Inside one game, the legs talk to each other.
Every leg of a same game parlay comes out of a single event, so the legs are not independent and the book cannot price them by multiplication. It reprices the combination from its own view of how those outcomes move together, then quotes one number. The gap between that number and the naive product of the legs contains a correlation adjustment and a further slice of margin, in a proportion nobody outside the book can see. This page shows how to measure that gap, why a generous looking price is usually a warning, and what to settle before you build one.
What a same game parlay is
A same game parlay, usually shortened to SGP, is one wager built from two or more selections inside a single game, and every selection has to win for the ticket to pay. That much it shares with any parlay. The difference is what happens to the price.
A parlay spread across three different games multiplies the legs' decimal prices, because a result in one stadium tells you nothing about a result in another. Inside one game that assumption breaks. The outcomes there are produced by the same drives, the same possessions and the same innings, so the book discards the multiplication, estimates the joint probability of the whole combination from its model of the game, and quotes a single price for it. The repricing is not a detail of the product. It is the product.
The number on your slip is therefore not built from the numbers printed beside it on the same screen, and no amount of arithmetic on the individual legs will reconstruct it. If you want the independent case first, what is a parlay covers how legs behave when they genuinely have nothing to do with each other.
What the repricing costs, in numbers
Take a football game, AAA @ BBB. One leg is the quarterback over a raised passing yardage number, priced +150. The second is his primary receiver over a raised receiving number, priced +180. The mechanism joining them is not subtle, since the same completions produce both totals, so the two legs rise and fall together.
Price them as though they were unrelated. A price of +150 is decimal 2.50 and +180 is decimal 2.80, and 2.50 × 2.80 is decimal 7.00, which is +600 and an implied chance of 1 divided by 7.00, or 14.3 percent. Now suppose the book offers decimal 5.20, or +420, an implied 19.2 percent. The haircut is 1 minus 5.20 divided by 7.00, which is 25.7 percent of the naive payout, gone before the game starts.
Two different things are hiding inside that 25.7 percent, and only one of them is a charge. The first is real information: the book is saying the joint chance is nearer 19.2 percent than 14.3 percent, which is exactly what positive correlation looks like when the same throws feed both lines. The second is margin, and 7.00 was never a fair benchmark either, since +150 and +180 each carry a cut of their own, so what you have measured is the correlation adjustment plus the margin charged on top of the margin you were paying anyway. On a two-way market you can add the two sides, watch them sum past 100 percent, and read the margin off directly, which is what juice and hold describes. Here there is no second side to add, and nothing on the screen tells you where one part ends and the other begins.
A longer price is a warning, not a gift
The two usual reactions to that haircut are both wrong. One says the whole gap is margin, so same game parlays are a scam. The other says the gap must be the correlation adjustment, so the price must be fair. Neither can be checked from outside the book, which makes both of them guesses dressed as conclusions.
The expensive version of the mistake is the one that arrives looking like good news. Take a basketball game and two legs priced +100 each, the under on an alternate total and a starter over a raised points number. Those pull against each other, because a slow, low scoring game hands every player fewer possessions to score in. Multiplied naively they are 2.00 × 2.00, decimal 4.00, or +300, an implied 25 percent. Suppose the slip comes back at +360, decimal 4.60, an implied 21.7 percent.
That extra payout is not a discount. Absent a promotional boost, margin can only push the price a book offers below fair, never above it, so a quote longer than the naive product is the book stating plainly that it thinks the combination is less likely than its parts multiplied. Unless you have a specific reason to believe that read is wrong, you are being paid more because you are less likely to collect, and you are paying margin on top of that. A same game parlay that pays unusually well is normally paying for negative correlation the builder never noticed.
Name the mechanism, then price it yourself
Only build a same game parlay when you can state the link between the legs in one plain sentence. In baseball, a starting pitcher over his strikeout number together with the game under is a real sentence: every strikeout is an out that cannot become a run, and a pitcher missing bats is usually avoiding damage as well. If the best sentence available is that you like both selections, there is no mechanism, and you are paying same game pricing on a combination with nothing in it to reprice.
Then do the arithmetic before you look at the offer. Multiply the legs' decimal prices yourself with the parlay calculator and work out the haircut as 1 minus the quoted decimal divided by your product. That percentage is the only handle this market gives you, and the test it supports runs in both directions. If the haircut is far larger than the link you just described in words, pass. If it comes back negative, meaning the quote is longer than your product, pass as well, because the book is pricing those legs as pulling against each other. What you are betting on either way is that the quoted joint price is wrong, which happens when you beat the legs on their own prices or when the correlation adjustment is off. What is never an edge by itself is the bare fact that the legs are related.
