Same-game parlays don't die where bettors think they do. The instinct is that three correlated legs is the danger zone — the point where the book's margin starts biting hard enough to matter. In practice, the math turns against you one leg later. Four legs is where same-game parlays stop being a slightly worse version of a straight bet and start behaving like a lottery ticket with a sportsbook logo on it.

Why three legs still feels survivable

Three legs is the sweet spot sportsbooks are happy to let you live in. Correlation pricing on a three-leg SGP — say a quarterback's passing yards, his touchdown count, and his team's moneyline — usually knocks maybe 8–14% off the fair combined price compared to multiplying the legs independently. On a typical build, that's the difference between +550 and +480. Annoying, but not fatal.

The reason is structural. With three legs you have three pairwise correlations to price. Books have gotten good at this. They have years of data on how a QB's yardage and his team's spread move together, and the models handle two-way interactions cleanly.

The fourth leg is where the model gets lazy

Add a fourth leg and the number of correlations doesn't grow linearly — it grows combinatorially. Three legs means three pairwise relationships. Four legs means six. Five means ten. Books price the easy pairs well and price the rest with a blunt haircut, because modeling every higher-order interaction isn't worth their engineering time.

That's the gap. You're not just paying a fourth leg's vig; you're paying an imprecise surcharge on the five relationships that leg creates with everything already in the slip. On a four-leg SGP, the effective hold frequently lands between 20% and 30%, versus roughly 5% on a single straight bet at -110. A $50 four-leg SGP carries an expected loss more than four times a $50 straight.

A concrete comparison

Take a four-leg NFL SGP priced at +900. Fair odds, if the legs were independent and the book took standard juice, might sit closer to +1,100. That ~18% price gap is the correlation tax. Now run the same four legs as two separate two-leg parlays. You'll usually get a better combined number, and you've capped your correlation exposure at the pairs the book actually knows how to price.

Where the collapse actually shows up

The collapse isn't a single bad beat. It's the distribution. A three-leg SGP at +480 wins often enough that your bankroll feels the rhythm — you hit one in five, roughly, and the swings are tolerable. A four-leg at +900 hits closer to one in ten. Your variance doubles, your hit rate halves, and the correlation tax you're paying grows at the same time.

Stack a few of those and the math compounds fast. Ten four-leg SGPs at $50 each is $500 staked. At a 25% effective hold, you're looking at an expected return of around $375 — before you account for the fact that most bettors pick legs that feel correlated but aren't, which pushes the real hold higher still.

The uncomfortable question

If four legs is where the pricing gets sloppy, why do books promote four-leg and five-leg SGPs so aggressively — boosted odds, insurance tokens, profit bumps? Either the boosts are priced to still leave them ahead after the correlation tax, or the tax is doing more work than the marketing admits. Both can be true. The boost is the hook; the fourth leg is the margin.

The real question isn't whether you can win a four-leg SGP. You can. It's whether you'd ever take the same four positions as separate bets if the parlay button didn't exist. If the answer is no, the fourth leg isn't a strategy — it's a mood.