Same-game parlays in the NHL: correlation and pricing explained

You build a three-leg ticket on one Habs game: Habs win, their top-line winger scores, over 5.5 goals. You multiply the odds in your head on the metro home, open the bet builder, and the price is way lower than your maths. The app is not broken. That gap is the whole story of same-game parlays (paris combinés sur un même match), and once you get it, you will read every SGP price differently.

What is a same-game parlay?

A regular parlay combines legs from different games. A same-game parlay, usually built in a "bet builder" tool, combines legs from one game: a moneyline, a total, a player prop, a goalie prop. All legs have to win for the ticket to pay.

In a regular parlay across separate games, the books simply multiply the decimal odds, because what happens in Montreal tonight has nothing to do with what happens in Vegas. Inside one game, the legs are often connected. That is called correlation, and it changes everything.

Why is a same-game parlay price lower than multiplying the legs?

Because the legs are not independent. If the Habs win, their best forward probably had a decent night. If he scores, the Habs are more likely to win. Each leg makes the other more likely, so the combination happens more often than straight multiplication assumes.

Let's put numbers on it. Two legs:

• Habs to win at 2.40 (implied 41.7%)
• Their top-line forward to score at 2.80 (implied 35.7%)

Multiply them and you get 6.72, which implies the pair hits 14.9% of the time.

Now suppose the true chances (with no margin) are: Habs win 40%, forward scores 33%. But in games the Habs win, he scores 45% of the time. The real chance of both happening is 40% × 45% = 18%, not 14.9%. Fair odds for the combo: 1 / 0.18 = 5.56.

So a book that paid 6.72 would be paying far above fair value on a combination that hits 18% of the time. Instead it might offer 4.90, which implies 20.4%: the fair 18% plus its margin. On a $20 ticket, that is $98 back instead of the $134.40 your head maths promised.

Positive and negative correlation, with NHL examples

Positive correlation is the common case in SGPs, because people build stories: "Habs win, captain scores, over hits." Stories are, by design, correlated. Negative correlation (legs that work against each other) gets a bigger price, but it is also less likely to hit, and books often limit those combos.

How do sportsbooks price correlation?

Most books run game simulations: thousands of virtual games using team and player models. They count how often your exact combination happens, turn that into a probability, then add their margin. You never see the simulation, just the final price.

This is why two books can give very different SGP prices for the same legs. Their models disagree on how strong the correlation is. A book that underestimates the link between a team winning and its star scoring will pay more than one that overestimates it. Comparing SGP prices across books is worth more than comparing single-game moneylines, because the disagreements are bigger.

Margin compounding: the hidden cost

Even without correlation, parlays cost more than singles because the margin stacks. If each leg carries roughly a 5% margin, a three-leg parlay pays about 1 / (1.05 × 1.05 × 1.05), or roughly 86% of the fair payout. At 8% per leg, which is common on props, it falls to about 79%.

Put simply: on a three-leg prop-heavy SGP, the book can be keeping around a fifth of the fair value before correlation is even considered. That is why SGPs are among the most profitable products for sportsbooks.

Worked example: a three-leg Habs ticket

Back to the ticket from the metro:

Multiplied: 2.40 × 2.80 × 1.85 = 12.43, which implies about 8.0%. A $10 ticket would return $124.32.

Now look at the links. A goal by the forward is also a goal toward the over. A Habs win often comes with Habs goals, which also feeds the over. Every leg pushes the others up. Say the builder offers 9.50, which implies about 10.5%. That price is not the book being stingy for no reason: the combination genuinely happens more often than 8%. Your $10 now returns $95.

The question to ask is not "why is it lower than my maths?" It is "is 10.5% too high or too low for how often this really happens?" Without your own simulation, that is very hard to answer, which is why the edge usually sits with the book.

Which combinations get blocked?

Some combinations are so tightly linked that books refuse them or reprice them heavily:

• Same team moneyline plus puck line -1.5. Winning by two already means winning.
• Player to score plus same player over 0.5 points. A goal is a point.
• Over and under on the same total. One cancels the other.
• Team to win plus team to win in regulation. One is a subset of the other.

If the bet builder greys out a leg, this is usually why.

What happens if one leg is voided?

Common situation: your SGP includes a player who gets scratched an hour before puck drop. Books handle this differently. Some remove the void leg and recalculate the price from the remaining legs. Others void the whole same-game parlay and return your stake. A few have special rules for player legs. Check the house rules for bet builders before you build one, especially if you like adding props on players who are game-time decisions.

Same-game parlay or a regular parlay across games?

A regular parlay with one leg from each of two different games has no correlation to worry about. Habs win at 2.40 plus a favourite in another game at 1.70 pays 2.40 × 1.70 = 4.08, exactly the multiplied price. You still pay the margin on each leg, so it is not free value, but at least the price is transparent and you can check it with a calculator.

How many legs is too many?

Every leg you add cuts the chance of cashing. With coin-flip legs at fair odds:

Real legs carry margin, so the true picture is a bit worse. A six-leg ticket that hits once in 64 tries can feel close every time, because you usually miss by one leg. That near-miss feeling is part of why big parlays are so sticky.

Are same-game parlays ever a good bet?

Honestly, rarely from a pure maths point of view. Stacked margins plus the book's control over correlation pricing make it hard to find value. Where people sometimes find it:

• Odds boosts on SGPs. A boost can offset part of the margin. Do the maths on the boosted price, not the headline.
• Books with weaker correlation models. Rare, and they adjust once they notice.

If you play SGPs for fun, treat them like a lottery ticket for the game you are already watching: small stakes, low expectations. For a fuller look at the legs themselves, see our NHL player props guide, and for odds conversions, the odds explainer.

SGP checklist

1. Convert each leg to implied probability before you open the builder.
2. Ask whether the legs help each other. If yes, expect a lower price.
3. Compare the builder price at two or three books. Our NHL betting sites for Montreal is a starting point.
4. Read the void rules for player legs.
5. Keep the stake to an amount you are fine losing, because most SGPs lose.

Keep the parlays small

Same-game parlays are built to be fun and to look like a big score. That is exactly why they are risky: the payouts are high, the hit rate is low, and it is easy to keep firing. Set a deposit limit, think about volatility when you size your stakes, and visit our responsible gambling page if you need a break or self-exclusion. Jeu: aide et référence is free, confidential and open 24/7 at 1 800 461-0140 (514 527-0140 in the Montreal area).