"Fade the team on the back-to-back" is one of the oldest NBA betting rules. It's also one of the most priced-in. Here is what the numbers show, from every regular-season game we have: 14,322 games at home arenas, 2014-15 through 2025-26.
A back-to-back here means a team also played the previous calendar day (Eastern time). It's known from the schedule days in advance, which is exactly why the market is good at pricing it.
The headline numbers
| Situation | Games | Home win rate | Avg. home margin |
|---|---|---|---|
| Neither team on a back-to-back | 9,735 | 56.2% | +2.1 |
| Only the home team on a back-to-back | 1,371 | 49.8% | -0.3 |
| Only the road team on a back-to-back | 2,477 | 61.0% | +3.9 |
| Both on a back-to-back | 739 | 58.6% | +2.4 |
Read it this way. A normal NBA home game is a 56% proposition for the home side. Put the home team on the second night of a back-to-back and it drops to a coin flip. Put the road team there and home rises to 61%. That's about a 2-point swing in scoring margin either way. For scale, the whole home-court advantage is about 2 points.
When both teams are tired, it roughly washes out.
Road back-to-backs are more common
Notice the counts. Road teams were on the short end almost twice as often as home teams (2,477 vs 1,371). The usual pattern is a home game one night, then a flight and a road game the next. That second night stacks travel on top of fatigue.
The league has been cutting them
The NBA has spent a decade reducing back-to-backs. In our data:
| Season | Back-to-back games per team |
|---|---|
| 2014-15 | 19.3 |
| 2016-17 | 16.4 |
| 2018-19 | 13.3 |
| 2021-22 | 14.1 |
| 2025-26 | 14.7 |
Fewer back-to-backs means fewer of these spots, but they haven't gone away. The average team still plays about 15 of them.
The effect on the home team has, if anything, grown. Before the 2020-21 season, home teams on a back-to-back still won 53.7% of the time. From 2021-22 on, that fell to 45.9%. Road teams on a back-to-back were hurt about the same in both eras. One likely reason: teams now rest regulars on the second night more than they used to. That is a player-availability effect wearing a schedule costume.
Home court itself has shrunk
A side finding: NBA home teams won 57.6% to 59.4% of games from 2014-15 through 2018-19. Since the 2020 bubble, most seasons have landed between 54.5% and 55.5% (2022-23 was an outlier at 58.1%). Our model's home-court setting (50 Elo points, about 57% between even teams) was fit to that recent reality, not the old one.
Is the back-to-back priced in?
Mostly, yes. Sportsbooks know the schedule better than anyone, and "tired team" is the first adjustment a bettor learns. The rest effect is real, but that doesn't make every rested side a bet. The question is always whether the price already moved enough.
That's how our NBA model uses it. The back-to-back is one input, worth 50 rating points (roughly 7 points of win probability between even teams). Our backtest chose that value because it predicted outcomes best. Then the model's number goes up against the actual moneyline. If the market moved further than the rest effect justifies, the tired team can be the value side.
The two mistakes to avoid:
- Auto-fading the back-to-back team. Books adjust the line; the bettor who fades blindly pays the adjusted price.
- Ignoring the second-order effect. The real swing on a back-to-back is often who sits. If a star's status is unclear at publish time, our card can drop the game entirely rather than guess.
How to use this
- Check the schedule, not just the injury report. Know which side played last night.
- Convert the moneyline to implied probability (how).
- Ask whether the gap between the two sides looks like 2 points of margin, or like a missing player.
- Bet flat. One spot is noise. Rest effects only show up over hundreds of games.
The free NBA Pick of the Day already does steps 1-3 before it posts. Results go on the NBA record.
Data: ESPN schedules and final scores, 2014-15 through 2025-26 regular seasons, excluding neutral-site games and the 2020 Orlando bubble. Win rates are descriptive, not a betting system.