MODEL vs MARKET

Our ML Model vs Real Money Odds

Our machine learning model predicts every F1 race. Kalshi is a CFTC-regulated prediction market where real money is at stake. Here is where we agree — and where we do not.

After each race, we score both: who was closer to reality?

Our Model

An algorithm. It analyses 52 data points on every driver and race: recent form, Elo ratings, circuit type, car reliability, teammate matchups. Then it simulates the race 10,000 times. If Russell wins in 4,110 simulations, we say 41.1%. Pure data. No opinion.

Kalshi

Real people with real money. Kalshi is a CFTC-regulated prediction market. Thousands of traders buy and sell contracts on F1 outcomes. If the price is 58 cents, the market thinks there is a 58% chance. Collective intelligence — sentiment, rumours, and conviction, all priced in.

Our model sees the data. The market sees the sentiment. Where they diverge is where the analysis gets interesting.

Coming soon: Polymarket comparison. Over $87M in F1 trading volume on the world's largest crypto prediction market — once their developer API covers sports, we will compare against both markets.

Full Comparison — Formula 1 Lenovo Grand Prix du Canada 2026

DriverOur ModelKalshiGapPole (Kalshi)WDC (Kalshi)
George Russell40.3%=29%
Kimi Antonelli22.9%=39%
Oscar Piastri12.5%=6%
Lando Norris10.3%=12%
Lewis Hamilton6.6%=2%
Charles Leclerc5.2%=4%
Max Verstappen0.7%=7%

Odds from Kalshi. Updated in real-time. Our model uses Elo ratings, 52 features, and 10,000 Monte Carlo simulations.

Constructors' Championship (Kalshi)

Mercedes

70%

McLaren

21%

Ferrari

7%

Red Bull Racing

2%

What is Kalshi?

Kalshi is a CFTC-regulated prediction market based in the United States. Traders buy and sell contracts on real-world events — including Formula 1 race outcomes. Prices reflect the crowd's collective belief about what will happen.

When Kalshi says Russell has a 58% chance of winning, it means traders have put real money behind that belief. Our model uses machine learning, not market sentiment. Where the two diverge is where the analysis gets interesting.