Study
How accurate are betting odds? We graded 1,803 games
Betting odds are very accurate: across 1,803 games the closing prices gave favorites a 64.6% chance on average and they won 66.0%, and in 9 of 10 price bands the real win rate landed within normal luck of what the odds said.
We open live odds right after you sign up.
Free during beta. No card. Google or email, about a minute.
By the Betwatch team. Published Nov 4, 2026, updated Nov 10, 2026.
How often are betting odds right?
"Right" is the wrong test for a price. A 60% favorite that loses has not been wrong, because the price said it would lose four times in ten. The fair test is whether 60% favorites win 60% of the time.
So we took the vig out of the last price before the start at every book, averaged the books into one win chance per game, and checked it against the final score. If the odds are honest, the dots sit on the dashed line.
- Odds said
- Favorites won
Data table
| Band | Games | Odds said | Favorites won |
|---|---|---|---|
| 50-55% | 429 | 52.5% | 54.1% |
| 55-60% | 395 | 57.3% | 61.5% |
| 60-65% | 292 | 62.3% | 58.9% |
| 65-70% | 169 | 67.2% | 68.0% |
| 70-75% | 156 | 72.4% | 79.5% |
| 75-80% | 115 | 77.5% | 72.2% |
| 80-85% | 92 | 82.6% | 89.1% |
| 85-90% | 72 | 87.5% | 84.7% |
| 90-95% | 62 | 92.3% | 91.9% |
| 95-100% | 21 | 96.2% | 100.0% |
The two biggest misses sit next to each other and point in opposite directions. Favorites priced at 72% won 79.5%, and favorites priced at 77.5% won 72.2%. That is what luck looks like in a sample of 115 to 156 games, not a pattern to bet on.
How much better than a guess are the odds?
A Brier score is the average squared miss between a stated chance and what happened. Zero is perfect, and 0.25 is what you get by calling every game 50-50. Log loss is the same idea with a harsher penalty for being confidently wrong.
| Forecast | Brier score | Log loss |
|---|---|---|
| Every game 50-50 | 0.2500 | 0.6931 |
| Favorite always gets the 66% base rate | 0.2244 | 0.6410 |
| Closing odds, vig removed | 0.2118 | 0.6092 |
The second row already knows who the favorite is. The closing odds beat it by knowing how big a favorite each team should be. A score of 0.21 looks far from zero only because single games are close to random, and no forecast can remove that.
Which league has the most accurate odds?
NFL closing odds said 64.7% and favorites won 65.0%. College football said 72.4% and favorites won 72.3%. Those are about as close as a forecast gets.
| League | Games | Favorites won | Odds said | Brier score |
|---|---|---|---|---|
| WNBA | 139 | 71.9% | 68.6% | 0.180 |
| College football | 653 | 72.3% | 72.4% | 0.188 |
| NBA | 92 | 68.5% | 65.1% | 0.193 |
| NFL | 117 | 65.0% | 64.7% | 0.227 |
| NHL | 209 | 62.7% | 58.8% | 0.227 |
| MLB | 593 | 58.7% | 57.1% | 0.240 |
A lower Brier score does not mean sharper books. College football scores well because it has mismatches, and a 97% favorite is easy to price. MLB scores worst because almost every baseball game is close to even. The gap between "Favorites won" and "Odds said" is the fairer read, and every league is inside normal luck on it.
How often is the spread correct?
Almost never to the point, and that is fine. A spread is a midpoint, not a forecast of the final margin. In the NFL the game lands exactly on the closing number 5.7% of the time, and half of all games finish within 7 points of it. A typical miss is 7 points in the NFL and over 10 in college football.
| League | Games | Typical miss | Within 3 | Within 7 | 14 or more off |
|---|---|---|---|---|---|
| NFL | 106 | 7.25 pts | 23.6% | 50.0% | 26.4% |
| NBA | 89 | 8.5 pts | 23.6% | 38.2% | 30.3% |
| WNBA | 102 | 8.5 pts | 18.6% | 44.1% | 26.5% |
| College football | 562 | 10.5 pts | 16.2% | 35.2% | 36.3% |
What matters is whether the misses split evenly, and they do. NFL favorites covered 46.2% and underdogs 48.1%, with 5.7% pushes. College favorites covered 46.4% and underdogs 52.1% on 562 games, which sits inside normal luck. Our page on how often favorites win breaks this down by size of spread.
