How has the NFL changed since 1999?
The NFL of 2025 passes only a little more than it did in 1999 (60% of plays against 58%) but far more efficiently: completions are up from 53% to 60% of attempts, expected points added per play has turned positive for the first time, turnovers have fallen from 3.8 a game to 2.3, and teams go for it on 23% of fourth downs, double the 1999 rate. Kickers make 86% of field goals instead of 78% and try three times as many from 50 yards. Home teams won 60% of games in 1999 and 54% in 2025.
- 60%of plays were passes in 2025, against 58% in 1999 and a peak of 62% in 2015
- 23%of fourth downs were go-for-it attempts in 2025, a record; 12% in 1999
- 86%of field goals made in 2025, up from 78%, with three times as many tries from 50 yards
- 54%home win rate in 2025; 60% in 1999 and 61% in 2003
Based on every regular-season play and game from 1999 through week 3 of 2026, 7,015 games in all. The 2026 season is partial and is left out of every trend quoted below. Explore the live data.
Has the NFL become a passing league?
Less than the talk suggests. Passes were 57.7% of plays in 1999, peaked at 62.1% in 2015 and were 60.2% in 2025. What changed is the pass itself: completions rose from 53.0% of attempts to 59.6% (61.1% in 2020), the average throw got shorter, from 8.7 yards past the line in 2011 to 7.7 in 2025, and the shotgun went from 12% of snaps to 66%. Quarterbacks also scramble more than ever, on 5.9% of dropbacks in 2025.
| Season | Pass rate | Completion | Yards per attempt | Shotgun | Sack rate |
|---|---|---|---|---|---|
| 1999 | 57.7% | 53.0% | 6.28 | 11.9% | 7.0% |
| 2004 | 55.4% | 55.5% | 6.55 | 13.0% | 6.8% |
| 2009 | 58.4% | 57.1% | 6.54 | 36.6% | 5.9% |
| 2014 | 61.0% | 58.5% | 6.74 | 60.6% | 6.1% |
| 2019 | 62.0% | 59.0% | 6.71 | 64.5% | 6.4% |
| 2024 | 60.4% | 60.5% | 6.61 | 70.3% | 6.5% |
| 2025 | 60.2% | 59.6% | 6.50 | 65.9% | 6.5% |
Depth of target peaked at 8.65 yards in 2011 and bottomed at 7.65 in 2024; deep attempts have stayed between 11% and 13% of throws throughout. The no-huddle peaked at 12.5% of plays in 2014. The report draws every one of these as a small multiple with the rule changes marked on it.
Are NFL offenses more efficient than they used to be?
Yes, and the turn came late. Expected points added per play was -0.056 in 1999 and still -0.006 in 2019; it went positive in 2024 (0.009) and stayed there in 2025 (0.007), with the 2020 season the high at 0.032. Success rate rose from 39.4% of plays to 44.0%. Running improved most: EPA per rush went from -0.110 to -0.022 and yards per carry from 3.89 to 4.35. Third downs convert at 39.9% against 37.0%, and red-zone trips end in a touchdown 19.2% of the time against 17.5%.
| Season | EPA/play | EPA/dropback | EPA/rush | Success | Yards/carry |
|---|---|---|---|---|---|
| 1999 | -0.056 | -0.005 | -0.110 | 39.4% | 3.89 |
| 2004 | -0.018 | 0.032 | -0.070 | 42.1% | 4.13 |
| 2009 | -0.020 | 0.030 | -0.065 | 42.0% | 4.24 |
| 2014 | -0.015 | 0.059 | -0.085 | 42.5% | 4.16 |
| 2019 | -0.006 | 0.059 | -0.068 | 43.0% | 4.31 |
| 2024 | 0.009 | 0.065 | -0.026 | 43.9% | 4.42 |
| 2025 | 0.007 | 0.056 | -0.022 | 44.0% | 4.35 |
Explosive plays have not moved: 8.5% of plays in 1999, 8.8% in 2025, with a high of 9.75% in 2018. The gain is in the ordinary down, not the big one.
