How do I calculate fair odds?

Divide 100 by the probability in percent. 40% gives 2.50; 60% gives 1.67; 20% gives 5.00.

Are your fair odds bookmaker odds?

No. They come from our model's probability and contain no margin. Bookmaker prices are usually shorter because they include one.

How do I get fair odds from bookmaker prices?

Convert each price to implied probability, divide each by the total so they sum to 100%, then take 1 ÷ each probability.

Do fair odds mean a bet will win?

No. Fair odds only express a probability as a price. A 60% outcome at fair odds of 1.67 still fails four times in ten.

What are fair odds?

By ProbaPredict Data Desk · Last reviewed Fri, 2 Oct 2026

Fair odds are the odds at which a bet would break even in the long run, with no bookmaker margin. They equal 100 divided by the probability: a 50% chance has fair odds of 2.00, 25% has 4.00. On ProbaPredict, "Fair odds (from our probability)" are 100 ÷ our model's percentage, not a bookmaker's price.

What makes odds "fair"?

A fair price is one where neither side has an edge. If an outcome happens 25% of the time and you bet 1 at odds of 4.00 every time, you lose 1 three times and win 3 once over four bets: net zero. So 4.00 is the fair price for 25%. The formula is simply fair decimal odds = 1 ÷ probability, or 100 ÷ the percentage.

How do we calculate our fair odds?

Every probability on ProbaPredict is shown with "Fair odds (from our probability)". That is 100 ÷ our model's percentage, rounded to two decimals. A 62% home win chance becomes 1.61. They are not bookmaker prices, and they carry no margin. Click any fair-odds figure on a match page to open it in the odds converter and see it in fractional and American form.

How do you get fair odds from a bookmaker?

Bookmaker odds include a margin, so their implied probabilities add up to more than 100%. To get fair odds, remove the margin: divide each implied probability by the total, then take 1 ÷ each result. The average 1X2 margin in the closing odds we hold is 5.3%:

Average bookmaker margin in closing odds, by league

League	Matches	1X2 margin	Over/Under 2.5 margin
Premier League	1,950	4.4%	4.9%
Bundesliga	1,566	4.8%	5.2%
La Liga	1,969	4.8%	5.2%
Serie A	1,950	4.9%	5.2%
Ligue 1	1,723	5.0%	5.3%
Championship	2,855	5.5%	5.6%
Eredivisie	1,593	5.7%	6.1%
Primeira Liga	1,592	6.2%	6.0%
Brasileirão Série A	2,177	6.3%	not in source

Source: ProbaPredict analysis of 17,375 matches, updated Fri, 2 Oct 2026. Market-average closing odds from football-data.co.uk. Margin = sum of implied probabilities − 100%.

Worked example 1: from our probability

Our model gives Over 2.5 goals 57%. Fair odds = 100 ÷ 57 = 1.75 (fractional 3/4, American -133). Under 2.5 is 43%, fair odds 2.33. Notice that 1 ÷ 1.75 + 1 ÷ 2.33 = 57% + 43% = 100%: fair odds always form a 100% book.

Worked example 2: from bookmaker odds

Bookmaker 1X2 prices: 1.95, 3.50, 4.20. Implied 51.3% + 28.6% + 23.8% = 103.7%. Margin-free: 49.5%, 27.6%, 23.0%. Fair odds: 2.02, 3.63, 4.36. The bookmaker's 1.95 is shorter than its own fair price of 2.02; that gap is the margin at work.

How do fair odds help?

They let you compare prices on a level footing. If a bookmaker offers more than our fair odds, its price implies a lower chance than our model does. The value bet calculator expresses that as an edge. Keep two things in mind: our model is an estimate (closing odds are slightly more accurate in our back-test, see our accuracy guide), and a positive edge does not mean the bet will win.

UK and US terminology

UK: "fair odds", "true odds" or "margin-free odds". US: "no-vig odds" or "fair line". All mean the same thing.

Common mistakes with fair odds

Confusing our fair odds with a bookmaker's offer. We don't publish bookmaker prices.
Thinking fair odds are a recommendation. They are a probability written as a price.
Leaving the margin in. 1 ÷ a bookmaker's price is not a fair probability.
Over-trusting precision. 1.61 vs 1.64 is within the model's uncertainty.

How precise are fair odds?

Fair odds inherit the uncertainty of the probability behind them. A 55% estimate might reasonably be 52% or 58%, which corresponds to fair odds anywhere from about 1.72 to 1.92. That's a wide range compared with the small differences between bookmakers' prices. So treat our fair odds as a central estimate rather than an exact line, and be cautious about reading meaning into gaps of a few hundredths. Our calibration tables, on the accuracy and methodology pages, show how well the underlying probabilities have held up in practice.

Why might our fair odds differ from the market?

Our fair odds come from a results-based statistical model. The market's prices reflect far more information: team news, injuries, line-ups, and the views of many bettors. When the two disagree, the market is more often right than our model, which is what our back-test shows. Sometimes the model is slow to notice a real change in a team; sometimes the market overreacts to a recent result.

Seeing the gap is still useful. It shows where a purely statistical view differs from the consensus, and it lets you check any price against a transparent, published baseline. Just remember that a difference is not a signal to bet, and that our fair odds come with their own uncertainty.

Frequently asked questions

How do I calculate fair odds?: Divide 100 by the probability in percent. 40% gives 2.50; 60% gives 1.67; 20% gives 5.00.
Are your fair odds bookmaker odds?: No. They come from our model's probability and contain no margin. Bookmaker prices are usually shorter because they include one.
How do I get fair odds from bookmaker prices?: Convert each price to implied probability, divide each by the total so they sum to 100%, then take 1 ÷ each probability.
Do fair odds mean a bet will win?: No. Fair odds only express a probability as a price. A 60% outcome at fair odds of 1.67 still fails four times in ten.

18+. Gambling can be addictive; only bet what you can afford to lose. Free help: BeGambleAware.org · GamStop.