Poisson Distribution for Football Predictions: Mathematical Edge
Updated April 2026 | By WinFulltime Team | 18 min read
The Poisson distribution is a mathematical model that predicts the probability of events occurring over a fixed time period. In football, it can predict goal distributions and help you find value in the odds.
What is Poisson Distribution?
Poisson models the probability of a given number of events (goals) occurring when the average rate is known. It assumes goals occur independently and at a constant average rate.
P(k goals) = (λ^k × e^-λ) / k!
λ = expected goals (average), k = number of goals, e = 2.718
Building Your Poisson Model
Step 1: Calculate Expected Goals (xG)
Use xG stats to determine each team's average goals per game. Sources: StatsBomb, Understat, SofaScore.
Example: Man City vs Arsenal
- Man City home xG: 2.1
- Arsenal away xG conceded: 1.3
- Adjusted xG for Man City: (2.1 + 1.3) / 2 = 1.7
Step 2: Calculate Probability for Each Score
Use the Poisson formula to calculate probabilities for 0, 1, 2, 3+ goals.
P(0 goals) = e^-1.7 × 1.7^0 / 0! = 18.3%
P(1 goal) = e^-1.7 × 1.7^1 / 1! = 31.1%
Step 3: Create Scoreboard Matrix
Calculate probabilities for every scoreline (0-0, 1-0, 2-1, etc.) by multiplying home goal probability × away goal probability.
Step 4: Derive Market Probabilities
Add up relevant scores to get probabilities for:
- 1X2: Sum win/draw/loss scores
- Over 2.5: Sum all scores with 3+ total goals
- BTTS: Sum scores where both teams score
Comparing Model to Bookmaker Odds
Your Model vs Bookmaker
| Outcome | Your Prob | Bookmaker | Value? |
|---|
| Over 2.5 | 58% | 1.83 (54.6%) | ✅ Yes |
| Man City win | 52% | 1.75 (57%) | ❌ No |
Improving Your Poisson Model
- Attack/Defense ratings - Adjust for home/away
- Recent form weighting - More weight to recent games
- Expected vs actual goals - Look for regression
- Starting lineups - Missing players affect xG
- Rest days - Fatigue matters
💡 Key Insight: The difference between good and great Poisson models is in the xG inputs. Better data = better predictions.
Tools for Poisson Betting
- Understat - Free xG data
- StatsBomb - Advanced xG metrics
- Football-data.co.uk - Historical stats
- Oddschecker - Compare your odds to bookmakers
Limitations of Poisson
- Assumes independence (goals don't affect each other)
- Doesn't account for key player absences
- Can miss tactical changes
- Historical data may not predict future
Verdict: Does Poisson Work?
Yes, for probability estimation. Poisson is the foundation of most professional betting models:
- ✅ Provides mathematical framework for predictions
- ✅ Can identify genuine value vs bookmaker odds
- ✅ Works well for goal markets
- ⚠️ Requires quality xG data
- ⚠️ Must be combined with proper bankroll management
Start with simple xG inputs and improve your model over time.