Implied odds after the sportsbook's margin has been stripped out, summing to exactly 100%.
Raw sportsbook prices include the vig. The implied probabilities of all outcomes sum to more than 100% — typically 104-106% on a two-way market, more on parlays and futures. No-vig fair odds strip that excess out and present the cleanest estimate of the market's actual belief about the probability of each outcome.
The math is straightforward. Compute the implied probability of each side. Divide each by the sum of all implied probabilities. The result is a set of probabilities that sum to exactly 100% — the market's no-vig fair probabilities. Convert each back to American or decimal odds and you have no-vig fair odds.
The use case is comparison. If a model thinks a side is 56% and the raw line is -110 (52.4% implied), the model has a 3.6-point edge on the raw price. But the no-vig fair probability of that side might be 54% — the other side's price tells you the market's actual estimate. The true edge is 2 points, not 3.6.
Most serious analytics-based bettors run their EV math against no-vig fair odds rather than raw prices. The CLV calculation does the same thing. Books that publish closing lines effectively publish the market's final no-vig probability estimate — beat that estimate often enough and you have positive expected value.
We publish a no-vig fair probability for every game in our coverage. Our edge calculations compare model probability to that fair probability, not to the offered line. Otherwise we'd count book margin as edge.