Parlays are the most-bet, most-misunderstood product in sports betting. The payout numbers look thrilling — a 6-leg parlay can pay 60-to-1 on $100 — and that's exactly why sportsbooks promote them so aggressively. The math underneath is brutal: each leg compounds the vig, and the breakeven win rate is much higher than most bettors realize.
This post walks through the formulas, the worked examples, and the rare cases where a parlay is actually positive expected value.
What a Parlay Is
A parlay (or "accumulator" in the UK) is a single bet that combines multiple individual selections. Every leg has to win for the parlay to pay. One losing leg kills the whole ticket.
The payout for the parlay is roughly the product of the individual payouts, which is why parlays look so tempting. Three coin-flip-priced legs at -110 each pays ~6-to-1 instead of the 1-to-1 each leg would pay individually.
The Math: Combined Decimal Odds
Working in decimal odds makes the math clean. Combined decimal odds for a parlay are the product of the individual legs:
combined_decimal = leg_1_decimal × leg_2_decimal × ... × leg_n_decimal
A three-leg parlay where every leg is -110 (decimal 1.91):
combined = 1.91 × 1.91 × 1.91 = 6.97
So your $100 bet returns $697 ($597 profit). That's how a six-figure parlay slip happens — chain enough +200 legs together and you get to four figures fast.
But "combined" implied probability is also multiplicative:
combined_implied_prob = (1 / combined_decimal)
So the parlay above has an implied probability of 1 / 6.97 = 14.3%. The book is saying "this combination only hits about one time in seven."
How Vig Compounds in a Parlay
This is the part most bettors miss. Each individual -110 leg has built-in vig — the implied probability is 52.4% even though the true probability of a coin-flip event is 50%. Across one leg the bookmaker's edge is small: about 2.4% per bet.
Across a parlay, that edge compounds multiplicatively. The true probability of three independent 50/50 events all hitting is 0.5³ = 12.5%. The implied probability the sportsbook is offering on the parlay is 14.3%. The gap is 1.8 percentage points — which sounds small, but it's a bigger relative edge than the bookmaker has on any individual leg.
In percentage terms:
- Per leg: 52.4% / 50% = 4.8% relative overround
- Three-leg parlay: 14.3% / 12.5% = 14.4% relative overround
- Five-leg parlay: ~25% relative overround
- Ten-leg parlay: ~57% relative overround
A ten-leg parlay at "fair" -110 lines is asking you to spot the bookmaker 57 cents on every dollar before any of your individual leg picks even matter. This is why same-game parlay slips with 8+ legs are essentially lottery tickets with worse expected value than actual lotteries.
Worked Example: Four-Leg NFL Parlay
You take four NFL games, all at -110, all sides you "like":
Leg 1: Patriots -3.5 at -110
Leg 2: Chiefs +1 at -110
Leg 3: Bills total over 47.5 at -110
Leg 4: 49ers -7 at -110
Combined decimal: 1.91^4 = 13.31
Combined American: +1231 (i.e., $100 wins $1231)
Combined implied prob: 7.5%
True prob if every leg is 50/50: 6.25%
Bookmaker edge on the combined ticket: 1.25 percentage points / 6.25% = 20% relative edge
Your $100 stake has an expected return of $100 × (true_prob × 13.31 + (1 - true_prob) × 0) = $100 × 0.0625 × 13.31 = $83.19.
Expected loss per parlay: $16.81 (about 17% of stake).
Compare this to flat-betting all four legs individually. Each leg costs you about 2.4% in vig. Total expected loss across four bets: 4 × 2.4% = 9.6% of total stake. The parlay nearly doubles your expected loss for the same four picks.
The Breakeven Win Rate
For a parlay to break even, your true win rate per leg has to be high enough that the product of true probabilities equals the implied parlay probability.
Breakeven for a $100 stake to return $100:
true_prob_per_leg = (1 / combined_decimal)^(1/n)
For the four-leg -110 parlay above:
true_prob_per_leg = (1 / 13.31)^(1/4) = 0.523
You need to be a 52.3% picker against -110 lines just to break even on every leg in a 4-leg parlay. Most public bettors are 49-51% pickers. The math doesn't work.
