GoStrategy

Open-source implementation of the classic Stratego board game. Built with a high-performance Go backend and a modern SvelteKit frontend, designed for both Human-vs-AI and AI-vs-AI experimentation.

Play the game here: https://gostrategy.dotsem.be

Or run it yourself locally: see "Running with Docker (Recommended)".

Project Overview

This project aims to provide a robust platform for playing GoStrategy while serving as a testing ground for various AI strategies. It features real-time gameplay via WebSockets, secure user authentication, and a declarative infrastructure managed via NixOS.

Tech Stack

• Backend: Go (Gin, Gorilla WebSocket)
• Frontend: SvelteKit 5 (Runes, Tailwind CSS)
• Database: PostgreSQL 15
• Infrastructure: NixOS, Docker, Cloudflare Tunnels
• Monitoring: Loki, Promtail

Getting Started

Prerequisites

• Docker and Docker Compose
• Bun (for local frontend development)
• Go 1.25+ (for local backend development)

Running with Docker (Recommended)

The easiest way to get the full stack running is using docker compose:

docker compose up

This will spin up the backend, frontend, and database containers. The app will be available at http://localhost:5000.

Local Development

If you prefer running services outside of Docker:

Backend:

cd code/backend
go run main.go --server

Frontend:

cd code/frontend
bun install
bun run dev

or use pnpm:

cd code/frontend
pnpm install
pnpm dev

Simulate AI vs AI Games

go run main.go --ai={ai1}:{ai2} --format md --logging=false --matches {n}

Herby you can replace {ai1} and {ai2} with any AI type available in the codebase. Replace n with the amount of matches you want to run.

Example:

go run main.go --ai=fafo:fato --format md --logging=false --matches 100

Architecture & Performance

High-Performance Concurrency

The backend uses an event-driven architecture with Go channels to manage game states. This ensures minimal mutex contention and high throughput for concurrent matches.

AI Experimentation

One of the core goals is the development of a comprehensive AI test suite. Current supported AI types:

• FAFO: Random move generator.
• FATO: Random move generator with basic memory to avoid loops.

Planned additions include Heuristic-based evaluation, MiniMax, and Monte Carlo Tree Search (MCTS). Detailed results of AI matches can be found in the AI Data folder.

Documentation

• Changelog - Project history and versioning.
• Roadmap - Future plans and hardening steps.
• API Documentation - Swagger UI (available when running locally).

Contributing

• Contributing Guide - Guidelines for contributing to the project.

License

This project is licensed under the MIT License - see the LICENSE file for details.