2025 NFL season · final
Turnover luck, 2025: who the fumbles bounced for
The Bears recovered 22 of the 30 fumbles in their games in 2025, 7.0 more than chance would give, worth +1.05 wins. The Lions recovered 9 of 32, 6.8 fewer than expected (−0.65 wins).
A loose ball goes back to the team that fumbled it 51.7% of the time, whoever fumbled it and wherever. Beating that rate doesn’t last: a team’s recovery rate one season correlates with the next at 0.09 (192 team-seasons since 2019). Luck in wins counts each bounce by the moment it came: a fumble recovered late in a tied game is worth most of a win, one in a blowout next to nothing. Regular season only. Methodology
Fumble recoveries by team, 2025
Every fumble in a team’s games, its own and its opponents’: how many it came away with, against how many a 51.7% bounce would give.
| # | Team | Fumbles | Recovered | Rate | Above expected | Luck in wins |
|---|---|---|---|---|---|---|
| 1 | 30 | 22 | 73% | +7.0 | +1.05 | |
| 2 | 31 | 20 | 65% | +4.6 | +0.65 | |
| 3 | 41 | 25 | 61% | +4.5 | +0.18 | |
| 4 | 31 | 20 | 65% | +4.4 | +0.63 | |
| 5 | 36 | 22 | 61% | +4.0 | +0.41 | |
| 6 | 38 | 23 | 61% | +4.0 | +0.35 | |
| 7 | 26 | 17 | 65% | +3.9 | +0.32 | |
| 8 | 28 | 17 | 61% | +2.9 | +0.10 | |
| 9 | 28 | 17 | 61% | +2.9 | +0.22 | |
| 10 | 25 | 15 | 60% | +2.5 | +0.25 | |
| 11 | 29 | 16 | 55% | +1.3 | +0.36 | |
| 12 | 26 | 14 | 54% | +1.2 | +0.12 | |
| 13 | 32 | 17 | 53% | +1.1 | +0.59 | |
| 14 | 24 | 13 | 54% | +1.0 | −0.38 | |
| 15 | 32 | 17 | 53% | +0.8 | −0.19 | |
| 16 | 23 | 12 | 52% | +0.5 | +0.05 | |
| 17 | 28 | 14 | 50% | +0.1 | +0.09 | |
| 18 | 31 | 15 | 48% | −0.4 | +0.07 | |
| 19 | 25 | 11 | 44% | −1.5 | −0.48 | |
| 20 | 35 | 16 | 46% | −1.5 | −0.27 | |
| 21 | 33 | 14 | 42% | −2.5 | −0.15 | |
| 22 | 37 | 16 | 43% | −2.5 | +0.01 | |
| 23 | 19 | 7 | 37% | −2.5 | −0.64 | |
| 24 | 31 | 13 | 42% | −2.6 | −0.23 | |
| 25 | 34 | 14 | 41% | −2.9 | +0.03 | |
| 26 | 30 | 12 | 40% | −2.9 | −0.20 | |
| 27 | 37 | 15 | 41% | −3.5 | −0.20 | |
| 28 | 29 | 11 | 38% | −3.5 | −0.42 | |
| 29 | 31 | 11 | 35% | −4.6 | −0.48 | |
| 30 | 31 | 11 | 35% | −4.6 | −0.56 | |
| 31 | 35 | 13 | 37% | −4.7 | −0.64 | |
| 32 | 32 | 9 | 28% | −6.8 | −0.65 |
Fumbles: every fumble in the team’s regular-season games where the ball was loose, its own and its opponents’. Above expected: recoveries minus the 51.7% each fumbler keeps on average. Luck in wins: each bounce’s swing in win probability at the moment it came, net of the opponent’s.
And interceptions?
Interceptions are only partly luck. FTN’s charting marks every throw a defender could have caught. How many of those a team throws is mostly its quarterback: it carries over from one season to the next at 0.39. Whether the defense holds on is closer to chance (0.02, 96 team-seasons), but within a season it runs with team quality, so our luck numbers leave interceptions out (methodology). Here are both sides, 2025.
| Team | Its interception-worthy throws | Picked | Opponents’ interception-worthy throws | Caught | Caught vs expected |
|---|---|---|---|---|---|
| 18 | 6 | 27 | 21 | +6.3 | |
| 15 | 8 | 16 | 12 | +3.3 | |
| 21 | 11 | 20 | 14 | +3.1 | |
| 15 | 5 | 20 | 14 | +3.1 | |
| 12 | 9 | 19 | 13 | +2.6 | |
| 20 | 10 | 14 | 10 | +2.4 | |
| 18 | 11 | 29 | 18 | +2.2 | |
| 14 | 6 | 22 | 14 | +2.0 | |
| 20 | 9 | 26 | 15 | +0.8 | |
| 11 | 6 | 17 | 10 | +0.7 | |
| 23 | 11 | 17 | 10 | +0.7 | |
| 18 | 8 | 12 | 7 | +0.5 | |
| 25 | 12 | 14 | 8 | +0.4 | |
| 18 | 9 | 5 | 3 | +0.3 | |
| 19 | 15 | 24 | 13 | −0.1 | |
| 20 | 14 | 13 | 7 | −0.1 | |
| 9 | 6 | 15 | 8 | −0.2 | |
| 24 | 16 | 12 | 6 | −0.5 | |
| 14 | 9 | 16 | 8 | −0.7 | |
| 10 | 5 | 18 | 9 | −0.8 | |
| 18 | 6 | 11 | 5 | −1.0 | |
| 26 | 16 | 17 | 8 | −1.2 | |
| 22 | 13 | 14 | 6 | −1.6 | |
| 14 | 10 | 16 | 7 | −1.7 | |
| 11 | 5 | 16 | 7 | −1.7 | |
| 22 | 11 | 18 | 8 | −1.8 | |
| 14 | 8 | 22 | 10 | −2.0 | |
| 17 | 13 | 15 | 6 | −2.2 | |
| 22 | 10 | 17 | 7 | −2.3 | |
| 19 | 8 | 21 | 9 | −2.4 | |
| 14 | 7 | 27 | 12 | −2.7 | |
| 16 | 12 | 9 | 0 | −4.9 |
Interception-worthy throws: FTN Data via nflverse. Expected catches use each defense’s own record on such throws, shrunk toward that season’s league rate.
Questions
Which NFL team had the most fumble luck in 2025?
The Bears: they recovered 22 of the 30 fumbles in their games (73%), 7.0 more than expected, worth +1.05 wins, the most in the league.
Is turnover luck real?
For fumbles, yes. A loose ball goes back to the team that fumbled it 51.7% of the time, and a team that beats that rate one season doesn't the next: the correlation between consecutive seasons is 0.09 over 192 team-seasons since 2019. Interceptions are different: how many interception-worthy passes a quarterback throws carries over (0.39), so part of a team's turnover margin is skill.
Do teams that recover a lot of fumbles keep doing it?
No. Recovery rates since 2019 carry over from one season to the next at 0.09, close to the zero a coin would give. Teams at the top of this table usually land near the middle a year later.
Other seasons: 2026 · 2024 · 2023 · 2022 · 2021 · 2020 · 2019
2025 luck rankings · Kicking luck · Injury luck · Regression candidates · Every kind of luck