GAMES / EXPERIMENTS
Battleships.
You sank my battleship. But can you prove it? Battleships with cryptographic commitments (commit-reveal).
Key facts
- Status
- Experimentas of October 2026
- Years
- 2024
- Role
- Design and code.
- Stack
- SvelteKit
- Fabric.js
- Go
- WebSocket
- SHA-256
- Highlight
- Two players · every cell is committed with a salted SHA-256 hash
Why I built it
In Battleship you have to trust the other player: you shoot, and they tell you if it is a hit or a miss.
This is a place to explore what can be verified without revealing the underlying information.
How it works
Once the ships are placed, each player sends a commitment for every cell of their board: a SHA-256 hash of the cell value, water or ship, with a random salt.
When a cell is shot, its owner reveals the value and the salt, and the other player checks them against the commitment. So a player cannot change the answer after the shot.
What it is not
The page on utile.space says "zero knowledge proof". To be exact, this is a commit-reveal scheme, not a zero-knowledge proof: nothing proves that a board follows the rules without revealing it.
The salts also come from Math.random(), which is not made for cryptography. A real zero-knowledge proof of the board would be a possible next step.
How it's built
The game is part of utile.space. The front end is a SvelteKit page that draws the boards with Fabric.js. The back end is in Go, in the utile.space API, and the players talk to it over a websocket.