A strong opinion can still be a poor wager when the sportsbook charges too much.
A bettor estimates a team has a 58% chance to win, then sees two sportsbooks offering -130 and -150. The team and matchup are identical, but the wagers are not. A 58% estimate converts to fair odds of roughly -138. At -130, the required break-even rate is about 56.5%, so the bettor’s estimate suggests a small edge. At -150, the break-even rate is 60%, making the price unattractive despite the same confident pick.
That distinction is central to understanding how betting odds work. Picking asks, “What is most likely to happen?” Betting asks, “Is that outcome more likely than the moneyline implies?” Favorites may win often and still be overpriced; underdogs may lose more often and still offer value. The goal is not simply to predict winners, but to compare an independent probability estimate with the sportsbook’s available line—while accounting for juice and the uncertainty in that estimate.
Prepare the probability correctly
Start by dividing the percentage by 100. A 55% estimate becomes 0.55, while 42.5% becomes 0.425. This decimal is a probability, not decimal betting odds.
The estimate must describe the sportsbook’s exact market. “The Yankees win” could mean the full game including extra innings, a five-inning moneyline, or a regulation-only market. Likewise, an NFL team covering -3 is different from simply winning the game.
Account for possible pushes
For a market that can push, separate the outcomes:
- Win probability: 0.52
- Loss probability: 0.44
- Push probability: 0.04
A push normally returns the stake, so it contributes neither profit nor loss. To price only the graded win-or-loss outcomes, use 0.52 ÷ (0.52 + 0.44) = 0.5417, or 54.17%.
Extra decimal places can improve the calculation, but they cannot improve the underlying forecast. A carefully converted guess remains a guess.
Convert probability to fair decimal odds
Use the probability as a decimal in this formula:
Fair decimal odds = 1 ÷ probability
For an estimated probability of 58%, first use 0.58:
1 ÷ 0.58 = 1.724
The result, 1.724, is the fair price. At those odds, a bettor would need to win 58% of comparable wagers to break even over time. Decimal odds also show the total potential return per $1 staked, so a $100 wager at 1.724 would return $172.40, including the original $100 stake.
“Fair” means the price contains no vig, juice, or built-in sportsbook profit margin. It reflects only the estimated probability. An actual sportsbook line will usually offer a less favorable payout once its margin is included.
Decimal odds are the cleanest intermediate format because the relationship between probability and payout is direct. They also make later comparisons easier, even when the sportsbook displays American odds. Keep the unrounded result during calculations, then round only for display; premature rounding can slightly distort a thin estimated edge.
Odds above 1.724 offer a better payout than the 58% estimate requires. Odds below 1.724 require a higher win rate to break even.
Convert fair probability to American odds
American odds use one formula for favorites and another for underdogs. Let p equal the decimal probability, such as 0.60 rather than 60.
- Above 50%:
American odds = -(p ÷ (1 - p)) × 100 - Below 50%:
American odds = ((1 - p) ÷ p) × 100 - Exactly 50%: even money, normally written as +100
Favorite example: 60%
For a 60% outcome:
-(0.60 ÷ 0.40) × 100 = -150
A fair 60% probability therefore converts to -150. At that price, a bettor risks $150 to win $100.
Underdog example: 40%
For a 40% outcome:
(0.60 ÷ 0.40) × 100 = +150
A fair 40% probability converts to +150. A $100 wager would return $150 in profit if it wins. The reverse calculation is covered in the guide to derive implied probability from American odds.
Rounding the price
Calculated odds may not land on a clean number. A result of -147.6 can be shown as -148, while some betting discussions round to common increments such as -150. Keep the unrounded figure when comparing a fair price with a sportsbook line; early rounding can make a small edge appear or disappear. Use rounded odds mainly for display, and state the convention when rounding to the nearest 5 or 10 cents.
Use a probability-to-odds reference table
The table below provides quick vig-free odds for several common probability estimates. Prices are rounded, so calculations using the original probability may differ slightly.
| Probability | Fair decimal | Fair American | Fair fractional |
|---|---|---|---|
| 33.3% | 3.00 | +200 | 2/1 |
| 40% | 2.50 | +150 | 3/2 |
| 50% | 2.00 | +100 (even) | 1/1 |
| 55% | 1.82 | -122 | 9/11 |
| 60% | 1.67 | -150 | 2/3 |
| 66.7% | 1.50 | -200 | 1/2 |
For fractional odds, convert the percentage to decimal form as p, then use:
Fractional odds = (1 − p) ÷ p
For example, a 60% estimate gives (1 − 0.60) ÷ 0.60 = 0.667, commonly written as 2/3.
