Same-Game Parlays: How Books Price Correlation Against You

The arithmetic every bettor learns first is that the fair price of a parlay of independent events is the product of the individual leg prices. Probability theory states it plainly: for independent events, \(P(A \cap B) = P(A)\times P(B)\). A bookmaker building a standard multi-game parlay starts from that rule and then whittles in a fixed hold. But a same‑game parlay is not priced that way. Caesars Sportsbook’s Same Game Parlay Guide confirms that SGP odds start from each leg’s individual price and are then adjusted by the sportsbook’s own correlation model, because legs from one game are not independent—one outcome can make another more or less likely. A bettor who multiplies the standalone odds of two legs from the same game and compares the product with the SGP ticket price will find a gap. That gap is the correlation tax, and the book never shows its working.

Why the fair price is the product of independent legs

A regular parlay across different games works because the events have no causal link. The coin‑flip of whether the Jets cover the spread in the early window really is independent of the quarterback’s passing‑yards total in the nightcap. Multiply the implied probabilities of the two legs and you get the joint probability, and from that you can derive the true odds. The product-of-legs rule is a baseline a bettor can calculate from the posted moneylines or totals without needing access to any internal model. DraftKings’ help center notes that a same‑game parlay “works like a regular parlay” in the sense that all selected outcomes must win for the bet to cash, and you need at least two selections from the same game on the slip to receive odds. The difference sits entirely in how the price is built. For a regular parlay, if you have two legs at –110 each, the fair parlay payout assuming a competitive market is around +264. If the same two outcomes appear in an SGP and the book quotes +230, the missing 34 cents are not slippage; they are the price of correlation the book has modeled and you have not.

What changes inside one game

Sportsbooks explicitly forbid you from pretending two legs from the same contest are independent. Caesars’ house rules say an SGP must be made from a single game and cannot be combined across multiple games. The operator’s guide goes further: legs in the same game are correlated, meaning one outcome can make another more or less likely. A wide receiver prop and his quarterback’s passing‑yards over are positively correlated; a running back’s rushing attempts and his team’s pass‑heavy game script are negatively correlated. The book’s pricing engine reads those relationships and discounts the payout accordingly. Bovada’s help section uses the phrase “dynamic, correlation‑based pricing.” The basket of eligible selections inside an SGP at Caesars includes player props, game props, spreads, totals, alternate lines, and some live wagers—a broad enough menu that a bettor can easily assemble legs that look attractive individually but whose joint probability is smaller than the product of the singles suggests. The model catches what the napkin math misses.

How books handle pushes, voids, and settlement

SGP settlement rules introduce another layer where correlation works in the book’s favor. If one leg of a regular parlay voids, the whole wager typically reduces to the remaining selections, but on a ticket built from correlated legs the mathematics of the adjustment is opaque. Caesars’ house rules state that if any leg of an SGP is made void or settles as a push, the odds are recalculated based on the remaining valid legs. DraftKings says the same: when an SGP contains a player who did not participate and legs become voided, the SGP adjusts the odds and settles on the remaining legs. The bettor cannot know in advance what payout that recalculation will produce, because the correlation model has already baked in the relationships among the now‑removed leg and the survivors. A void can significantly alter the payout, and the book does not spreadsheet that change. The result is a settlement that often feels less generous than the player hoped—because the original price assumed a dependency that no longer exists, and the book’s recalibration is one‑sided.

What can be stated, and what cannot

The primary sources the sportsbooks publish confirm that correlation‑based adjustment is real and mechanical. Caesars says SGP odds start from each leg’s individual price and are then adjusted by the sportsbook’s correlation model. Bovada describes dynamic, correlation‑based pricing. DraftKings and Caesars allow player props, game props, and market‑type combinations in the same ticket, so the model must handle multiple classes of event. What the public documents do not give is the formula. No operator rulebook, regulator filing, or published API documentation contains a mathematical expression for pricing correlated legs. A search of sportsbook house rules and help centers turned up descriptions of the adjustment, not an equation. The same gap exists for house‑edge figures: no official percentage increase, tax, or hold rate specifically tied to same‑game parlays appears in a primary source. The industry has chosen to keep the correlation engine inside a black box. A bettor who wants to judge whether a boosted SGP is value does not need the internal model, however. Multiply the fair prices of the legs as if they were independent. Compare that product with the SGP payout the book is offering. The spread between the two tells you what the correlation model is charging you—even if the book never prints the number.