How to Convert Decimal Odds Into Implied Probability and Fair Odds
Published July 17, 2026 · 5 min read
Decimal odds look like payout numbers, but underneath they are probabilities. Once that connection becomes clear, prices such as 1.50, 2.10 or 4.00 become much easier to compare with football data.
Two formulas do most of the work. One converts a market price into implied probability. The other converts an estimated probability back into fair odds.
Decimal odds to implied probability
Use this formula:
Implied probability = (1 / decimal odds) x 100
At odds of 2.00:
1 / 2.00 = 0.50, or 50%
At odds of 1.50:
1 / 1.50 = 0.6667, or 66.7%
At odds of 3.20:
1 / 3.20 = 0.3125, or 31.25%
Decimal odds
Raw implied probability
1.25
80.0%
1.50
66.7%
1.80
55.6%
2.00
Frequently asked questions
How do you convert decimal odds to probability?
Divide 1 by the decimal odds and multiply by 100. Odds of 2.00 imply 50%, while odds of 1.50 imply about 66.7%.
How do you calculate fair odds?
Divide 1 by your estimated probability expressed as a decimal. A 40% probability produces fair odds of 2.50.
Why do implied probabilities add up to more than 100%?
Bookmakers build a margin into their prices. The combined implied probability above 100% is commonly called the overround.
Do shorter odds mean an outcome is certain?
No. Short odds indicate a higher implied probability, not certainty. Upsets and unexpected match events remain possible.
50.0%
2.50
40.0%
3.00
33.3%
4.00
25.0%
5.00
20.0%
The word raw matters because the bookmaker's margin is still inside those probabilities.
Probability to fair odds
The reverse formula is:
Fair decimal odds = 1 / estimated probability
Write the probability as a decimal. If your estimate is 40%, use 0.40:
1 / 0.40 = 2.50
If the estimate is 55%:
1 / 0.55 = 1.82
Fair odds describe the break-even price before bookmaker margin. They are not a guarantee that an estimate is correct. The quality of the output depends entirely on the quality of the probability input.
Why a three-way market exceeds 100%
Imagine a football 1X2 market with these prices:
Outcome
Odds
Implied probability
Home win
2.00
50.0%
Draw
3.40
29.4%
Away win
4.00
25.0%
Total
104.4%
The outcomes are mutually exclusive, so a truly fair set of probabilities must total 100%. The extra 4.4 percentage points represent the market overround in this simplified example. It is one way of seeing the bookmaker's built-in margin.
To estimate margin-free probabilities, divide each raw probability by the total:
Home: 50.0 / 104.4 = about 47.9%
Draw: 29.4 / 104.4 = about 28.2%
Away: 25.0 / 104.4 = about 23.9%
These values now total approximately 100%. Converting them back to prices produces a rough no-margin market.
Actual odds versus fair odds
Suppose a statistical model estimates a 52% probability for an outcome. The fair price is:
1 / 0.52 = 1.92
If the available decimal odds are 2.10, the market price is higher than the model's fair price. If the available odds are 1.70, the market is shorter than the model estimate.
This comparison identifies disagreement. It does not prove that the model is right. Team news may be missing, the sample may be poor or the market may know something the model does not. Fair-versus-actual should begin further research, not end it.
Price movement changes the comparison
Odds are not fixed. A price can move after lineup news, market activity or a change in available information. If your fair estimate stays at 1.92 while the actual price moves from 2.10 to 1.85, the original disagreement has disappeared.
Record the time and source of every price used in analysis. Comparing a morning model with a closing price while pretending both were available together creates a false historical test. Backtesting needs the odds that genuinely existed at the recorded moment.
Break-even rate and recorded results
The implied probability is also a rough break-even rate before practical complications. Odds of 2.10 correspond to 47.6%. Across a large independent sample at the same price, a strike rate above that level would be required to show a positive return before considering variation and execution.
That is why win rate alone can mislead. A 70% success rate at very short odds may perform worse than a 45% success rate at larger prices. Price and frequency must be read together.
Probability is not expected return
Probability estimates how often an outcome would occur across comparable repetitions. Expected return combines that chance with the offered price. A likely outcome can still be poorly priced, while a less likely outcome can create a larger disagreement between model and market.
This is also why labels such as favourite and underdog are incomplete. They describe relative market probability, not whether the listed odds are attractive or whether the underlying estimate is reliable.
Closing price is a useful review point
After a match, compare the recorded price with the closing market. A method that repeatedly identifies prices which later shorten may be detecting information before the broader market, even when individual results vary. The comparison is diagnostic, not proof of future performance. It also requires consistent timestamps and the same settlement rules.
Common calculation mistakes
Treating the raw probability as fair
The simple 1/odds formula includes market margin. Remove the overround when comparing all outcomes in the same market.
Mixing decimal and percentage formats
Use 0.40 in the fair-odds formula, not 40. Dividing 1 by 40 produces the wrong result.
Comparing different market rules
A 90-minute result, draw-no-bet line and qualification market settle differently. Prices cannot be compared until the settlement rules match.
Assuming precision equals accuracy
A model may output 53.7%, but extra decimal places do not make the estimate reliable. Sample quality, inputs and assumptions matter more.
A practical workflow
Convert the available odds into raw probabilities. Check the total market overround. Remove the margin if you need a cleaner market baseline. Then compare that baseline with your statistically derived estimate.
Investigate large disagreements using team news, venue splits, recent performance, xG and sample size. Finally, record the original price and outcome if you are testing the method historically.
Odds are not a verdict. They are a compact market estimate. Understanding the maths makes that estimate easier to question intelligently.
