Nano Platformer
A 12-Stage 2D Platformer Game with Upgrades, Boss Fights & Adaptive Difficulty
Deliverables / Skills Utilized
Python
Pygame
JSON
Sprite Animation
🔍 The Problem
Wanted to build a complete game from scratch—not a tutorial clone, but a fully-featured platformer with real game design: progression systems, enemy AI, boss fights, and player-driven difficulty.
The goal was to learn game development fundamentals by shipping something end-to-end: physics, state machines, animation, game balance, and persistence.
💡 The Solution
A 12-stage 2D platformer with a full campaign arc: escalating enemy types, a shop/upgrade system, a multi-phase boss fight, and an adaptive assist system that adjusts difficulty based on player performance.
Built entirely in Python with Pygame—no game engine, no shortcuts.
🛠️ Technical Implementation
⚙️ Physics Engine
- Custom gravity and acceleration system with configurable constants
- AABB collision detection for platforms, walls, and entities
- Jump mechanics with variable height (tap vs hold) for precise control
- Friction and air resistance for natural-feeling movement
👾 Enemy AI System
- Walkers: Patrol paths with edge detection and turn behavior
- Shooters: Ranged enemies with projectile spawning and cooldown timers
- Boss: Multi-phase fight with burst-fire patterns, health thresholds, and enrage mechanics
- State machine architecture for clean AI behavior transitions
🛒 Upgrade System
- Vitality: Max HP increases and passive health regen
- Blade: Damage multiplier and attack range extensions
- Tempo: Movement speed and jump height boosts
- 3 tiers per upgrade path, purchased with in-game currency from defeated enemies
🎚️ Adaptive Difficulty
- Tracks player deaths per stage and cumulative performance
- Activates assist mode after repeated failures: extra HP, damage reduction, extended checkpoints
- Scales back assistance when player improves—never feels patronizing
📁 JSON-Based Level System
- Each stage defined as a JSON file: platforms, enemy spawns, collectibles, and triggers
- Hot-reloadable during development for rapid level design iteration
- Supports custom tilesets and background layers per stage
💾 Persistent Save System
- Player progress, upgrades, and currency saved to local JSON
- Checkpoint system within stages for respawn on death
- Full campaign state preserved across sessions
🗺️ Campaign & Boss Fights
🎯 Key Features
- 🎮 12-Level Campaign: Progressive difficulty from tutorial to boss fight
- 👾 3 Enemy Types + Boss: Walkers, shooters, and a multi-phase boss encounter
- 🛒 Shop & Upgrades: 3 upgrade paths (Vitality/Blade/Tempo) with 3 tiers each
- 🎚️ Adaptive Assist: Dynamic difficulty adjustment based on player performance
- ✅ Checkpoint/Respawn: Mid-stage checkpoints with instant respawn
- 🎨 Sprite Animation: AI-generated character and environment assets with frame-by-frame animation
- 💾 Persistent Save: Full progress saved across sessions
📊 Technical Highlights
| Metric | Detail |
|---|---|
| 🎮 Campaign Stages | 12 |
| 👾 Enemy Types | 3 + Boss (Walker, Shooter, Boss) |
| 🛒 Upgrade Paths | 3 (Vitality, Blade, Tempo) |
| 🎚️ Upgrade Tiers | 3 per path (9 total) |
| 💾 Save Format | JSON (portable, human-readable) |
| 🖼️ Art Pipeline | AI-generated sprites + manual cleanup |
🎯 Why This Matters
This project demonstrates:
- ⚙️ Game Physics: Custom gravity, collision detection, and movement systems from scratch
- 🤖 State Machines: Clean enemy AI architecture with behavior transitions
- ⚖️ Game Balance: Upgrade pricing, enemy scaling, and adaptive difficulty tuning
- 🎨 AI Art Pipeline: Generating game assets with AI tools and integrating into sprite sheets
- 🏗️ Systems Design: Interconnected systems (combat, economy, progression) working together
🔧 Stack
- Python: Core game logic and systems
- Pygame: Rendering, input handling, and game loop
- JSON: Level definitions, save data, and configuration
- Sprite Animation: Frame-by-frame animation with AI-generated assets