The No-Vig Ritual Before Every Argument

Two bettors argued for ten minutes. The book won during the argument. #

A friendly Sunday debate. One bettor likes the home side at -110; the other likes the road side, also at -110. They argue for ten minutes. Both believe the other has bad reasoning. Neither has yet removed the house edge from the discussion. The book is currently holding 4.55% on the market and is content to let the conversation continue indefinitely. Every minute spent comparing personal probabilities to posted implied probabilities is a minute spent comparing apples to apples wearing the book’s hat.

The ritual is short, repeatable, and not interesting to look at. Convert both sides of the market to implied probability. Remove the hold by normalizing to 100%. The remaining number is the no-vig fair price — the cleaner market opinion. Then, and only then, compare your personal probability to it. If your number beats the no-vig number by enough to cover the vig and your uncertainty, you have a bet. If not, you have an argument. The argument is free; the bet is not.

What "no-vig" actually is #

Posted American odds imply a probability. A -110 line implies 52.38%. A +110 line implies 47.62%. The two implied probabilities for the same market sum to more than 100%, because both numbers include a slice of the book’s hold. The no-vig conversion removes the hold by normalizing: divide each implied probability by their sum. The result is the cleaner two-sided estimate of true probability, as the book would set it without margin.

A -110/-110 market converts to a no-vig 50/50. A -150/+130 market converts to a no-vig 57.6% / 42.4%, not 60.0% / 43.5%. The difference is not cosmetic. A bettor who estimates the home side at 58% true probability has a +EV bet against the no-vig number (57.6%) but a -EV bet against the posted number (60.0%). The same opinion is profitable or unprofitable depending on which number you compared against.

Hold widens as the market gets more exotic #

Spread and total markets at -110/-110 carry the standard 4.55% hold. Player props carry a median 6.1% hold. Alternate spreads carry 7.4%. Three-leg same-game parlays carry roughly 14.2%. The math gets worse the further the bettor walks from the main markets, and the bettor walks further every season because the apps surface the exotic markets at the top of the page. The hold tax is highest precisely where the book displays the bets most prominently.

The two-step ritual #

Step one: capture both sides of the market from one of the lower-margin books (Pinnacle if available, Circa, or a no-vig calculator fed posted American odds). Step two: convert to no-vig and compare your personal probability against that. Anything less is a debate with the sportsbook’s margin still on the table. The ritual takes about thirty seconds with a calculator, which is roughly twenty-eight seconds less than the average pregame argument between two bettors who already agree on the team and disagree on the price.

A small worked example, since the calculator is rude until you have used it once. NFL game posted at home -150 / road +130. Sum of implied probabilities: 60.0% + 43.5% = 103.5%. No-vig home probability: 60.0 / 103.5 = 57.97%. No-vig road probability: 43.5 / 103.5 = 42.03%. Hold: 3.5%. The bettor who estimated the home side at 60% true probability does not have a +EV bet at -150; the bettor only has a +EV bet if the personal estimate beats 57.97% by enough to cover the vig already paid in the price.

The no-vig ritual will not make you brilliant. It will make you less available to obvious nonsense, which is a respectable first step in a business full of louder alternatives.

Where the ritual matters most #

Player props. Alt spreads. Live markets. Anywhere the hold is above 6%, the gap between implied probability and no-vig fair probability is large enough to flip a bet from +EV to -EV without changing the underlying opinion at all. A bettor who estimates a 56% true win probability on a player prop priced at -125 has a +EV bet against the implied 55.6% but a -EV bet against the no-vig 53.8%. The ritual is the only way to catch the difference before the bet is placed.

The ritual also matters most for the bettor with a thin edge. A bettor whose edges average 3-4% over fair probability has very little room to absorb hold without giving back the entire bet. A bettor whose edges average 8-10% has more margin for error but still loses meaningful EV per ticket by skipping the conversion. There is no edge tier at which the no-vig ritual is optional. It only becomes more or less expensive to skip.

Tools and minimal infrastructure #

A no-vig calculator is not optional infrastructure. Either a spreadsheet with the two-way normalization formula (=implied_a / (implied_a + implied_b)) or a phone-friendly web calculator works. The Desk publishes one. So does Pinnacle. So does any reputable bankroll-tools site. The point is to have the calculator open when the bet is being considered, not after the ticket has cleared. The ritual is preventive, not diagnostic.

A second piece of infrastructure: a habit of reading the no-vig number aloud before placing the bet. This sounds absurd. It is also the cheapest available defense against the bettor’s own pattern matching. A number that looked fine when it lived inside the bettor’s head will, half the time, sound a little off when said out loud against the no-vig fair price. That half-second of reconsideration is worth more over a season than most of the analytical work that preceded it.

Common mistakes that survive the ritual #

Even after the no-vig conversion is computed correctly, two mistakes persist. The first is treating the no-vig fair price as if it were the true probability. It is not. The no-vig fair price is the book’s two-sided estimate of true probability after the book has removed its own margin. The book’s estimate is approximately efficient on liquid markets and meaningfully inefficient on thin markets, but in neither case is it identical to truth. A bettor who beats the no-vig number by half a point still has a real edge; a bettor who matches the no-vig number has no edge at all and is paying the vig for nothing.

The second persistent mistake is computing the no-vig from a single book rather than from the sharpest available book. A retail book with a high hold will produce a no-vig estimate biased toward the book’s liability side, because the retail book is not in the price-discovery business; it is in the customer-acquisition business. The cleanest no-vig conversions come from Pinnacle, Circa, or — when those are not available — the median of three or more retail books. Using the no-vig from a single retail book is a procedural improvement over using the posted implied probability, but it is not a complete fix.

A third subtlety, for the bettor who has internalized the first two: the no-vig conversion assumes a two-sided market. Three-way markets (regulation moneyline, draw included) require a different normalization, and many bettors apply the two-sided formula to them anyway. The result is a no-vig estimate that is systematically off by 1-3 percentage points. The fix is the three-sided normalization: divide each implied probability by the sum of all three, not by the sum of the two main sides. This matters for soccer, hockey three-way lines, and any NFL market that includes a tie possibility.

The closing argument #

The no-vig ritual is mechanical, fast, and not glamorous. That is the entire reason it works. It removes the most common pricing error in retail betting — comparing a personal probability to a posted implied probability — and replaces it with the slightly less common but still pervasive error of comparing a personal probability to a no-vig fair probability that has been correctly computed. The second error is much smaller than the first.

Do the conversion before the argument. Take the result seriously. If your number does not beat the no-vig number, the debate is academic and the wallet is the silent winner of the discussion — usually for the other party. The arithmetic does not need to be flattered. It needs to be used.