Odds explained

How to calculate implied probability from football odds

Divide one by decimal odds to find the raw implied probability. For a complete football market, the raw probabilities usually exceed 100% because the prices include bookmaker margin.

· 7 min read

Convert decimal odds with one division

Implied probability = 1 ÷ decimal odds

Multiply the decimal result by 100 to express it as a percentage. This is the raw probability implied by that one price.

The formula translates a price into a break-even rate before accounting for bookmaker margin. It does not prove the outcome has that true chance.

Examples at 2.00 and 2.50

Why a three-way football market exceeds 100%

Match Result has home, draw and away outcomes. Imagine prices of 2.00, 3.60 and 4.00. Their raw implied probabilities are 50.0%, 27.8% and 25.0%, which total 102.8%. The extra 2.8 percentage points are the overround in this simplified view.

Overround = sum of raw implied probabilities − 100%

A positive total shows that the quoted outcome prices cannot all be fair probabilities at the same time.

A simple way to estimate fair market probabilities

Fair probability = raw implied probability ÷ total raw probability

Proportional normalisation removes the overround by scaling every outcome by the same factor.

Compare the market with an independent model

A model probability and a normalised market probability are two estimates built from different evidence. Their gap can flag a decision for investigation, but only if the model was validated and the quoted price was current.

Common calculation mistakes

  • Dividing decimal odds by one instead of one by the odds.
  • Treating one raw implied probability as margin-free.
  • Mixing prices captured at different times.
  • Comparing probabilities for different markets or settlement rules.
  • Assuming a price difference guarantees a profitable result.

Limits and responsible use

Proportional normalisation is convenient, not a unique truth about how margin is distributed. Prices move, markets vary in liquidity and every independent model can be wrong.

Sources and further reading