Football Predict Poisson Distribution

The Poisson distribution approach for football predictions is a common method used to predict outcomes such as match scores, total goals, and other goal-based markets. The process, while mathematically driven, offers bettors a way to create a predictive model that provides their own calculated probabilities and odds. Here’s a summary of the key steps to create such a model:

1. Understanding Poisson Distribution:

• Poisson distribution helps calculate the probability of a given number of events (goals) happening within a fixed period (match duration), based on the known average.
• This approach works well for goal-based markets like match outcomes, over/under goals, both teams to score, and correct score markets.

2. Building the Model:

• Step 1: Choose a league. Start with one league you know well to simplify testing.
• Step 2: Gather historical data. Use sites like Soccerway to collect data (goals scored/conceded by teams).
• Step 3: Calculate averages. Find league averages for goals scored at home and away, then calculate individual team averages for goals scored and conceded.
• Step 4: Compute strengths. Use the league averages to determine each team’s attacking and defensive strength at home and away.
• Step 5: Compute goal expectancy. For any match, multiply the attacking strength of one team by the defensive weakness of the opponent to get the goal expectancy.

For instance, if Team A has an attacking strength of 1.2 at home and Team B has a defensive strength of 1.07 away, the goal expectancy for Team A is:Team A’s Goal Expectancy=1.2×1.07×league average home goals\text{Team A’s Goal Expectancy} = 1.2 \times 1.07 \times \text{league average home goals}Team A’s Goal Expectancy=1.2×1.07×league average home goals

3. Calculating Probabilities Using Poisson:

• Step 6: Use Poisson in Excel. Once you have goal expectancy values for each team, use Excel’s Poisson function to calculate the probability of each possible scoreline. Poisson(x,goal expectancy,FALSE)\text{Poisson}(x, \text{goal expectancy}, \text{FALSE})Poisson(x,goal expectancy,FALSE) The result gives you the probability of each team scoring exactly ‘x’ goals.
• Step 7: Create odds. By summing up the probabilities for different outcomes (e.g., all home wins or under 2.5 goals), you can create your own odds for the event. Convert probability into decimal odds using: Decimal Odds=100Probability (%)\text{Decimal Odds} = \frac{100}{\text{Probability (\%)}}Decimal Odds=Probability (%)100​

4. Finding Value:

• Step 10: Compare with bookies’ odds. The key to value betting is finding where your odds are more favorable than those offered by bookmakers. Place bets when your calculated odds suggest a higher chance of winning than the bookmaker’s implied probability.

5. Model Refinement:

• Adjust and improve. As games progress, continually update your model with new results, removing outdated data if necessary. Refine your inputs and tweak your model as needed.
• Account for external factors. While Poisson doesn’t take injuries, weather, or tactical changes into account, these factors are crucial in improving the accuracy of your predictions.

6. Consider Limitations:

• Historic bias: Past results may not always accurately predict future performance (e.g., player changes, manager changes).
• Objective limitations: The model doesn’t account for real-time match events (like red cards or in-game tactics), which can drastically affect outcomes.
• Zero inflation: The Poisson distribution tends to underestimate the probability of 0-0 draws and other low-score games. Adjust for this by using methods like zero-inflation to correct bias.

By following these steps and refining your model, you can potentially identify value betting opportunities based on your own odds predictions. However, it’s important to recognize the limitations and continually adapt your approach as new data emerges.