A parlay requires every selection to win. If each leg is an even-money proposition, two legs land about a quarter of the time, three about an eighth, four about a sixteenth. The payout rises, but not as fast as the odds against you.

The intuition that trips people up is that each leg feels individually reasonable. Four bets you are 55% confident in do not make a 55% parlay — they make roughly a 9% one. Confidence does not add across legs; it multiplies.

Each leg carries the sportsbook's margin, and a parlay compounds those margins rather than averaging them. This is the structural reason parlays are worse value than the equivalent single bets, and why books promote them so heavily.

It is not an argument that nobody should ever place one. It is an argument for knowing that the price of the entertainment is real and grows with each leg added.

Some legs are not independent. A quarterback's passing yards and his top receiver's receiving yards move together — if one clears, the other is much more likely to. Books know this, which is why genuinely correlated parlays are often restricted or repriced.

Correlation cuts both ways: it improves your chances when the game breaks your way, and it means a single bad game sinks multiple legs at once. A parlay built entirely from one game is far more all-or-nothing than its leg count suggests.

A standard parlay — sometimes called a power play — needs every leg. A flex play, offered by DFS pick'em operators, still pays at a reduced multiplier when one leg misses, and on larger slips when two do.

Our builder supports both, and grades saved parlays accordingly, because a slip that went 2-for-3 is a loss under one and a win under the other. The tiers are not simply "miss one": a five or six-pick flex can survive two misses, which is why we grade against the actual tier table rather than a rule of thumb.