Poisson Distribution in Football: A Mathematical Approach to Predictions
Understand how the Poisson distribution powers football goal prediction, from calculating expected goals to generating accurate scoreline probabilities.
Admin
The Mathematics Behind Goal Prediction
The Poisson distribution is a probability formula that calculates the likelihood of a given number of events occurring in a fixed interval. In football, the "events" are goals and the "interval" is a 90-minute match. This mathematical foundation powers the most accurate goal prediction models in the world.
Calculating Expected Goals (Lambda)
The key input to the Poisson formula is lambda — the expected number of goals for each team. To calculate this, we combine:
- Team attacking strength: Goals scored divided by the league average
- Opponent defensive weakness: Goals conceded divided by the league average
- League scoring average: The baseline rate of goals in that competition
For example, if a home team has an attacking strength of 1.3 and the away team has a defensive weakness of 1.1, with a league average of 1.4 goals per home team, then lambda = 1.3 x 1.1 x 1.4 = 2.0 expected home goals.
From Lambda to Probabilities
Once you have lambda for each team, the Poisson formula generates probabilities for 0, 1, 2, 3, 4, and 5+ goals. Multiply the home and away probabilities to get exact scoreline probabilities. Sum the relevant cells to get Over/Under and BTTS probabilities.
The Dixon-Coles Correction
Pure Poisson assumes home and away goals are independent, but in practice, low-scoring outcomes (0-0, 1-0, 0-1, 1-1) are slightly correlated. The Dixon-Coles correction adjusts for this, improving draw probability estimates by 1-3 percentage points. Our platform applies this correction automatically.
Limitations of Poisson
Poisson works best for league matches with stable team strength. It can struggle with cup ties, matches involving significant team changes (transfer windows), and situations where motivation heavily skews effort levels. That is why we combine it with ELO and form analysis in our ensemble.
Practical Application
You do not need to calculate Poisson probabilities by hand. Our AI does it for every match, incorporating xG data, injury adjustments, and the Dixon-Coles correction. Check the predictions page to see Poisson-informed picks for today's fixtures.