How to devig odds
The math for removing the vig from betting odds to get fair probabilities, with a worked example and API support.
Convert both sides of a two-sided market to implied probability, then divide each by their sum. At -110/-110 each side implies 52.4 percent, summing to 104.8, so the fair probability is 50 percent and the extra 4.8 points are the vig. ParlayAPI's consensus endpoint returns no-vig fair lines from real book prices.
More detail
The division step is the multiplicative devig, the standard baseline method. It assumes the book spreads its margin proportionally across outcomes, which is a reasonable default for near-even two-sided markets. For heavy favorites, methods like the power devig can model the favorite-longshot bias better, but the multiplicative method is where every implementation should start.
Devigging matters because a fair probability is the reference point for everything downstream: expected value calculations compare a soft book's price against a devigged sharp price, and no-vig closing lines are the honest benchmark for closing line value.
GET /v1/sports/{sport_key}/consensus computes no-vig fair lines across books server side, always from real published prices. To experiment by hand, the free calculator at /tools/devig-calculator shows each step of the arithmetic.
Endpoint
GET /v1/sports/{sport_key}/consensus
curl 'https://parlay-api.com/v1/sports/basketball_nba/consensus?markets=h2h' -H 'X-API-Key: YOUR_KEY'
Why developers use it
- No-vig fair lines computed from real published prices
- Free step-by-step devig calculator at /tools/devig-calculator
- Feeds EV and CLV workflows on the same key
FAQ
What is vig?
The bookmaker's margin built into prices. When implied probabilities on a market sum above 100 percent, the excess is the vig. At -110/-110 the sum is 104.8 percent, so the vig is 4.8 points.
Which devig method should I use?
Start with multiplicative: divide each implied probability by the sum. Consider power or shin methods for markets with heavy favorites, where the proportional assumption is weakest.
Can the API devig for me?
Yes. GET /v1/sports/{sport_key}/consensus returns no-vig fair lines computed from real book prices, so your model consumes fair probabilities without reimplementing the math.
Related
/glossary/devig/tools/devig-calculator/answers/positive-ev-api/glossary/no-vig-price