Understanding and Calculating Betting Edge

Edge in betting is an estimated probability advantage at a specific price. In this guide, it means your estimate of an outcome’s probability minus the price-implied break-even probability. The result is measured in percentage points, not as a percentage profit.

For example, odds of +150 imply a break-even probability of 40%. If you estimate the outcome’s probability at 44%, your edge is 4 percentage points. This is a hypothetical calculation, not a recommendation to bet: your estimate could be wrong.

Edge is distinct from expected value (EV). The probability gap tells you how far your estimate is above or below the break-even threshold. EV is the average net result under the model, while EV yield expresses that result as a percentage of the stake.

How do you convert odds into break-even probability?

Break-even probability is the outcome probability at which a bet has zero EV at the stated price, assuming ordinary win-or-loss settlement. The implied probability derived from a single price is not automatically a no-vig probability or the true probability of the outcome.

For decimal odds d:

Decimal odds express the total return per unit staked, including the returned stake. Net profit on a win therefore uses d − 1, not d.

For positive American odds +A:

For negative American odds -A, where A is the unsigned magnitude:

Multiply the resulting fraction by 100% to express it as a percentage.

• At +150: 100/250 = 40%.
• At -110: 110/210 = 52.38095…%.
• American odds of -110 correspond to decimal odds of 1 + 100/110 = 1.90909….

The ellipsis indicates a continuing decimal. Use unrounded values in calculations: rounding the odds too early can slightly change the result. These formulas describe a price; they do not predict an event’s outcome.

For a separate check, use the implied probability calculator or odds converter.

How do you calculate edge and EV?

Edge is a probability gap, while EV is the average net monetary result under a stated model. To calculate them, you need the odds, your estimate of the outcome probability, and the stake.

If p is your probability estimate expressed as a fraction, edge in percentage points is:

For a bet with two settlement outcomes—win or loss—expected value is:

Here, s is the stake and d is the decimal odds. A win produces net profit of s × (d − 1); a loss produces a net result of −s. The returned stake is not profit.

EV yield is net EV divided by the stake:

Multiply this fraction by 100% to express EV yield as a percentage. Monetary EV uses the same currency as the stake.

This formula assumes fixed odds, loss of the entire stake on a losing bet, and return of the stake plus profit on a winning bet. It does not automatically account for pushes, voids, commission, taxes, dead heats, or other payout outcomes. For those cases, list every possible settlement outcome and its net result, then weight each result by its probability.

Example: -110 with an estimated probability of 55%

At -110 with a probability estimate of 55%, the edge is 2.61905 percentage points, EV yield is 5%, and modeled net EV is \$5 per \$100 risked.

The probability gap is 55 − 52.38095…. The EV calculation uses unrounded decimal odds; edge is displayed rounded to five decimal places. The results of 5% and \$5 describe the model’s average net result, not guaranteed profit or the payout from a single bet. They assume the binary win-or-loss settlement described above.

Example: +150 with an estimated probability of 44%

At +150 with a probability estimate of 44%, the edge is 4 percentage points, EV yield is 10%, and modeled net EV is \$10 per \$100 risked.

The monetary unit in the final formula is dollars: the result is \$10 for a stake of \$100. These values require no rounding and assume binary win-or-loss settlement. A winning bet’s total return includes the returned stake, but \$10 is not that payout—it is the average net result weighted by the probabilities of winning and losing.

Both examples use hypothetical probability estimates. They are not actual bets, observations, or betting advice.

What do positive, zero, and negative results mean?

A positive edge means your probability estimate is above the break-even threshold; zero means it equals the threshold; a negative edge means it is below the threshold. In the binary model described here, edge and EV have the same sign.

A value bet is a bet considered to have positive EV based on a probability estimate. That label does not establish that the estimate is correct.

Positive EV does not guarantee an individual win, realized profit, or an accurate prediction. Correct arithmetic cannot fix an incorrect probability estimate or a mismatch between actual settlement rules and the model’s assumptions.

How does edge differ from bookmaker margin?

