Can You Predict Soccer Match Outcomes with a Simple Poisson Model?

Can we predict the result of a football match before it starts?

Using a probability‑based model that incorporates each team's offensive and defensive strengths and a home‑field factor, we can estimate likely outcomes for any pair of teams.

League Data

The league consists of 20 teams (A–T) with known points, goals scored and goals conceded.

Calculating Averages

Average goals per team = total goals ÷ 20 = 45.95. Average goals conceded per team = total conceded ÷ 20 = 45.95.

Offensive and Defensive Indices

Offensive index = team goals ÷ average goals. Defensive index = team conceded ÷ average conceded.

For example:

Team A offensive = 67 / 45.95 = 1.458

Team A defensive = 24 / 45.95 = 0.522

Team B offensive = 74 / 45.95 = 1.610

Team B defensive = 26 / 45.95 = 0.566

Home‑Away Adjustment

Home factor = 1.36, away factor = 1.06.

Expected goals:

A (home) vs B = 1.36 × 1.458 × 0.566 = 1.122

B (away) vs A = 1.06 × 1.610 × 0.522 = 0.892

Using the Poisson Distribution

The Poisson formula P(k; λ) = (e⁻ˡ λᵏ) / k! gives the probability of scoring k goals when the expected number is λ .

Applying it to the expected goals above yields probability tables for 0, 1, 2, … goals for each team.

Combining the two distributions provides the probability of each possible scoreline.

Predicted Score Probabilities

0‑0: 13.35%

1‑0: 14.98%

2‑0: 8.40%

0‑1: 11.90%

1‑1: 13.36%

2‑1: 7.49%

0‑2: 5.31%

1‑2: 5.95%

2‑2: 3.34%

The most likely result is a 1‑0 win for the home team, followed by 0‑0 and 1‑1 draws.

While no model can guarantee perfect accuracy due to unpredictable factors (weather, injuries, red cards), this statistical method offers a data‑driven way to assess match outcomes.