game-balance

name	game-balance
description	Analyzes a game system against design principles — economy health metrics, difficulty curve, reward schedules, and progression cost curves. Flags imbalances with severity levels. Use after creating an MVP plan with game-architect. Extract the analysis scope (economy/difficulty/progression/rewards/all) from the user's message.

Game Balance — Balance Analyzer

Prerequisites

• Requires docs/mvp-first-draft.md in the project (run game-architect first)

---

Phase 0 — Context Load (silent, mandatory)

Execute in this exact order:

1. READ docs/mvp-first-draft.md
   • If it does NOT exist: output ⛔ Cannot proceed. docs/mvp-first-draft.md not found. Run the game-architect skill first to generate the MVP plan before running a balance analysis. and stop.
2. Determine scope from the user's message:
   • If no scope specified or "all": analyze all systems (Phases 1–4)
   • If "economy": run Phase 1 only
   • If "difficulty": run Phase 2 only
   • If "progression": run Phases 2 + 3
   • If "rewards": run Phases 1 + 4
3. Load skills based on scope:
   • economy or all → Apply the game-economy-design skill
   • difficulty or progression or all → Apply the game-design-fundamentals skill
   • progression or all → Apply the skill-progression-trees skill
4. READ docs/build-registry.md if it exists
5. READ docs/quest-registry.md if it exists (for reward cross-reference in Phase 4)

Extract from docs/mvp-first-draft.md:

• Game genre and core loop description
• Currency types and economy parameters (faucets, sinks, amounts)
• Difficulty settings and intended player progression
• Progression system details (skill trees, level gates, cost tables)
• Reward tables and schedules

Phase 1 — Economy Health Check

Skip this phase entirely if scope is "difficulty" or "progression".

Apply the Economy Health Metrics table from game-economy-design. For each metric, find the relevant value in docs/mvp-first-draft.md (or flag as "not defined"):

Metric	Target Range	Found in Plan	Status
Daily currency earn/spend ratio	0.8–1.2	[value or "not defined"]	🟢/🟡/🔴
Median player currency held	2–5× avg purchase cost	[value or "not defined"]	🟢/🟡/🔴
Premium conversion rate	2–5% of active players	[value or "not defined"]	🟢/🟡/🔴
Time to first meaningful purchase	5–15 minutes	[value or "not defined"]	🟢/🟡/🔴

Status legend:

• 🟢 Within target range
• 🟡 Outside range but recoverable — specific fix suggested
• 🔴 Critical imbalance or not defined at all — blocks healthy launch

Also check for these anti-patterns from the skill:

• Hyperinflation — unlimited currency faucets with no corresponding sinks → 🔴 CRITICAL
• Pay-to-win — premium currency or IAP that grants direct gameplay power → 🔴 CRITICAL
• Currency confusion — more than 3 currency types without clear differentiation → 🟡 WARNING
• Dead economy — currency accumulates with nothing interesting to buy → 🟡 WARNING
• Early abundance — excessive rewards in first 10 minutes that devalue later rewards → 🟡 WARNING
• Exploitable loops — crafting or trading paths that generate infinite value → 🔴 CRITICAL

For each flagged issue, cite the specific skill principle: e.g., [→ game-economy-design: every faucet needs a sink]

Phase 2 — Difficulty & Flow Channel Check

Skip this phase entirely if scope is "economy" or "rewards".

Apply the flow channel framework from game-design-fundamentals (Csikszentmihalyi principle: challenge must scale with skill).

Analyze the intended difficulty curve from mvp-first-draft.md against the four progression phases:

Phase	Expected Pattern	Found in Plan	Assessment
Early game	Rapid unlocks, teaching mechanics, high success rate	[from plan]	🟢/🟡/🔴
Mid game	Meaningful choices, branching paths, ~65% success rate	[from plan]	🟢/🟡/🔴
Late game	Mastery optimization, meaningful challenge	[from plan]	🟢/🟡/🔴
Endgame	Social/competitive/infinite scaling	[from plan]	🟢/🟡/🔴

Check for these anti-patterns from the skill:

• Flat difficulty curve — no meaningful challenge increase → 🔴 (players master and leave)
• Punishing failure — harsh penalties that discourage experimentation → 🟡
• Tutorial as separate experience — tutorial feels disconnected from core gameplay → 🟡
• Single player-type focus — designed exclusively for one Bartle type → 🟡
• Boredom cliff — player masters the game too quickly → 🔴

The getDifficultyMultiplier formula from the skill: 1 + (playerSuccessRate - targetSuccessRate) * 0.3 — note if the plan accounts for adaptive difficulty or not.

