Sportsbooks price same-game parlays differently from regular parlays because the legs are not independent. When you bet a quarterback over passing yards and his tight end to score a touchdown, those outcomes move together: a big passing day makes the touchdown more likely. The true combined probability is higher than naive multiplication suggests, so the fair odds should be lower, and the book applies a correlation discount to account for it. That discount, the correlation tax, is why a same-game parlay pays less than the same legs would as a regular parlay, and why the house edge on these tickets runs far higher than on straight bets.
The math the book does that you do not see
A regular parlay multiplies the implied probabilities of independent legs. Two -110 legs combine to 1.909 times 1.909, or 3.644 in decimal odds, about +264. A same-game parlay does not do that, because the legs are not independent. The correct formula is the conditional one: the probability of both events is the probability of the first times the probability of the second given the first. If the quarterback going over makes the touchdown 68 percent likely instead of the 55 percent the standalone odds imply, the fair price is +167, not +231, and the book might offer +150 after margin. That gap between +167 and +150 is the house's cut, not evidence the bet is bad, but the gap between +231 and +167 is the correlation tax, and it is the bigger number.
The books do not multiply displayed odds to build these prices. They estimate a joint probability first, using historical joint frequencies, game simulations, conditional probability models, or copula-style dependence models, then layer margin on top. That two-step process, correlation adjustment followed by margin, is why a same-game parlay often looks shorter than a naive calculation would predict, even before the house edge is added. And the books do not publish their correlation models, so there is no way to verify whether a given ticket is priced fairly. Independent analysis consistently shows they err on the side of caution, which is their favor.
The scale of the edge is what surprises most bettors. A two-leg same-game parlay at reasonable prices might carry a house edge around 10 to 15 percent, while an eight-leg ticket on highly correlated props can carry an effective edge of 30 percent or more. New Jersey gaming data shows sportsbooks holding about 24 percent on parlays against 4 to 5 percent on straight bets. Same-game parlays became widely available around 2019 and 2020, driven by FanDuel, DraftKings, and BetMGM, and they are now the most heavily marketed product at major books, for exactly this reason.
Which legs get taxed hardest
The more tightly correlated your legs, the steeper the discount. Moneyline plus team total plus a player from that team is the classic over-correlated ticket: winning usually requires scoring points, scoring points usually involves the skill players, and the book knows all three legs rise and fall together. The less correlated your picks, the closer the price gets to regular parlay math. A leg that runs against the others, a defensive prop in a game you expect to go over, for example, can even carry negative correlation, which the book prices differently again.
Most major books cap same-game parlays at 10 to 15 legs and restrict certain combinations outright, such as a team's moneyline and its own spread. They also cap bets or reject combinations they consider too favorable to the bettor, and they reprice when a leg is voided or graded a push. None of this is hidden. It is in the house rules, and it is the other half of the pricing: the book does not just discount your correlation, it refuses the combinations where its model might be wrong.
The practical consequence is that comparing odds across books matters more for same-game parlays than for straight bets, because correlation pricing varies by operator. The same ticket can be priced 50 to 100 points apart across books, which is a wider gap than you will ever see on a single game line. If you are going to bet these at all, the first edge to take is the shopping, not the picking.
How to bet them without paying the full tax
First, count the correlation before you count the payout. If two legs would usually win or lose together, the price you see already reflects that, and the number that matters is the hold, not the headline odds.
Second, keep the leg count low. Every leg you add is another correlation adjustment and another layer of margin, and the edge compounds faster than the payout does. A two-leg ticket at a 10 to 15 percent edge is a bad bet; an eight-leg ticket at 30 percent is a donation. Third, use the profit boosts books offer on specific combinations, but read the settlement fine print, because a boost on a heavily discounted ticket is still a discounted ticket. And if you want the upside of a parlay without the correlation tax, build it across games, where the legs are actually independent and the math is honest.
Frequently asked questions
What is the correlation tax on a same-game parlay?
The discount the book applies because your legs are related. Correlated outcomes have a higher combined probability than independent math suggests, so the fair odds are lower. The book estimates the joint probability, then layers margin on top. On a two-leg ticket the edge runs about 10 to 15 percent; on an eight-leg correlated ticket it can exceed 30 percent.
Why do same-game parlays pay less than regular parlays?
A regular parlay multiplies independent probabilities. A same-game parlay prices the legs conditionally, because one leg changes the odds of the others. The book adjusts the combined price downward to reflect that dependency, so the payout is lower than the same legs would pay as a regular parlay.
How many legs can a same-game parlay have?
Most major sportsbooks require a minimum of two legs and cap same-game parlays at 10 to 15 legs, depending on the operator. Certain combinations are restricted outright, such as a team's moneyline and its own spread.
Should I shop same-game parlay odds across books?
Yes, more than for any other bet type. Correlation pricing varies by operator, and the same ticket can be priced 50 to 100 points apart across books. That gap is wider than you will see on a single game line, and it is the easiest edge available on these tickets.