Level editor - Mod support - Localization ## Scope Management ### The 30% Rule Whatever scope you think the MLP needs, cut 30%. Then you're close to right. **Why:** First-time scoping for games is always optimistic because: - "Simple" mechanics have hidden complexity (collision, edge cases, feel) - Fun requires iteration time you didn't budget for - Integration between systems takes longer than building them separately - Playtest feedback will change your priorities ### Ruthless Cutting Heuristics Ask these questions about every feature. If any answer is "yes," cut it. | Question | If Yes → Cut | |----------|-------------| | Can the core loop function without this? | Cut | | Is this adding content rather than mechanics? | Defer | | Does this require a new system/architecture? | Cut from MLP | | Am I adding this because other games have it? | Cut | | Would a player miss this in a 5-minute playtest? | If no → Cut | | Does this feature serve MY game's core loop? | If no → Cut | ### Scope Creep Signals Watch for these during planning: - "While we're at it, we could also..." → Stop. Separate issue. - "This would only take a day" → It won't. Defer it. - "Players expect this in [genre]" → Maybe. After MLP. - Feature list growing between planning sessions → Freeze and re-evaluate. - Excitement about a feature unrelated to core loop → Parking lot it. - "We need this for the trailer/demo" → Marketing scope ≠ MLP scope.

Genre-Specific Scoping Tips: Platformer

**MLP must have:** Run, jump, land on platforms, one hazard type, one level. **Common over-scope:** Wall jump, double jump, dash, multiple worlds. Pick ONE movement verb beyond run+jump for MLP. **Tracer bullet:** Character moves left/right, jumps, lands on a platform, falls into a pit, respawns. **Hidden complexity:** Jump feel (coyote time, jump buffering, variable height) can consume as much time as building 5 levels. Budget for it.

Genre-Specific Scoping Tips: Roguelike

**MLP must have:** One floor, one enemy type, one item/upgrade, permadeath + restart. **Common over-scope:** Meta-progression, 10+ enemy types, complex inventory, procedural generation before fun is proven. Handcraft your first 3 rooms. **Tracer bullet:** Enter room, fight enemy, pick up item, die, restart with nothing. **Hidden complexity:** Procedural generation that feels good (not random) is a project unto itself. Use handcrafted rooms for MLP.

Genre-Specific Scoping Tips: Puzzle Game

**MLP must have:** One puzzle mechanic, 5-10 handcrafted puzzles, clear win state per puzzle. **Common over-scope:** Procedural puzzle generation, hint system, multiple mechanics in MLP. **Tracer bullet:** Player sees puzzle, manipulates elements, puzzle solves, next puzzle loads. **Hidden complexity:** Difficulty curves in puzzles require extensive playtesting. Your "easy" puzzle won't be easy for everyone.

Genre-Specific Scoping Tips: Tower Defense

**MLP must have:** One enemy path, one tower type, waves, win/lose condition. **Common over-scope:** Tech trees, multiple maps, hero units, special abilities. **Tracer bullet:** Enemy walks path, player places tower, tower shoots enemy, enemy dies or reaches end. **Hidden complexity:** Balancing tower cost vs enemy health vs wave difficulty is iterative. Don't try to get numbers right in planning.

## Post-MLP Iteration Phases | Phase | Focus | Duration Target | |-------|-------|-----------------| | **MLP** | Core loop + 3 features | Build as fast as possible | | **Fun Polish** | Juice, game feel, feedback | Until core loop feels good | | **Content Alpha** | Variety (enemies, levels, items) | Until 15-30 min play session | | **Systems** | Progression, meta-game, saves | Until players want to return | | **Content Beta** | Full content pipeline | Until target play length | | **Polish** | UI, audio, accessibility | Until shippable | **Rule:** Don't advance to the next phase until the current phase's goal is met. Especially don't skip Fun Polish. ## Checklists ### Before Scoping - [ ] Core loop defined in one sentence - [ ] Core loop validated with checklist above - [ ] Genre identified (affects scope expectations) - [ ] Target play session length defined (5 min? 30 min? 2 hours?) - [ ] Platform/input constraints known (keyboard? controller? touch?) - [ ] Solo dev or team? (drastically affects achievable scope) ### Scope Review - [ ] MLP has ≤3 features beyond core loop - [ ] 30% rule applied (original scope reduced) - [ ] Every feature passes the cutting heuristics test - [ ] Tracer bullet target defined (first thing to build) - [ ] No features that require "new systems" to support - [ ] No content-heavy items in MLP (those come later) - [ ] Scope fits target timeline (be honest) - [ ] Playtest plan: who will test, when, what you're testing for ## See Also - **game-design-frameworks** — MDA, core loop theory, player motivation (the "why" behind scoping decisions) - **game-antipatterns** — Common planning mistakes to avoid during scoping

game-scoping

Scan your own AI components