For each flagged issue, cite the skill: [→ game-design-fundamentals: flow channel — challenge must scale with skill]

Phase 3 — Progression Cost Curve Check

Skip this phase entirely if scope is "economy", "difficulty", or "rewards".

Apply the cost scaling principles from skill-progression-trees.

Expected formula: cost = baseCost × tier^exponent where exponent should be 1.5–2.0.

Analyze the plan's progression costs:

Progression System	Cost Structure in Plan	Expected Formula Applied?	Issue
[system 1 from plan]	[flat / linear / exponential]	🟢/🟡/🔴	[if flagged]
[system 2 from plan]	...	...	...

Check for these anti-patterns from the skill:

• Flat cost curves — every node/level costs the same → 🟡 (endgame nodes become trivially cheap as income scales)
• Missing respec mechanic — players locked into bad choices permanently → 🟡
• All nodes visible upfront — full tree shown immediately, spoiling discovery → 🟡
• Client-side tree state — plan mentions client-side unlock validation → 🔴 CRITICAL
• Hardcoded effects — effects in code instead of data config → 🟡

For each flagged issue, cite the skill: [→ skill-progression-trees: cost = baseCost * tier^exponent]

Phase 4 — Reward Schedule Check

Skip this phase entirely if scope is "difficulty" or "progression".

Apply reward schedule principles from game-design-fundamentals (variable ratio > fixed ratio — B.F. Skinner principle).

Analyze the reward structure:

Reward Type	Schedule in Plan	Variable?	Assessment
Quest rewards	[fixed / variable]	[yes/no]	🟢/🟡/🔴
Daily login	[fixed / variable]	[yes/no]	🟢/🟡/🔴
Loot drops	[fixed / variable]	[yes/no]	🟢/🟡/🔴
Achievement rewards	[fixed / variable]	[yes/no]	🟢/🟡/🔴

If docs/quest-registry.md exists, cross-reference quest reward amounts against the economy metrics from Phase 1.

Check for these anti-patterns:

• Reward inflation — too many rewards too fast; all future rewards feel devalued → 🟡
• Predictable schedules — fixed ratio rewards sustain engagement much less than variable → 🟡
• Nothing to spend on — rewards accumulate with no meaningful sink → connects to Phase 1 dead economy → 🔴

For each flagged issue, cite the skill: [→ game-design-fundamentals: variable ratio reward schedules sustain engagement longer]

Phase 5 — Balance Report

Output the complete report:

# Balance Report — [game name from mvp-first-draft.md] — [current date]

## Executive Summary
🔴 Critical issues: [count]
🟡 Warnings: [count]
🟢 Healthy systems: [count]

[If 0 critical issues]: ✅ No blockers found — safe to proceed with current design.
[If critical issues exist]: ⛔ [count] critical issue(s) must be resolved before launch.

---

## Economy Health [only if analyzed]
[metrics table from Phase 1]
[anti-pattern findings]

## Difficulty Curve [only if analyzed]
[phase table from Phase 2]
[anti-pattern findings]

## Progression Costs [only if analyzed]
[table from Phase 3]
[anti-pattern findings]

## Reward Schedules [only if analyzed]
[table from Phase 4]
[anti-pattern findings]

---

## Recommended Actions

### 🔴 Critical (resolve before building)
1. [Issue description] — [specific fix] `[→ skill-name: principle]`
...

### 🟡 Warnings (address before launch)
1. [Issue description] — [specific fix] `[→ skill-name: principle]`
...

### 🟢 Notes (optional improvements)
1. [Suggestion] `[→ skill-name: principle]`
...

---

Hard Constraints

1. Always cite the skill for every flagged issue — no undocumented opinions
2. Never flag "not defined" as healthy — undefined parameters are 🔴 by default (you can't launch without them)
3. Pay-to-win is always 🔴 CRITICAL — no exceptions
4. Client-side game state is always 🔴 CRITICAL — server-authoritative is non-negotiable
5. Provide a specific fix for every 🔴 and 🟡 finding — do not just flag without recommendation