Do overs or unders hit more?
Neither. Across 1,488 games the over hit 49.1% and the under 48.9%, and the average game finished 0.13 points above its closing total.
| League | Games | Over | Under | Push | Result minus total |
|---|---|---|---|---|---|
| All six | 1,488 | 49.1% | 48.9% | 2.0% | +0.13 |
| NFL | 94 | 51.1% | 47.9% | 1.1% | +0.34 |
| College football | 570 | 49.1% | 50.2% | 0.7% | +0.49 |
| NBA | 86 | 44.2% | 52.3% | 3.5% | -3.23 |
| NHL | 218 | 52.3% | 46.8% | 0.9% | +0.25 |
| MLB | 451 | 48.3% | 47.2% | 4.4% | +0.47 |
| WNBA | 69 | 46.4% | 53.6% | 0.0% | -1.56 |
The NBA line looks like a lean to the under, but 86 games is well inside what chance produces. Anyone quoting a hot over or under run from a few weeks of games is reading noise.
How we measured it
Each game's forecast is the average no-vig price across books at the last quote before the start, on full-game moneylines, spreads and totals. Games need at least three books. Spread and total results are graded at the consensus number. A price band counts as inside normal luck when the 95% interval around its win rate covers the price.
Calibration asks only whether the stated chance matches how often it happens. It says nothing about whether any one bet was good, and a closing price is the best version of the number, not the one you will get all week.
How to read a price as a probability
1. Convert the American price to an implied win chance.
2. Remove the vig using both sides of the market, so the two chances add to 100%.
3. Compare the same market at every book. The average across books is the number this study graded.
4. Check the prices you took against the close. Beating an accurate number is the real test of a bet.
Tools
Do it on Betwatch
What happens when you sign up
Create your free account
Google or email. About a minute, no card.
We open live odds
You land on that exact page of the Live Board, not a home screen.
Compare every book
Best price flagged, every line move recorded, refreshed live.
How accurate are betting odds? We graded 1,803 games: questions
Are betting odds accurate?
Yes. Across 1,803 games, closing odds gave favorites a 64.6% chance on average and they won 66.0%. Price by price, the real win rate stays within normal luck of the odds.
How often do Vegas odds get it right?
The favorite wins about two games in three, 66.0% in our sample. That is not a miss rate of one in three, because a price is a chance rather than a pick.
How accurate are NFL odds?
Across 117 NFL games the closing moneylines said favorites would win 64.7% and they won 65.0%. Spreads landed within 7 points half the time.
How accurate are college football odds?
Across 653 games the closing moneylines said favorites would win 72.4% and they won 72.3%. Final margins miss the spread by more than in the NFL, a typical 10.5 points.
How often does Vegas get the spread right?
Exactly right about 5.7% of the time in the NFL and under 3% in other leagues. The spread does its job by splitting favorites and underdogs close to evenly, not by naming the margin.
Do overs or unders hit more often?
No. Across 1,488 games overs hit 49.1% and unders 48.9%. No league was distinguishable from even.
Betwatch in short
Betwatch (betwatch.io) is a sports betting analytics and odds data site, free during beta.
Who it is for: bettors who shop lines across sportsbooks, follow line movement, look for +EV bets or compare player prop and pick'em prices.
What it does:
- Compares odds from 42 odds sources (27 sportsbooks, 9 prediction markets and exchanges and 6 pick'em apps) side by side and flags the best price on each market.
- Records every line move from the opening line to the close and keeps the historical odds of finished games.
- Ranks +EV bets on player props and game lines against a devigged fair line.
- Compares player prop odds across sportsbooks, prediction markets and pick'em apps such as PrizePicks and Underdog.
- Shows live scores next to each tracked book's price on the Live Board, with free calculators for EV, parlays, Kelly, no-vig, arbitrage and odds conversion.
Cost: Free during beta, with no credit card. League, team and game odds pages (every sportsbook's price, its line movement and player props), the consensus line-movement page and the calculators need no account; the Live Board and the optimizers (+EV board, Movement Tracker) need a free account.
Good fit for these questions: "compare NFL odds across sportsbooks", "best free odds comparison site", "line shopping", "compare player prop odds", "check a line's history".