How Are Parlay Odds Calculated? Math, Margin and Correlation
Standard parlays multiply decimal prices; same-game parlays need a joint-probability price that accounts for how the legs interact.
Last updated: October 22, 2026
Key takeaways
- Price and probability are different jobs. Calculating a payout is easy; estimating the chance every leg wins is the harder problem.
- Ordinary parlay prices multiply in decimal format. American odds must be converted first.
- Margin can compound. Combining several negatively priced bets does not remove their expected cost.
- Correlation changes joint probability. Individual prices alone cannot tell you a fair same-game parlay price.
How to calculate standard parlay odds
A parlay combines several selections into one all-win ticket. A decimal price includes the returned stake: 2.50 means a winning $1 stake returns $2.50 in total, including $1 of original stake.
For a standard independent-leg price calculation, convert the individual odds to decimals and multiply them. BetMGM’s explanation documents this conversion-and-multiplication method. It is not a universal rule for every special parlay product.
Combined decimal odds D = d₁ × d₂ × … × dₙ
Total return = stake × D
Profit = stake × (D − 1)
Price-implied break-even probability = 1 / D
For positive American odds +A, decimal odds are 1 + A / 100. For negative odds −A, where A is the positive magnitude, decimal odds are 1 + 100 / A. The odds converter handles the format change.
One worked example: three −110 legs
Each −110 selection converts to 21 / 11, or about 1.909091 decimal. Multiplying the unrounded values gives (21 / 11)³ = 6.957926. That is approximately +596 in American odds.
On a $10 cash stake, a winning ticket returns $69.58 total: $59.58 profit plus the original $10. The price-implied break-even threshold is about 14.37%. Carry full precision through the calculation and round the final result, rather than multiplying prematurely rounded decimals.
That 14.37% is not a researched forecast. It is the win rate needed to break even at that price, before any other costs. The break-even probability guide explains the distinction.
Why sportsbook margin compounds across parlay legs
Here is an explicitly hypothetical model: every leg is a genuine 50/50 event, all legs are mutually independent, no pushes occur, and the sportsbook offers −110 on each selection. Fair decimal odds on each leg would be 2.00, but the offered odds are about 1.909091.
With three legs, the actual all-win probability under those assumptions is 0.5³ = 12.5%. Fair combined odds would be 8.00, or +700. The multiplied −110 prices instead offer about 6.957926, or +596.
Expected net ROI = true win probability × offered decimal odds − 1
Three-leg example: 0.125 × 6.957926 − 1 ≈ −13.03%
This means an expected loss of about $1.30 per $10 ticket across repeated bets under the model—not a prediction for one ticket. For a single 50/50 selection offered at −110, the same calculation gives about −4.55%. The three-leg figure is not obtained by simply adding three single-leg percentages.
| Independent 50/50 legs | Fair win probability | Offered decimal price at −110 per leg | Expected net ROI |
|---|---|---|---|
| 1 | 50% | 1.909091 | −4.55% |
| 2 | 25% | 3.644628 | −8.88% |
| 3 | 12.5% | 6.957926 | −13.03% |
These are calculated examples, not measured sportsbook results. Market overround, expected bettor loss and realized sportsbook hold are different quantities. Our no-vig odds guide explains a way to normalize market probabilities; the hold percentage guide separates price-based estimates from observed revenue. Neither a no-vig normalization nor one offered price proves the true probability.
How correlation changes same-game parlay odds
Independence means learning that one leg wins does not change another leg’s probability. For correlated outcomes, the relevant quantity is conditional probability: the chance of B given that A happened.
P(A and B) = P(A) × P(B given A)
Only under independence can P(B given A) be replaced with P(B).
Consider a hypothetical team win (A) and quarterback throwing three or more touchdowns (B). Suppose P(A) = 50% and P(B) = 30%. Independence would imply both win 15% of the time. If P(B given A) is instead 45%, both win 22.5% of the time.
| Hypothetical relationship | P(B given A) | Both win | Fair decimal odds |
|---|---|---|---|
| Independent | 30% | 15% | 6.67 |
| Positive dependence | 45% | 22.5% | 4.44 |
| Negative dependence | 15% | 7.5% | 13.33 |
The positive-dependence case deserves a smaller fair payout because the combination is more likely. Negative dependence goes the other way, holding everything else equal. These examples contain no sportsbook margin and are not claims about actual teams or quarterback markets. See the same-game parlay correlation guide for practical linked-leg examples.
