METHODOLOGY

How our football prediction models work.

1. Data and context

We combine match-level statistics, expected-goal signals, market prices and contextual motive profiles. Missing or low-quality inputs reduce confidence rather than being hidden.

2. Prediction layer

The system ranks markets and selections using repeatable rules. The published pick, odds and confidence form the prediction record.

3. AI explanation layer

AI turns structured evidence into readable analysis. It explains the pick; it does not replace the underlying scoring logic or invent bookmaker odds.

4. Evaluation

Public performance uses settled free picks. Hit rate excludes voids. ROI uses a flat one-unit stake and the odds recorded with each pick.

5. Limits

Football is volatile. Historical performance does not guarantee future returns, and a high-confidence label is not certainty.