Short answer: The Poisson distribution turns the expected number of goals for each team into the probability of every possible score. From those score probabilities you can calculate fair odds for the match result, over/under, both teams to score and correct score – and compare them with the bookmaker's prices to find value bets.
Why Poisson works for football
Goals are rare, roughly independent events spread across 90 minutes – exactly the situation the Poisson distribution describes. Given an average (λ, "lambda"), it gives the probability of 0, 1, 2, 3… events:
P(k goals) = λ^k × e^(−λ) ÷ k!
You do not need to calculate this by hand. In Excel or Google Sheets: =POISSON.DIST(k, λ, FALSE).
Step 1: get the league averages
You need the average goals per match scored by home teams and by away teams in the league. In our example:
- League average home goals: 1.55
- League average away goals: 1.20
Step 2: calculate attack and defence strength
Compare each team with the league average:
| Rating | Formula | Example | Result |
|---|---|---|---|
| Home team attack | Home goals scored per game ÷ 1.55 | 1.90 ÷ 1.55 | 1.226 |
| Away team defence | Away goals conceded per game ÷ 1.55 | 1.70 ÷ 1.55 | 1.097 |
| Away team attack | Away goals scored per game ÷ 1.20 | 1.00 ÷ 1.20 | 0.833 |
| Home team defence | Home goals conceded per game ÷ 1.20 | 0.95 ÷ 1.20 | 0.792 |
A rating above 1 means better than average in attack – or worse than average in defence, because the team concedes more.
Step 3: expected goals for the match
Home expected goals = home attack × away defence × league home average Away expected goals = away attack × home defence × league away average
- Home: 1.226 × 1.097 × 1.55 = 2.08
- Away: 0.833 × 0.792 × 1.20 = 0.79
Step 4: probability of each team's goal count
| Goals | Home team (λ = 2.08) | Away team (λ = 0.79) |
|---|---|---|
| 0 | 12.4% | 45.3% |
| 1 | 25.9% | 35.9% |
| 2 | 27.0% | 14.2% |
| 3 | 18.8% | 3.7% |
| 4 | 9.8% | 0.7% |
Step 5: score probabilities
Multiply the two columns. The probability of a 2–1 is 27.0% × 35.9% = 9.7%. The most likely scores in this match:
| Score | Probability |
|---|---|
| 2–0 | 12.2% |
| 1–0 | 11.8% |
| 2–1 | 9.7% |
| 1–1 | 9.3% |
| 3–0 | 8.5% |
The full grid powers correct score betting.
Step 6: from scores to markets
Add up the relevant scores in the grid:
| Market | Probability | Fair odds |
|---|---|---|
| Home win | 67.3% | 1.49 |
| Draw | 19.6% | 5.11 |
| Away win | 13.1% | 7.62 |
| Over 2.5 goals | 54.8% | 1.82 |
| Both teams to score – yes | 47.9% | 2.09 |
| Home -1.5 | 42.8% | 2.34 |
If a bookmaker offers the home win at 1.60 or over 2.5 at 1.95, those prices are above the model's fair odds – candidates for a bet. See the related guides on 1X2, over/under, BTTS and -1.5 handicaps.
How to make the model better
- Use expected goals instead of goals. xG and xGA are more stable than actual goals and predict better.
- Weight recent matches more – form, injuries and new coaches change team strength.
- Use enough data – at least 10–15 matches per team; early-season numbers are noisy.
- Adjust for low-scoring draws. The basic model slightly underestimates 0–0 and 1–1; the Dixon–Coles correction fixes this.
- Add context the numbers miss: suspensions, rotation before European matches, motivation at the end of the season.
Limitations
Poisson assumes the two teams' goals are independent and that scoring rates are constant through the match. In reality, a team that falls behind attacks more, and a team protecting a lead sits deeper. The model is a strong starting point, not a crystal ball – always compare its output with the market and stake sensibly (bankroll management).
FAQ
What is the Poisson distribution in football?
A statistical formula that turns an average number of goals into the probability of each exact number of goals, used to predict scores and fair odds.
How accurate are Poisson football predictions?
They are a solid baseline, often close to market prices for major leagues. Accuracy improves with xG inputs, recent-form weighting and a low-score correction.
What is the Poisson formula in Excel?
=POISSON.DIST(goals, expected_goals, FALSE) gives the probability of exactly that number of goals.
Can Poisson predict over 2.5 goals?
Yes. Add up the probabilities of all scores with three or more goals, or use 1 minus the probability of 0, 1 and 2 total goals.