BoxBox

Inspiration

A Formula One Grand Prix is often a math equation at 200 mph. The thrill of guessing whether a team's pit strategy will work out is something we wanted to incorporate with the data. We created an app that simulates how different pit strategies affect the lap times over the course of a race

What it does

BoxBox puts the pit wall in your hands. Users can define the total race distance, select a multi-stop strategy, and pinpoint the exact lap for every trip down the pit lane. The app instantly simulates the entire Grand Prix, projecting lap-by-lap telemetry so you can stress-test different strategies and shave those crucial seconds off your total race time.

How we built it

The backbone of BoxBox is a robust predictive model trained on a massive historical dataset spanning every Grand Prix from 1950 to 2024. We implemented a Random Forest Algorithm to parse the complex relationships between tire age, compound degradation, and race-lap progression. By running a plethra of parallel simulations and averaging the outcomes, the model filters out "noise" to deliver highly accurate lap-time forecasts.

Accomplishments that we're proud of

We are incredibly proud of our core engine, which processes a massive historical dataset to deliver a robust predictive model. By analyzing performance trends across hundreds of Grand Prix starts, we’ve moved beyond simple estimates to high-fidelity lap time simulations.

We successfully balanced complex data with a "driver-first" UI. We’re proud of building an interface that handles sophisticated telemetry behind the scenes while remaining intuitive and accessible for the end user.

Our model doesn't just look at one lap; it understands the "long game." We’ve created a system capable of identifying subtle performance patterns across different tracks and seasons, giving users a truly panoramic view of race strategy.