True Probability Calculator: How to Strip the Vig and Find Real Odds

Every sportsbook price lies to you by the same amount

A sportsbook line is not a probability estimate. It is a probability estimate plus a tax. That tax is the vig. A -110/-110 market implies 52.38% on each side, totaling 104.76%. The extra 4.76% is what the book keeps. The true probability of each side is closer to 50%, not 52.38%.

The true probability calculator strips that tax. You input the odds from both sides of a market, choose a de-vig method, and get the actual implied probability the line represents. This number is what you compare against your own estimate to find expected value. Without it, you are comparing your probability against a biased number and overestimating how much edge you have.

Run any two-sided or multi-outcome market through the de-vig calculator to get the true probability instantly. The tool supports all 7 methods covered below and handles markets with 2 to 20+ outcomes.

Step 1Get odds from both sides

→

Step 2Choose a de-vig method

→

Step 3Calculate true probabilities

→

Step 4Compare to your estimate

→

Step 5Calculate EV if edge exists

→

Step 6Size the bet with Kelly

How implied probability becomes true probability

Implied probability is what you get when you convert odds directly. -150 implies 60%. +130 implies 43.48%. Add them: 103.48%. That sum exceeding 100% is the overround, and it equals the book's margin on that market.

True probability removes the overround so both sides sum to exactly 100%. The question is how to distribute that removal. Different methods make different assumptions about where the vig is hiding.

Worked example: -150/+130 market. Implied probabilities: 60% and 43.48%. Total: 103.48%. Overround: 3.48%.

Using the simplest method (multiplicative), divide each side by the total:

• Favorite true probability: 60% / 103.48% = 57.98%
• Underdog true probability: 43.48% / 103.48% = 42.02%

Now the sides sum to 100%. The favorite's true probability is 57.98%, not the 60% the line suggested. That 2.02 percentage point difference is the vig embedded in the favorite's price. Use the odds converter to translate these probabilities back into any odds format.

This matters because if your model says the favorite wins 59% of the time, your edge against the true line is only 1.02%, not the 1% you would calculate against raw implied odds. Small differences like this determine whether a bet is +EV or -EV.

The 7 de-vig methods and when each one matters

Not all vig removal is equal. The method you choose changes the output probability, sometimes by 1-2 percentage points. On tight edges, that is the difference between a profitable bet and a losing one.

1. Multiplicative (Basic)

Divides each implied probability by the total overround. Simple and fast. Assumes the book adds vig proportionally to each side.

Best for: Quick estimates, evenly priced markets (-110/-110), casual analysis.

Formula: True P = Implied P / Sum of all implied P

2. Additive

Subtracts an equal amount from each side. If the overround is 4%, subtract 2% from each side of a two-outcome market.

Best for: Markets where both sides have similar odds. Less accurate when one side is a heavy favorite.

Formula: True P = Implied P - (Overround / number of outcomes)

3. Power Method

Raises each implied probability to a power that forces the total to 100%. More mathematically rigorous than multiplicative for lopsided markets.

Best for: Markets with heavy favorites (-300 or beyond) where vig distribution is likely uneven.

4. Shin Method

Based on academic research by Hyun Song Shin. Assumes some fraction of bettors have inside information, and the book adjusts prices to protect against them. The vig is distributed more heavily toward longshots because informed bettors tend to bet favorites.

Best for: Horse racing, futures markets with many outcomes, any market where insider information is plausible. This is the method sharp bettors trust most for multi-outcome markets.

5. Worst-Case

Assigns the maximum possible probability to each side independently. Conservative. Gives you the least favorable true probability for your side.

Best for: Risk management. If the bet is still +EV under worst-case de-vig, the edge is robust.

6. Logarithmic

Uses logarithmic scaling to distribute the overround. Falls between multiplicative and power in terms of assumptions about vig distribution.

Best for: Academic work and model calibration where you want a theoretically grounded middle ground.

7. Prorated (Margin-Weighted)

Distributes vig proportionally to each side's distance from 50%. Sides closer to even absorb more vig. Sides near the extremes absorb less.

Best for: Markets where you believe the book protects its margin more on the competitive side of the line.

The de-vig calculator runs all 7 methods simultaneously so you can compare outputs. For most two-outcome sports markets, multiplicative and power give nearly identical results. The differences widen on lopsided lines and multi-outcome markets. For a deeper dive into how vig works and why it exists, read the complete vig guide.

Worked examples across different market types

NFL spread: -110/-110

Both sides imply 52.38%. Total: 104.76%.

On a standard -110/-110 market, every method agrees: true probability is 50/50. The book's edge is entirely in the 4.76% overround. This is the simplest case. Check what the book keeps with the hold calculator.

NBA moneyline: -180/+155

Implied: 64.29% and 39.22%. Total: 103.51%.

