How implied probability works
Odds are prices. Implied probability is the break-even frequency encoded by that price, before you decide whether the price is informative or fair.
Reviewed
From price to probability
American odds describe profit relative to a $100 reference. Positive odds show the profit on a $100 stake; negative odds show the stake required to win $100. Decimal odds show the total return for each unit staked. Each can be converted to the same probability scale.
For +150, the formula is 100 / (150 + 100), or 40%. For -150, it is 150 / (150 + 100), or 60%. Decimal 2.50 is 1 / 2.50, also 40%.
Break-even is not a forecast
If you repeatedly accepted +150 on independent events, ignoring practical costs, you would need to win 40% of them to break even. That statement describes the price. It does not establish that a specific event has a 40% real-world chance. To make that claim, you need a separate, well-calibrated estimate and matching settlement rules.
Why markets exceed 100%
Convert both sides of a typical market and the probabilities usually sum above 100%. Two sides at -110 imply 52.38% each, totaling 104.76%. The excess is the sportsbook's displayed margin under a simple proportional interpretation. Comparing one raw side directly with a normalized probability mixes two different scales.
Common mistakes
- Treating 60% as certainty. It still implies failure four times in ten.
- Ignoring vig when comparing opposite sides of one market.
- Comparing prices from different timestamps after news has moved the market.
- Mixing markets with different overtime or settlement rules.
- Assuming a shorter price always means a better opportunity; price and probability must be assessed together.
A reproducible workflow
- Record the exact market, price, sportsbook, and timestamp.
- Convert every mutually exclusive outcome to raw implied probability.
- Inspect the total overround and normalize if a no-vig estimate is required.
- Keep the raw price and normalized estimate separately labeled.
- State uncertainty and avoid turning a distribution into a guarantee.
Try the arithmetic in the implied-probability calculator, then normalize a two-way market with the no-vig calculator.