Mathematical Predictions

How Our Mathematical Model Works

Our prediction engine combines multiple statistical approaches. At its core is a Poisson distribution model that estimates the probability of each possible scoreline based on each team's attacking and defensive strength. These strength ratings are derived from historical match data, weighted towards recent performance.

The model also incorporates expected goals (xG) data, home advantage factors from our home/away statistics, and current form trends. The output is a probability distribution for match outcomes (home win, draw, away win) and predicted goals, which we compare against bookmaker odds to identify value.

Reading the Prediction Bars

Each match card shows a win probability bar split into three segments: home win (green), draw (grey), and away win. The wider a segment, the higher the probability. These probabilities always sum to 100%.

Click any match to view the full fixture page with detailed statistics, head-to-head data, and available odds. To identify value, compare our probability with the bookmaker's implied probability — use our odds converter to translate decimal or fractional odds into a percentage. If our model gives Home at 60% but the odds imply 50%, the home win may be a value bet.

Model Limitations

No model captures everything. Ours does not account for individual player absences (injuries, suspensions), managerial changes, or motivational factors like relegation battles or title run-ins. It also performs best in leagues with sufficient historical data — early-season predictions or matches in smaller leagues carry wider uncertainty margins. Always cross-reference with our full statistics and your own football knowledge for the most informed decision.