BY 
I built projections for a living for years, and the fastest way to lose my attention is still a screenshot of somebody's ten-leg parlay. Sportsbooks promote parlays harder than any other bet type. That alone deserves your suspicion.
The marketing writes itself, small stake and big payout. The arithmetic underneath is the least favorable math in mainstream betting, so I want to work through it with you properly, with real numbers you can check yourself.
Start with the single bet baseline
A standard -110 straight bet implies a 52.38% win probability. You can verify that yourself: 110 divided by 210. Price both sides of a game at -110 and the implied probabilities sum to 104.76%, and that extra 4.76% is the book's margin, the vig.
In expected-value terms, a coin-flip bettor at -110 loses about 4.5 cents per dollar wagered. Write that number down. It's the baseline tax on ordinary betting, and everything that follows gets measured against it.
Now the parlay, leg by leg
A parlay needs every leg to win, and the payout compounds the decimal odds. Take two independent -110 legs, each worth 1.909 in decimal odds.
The payout: 1.909 times 1.909 equals 3.645. Bet $100, collect $364.50, for a profit of $264.50. The book is paying you about +264.
The true odds: if each leg is a fair coin flip, both hit 25% of the time, and the fair payout on a 25% event is +300. You're getting +264 for a +300 proposition.
The expected return: 0.25 times 3.645 equals 0.911 per dollar, so the hold is roughly 8.9%, nearly double the 4.5% you'd pay on a straight bet. Add a third -110 leg and the payout becomes 1.909 cubed, about 6.96, or +596, against true odds of +700 at 12.5%. The hold climbs toward 13%.
The pattern generalizes. Every leg you add compounds the book's margin. That's why the promos push six-leg same-game specials and never two-team spreads.
A quick sanity check before we move on. If the two-leg number surprised you, that's normal, because hold percentages hide inside payouts rather than sitting next to them on the bet slip. The book never shows you the fair price. You have to compute it yourself, which takes about thirty seconds once you know the steps, and I promise the habit pays for itself.
But what if you really do have an edge?
This is the counterintuitive part, and I want to be fair to parlays here. If you genuinely pick winners above break-even, parlays compound your edge too. A true 55% picker hits two legs 30.25% of the time (0.55 squared), and against that +264 payout the bet is EV-positive. So why do sharp bettors still overwhelmingly bet straight? Three reasons, each practical rather than theoretical.
- Variance explodes. Hitting 30% of your bets at +264 produces brutal losing streaks, and standard bankroll sizing at 1% to 2% units gets much harder to hold when your win rate drops by a third and drawdowns stretch out.
- Edge estimates are noisy. If your true rate is 52% rather than 55%, singles are barely losing while parlays are bleeding. Parlays punish overconfidence at compound rates.
- Bankroll growth prefers frequency. Steady singles compound a roll more reliably than lumpy parlay hits, which is the same logic that makes quarter-Kelly a ceiling rather than full Kelly a target.
The correlation exception
There's one intellectually honest case for parlays, and it's correlation. When leg outcomes are positively related, the true joint probability is higher than naive multiplication suggests, and the payout can exceed fair value.
Picture a cold-weather under paired with a low team total, or a spread and total combination that both depend on a blowout script. Books know this, which is why genuinely correlated combinations get restricted or repriced in same-game parlay engines. Correlation-aware bettors still find soft spots.
My honest teacher's warning: this is expert territory. If you can't articulate why the legs are correlated and roughly by how much, what you're holding is a lottery ticket with extra steps.
What verified records actually show
Spend some time with the profiles on the CAPTRACKER leaderboard. It covers over 900 handicappers, and each pick is timestamped and settled automatically against ESPN data. Nothing gets edited or deleted after the fact. The verified-record culture there runs overwhelmingly on straight bets.
Units won on singles is the currency of credibility in that world. Parlay screenshots are the currency of marketing. That split exists because of the math in this article, absorbed by people whose records are public and immutable. The settlement process is documented on the methodology page if you want to check my homework.
My practical rules
- Default to straight bets. An actual edge tends to look boring on a ledger.
- If you parlay for fun, size it at a fraction of a unit and record it honestly. Entertainment is a fine reason to bet small. It's a terrible reason to bet big.
- Never chase a drawdown with a parlay. You'd be stacking the book's compounded hold on top of your own tilt, and the math punishes both.
- Before tailing anyone's parlay, check their singles record on the daily feed. If the base rate isn't verified and profitable, the parlay is fiction squared.
One more thing, since I promised you teacher energy. Redo the two-leg calculation yourself tonight, with a calculator, slowly. Numbers you've derived stick in a way that numbers you've read never do. That habit alone will save you more money than any pick I could give you.