Once you understand the basic Poisson distribution, applying it to specific betting markets is straightforward. This guide walks you through calculating over/under probabilities and correct score odds using xG data.
Building Your Probability Grid
The foundation is a simple grid showing all possible scorelines and their probabilities.
Step 1: Set up your grid Create a table with possible home scores (rows) and away scores (columns), typically ranging from 0-5 for each team.
Step 2: Get xG values Determine home and away team xG. Use recent averages (last 10-15 matches) or full-season data depending on whether you're emphasising form or underlying quality.
Step 3: Calculate cell probabilities Use the Poisson formula for each scoreline. In Excel/Google Sheets, this looks like:
=POISSON(home_goals, home_xg, FALSE) * POISSON(away_goals, away_xg, FALSE)
For example, the probability of a 2-1 scoreline with home xG of 1.5 and away xG of 0.9:
=POISSON(2, 1.5, FALSE) * POISSON(1, 0.9, FALSE) = 0.258 * 0.368 = 0.095 (9.5%)
Step 4: Verify totals Sum all probabilities should equal approximately 100%. Minor rounding is normal.
Practical Example Grid
Let me show a simple grid. Assume Team A (home) has 1.6 xG and Team B (away) has 1.0 xG.
Your grid might look like:
Away 0 Away 1 Away 2 Away 3
Home 0 6.2% 6.2% 3.1% 1.0%
Home 1 9.9% 9.9% 5.0% 1.6%
Home 2 7.9% 7.9% 3.9% 1.3%
Home 3 4.2% 4.2% 2.1% 0.7%
From this grid, you can derive:
- Home wins: Sum all cells where Home score > Away score = 36.8%
- Draws: Sum diagonal cells = 24.0%
- Away wins: Sum all cells where Away score > Home score = 14.3%
- Over 2.5: Sum all cells where Home + Away > 2.5 = 52.1%
- Under 2.5: Sum all cells where Home + Away ≤ 2.5 = 47.9%
Calculating Over/Under Goals
Over 2.5 Goals
Sum probabilities for scorelines where combined goals exceed 2.5 (meaning 3 or more total goals):
From our example: 1-2, 1-3, 2-1, 2-2, 2-3, 3-0, 3-1, 3-2, etc.
Sum these probabilities to get over 2.5 likelihood.
Over 1.5 Goals
Sum all scorelines with 2+ total goals. This is typically higher than over 2.5 since it includes 0-2, 1-1, 2-0, etc.
Under Totals
Subtract your over probability from 100% to get under probability.
Converting to Odds
If your calculation shows 52.1% probability for over 2.5, the fair odds are:
Fair odds = 100% / 52.1% = 1.92
If bookmakers offer 1.95, there's marginal value on over 2.5. If they offer 1.85, the bet is not worth taking.
Correct Score Betting
Poisson directly generates correct score probabilities, making this the most transparent application.
From our grid, 2-1 (home win 2-1) appears at 7.9%. If bookmakers offer 9.0, this is value. If they offer 6.5, it's not.
However, note these probabilities in isolation. If you're placing multiple correct score bets in the same match, calculate the total value across all selections.
Most Likely Scorelines
In our example:
- 1-0: 9.9%
- 1-1: 9.9%
- 2-0: 7.9%
- 0-0: 6.2%
- 2-1: 7.9%
The most likely scorelines cluster around these, with everything else being 5% or less.
Advanced Adjustments
Home Advantage Adjustment
Rather than using raw xG, some bettors adjust for home advantage:
- Add 0.3 to 0.4 to home team xG
- Subtract 0.2 to 0.3 from away team xG
This reflects that home teams tend to outperform their underlying xG.
With our example, adjusted xG might be 1.9 (home) and 0.8 (away), which generates slightly higher home win probability.
Draw Adjustment
Some teams naturally produce more draws. If a team has 15 draws in 38 matches, they're 39% of the time a draw. Poisson typically underestimates this.
Apply a small upward adjustment to draw probability if the team is draw-prone.
Correlation Adjustment
Poisson assumes independence. When one team scores, the other is slightly more likely to score in response. This correlation adjustment slightly increases 1-1 and 2-2 probabilities whilst decreasing 0-0 and 3-0 probabilities.
Most bettors don't apply this refinement, but it improves accuracy slightly.
Building Your Calculator
Spreadsheet Approach
Use a formula-driven spreadsheet with:
- Input cells for home xG and away xG
- Grid of scoreline probabilities using POISSON functions
- Derived calculations for win/draw/loss, over/under at various thresholds, correct scores
- Comparison cells where you input bookmaker odds and calculate edge
This takes 30 minutes to build and provides a repeatable tool for every match.
Online Tools
Several websites offer free Poisson calculators. Search for "Poisson calculator football" to find them. These are useful for quick checks but less flexible than building your own.
Common Calculation Mistakes
Using 0 or 1 for xG values: Teams with 0.5 xG have roughly 60% probability of scoring 0 goals and 30% for 1 goal. Using incorrect values generates dramatically different results.
Forgetting to multiply: To get the probability of a specific scoreline, multiply the home Poisson probability by the away Poisson probability.
Rounding prematurely: Round only at the final step. Rounding intermediate probabilities compounds errors.
Using wrong base xG: xG from a single match is noisy. Use rolling averages of 5-15 matches for reliable input.
Forgetting context: A 0.1% probability scoreline (5-5 with teams averaging 1.5 xG) might still happen. Don't use Poisson as gospel.
Testing Your Model
To validate your Poisson calculator:
- Run Poisson on 20-30 recent matches
- Compare predicted win/draw/loss percentages to actual results
- Calculate how often over 2.5 actually happened versus your prediction
- Note any systematic biases (Poisson always overestimating draws, etc.)
- Adjust for these biases going forward
This calibration process improves accuracy substantially.
Practical Betting Application
Before betting:
- Calculate Poisson probabilities for the match
- Identify which markets offer value (bookmaker odds don't align with your probability)
- Focus only on markets with 5%+ edge
- Ensure your kelly fraction bankroll management (never risk more than the edge suggests)
In Summary
- Poisson calculations provide direct, objective probabilities for over/under goals and correct scores.
- Building a spreadsheet tool takes minimal time and provides repeatable analysis for every match.
- The key is using accurate xG inputs and understanding where value exists relative to bookmaker odds.
- Start simple with just over 2.5 and under 2.5 calculations.
- Once comfortable, expand to correct scores and Asian handicaps.
FAQs
What xG should I use if teams haven't played recently? Use their average xG across the full season. Alternatively, use their average against similar-quality opposition if this season's schedule is skewed.
How accurate is Poisson overall? Very accurate for distribution shape across many matches. Less accurate for individual matches due to variance. Use for edge identification, not match prediction.
Should I weight recent form more than season average in xG? Depends on your preferences. Recent form (last 5-10 matches) better captures current state but with higher noise. Full season is more stable but less current. Many bettors use a blend.
How much does home advantage change Poisson? A 0.3 xG adjustment changes the probability distribution noticeably. Home win probability increases 3-5 percentage points depending on the specific match.
Can I apply Poisson to leagues other than the Premier League? Yes, Poisson works across all leagues. The xG calculation methods are similar across leagues, though some variation exists.
What's the best threshold for over/under in your model? Test against historical data. Most matches fall in the 2-3 total goals range, making over 2.5 and under 2.5 the most commonly traded. Other thresholds can offer value depending on specific matches.