Your edge concerns one outcome at a specific price, whereas bookmaker margin is reflected in prices across the whole market. Overround is the amount by which the sum of implied probabilities for all mutually exclusive outcomes covering the market exceeds 100%.

For two opposite outcomes, each priced at -110:

The overround is shown rounded to four decimal places. It measures market pricing, not the bookmaker’s realized hold—the share of betting turnover actually retained. Realized hold and profit depend on the distribution of bets accepted and their settlement results.

To check the whole market, use the margin calculator. A single price is not enough to determine a market’s overround.

How do you calculate a no-vig probability estimate?

A no-vig probability estimate requires odds for all outcomes and a stated method of removing margin, also called devigging. It is not automatically the true probability.

One method is proportional normalization: divide each raw implied probability by the sum of all implied probabilities.

Here, p_i is the implied probability derived from an individual outcome’s price, and q_i is its normalized no-vig estimate. This method makes the estimates sum to one, but it does not prove they are correct. Other devigging methods may produce different results.

For more detail, see the no-vig calculator guide.

What is a good betting edge?

There is no universal threshold for a “good” edge. A positive gap is only as useful as the probability estimate behind it and the accuracy of the settlement assumptions. The number alone does not tell you how reliable the calculation is.

When interpreting a result, check:

• Uncertainty in your estimate. An inaccurate probability estimate can erase an apparent positive edge.
• The specific price. A change in odds changes both break-even probability and EV.
• Costs and settlement rules. Commission and additional payout outcomes can change the net result.
• The units. An edge of 4 percentage points is not the same as an EV yield of 4%.

Assess the basis for your probability estimate, not just the size of the edge. A calculator can check the arithmetic, but it cannot establish the quality of your prediction.

How do you check the calculation with a calculator?

You can check the arithmetic in the value bet calculator. In its user-entered probability mode, you supply one price in a selected odds format, one probability estimate, and one stake. The tool calculates results from those inputs.

1. Open the calculator using the link above.
2. Select the mode for entering your own probability estimate.
3. Select decimal odds and enter the inputs. For the +150 example, enter decimal odds of 2.5, a probability estimate of 44%, and a stake of 100 in the corresponding fields.
   The screenshots below show a separate example: decimal odds of 2.10, an estimated probability of 52%, a stake of \$100, and a bankroll of \$1,000.
   Decimal odds of 2.10 and estimated probability of 52% The odds and probability fields were captured in the local build; the live production interface was not verified.
   In this separate example, the stake and bankroll fields contain \$100 and \$1,000, respectively.
   Stake of $100 and bankroll of $1,000 The stake and bankroll fields were captured in the local build; the live production interface was not verified.
4. Check the calculated output for those inputs.
5. Interpret the result. The reference values are an edge of 4 percentage points, EV yield of 10%, and net EV of 10 monetary units. If the stake represents \$100, monetary EV is \$10. These values require no rounding; display precision does not change their meaning.
   EV yield of +9.20% and modeled net EV of +$9.20 Captured in the local build with the live production interface unverified, this screenshot shows +9.20% EV yield and +\$9.20 modeled net EV on a \$100 stake for the separate example with odds of 2.10 and estimated probability of 52%, not the percentage-point edge gap.
6. Account for the limitations. The calculator does not retrieve live prices or validate the probability you enter. The simple binary calculation also does not automatically cover commission, voids, pushes, dead heats, or other payout outcomes.

If the output differs from your manual check, first compare the selected odds format, probability input, stake, and rounding. Do not treat a positive result as confirmation of your prediction: the calculator checks the mathematical consequences of your inputs.

What are these formulas based on?

These formulas convert odds into implied probability and calculate expected value by weighting net outcomes by their probabilities. Their applicability is limited by the settlement assumptions stated above.

• Smarkets’ explanation of implied probability covers converting decimal odds and American odds into probability.
• Pinnacle’s explanation of expected value covers calculating EV using the probabilities of winning and losing.

The key takeaway: edge is an estimate at a particular price, not a promise of profit. Distinguish the probability gap from EV, then check both the arithmetic and the assumptions behind it.