Formula 0

Simulating competition itself — not just vehicles.

Comment

Inspiration

Formula 1 teams, autonomous racing labs, and drone leagues all rely heavily on telemetry and simulations - yet most systems today simulate individual vehicles, not competition itself.

I wanted to build a platform that could model competitive mobility: many agents, live events, and dynamic leaderboards - where even past races become living simulations.

up against the legends - 1976 Niki Lauda,1989 Ayrton Senna,2002 Michael Schumacher,2005 Fernando Alonso,2013 Sebastian Vettel, 2018 Lewis Hamilton,2023 Max Verstappen etc.

Inspired by:

  • F1 strategy and telemetry analysis
  • Multi-agent simulation research (ProrokLab, Cambridge)
  • The precision and chaos of real motorsport competition

The vision was simple but bold:

Create a simulator where historical races can come alive, and new agents (AI or humans) can compete against them — in real time.

🧩 Challenges Faced

  • Synchronizing real telemetry data with simulated events (time-step mismatches, missing entries).
  • Building a generalized agent model that can handle multiple domains (cars, drones, or supply-chain entities).
  • Optimizing for performance - ensuring 100+ agents could simulate in real-time on modest hardware.
  • Designing a universal event schema that supports diverse actions like pit stops, checkpoints, or drone path corrections.
  • Balancing accuracy vs. performance when replaying historical data alongside live simulations.

Applications

  • Formula 1 and Formula E race strategy sandbox
  • Driver benchmarking using “ghost data” from past races
  • AI agent competition against real-world data

Future Work

  • Integration with reinforcement learning agents for adaptive racing strategies.
  • Multi-UI leaderboard with real-time visual overlays (Map + Leaderboard + Event Feed).
  • Expansion to drone racing, delivery logistics, and autonomous fleets.
  • Developing a public API endpoint for developers to plug in their own “agents” to compete in simulated races.

Acknowledgments

Special thanks to:

  • ProrokLab, University of Cambridge — for open-sourcing the multi-agent simulator framework.
  • OpenF1 Contributors — for providing public telemetry data access.
  • The F1 Tech Community — for inspiring the concept of competitive mobility simulation.

“Replay the past. Compete in the present. Predict the future.”

Share this project:

Updates