Two habits follow from the fact that the number cannot be audited. Stake it below what you would put on a single side, since a market you cannot de-vig deserves less trust than a main side holding a measurable 4.55 percent at -110 or a two-way prop holding 6.52 percent at -115. Then shop the ticket, because correlation models disagree far more than game lines do, and be honest that this tells you which book pays most rather than which one is closest to fair. Pick'em apps such as PrizePicks and Underdog meet the same problem from the other direction and often handle it with rules about which same game selections may be combined, rather than with a price.
The product that hides its own price
The same game parlay is the most heavily promoted ticket on every app, and it is also the one market where you can least verify what you are being charged. Those two facts are related. In every other market a bad price eventually gives itself away, because you can de-vig it or set it beside a fair number, and here there is no such brake: a long run of overpriced tickets feels exactly like a long run of fairly priced ones. Nothing in the quote will tell you when to stop, so the limit has to be a rule you wrote down before the slip existed, and if you do not have one, our responsible gambling page is a better use of ten minutes than another build.
Same game parlays, answered
What is a same game parlay?
A same game parlay, or SGP, is a single wager built from two or more selections inside one game, and every selection still has to win for the ticket to pay. Because the legs come out of the same event, their outcomes are related rather than independent, so the sportsbook prices the combination as one unit instead of multiplying the individual prices. That repricing is what separates it from an ordinary parlay across several games.
Why do sportsbooks reprice a same game parlay instead of multiplying the odds?
Multiplying leg prices assumes the legs are independent, and legs inside one game rarely are. A quarterback's passing yardage and his primary receiver's yardage are produced by the same completions, so the chance of both landing is higher than the two prices multiplied together suggest. The book replaces that multiplication with a joint estimate drawn from its own model of the game and quotes one price for the whole ticket.
Why does adding a leg to a same game parlay sometimes barely change the price?
When a new leg is close to guaranteed by the legs already on the slip, it adds almost no risk, so it adds almost no payout. Pairing a receiver going over his yardage with his team clearing a low alternate team total is close to betting the same thing twice. The price moves a little because the joint probability moves a little, which is the pricing working correctly rather than the book being stingy.
What happens if one leg of a same game parlay is voided?
The voided leg does not simply divide out of the price, because the legs that remain have to be re-estimated as a fresh same game combination and whatever correlation the removed leg carried disappears with it. That is why the settlement price on a shortened ticket often looks nothing like the number you get by unwinding one leg yourself. Operators differ on which voids shorten a ticket and which cancel it outright, so the house rules on that are worth reading before the slip is placed rather than after.
What does correlation mean in sports betting?
Correlation describes whether two outcomes tend to happen together, pull against each other, or have nothing to do with one another. Positively correlated legs make a combination more likely than multiplying its parts suggests, and negatively correlated legs make it less likely. Sportsbooks price same game combinations from an estimate of that relationship, which is why two selections cost one amount separately and a different amount together.
Why do some same game parlays pay far more than the multiplied odds suggest?
A quote longer than the product of its legs is information rather than generosity: the book is telling you those legs pull against each other, so the combination lands less often than multiplying suggests. The pricing works the same way in reverse, which is why positively correlated legs come back shorter than the product. Unless a promotional boost is attached to the ticket, treat an unusually large payout on legs drawn from one game as a warning sign rather than a value sign.
Do all sportsbooks price the same game parlay the same way?
They do not, because each one applies its own model of how the legs relate, and those models disagree far more than the models behind a straightforward game line do. The identical slip can come back at meaningfully different prices on two apps on the same afternoon. Shopping a same game parlay is therefore worth more than shopping a straight side, with one limit attached: the best quote you find raises what you are paid without ever proving what you were charged.
Can you remove the vig from a same game parlay?
You cannot, because de-vigging needs the price of every outcome in a market so they can be rescaled to sum to 100 percent, and a same game parlay is quoted as one side with no opposite side to weigh it against. The nearest substitute is to multiply the individual leg prices yourself and measure the gap between that product and the quote. Keep in mind that the product is not a fair number either, since every leg price already carries its own margin, so the gap you measure is the correlation adjustment plus whatever the book charges over and above the margin you were paying anyway. No public number splits those pieces apart.
The only number you can check is the one you work out
A same game parlay arrives as one price with nothing beside it to check that price against, so the screen will never grade it for you. Multiply the legs the naive way, hold the quote up against that product, and you at least know the size of what you cannot see. Our projections tell you whether those legs were worth combining in the first place.