Same-game parlays don't fall apart at the third leg. They fall apart at the fourth. Pull the leg-by-leg hit rates from a few thousand SGPs and the pattern is stubborn: legs one through three clear at roughly 72–78% each, then leg four drops into the low 60s and leg five into the 50s. Correlation doesn't save you — it just hides the damage until the fourth leg shows up. And by the time it does, the cash-out button is usually grey.

The math nobody puts on the slip

Bookmakers price each leg with a margin baked in. Stack five legs at a 5% hold each and the compounded margin isn't 25% — it's closer to 22.6% after you account for the correlation discount the book applies to same-game markets. That discount is real, but it's not a gift. It's the book paying you back a fraction of what it removed.

Here's the uncomfortable part. On a standard 4-leg SGP where each leg prices at -150, the implied probability per leg is 60%. Multiply it out and you get 12.96% — a fair price of about +672. Most books post +600 to +640. The gap between +600 and +672 is your effective tax, and it lands hardest on the fourth leg because that's where the book's correlation model has the least data to work with.

Legs one through three are almost always the same markets: moneyline, a spread, maybe an over. Leg four is where bettors start reaching — a player prop, a first-scorer, a corner count. Those markets carry wider margins (often 8–12% versus 4–5% on main lines), so the fourth leg is where the pricing gets genuinely ugly.

Why cash-out disappears at the worst moment

Cash-out offers are generated by a live model that re-prices your position every few seconds. When three legs are home and one is pending, the model can value the ticket cleanly. When the fourth leg is live — say, a player needing one more reception in the fourth quarter — the model widens its uncertainty band. Most books respond by pulling the offer entirely rather than quoting a number they can't defend.

You'll see this most on:

• Player props tied to in-game events (targets, tackles, threes)
• Any leg involving a player who's questionable to return
• Markets where the underlying feed lags the broadcast by 5–15 seconds

The practical result: your cash-out window is open during legs one through three and closes roughly when you need it most. That's not a bug. It's the model protecting the book from a stale quote.

The 4-leg threshold isn't arbitrary

Books started tightening SGP pricing on 4+ leg tickets around 2022, when state regulators began requiring them to disclose correlation assumptions. Before that, a 4-leg SGP was often priced as if the legs were independent — which meant the book was either overcharging or undercharging, depending on the correlation. Once the math got audited, the fourth leg became the line in the sand.

What this means if you're building these

If you're going to bet SGPs, treat the fourth leg as the decision point, not the third. That's where the pricing turns against you and where the exit disappears. Two or three correlated legs is a defensible bet. Four is a bet on the book's model being wrong about correlation — and their model has more data than yours.

The open question is whether cash-out ever comes back for live fourth legs, or whether books quietly retire the feature for anything past three. Watch the next round of state gaming reports. If cash-out utilization keeps climbing while SGP leg counts stay flat, you'll have your answer.