A quick sign check can catch conversion errors: probabilities above 50% should produce negative American odds, while probabilities below 50% should produce positive odds. At exactly 50%, the fair price is +100, also called even money. These are fair prices; an actual sportsbook line usually includes vig or juice.
Compare the sportsbook price with the fair line
A 55% win probability produces fair American odds of about -122. That number becomes the bettor’s personal fair line for evaluating the sportsbook’s offer.
| Sportsbook line | Decimal odds | Assessment |
|---|---|---|
| -110 | 1.909 | Favorable price |
| -122 | 1.820 | Roughly fair |
| -130 | 1.769 | Unfavorable price |
At -110, a $110 wager returns $100 in profit if it wins. That is cheaper than laying $122 to win $100 at the estimated fair price, so the offered line carries a theoretical edge. At -130, the sportsbook requires too much risk for the same $100 profit.
Avoid the directional trap
Negative American odds can look backward at first: -110 is better than -130 because less money must be risked to win $100. With positive odds, the direction reverses: +130 is better than +110 because the same $100 wager produces more profit.
When the signs or comparisons become confusing, convert both prices to decimal odds. The higher decimal number always pays more for the same stake, making it easier to compare a sportsbook price with fair odds.
A personal fair line is best treated as the maximum acceptable price, not a prediction of the closing line or the game result. Any perceived edge still depends on the quality of the original 55% estimate.
Separate probability edge from expected value
At -110, a bettor risks $110 to win $100. The sportsbook’s break-even probability is:
110 ÷ (110 + 100) = 52.38%
If an independent estimate gives the wager a 55% chance of winning, the probability edge is:
55% − 52.38% = 2.62 percentage points
That is not the same as expected value. Probability edge compares two probabilities; expected value (EV) measures the average return implied by those probabilities and the offered payout.
| Measure | Calculation | Result |
|---|---|---|
| Probability edge | 55% − 52.38% | 2.62 points |
| EV on $110 risked | (0.55 × $100) − (0.45 × $110) | $5.50 |
| EV per $1 risked | $5.50 ÷ $110 | $0.05 |
The wager therefore has about 5% EV per dollar risked, assuming the 55% forecast is accurate. EV describes a long-run mathematical average, not the expected result of one bet.
Account for juice and uncertainty
Sportsbook juice affects the quoted break-even probability. In a two-sided market priced -110 on both outcomes, the implied probabilities total 104.76%; that excess is the overround. Removing it produces a no-vig market estimate of 50% for each side.
A no-vig probability is not an independent forecast. It is derived from sportsbook prices and mainly describes how the market prices the two outcomes after removing margin. A separate estimate can still be wrong because of uncertain inputs, injuries, lineup news, or model error. Thin edges deserve particular caution: if the true probability is only 52%, a -110 wager has negative EV despite the original 55% estimate.
Stress-Test the Betting Price
What could make the estimate wrong?
Check injuries, lineup assumptions, sample quality, and whether the model reflects the exact market. A precise percentage can still rest on weak inputs.
How should sensitivity be tested?
Recalculate fair odds after moving the estimate one and two percentage points each way. If value disappears quickly, the wager has little margin for error.
When should value be rechecked?
Compare prices across sportsbooks, then rerun the math after line movement. Also account for stake limits and promotion rules, including minimum odds, eligible wagers, and expiry.
What should be recorded?
Log the probability estimate, fair line, wagered price, and closing line. This makes it easier to measure closing line value and judge the process rather than short-term results.
Make the Price Earn the Bet
Start with a precise wager definition, estimate its fair probability, and convert that estimate into a no-vig price. Compare the sportsbook’s line, calculate the probability edge and EV, then stress-test the assumptions. If the edge disappears after reasonable adjustments or misses a preset threshold, pass.
The math only prices an opinion; it does not prove the opinion is right. Keep stakes modest, use fixed bankroll limits, and never increase wagers to chase losses.
