Correct Score Betting: Settlement and Probability

Exact means exact

The order is home-away unless the product states another convention. The qualifying period must also match. Betfair's football rules provide one operator example of regulation-time football settlement.

Illustrative independent-Poisson calculation

Suppose a deliberately simplified model uses home goal mean 1.6 and away goal mean 1.1. The OpenStax Poisson chapter gives the probability mass function for a count under its stated assumptions.

Illustrative component probabilities are:

• P(home scores 2) = 0.2584
• P(away scores 1) = 0.3662

If the two team counts are treated as independent:

Probability for a 2-1 score = 0.2584 * 0.3662 = 0.0946

The illustrative fair decimal price before any margin is:

Fair price = 1 / 0.0946 = 10.5708

Independence is an assumption, not a football fact. OpenStax explains the independence condition.

Low-score dependence and validation

Dixon and Coles proposed a football score model that modifies selected low-scoring probabilities (Dixon and Coles, 1997). Its existence does not validate the illustrative means or produce a timeless score forecast. Current inputs, time ordering, calibration, and complete held-out evaluation are still required.

Account for the full score field

Suppose the listed score probabilities sum to 0.88. The residual outcome mass is:

Residual probability = 1 - 0.88 = 0.12

That 12% might sit in higher scores or an Other selection. Removing it and renormalising the visible scores changes the probability question. Compare prices only after the field and residual handling are explicit.

Return and sensitivity

For an illustrative GBP 5 stake at decimal odds 12.00:

Gross return = 5 * 12 = 60

Net profit = 60 - 5 = 55

At model probability 0.0946, one-unit expected value follows the probability-weighted method in OpenStax:

EV = (0.0946 * 11) - (0.9054 * 1) = 0.1352

At probability 0.075, the same price is negative:

EV = (0.075 * 11) - (0.925 * 1) = -0.10

Small probability changes have a large effect at long decimal prices.

Audit checklist

1. Confirm home-away score order and settlement period.
2. Preserve every displayed score plus Other outcomes.
3. State the goal model, parameter timestamp, and dependence treatment.
4. Verify that probabilities across the complete score grid sum to one.
5. Test calibration and proper scores on later fixtures.
6. Use the accepted price and stake, including any dead-heat or promotional rule.

Test the full score grid

Set a maximum modelled score for each team and retain an explicit tail cell for every omitted higher score. Verify that all displayed cells plus the tail sum to one. If the grid is normalised after truncation without the tail, low scores receive too much probability.

Run symmetry tests with equal team parameters: P(1-0) should equal P(0-1), P(2-1) should equal P(1-2), and home and away score marginals should match. With unequal parameters, confirm that summing every cell in a row recovers the home-goal marginal and summing every cell in a column recovers the away-goal marginal.

Check the official score, qualifying period, awarded-result rule, and accepted market field after settlement. Preserve an Other result even when it loses. A correct-score evaluation that records only named selections cannot demonstrate whether the model allocated probability to the complete outcome space.

Next step

Use Correct Score Betting for the next part of this topic.

Continue learning

• Next guide: Double Chance Betting
• Related guide: Draw No Bet Explained
• Definition: Correct Score Betting

Assumptions and limitations

The means, probabilities, and prices are illustrative. DraftKings' soccer rules and Betfair are operator examples. Awarded matches, abandoned fixtures, extra time, and scorecast fallbacks require the accepted product wording.