A sharper number: even a strong handicapper who hits 55% against the closing line — which would make them genuinely sharp — only barely beats a three-leg parlay (breakeven there is ~54%). Add a fourth leg and most "sharps" are now losing in expectation.
When (If Ever) Is a Parlay +EV?
Three structural cases:
1. Correlated legs (same-game parlays). If the legs aren't independent, the math changes. A "QB throws for 300+ yards AND team wins" parlay has positive correlation — if the QB throws for 300 yards, the team probably won. Sportsbooks try to price out the correlation, but they often underestimate it on niche markets. SGP boosts and same-game parlay promotions are sometimes +EV for that reason. This is also why books cap SGP payouts and limit certain combinations.
2. Boosted odds promotions. When a sportsbook offers a 25% parlay boost or a "5+ leg parlay pays 2x" promotion, the math can flip positive on what would otherwise be a losing ticket. These are loss-leaders to acquire new bettors and the EV is the promotion's expected cost to the book. Take them.
3. Genuinely independent legs where each individual leg is +EV. If you have a 5% edge on every individual leg, the EV compounds the same way the payouts do. Three independent +5% edges combined = roughly a +15% edge on the parlay (slightly less due to vig). This is the only honest +EV parlay strategy and it requires having real edge on multiple independent events at once — which is much rarer than people think.
If you don't have one of those three things, you don't have a parlay strategy. You have a tax on excitement.
Parlays vs Prediction Markets
Polymarket and Kalshi don't natively offer parlays the way US sportsbooks do — both platforms are structured as individual binary outcomes. Kalshi has experimented with "combo orders" that approximate parlays, but with much smaller margins than DraftKings/FanDuel parlay slips.
The structural advantage of trading individual contracts is that each one is priced cleanly (price = probability, no vig) and you can size each one with the Kelly Criterion according to your actual edge. You also get to compound positions through correlation (positions that move together because they share underlying drivers) without paying the bookmaker's correlation tax.
A handful of independent +EV positions on Polymarket is mathematically equivalent to a +EV parlay — but you pay no compounded vig and you can exit any individual position mid-game if conditions change. A traditional parlay is locked from the first kickoff until the last whistle. That illiquidity itself has a cost the book gets to keep.
A Python Snippet
If you want to plug numbers in yourself:
def parlay_decimal(leg_decimals):
result = 1.0
for d in leg_decimals:
result *= d
return result
def parlay_implied_prob(leg_decimals):
return 1 / parlay_decimal(leg_decimals)
def parlay_ev(leg_decimals, true_probs, stake=100):
combined = parlay_decimal(leg_decimals)
win_prob = 1.0
for p in true_probs:
win_prob *= p
expected_return = win_prob * stake * combined
return expected_return - stake
def parlay_breakeven_per_leg(leg_decimals):
combined = parlay_decimal(leg_decimals)
return (1 / combined) ** (1 / len(leg_decimals))
Ten lines of math cover almost every parlay decision.
The Bottom Line
The reason parlays exist is that the human brain has trouble multiplying probabilities. A 50% × 50% × 50% × 50% picture feels like "I'll win at least one of these" and not "I have a 6% chance of winning the whole thing." Sportsbooks price exactly into that gap.
If you want the variance of a parlay without the compounding vig, take multiple independent +EV positions on a prediction market and let them compound through correlation rather than through the bookmaker's margin. If you want the entertainment of a parlay and you've decided the cost is acceptable, take a boosted promo and stop after one or two legs.
The math doesn't have to be brutal. It's only brutal because parlays are designed to weaponize the gap between how humans feel about probability and how probability actually works.
Free interactive parlay calculator — type any number of legs, see combined odds, payout, breakeven win rate, and expected value side-by-side. Pair with the Kelly Criterion calculator, the odds converter, and the fair value calculator. Related reading: can you parlay on Kalshi, how prediction markets work.