What the numbers mean
Every price on a sportsbook hides two things: a probability and a fee. This page shows both, with examples worked out by the same code that prices your slip, so nothing here can be wrong without the slip being wrong too.
American odds
Every price on the board is in American odds. A negative number, like −110, is the stake needed to win 100 credits: bet 110, win 100, get 210.00 back. A positive number, like +150, is the profit on a 100 stake: bet 100, win 150, get 250.00 back. A plus price is an underdog. When both sides are minus, as at −110/−110, the book thinks they are about even, and the minus is its fee.
Worked example
−110: stake 110 to win 100, so 10.00 returns 19.09 (1.909 per credit).
+150: win 150 on 100, so 10.00 returns 25.00.
Implied probability
Turn a price into a chance and you get its implied probability: for −110 it is 110 ÷ 210 = 52.4%. The catch is that the two sides of a −110/−110 market imply 52.4% each, 104.8% together. The extra 4.8% isn't a probability; it's the book's margin, hidden inside both prices.
Worked example
−110 implies 52.4%. +150 implies 40.0%.
Margin
Add up the implied probabilities of every side of a market. A fair market totals exactly 100%. Real markets total more: −110/−110 totals 104.8%, so the margin is 4.8%. The margin is the book's fee, built into the price, so you pay it on every bet, including the ones you win. What it costs you per credit staked is the house edge.
Worked example
−110 and −110: 52.4% + 52.4% = 104.8%. The margin is 4.8%.
Fair chance
Removing the margin gives each side its fair chance: divide each implied probability by their total. For −110/−110, 52.4% ÷ 104.8% = 50.0%. For a price to be worth taking, you need the real chance to be better than the fair chance, not just better than the implied one.
Worked example
Remove the margin and each side is 50.0%. You need to beat that, not 52.4%, for −110 to be worth taking.
House's cut
A margin-free book paying fair odds on a 50% chance would return 20.00 on a 10-credit stake. At −110 you get 19.09. The 0.91 difference is the house's cut of your winning bet.
Worked example: 10 credits on a −110 single
A fair book would pay 20.00.
This one pays 19.09.
The house keeps 0.91 of your win: an edge of 4.5% on every credit staked.
House edge
The house edge is the expected loss per credit staked if the fair chances are right: one minus the fair chance times what the slip pays. It is smaller than the margin (4.5% against 4.8% at −110/−110) because the margin is measured on chances and the edge on money. Parlays multiply it: four −110 legs carry 17.0%.
Parlay
A parlay multiplies the decimal odds of each leg into one price, and pays only if every leg wins. Each leg carries the book's margin, so the margins multiply too: four −110 legs carry a 17.0% house edge against 4.5% for one. The big payout isn't generosity; it's four margins stacked.
Worked example: 10 credits, −110 legs
| Legs | Fair chance | Fair payout | This book pays | House edge |
|---|---|---|---|---|
| 1 | 50.0% | 20.00 | 19.09 | 4.5% |
| 2 | 25.0% | 40.00 | 36.44 | 8.9% |
| 4 | 6.3% | 160.00 | 132.83 | 17.0% |
Four legs pay 132.83 where a fair book would pay 160.00: the house’s cut grows from 4.5% to 17.0%.
Push
If the final score lands exactly on the spread or total, nobody wins: the bet pushes and your stake comes back. In a parlay, a pushed leg drops out and the parlay pays on the remaining legs.
Void
A leg is void when its game is canceled or postponed, or has no final result 72 hours after the start time you bet on. Void legs refund their stake; in a parlay they drop out like a push.
Return on stakes
Net result (credits back minus credits staked) divided by total stakes, over settled bets. A random bettor at −110 prices trends towards -4.5%, the house edge that the 4.8% margin produces. Beating 0% over many bets needs skill or luck; which one is what the closing edge measures.
Bankroll return
Everything you hold now, counting open bets at their stake, compared with your starting 500 credits. Doing nothing would have kept all 500, so any bettor below 500 would have been better off not betting.
Closing line
Prices move until kickoff as bettors and the book react to news and to each other. The last price before the game starts, the closing line, is the most informed price there is. Consistently getting a better price than the fair closing price is the strongest sign of real skill, because it means you knew something before the market did.
Closing edge
For each pick, we take the closing line from the same book and market, remove its margin, and compare: your decimal odds × the fair closing chance − 1. Positive is an edge. Negative means your price was worse than the fair closing price, which happens even when the line never moves, because your price included the margin: a −110 pick on a market that closes at −110/−110 has an edge of -4.5%. Averaged over many picks, this predicts long-run results far better than a win-loss record.
Random bettor
Imagine someone who placed every bet you did, same games, markets, prices and stakes, but chose each side with a coin. We work out exactly what they would expect to have, averaging over every combination of sides. If you're ahead of them, your picks did something; if not, your result is what random picks produce at those prices.
Calibration
When you place a single you can say how sure you are. Group those bets by what you said and compare with how often they won. Well-calibrated people win about as often as they say. Over-confidence is one of the most common findings in research on prediction.