Same-Game Parlays in Basketball: Why Your Payout Gets Repriced, Not Multiplied

Same-game parlays are treated as correlated bets, not independent ones, which means the odds you see on your ticket account for how basketball outcomes move together. That is why the payout on a three-leg parlay inside one game never equals the simple product of the individual leg prices, and why sportsbooks can settle these tickets differently when a leg voids or pushes.

Correlation, Not Multiplication

Books state it plainly in their own rules: same-game parlay legs are correlated, meaning one outcome can make another more or less likely. The book gives the example of a quarterback's passing yards and the game total going over, both of which are statistically linked. While this example comes from football, the principle applies to basketball: outcomes within one game can influence each other.

Uncorrelated legs are bets that do not meaningfully affect each other's probability. In basketball, these combinations are harder to find. Even seemingly separate legs can be connected through game dynamics. The correlation model adjusts for these relationships, producing a payout lower than the raw multiplication would suggest.

The margin compounds across legs. When the book reprices for correlation, it is layering the correlation adjustment on top. The result is a ticket that pays less than the naïve math implies, not because the book is cheating, but because the legs were never priced as independent events.

What Books Say Happens to the Ticket

Market rules are consistent on settlement: a same-game parlay wins only if every leg wins, with voided or pushed legs removed and the wager repriced on the remaining legs at the odds available when the bet was placed. If all legs are voided or pushed, the entire same-game parlay is voided. The language repeats from book to book: a voided leg is removed and the payout recalculated.

TheScore Bet adds another wrinkle. If all remaining selections in a pre-game same-game parlay are wins after a push or void, the wager is graded as a win with recalculated odds and payout reflecting the remaining legs. This confirms that the original price was provisional, dependent on the full leg structure, and that the book maintains the right to reprice when that structure changes.

These rules matter because they reveal the mechanics behind the scenes. The ticket you place is not a fixed contract at fixed odds; it is a dynamic instrument that the book can adjust if the conditions shift. A player injury that voids a prop, a late scratch that removes a points line, these events trigger repricing, not a simple removal with the remaining legs locked at their original prices.

How Boosts Fit on Top

Apps often promote same-game parlays with odds boosts — a promotion that increases the reward on a winning bet by giving back part of the vig. The same source advises that a boost on an SGP should signal the bettor to build a parlay with odds as close to -110 as possible, suggesting the boost is applied to an already-repriced ticket, not to the raw multiplication.

This distinction matters for anyone trying to reverse-engineer the value. The boost is a promotional layer, not a reset of the underlying correlation pricing. The book has already run its model, adjusted for how the legs interact, and produced a number. The boost then improves that number by some percentage or fixed amount. Knowing this, the sharp play is to find legs that are less correlated than the model assumes, not to chase long-shot combinations that look appealing before the boost is applied.

What the Customer Can Infer from the Rules

The rules themselves are the best guide to what is actually happening. When one book calls legs correlated and another describes repricing on removal, they are describing the same process in different words. The correlation model reduces the payout to account for the increased probability that multiple events in one game will occur together. The repricing rule confirms that the original number was not a simple product but a modeled estimate.

One caveat matters: because each leg in a same-game parlay is correlated in some form, there is no uniform answer when a leg is removed and the odds have to be adjusted. This means bettors cannot easily predict the repriced payout without access to the book's internal model. The transparency ends at the rule description; the math itself is proprietary.

For basketball specifically, the implications are clear. Some outcomes within a game are positively or negatively correlated. The book's model prices these relationships down, even if the bettor sees only the final number.

Voids, Pushes, and Settlement

The settlement mechanics are where the correlation pricing becomes visible. Every book describes the same process: voided or pushed legs come off and the remaining ticket is repriced. This is only possible because the original price was never a fixed multiplication. If it had been, removal of one leg would simply leave the others at their original odds, with the payout adjusted proportionally. Instead, the book recalculates, implying that the legs were priced as a bundle, not as separable components.

TheScore Bet's rule is particularly telling: if all remaining selections are wins after a push or void, the wager is graded as a win with recalculated odds. This means the bettor does not automatically receive the original projected payout minus the voided leg; they receive a new payout based on a new calculation. The exact methodology is not disclosed, but the direction is clear: the book is protecting itself against the changed probability structure of the modified ticket.

DraftKings goes further, stating that if all legs are voided or pushed, the entire same-game parlay is voided. This suggests that the book views the ticket as a single instrument, not a collection of separable bets, and that the correlation pricing is fundamental to its existence. Without at least one leg remaining, there is nothing to reprice, and the bet collapses entirely.

The repricing rule is the clearest proof that same-game parlays are modeled, not multiplied, and that the payout you see reflects a correlation adjustment you cannot independently verify.