cocos-game-balance-director

Cocos Game Balance Director

Overview

Use this skill to make game numbers intentional, tunable, and testable. The output should define player-facing balance goals first, then translate them into Cocos-friendly constants, tables, formulas, and playtest checks.

This skill does not guess final production values. It creates an initial tuning model and identifies what playtest evidence should change it.

When the user asks for starting values, always provide concrete v0 tuning assumptions. Label them as defaults to test, not final production truth.

Core Workflow

1. Define balance intent.
   • State target session length, expected success rate, difficulty feel, reward cadence, and player skill assumptions.
2. Identify tunable systems.
   • Combat, movement, timers, scoring, spawns, upgrades, economy, level pressure, content unlocks, random events, crits, loot, drops, affixes, or procedural selection.
   • Keep first-slice numbers small enough to understand and change quickly.
3. Choose primary metrics.
   • Use metrics such as TTK, time to fail, score per minute, currency per run, upgrade runs required, hit frequency, spawn pressure, and recovery margin.
4. Build formulas and tables.
   • Prefer explicit v0 tables for early prototypes.
   • Use formulas only when the progression shape is more important than individual authored values.
   • For random systems, prefer probability tables, weighted tables, seeded selection rules, expected-value formulas, and variance guardrails over inline probability checks.
   • Include default, minimum, maximum, unit, confidence, and tuning rationale for important values.
5. Add guardrails.
   • Define min/max values, caps, cooldowns, pity rules, streak prevention, spawn budgets, duplicate prevention, and runaway prevention.
6. Hand off implementation.
   • Specify where Cocos should store values: serialized properties, JSON/config files, TypeScript constants, or spreadsheet exports.
   • Call out values that designers should tune without code edits.