Now the methods diverge. Shin gives the underdog the highest true probability (38.13%) because it assumes the book pads longshot prices more. If your model says the underdog wins 40% of the time, your edge ranges from 1.87% (Shin) to 2.53% (additive) depending on method. That is a meaningful spread when you are sizing with Kelly.

Futures market: 8 outcomes

In a multi-outcome market like "Who wins the division?" with 8 teams, the overround can reach 20-40%. Here the Shin method becomes critical. It properly accounts for the fact that longshots (teams priced at $0.03-$0.05) carry disproportionate vig. Multiplicative de-vig on a 30% overround futures market can overestimate longshot true probabilities by 2-5 percentage points.

Always use the de-vig calculator for multi-outcome markets. Manual calculation with 8+ outcomes is error-prone and the method choice genuinely changes the output.

Why method choice changes your EV calculation

Here is the practical impact. You find an NBA underdog at +155. Your model says 40% true probability.

Using multiplicative de-vig (37.89% true line):

• Your edge: 40% - 37.89% = 2.11%
• EV per dollar: (0.40 x $2.55) - (0.60 x $1.00) = $1.02 - $0.60 = +$0.42 per $1 risked

Using Shin de-vig (38.13% true line):

• Your edge: 40% - 38.13% = 1.87%
• EV per dollar: same calculation, but the confidence interval around your edge is tighter

The EV calculation itself does not change. What changes is your assessment of how much edge you actually have, which affects position sizing. Overestimating edge by even 0.5% means you size too large, which increases variance and drawdown risk.

Run the full pipeline: de-vig the line with the de-vig calculator, then plug your true probability estimate and the market odds into the EV calculator. The EV calculator shows whether the bet clears your minimum edge threshold after accounting for the true line.

True probability in prediction markets

Prediction markets do not have traditional vig. Instead, you pay fees (percentage of profit, per-contract charges) and spread costs (the gap between bid and ask). But the concept of true probability still applies.

On Kalshi, if Yes trades at $0.55 and No trades at $0.47, the sum is $1.02. That 2-cent spread functions like a 2% overround. The true probability of Yes is approximately 55/102 = 53.92%, not 55%.

On Polymarket, tight spreads on liquid markets mean the posted price is closer to true probability. But on illiquid markets, spreads widen to 5-10 cents, creating significant distortion. The fee calculator handles the platform-specific math.

For cross-platform analysis, de-vigging sportsbook lines gives you a "sharp" true probability that you can compare directly against prediction market prices. If the de-vigged NFL line implies 62% and Kalshi prices the same outcome at $0.58, that is a 4-cent edge on the prediction market side. The arbitrage calculator checks whether this gap is exploitable across platforms. Read the full framework in the cross-platform edge guide.

Common mistakes when calculating true probability

Using raw implied probability as your baseline. This is the most common error. Comparing your 55% estimate against a -130 line (implied 56.52%) overstates your edge because the 56.52% includes vig. De-vig first, then compare.

Ignoring method differences on lopsided lines. On -110/-110, it does not matter. On -400/+320, the method you choose can shift true probability by 1-2 points. If your edge is thin, run multiple methods and use the least favorable one as your baseline.

De-vigging one side only. You need both sides of the market to calculate the overround. A -150 moneyline by itself does not tell you the true probability. You need the +130 on the other side to compute the total implied probability and remove the vig.

Applying two-outcome de-vig to multi-outcome markets. A 6-team futures market needs a method that handles 6 outcomes simultaneously. The Shin method was designed for this. Multiplicative works but becomes less accurate as outcomes increase.

Forgetting that true probability is still an estimate. De-vigging tells you what the market thinks after removing the book's margin. It does not tell you the actual probability. The market can be wrong. Your job is to find spots where it is wrong by enough to clear fees. That is the definition of value betting.

Frequently asked questions

What is true probability in sports betting?

True probability is the market's implied probability after removing the sportsbook's vig (margin). A -110/-110 line implies 52.38% per side, but the true probability is 50% each. The extra 2.38% per side is the book's cut.

How do you calculate true probability from odds?

Convert both sides to implied probabilities, sum them to find the overround, then divide each side by the total. For -150/+130: implied 60% + 43.48% = 103.48%. True probability of favorite: 60/103.48 = 57.98%.

Which de-vig method is most accurate?

The Shin method is most accurate for multi-outcome markets and lopsided lines because it accounts for informed bettor behavior. For standard -110/-110 spreads, all methods give identical results (50/50).

Why does true probability matter for finding +EV bets?

Without de-vigging, you compare your probability estimate against a biased number that includes the book's margin. This makes every bet look worse than it is and causes you to miss profitable opportunities or miscalculate position sizes.

Can you calculate true probability for prediction markets?

Yes. On prediction markets, the bid-ask spread functions like vig. If Yes is $0.55 and No is $0.47 (sum $1.02), the true Yes probability is approximately 53.9%, not 55%. The 2% overround is the spread cost.