.png)
2 days ago
1
Repository files navigation
A reinforcement-learning framework that learns how to exploit an opponent’s weaknesses and outputs match-specific tactical game-plans.
RL-Tactics helps coaching staffs answer the question ➡️ “What is the best way to line up and press against Team X next Saturday?”
It does so by training a reinforcement-learning (RL) agent in a simulation environment and surfacing match-specific tactical plans ranked by expected goal differential (xGD).
- Version 0 (current) — we’re at v0; core research sandbox with continuous improvements underway.
- Version 0.1 — first prototype of the RL model that can generate basic tactic recommendations.
- Version 1.0 — final public release with refined model, full documentation, and a polished web interface.
- Fork the repo & create your feature branch (git checkout -b feat/awesome-idea).
- Commit using Conventional Commits & open a PR.
- Ensure pytest & pre-commit hooks pass.
All contributors must follow our Code of Conduct.
Full documentation lives in the docs/ folder:
- Getting Started
- Usage Guide
- Architecture
- Pitch Modeling
- Team Classifier
- Model Files
- Output Files
- Data Format
- Developer Guide
This project is licensed under the MIT License — see LICENSE for details.
