The vig (or juice) is the bookmaker's margin, built into the prices. Take it out and what is left is the market's fair view of each outcome: the no-vig, or fair, price.
The overround
Add up the implied probabilities of every outcome in a market. Anything over 100% is the overround.
- −110 on both sides: 52.38% + 52.38% = 104.76%, an overround of 4.76%. If the book takes equal money on both sides, it keeps about 4.55% of everything staked whichever side wins.
- −150 and +130: 60.00% + 43.48% = 103.48%, an overround of 3.48%.
The smaller the overround, the less every bet in that market costs you.
Removing the vig
Multiplicative (Unitley's default): divide each implied probability by the total, so they add to 100%. For −150 / +130:
- favorite: 60.00% ÷ 103.48% = 57.98%
- underdog: 43.48% ÷ 103.48% = 42.02%
Those fair chances are fair odds of about −138 and +138.
Power: raise each implied probability to the same power, chosen so they add to 100%. This takes more of the margin off the longshot than off the favorite, in line with the favorite-longshot bias: across many markets, longshots are priced worse than favorites. For −150 / +130 it gives 58.40% and 41.60%. The difference grows with the gap between the prices: for −400 / +300 multiplicative gives 76.19% and 23.81%, power 78.24% and 21.76%.
Neither method is exact. For close two-way markets they agree to within a point; for lopsided ones, the power method is usually the better guess.
Why it matters
- The no-vig chance is the market's best estimate. An edge is your estimate against that, not against the taxed price.
- The vig-included implied probability is still your break-even. To make money, your chance has to beat that.
- A no-vig line built from several books (a consensus) is a stronger estimate than one book's.
In Unitley
calc_no_vig_odds takes every price in one market and returns the overround and each outcome's fair chance and fair odds, with method set to multiplicative or power. get_game_odds and get_player_props already include a no-vig consensus across books, and the EV against it.
Names set like this are Unitley’s tools: what your AI agent, or the analyst in the web app, calls to get these numbers. See the docs