A browser-based memory game built around random graph traversal — where both you and your AI opponents navigate the same randomly generated maze, but only you know the rules.
You start in the first room of a randomly generated maze. Your goal: find and open the final door, hidden somewhere at stage 12. Move between rooms by clicking doors, backtrack through portals, and collect one stone per stage — the final door won't open until you have all 12.
The catch: you're not alone. A configurable number of random walkers (0–10) are racing through the same maze simultaneously. They don't need to collect stones — they just need to reach the final door before you do.
The core design question behind this project was: what graph structure makes a random walk take the longest to reach the exit?
This is a well-studied problem in the theory of random walks on graphs. The worst-case graph for a random walker — the one that maximizes expected hitting time to the target — is related to structures with bottleneck topologies (such as the lollipop graph). In practice, random trees were chosen as the layout for each of the 12 stages, providing a good balance between theoretical difficulty and playable structure: trees have no cycles, so a random walker can easily get "trapped" exploring dead-end branches far from the exit.
Each stage generates a fresh random tree, with rooms as nodes and doors as edges. The competing randys perform true random walks on this structure in real time — meaning the graph layout directly and measurably affects how hard the game is for both you and them.
This project began as a Python desktop application (radgame), where the graph structure and random-walk mechanics were first developed and analyzed. It was later reimagined as a web application using HTML, CSS, and JavaScript, with significant enhancements to gameplay, visuals, and interactivity added along the way.
- JavaScript — game engine, graph generation, random walk simulation
- HTML / CSS — UI and responsive layout
- Graph algorithms — random tree generation, random walk dynamics
No build step required. Just clone and open:
git clone https://github.com/avikpln/rad.git
cd rad
open index.html # or drag into your browserrad/
├── css/
│ └── styles.css
├── design/ # Design assets and UML diagrams
│ ├── game.bmp
│ ├── game.gaphor
│ ├── gui.gaphor
├── images/
│ ├── favicon.ico
│ └── screenshot.png
├── meta/ # Project metadata
│ ├── academy
│ ├── acknowledgements
│ ├── hierarchy
│ └── version
├── quotes/ # Quote database
│ ├── JakubPetriska.zip
│ ├── quotes.csv
│ └── quotes.py
├── script/
│ ├── game/ # RAD game logic
│ │ ├── test/ # Game testing utilities
│ │ ├── cell.js
│ │ ├── direction.js
│ │ ├── door.js
│ │ ├── element.js
│ │ ├── game-state.js
│ │ ├── game.js
│ │ ├── location.js
│ │ ├── player.js
│ │ ├── room.js
│ │ └── stone.js
│ ├── gui/ # Graphical user interface
│ ├── library/ # Shared utilities
│ └── main.js
├── sound/ # Sound effects and voice clips
│ ├── complete.mp3
│ ├── ...
│ └── wow-who-are-you.mp3
├── index.html
├── LICENSE
├── README.md
├── TODO.md
└── zipit.bat