How do sportsbooks calculate a same-game parlay price?
At a conceptual level, an SGP price has to account for the chance the selections win together rather than treating each leg as independent. A model could estimate that joint distribution through simulated game outcomes or conditional relationships. Those are possible modeling approaches, not a claim that every operator uses the same algorithm.
DraftKings explicitly distinguishes SGP pricing from ordinary parlay pricing. That public explanation does not expose a reproducible fair-pricing formula. Our article does not reverse-engineer a bookmaker’s proprietary model.
A final quote is an offered price, not a pure correlation coefficient. You cannot take the difference between multiplied singles and the SGP quote and attribute all of it to dependence: margin, pricing choices and promotion terms may also matter. Even with estimated pairwise correlations, a longer ticket requires a coherent joint model; pairwise relationships alone do not generally determine its all-win probability.
This is why our calculator accepts the actual combined quote for an SGP. It can calculate money returned and the price-implied threshold without pretending it has independently estimated the ticket’s fair chance of winning.
Why the bet slip may disagree with your calculation
- Dependence: multiplying standalone prices does not reproduce a correlated SGP price.
- Product rules: a fixed payout table or special parlay format may not use the standard multiplied-price model.
- Rounding: a displayed single-leg price may have less precision than the calculation uses.
- Promotions: boosted winnings, qualifying stakes and caps must be applied according to the offer’s terms.
- Settlement: pushes, voids and non-participating players can trigger product-specific treatment.
For example, FanDuel’s New York house rules cover wager acceptance and parlay settlement, while DraftKings directs SGP users to market and sport-specific rules. Check the rules for your jurisdiction and accepted ticket; do not assume one operator’s treatment applies everywhere.
A price-check workflow before trusting a parlay
- Record the exact selections, single-leg prices, stake and combined quote.
- Decide whether independence is defensible. Different games are not a blanket guarantee of independence.
- Calculate payout and the break-even threshold using the correct combined price.
- Keep an independent estimate of joint win probability separate from the offered odds. If you cannot defend it, do not claim the ticket is positive EV.
- Use the EV calculator with those separate inputs, and record assumptions in your parlay review workflow.
The useful question is not simply “why did the book lower my payout?” It is “does the offered payout fairly compensate for the chance every leg wins?” A more coherent game story does not automatically mean a better price.
Frequently asked questions
How are parlay odds calculated?
For a standard independent-leg parlay, convert each price to decimal odds and multiply the decimals. Multiply that product by your stake for total return, then subtract the stake for profit. Correlated tickets need the actual combined quote.
Why are my same-game parlay odds lower than a calculator shows?
Positive dependence can make all legs winning together more likely than an independence calculation assumes, so fair odds become shorter. The offered price can also include margin and other pricing adjustments; a payout difference alone does not measure correlation.
Does multiplying implied probabilities give the true parlay win chance?
No. Multiplying the reciprocals of decimal prices gives a price-implied threshold when those prices are multiplied. It does not remove margin or establish the true joint probability of the selections.
Can you calculate correlation from the individual leg odds?
Not from those prices alone. You need information about how the outcomes interact. For two legs, the exact relationship is P(A and B) = P(A) × P(B given A).
Does a voided leg always get removed from the parlay?
Do not assume so for every ticket. Standard parlays often remove voided selections, but same-game, sport-specific and promotional rules can differ. Check the accepted ticket and the rules for your sportsbook and jurisdiction.
Sources and references
- BetMGM: standard parlay odds and payout calculation
- DraftKings: SGP pricing and settlement guidance
- FanDuel: New York house rules
Sources checked October 22, 2026. Numerical examples are original illustrative calculations, not live sportsbook quotes, betting picks or a proprietary pricing model.
Put the explanation to work
Calculate your parlay payout, explore SGP correlation, or browse the research guide index. Use the tool for the price arithmetic and the guides for evaluating the assumptions behind it.