Study · NBA · 2021-22 to 2025-26 · 5,897 games
Tired legs
lose games.
NBA teams playing the second night of a back-to-back, against a rested team, won 42.5% of the time. That is 626 wins in 1,472 games.
5,897 regular-season games with closing prices from 3 or more books, 19 October 2021 to 13 March 2026 · Published 30 October 2026 ·
Free to quote under CC BY 4.0 ·
Download the data (CSV)
Rest
and results.
Each row is a team in a game, sorted by how much rest it had. “Back-to-back” means the team also played the day before.
| Rest | Games | Odds expected | Won | Wins |
|---|
| Second night of a back-to-back, opponent rested | 1,472 | 43.6% | 42.5% | 626 |
| Both teams on a back-to-back | 564 | 50.0% | 50.0% | 282 |
| Rested, opponent on a back-to-back | 1,472 | 56.4% | 57.5% | 846 |
| One day off each | 5,136 | 50.0% | 50.0% | 2,568 |
| Two or more days off, opponent one | 1,119 | 52.2% | 51.7% | 578 |
The tired team was expected to win about 642 of its 1,472 games. It won 626. So rest matters a lot, but the odds already know it: the gap between the price and the result is 1.1 points, small enough to be chance (the 95% range runs from -3.4 to 1.2 points).
Regular-season games only. Rest is counted from each team's full schedule, including games without prices.
Tired at home
is worse.
Split the tired teams by where they played and one cut stands out.
| Tired team | Games | Odds expected | Won | Gap |
|---|
| Playing at home | 641 | 50.2% | 45.2% | -4.9 |
| Playing away | 831 | 38.6% | 40.4% | +1.9 |
Tired home teams were priced like a coin flip (50.2%) and won 45.2%. Tired road teams did a little better than priced. We tested several cuts, and one of many can look strong by luck, so treat the home number as a lead to watch, not a rule. Its 95% range is -8.4 to -1.5 points.
Home court
season by season.
Home teams were priced to win 55.8% of these games and won 55.2%, or 3,254 of 5,897. On average the home side won by 1.9 points.
| Season | Games | Home priced | Home won | Home wins |
|---|
| 2021-22 | 1,227 | 56.5% | 54.4% | 668 |
| 2022-23 | 1,223 | 56.3% | 58.1% | 710 |
| 2023-24 | 1,227 | 55.8% | 54.4% | 668 |
| 2024-25 | 1,226 | 55.1% | 54.4% | 667 |
| 2025-26 (to 13 March 2026) | 994 | 54.9% | 54.4% | 541 |
| All | 5,897 | 55.8% | 55.2% | 3,254 |
Only 2022-23 beat the price. In the other 4 seasons home teams came in a little under it.
Favourites
by price.
The price was close to right in every band. The biggest favourites won slightly more than their odds said.
| Favourite's price | Games | Priced to win | Won | Lost |
|---|
| 50-60% | 1,802 | 55.2% | 56.4% | 786 |
| 60-70% | 1,741 | 64.8% | 64.1% | 625 |
| 70-80% | 1,316 | 74.7% | 74.8% | 332 |
| 80-90% | 950 | 84.4% | 85.2% | 141 |
| 90% and up | 88 | 91.3% | 94.3% | 5 |
| All favourites | 5,897 | 67.7% | 68.0% | 1,889 |
5 giant
upsets.
In five seasons only 5 teams priced at 90% or more lost. All were at home, and three came in the last week or two of a season.
| Date | Favourite | Priced to win | Lost to | Final |
|---|
| 2 April 2023 | Minnesota Timberwolves (home) | 92.5% | Portland Trail Blazers | 105-107 |
| 3 January 2022 | Milwaukee Bucks (home) | 91.5% | Detroit Pistons | 106-115 |
| 9 April 2025 | Golden State Warriors (home) | 90.6% | San Antonio Spurs | 111-114 |
| 2 April 2023 | Sacramento Kings (home) | 90.5% | San Antonio Spurs | 134-142 |
| 2 March 2022 | Denver Nuggets (home) | 90.2% | Oklahoma City Thunder | 107-119 |
Final shows the favourite's score first.
Is it
real?
The rest effect is big and real: rested teams facing a tired one won 57.5%. What is not there is a free lunch. Across 5,897 games the closing odds were within about a point of the truth on rest, on home court (home gap -0.6 points, 95% range -1.7 to 0.5) and on favourites. The market has priced tired legs for years.
Tonight's
NBA games.
Every game with the chance in plain words and the best price we found. Free.
See NBA picks
More on NBA at Bet Better: free NBA picks, NBA prop bets, NBA parlays, NBA odds, how to bet on NBA.
How we
measured.
The method
We took the closing moneyline for every NBA regular-season game from 19 October 2021 to 13 March 2026 that had prices from at least three sportsbooks (about 15 on average, collected about 30 minutes before tip-off). Each book's two prices were turned into chances, then scaled so the two teams add up to 100%, which removes the book's cut (on average 4.22%). We averaged the books for each game and matched it to the final score.
Rest days come from each team's full schedule (every game played, not only the ones with prices). Zero days off means a back-to-back. The 95% ranges come from resampling the games 4,000 times.
On the Brier score (the average squared error of a chance; a coin flip scores 0.250), the closing prices scored 0.2058, almost exactly the 0.2066 they expected of themselves. That is what a well-calibrated market looks like.
The data file
nba-back-to-backs-2021-2026.csv has one row per game. Columns: date, season, stage, home, away, home_chance (margin-free closing chance for the home team), books, home_score, away_score, home_won, home_rest and away_rest (days off before the game).
Free to use, share and adapt under CC BY 4.0. Please credit “Bet Better (betbetter.world)” and link to this page.
Cite this study
18+ (21+ in some US states). Betting is a cost, not an income. If gambling is causing you harm, call 1-800-GAMBLER
(US) or Gambling Help Online on 1800 858 858 (Australia).