Do NFL teams go for it on fourth down more than they used to?
Twice as often. Teams went for it on 11.9% of fourth downs in 1999, hit a low of 11.1% in 2011 and reached 23.2% in 2025, the highest rate in the data. Turnovers have fallen from 3.81 a game to 2.29, mostly because interceptions dropped from 3.1% of attempts to 2.0%. Penalties have barely moved, 13.1 accepted flags a game in 1999 and 12.7 in 2025, while touchdowns rose from 4.65 a game to 5.13.
| Season | Fourth-down go rate | Interceptions | Turnovers a game | Penalties a game |
|---|---|---|---|---|
| 1999 | 11.9% | 3.1% | 3.81 | 13.1 |
| 2004 | 12.0% | 3.0% | 3.52 | 13.9 |
| 2009 | 14.4% | 2.9% | 3.40 | 11.8 |
| 2014 | 12.0% | 2.4% | 2.94 | 13.2 |
| 2019 | 16.2% | 2.1% | 2.77 | 13.5 |
| 2024 | 20.0% | 2.0% | 2.41 | 12.9 |
| 2025 | 23.2% | 2.0% | 2.29 | 12.7 |
The flag-happiest season was 2005; 2009 was the quietest at 11.8 a game. The 2026 season has started at 14.4 flags a game over 48 games, which is the kind of number three weeks produce and a full season rarely keeps.
How has NFL kicking changed?
Kickers are far better and asked to do far more. Field goals were made at 77.8% in 1999 and 85.6% in 2025; attempts from 50 yards or more went from 0.34 a game to 0.98, nearly three times as many, and those are now made 68.8% of the time against 48.8%. The extra point went the other way after the 2015 move to the 15-yard line: 98.9% in 1999, 93.9% in 2019, 95.9% in 2025. Punts fell from 10.4 a game to 7.1. Touchbacks were 10% of kickoffs in 1999, 64% in 2024 and 21% in 2025 under the new kickoff.
| Season | FG made | 50+ tries a game | 50+ made | PAT made | Touchbacks |
|---|---|---|---|---|---|
| 1999 | 77.8% | 0.34 | 48.8% | 98.9% | 10.1% |
| 2004 | 80.8% | 0.36 | 58.2% | 99.0% | 8.5% |
| 2009 | 81.3% | 0.41 | 52.9% | 98.1% | 16.1% |
| 2014 | 84.0% | 0.60 | 61.0% | 99.3% | 50.3% |
| 2019 | 81.6% | 0.57 | 57.9% | 93.9% | 60.9% |
| 2024 | 84.0% | 1.03 | 69.9% | 95.8% | 64.2% |
| 2025 | 85.6% | 0.98 | 68.8% | 95.9% | 20.7% |
Two-point tries went from a third of a game to half a game once the extra point got harder. The 2025 kickoff rule, which moved the touchback to the 35, cut touchbacks to a fifth of kickoffs in one season.
Is NFL home-field advantage shrinking?
Yes. Home teams won 59.7% of games in 1999 and 61.3% in 2003, the high; they won 49.8% in the empty stadiums of 2020 and 53.9% in 2025, and have not been back above 58% since 2016. Against the spread the home side covers about half the time in any season, 49.0% on average, with a high of 52.8% in 2013 and a low of 43.1% in 2019. Scoring rose from 41.6 points a game to 46.0 (49.6 in 2020), and the share of games decided by eight points or fewer rose from 50% to 53%, with 57.6% in 2022 the most ever.
| Season | Points a game | Home win rate | Home cover rate | One-score games |
|---|---|---|---|---|
| 1999 | 41.6 | 59.7% | 51.0% | 50.0% |
| 2004 | 43.0 | 56.6% | 46.4% | 47.3% |
| 2009 | 42.9 | 57.0% | 46.4% | 46.9% |
| 2014 | 45.2 | 56.9% | 49.2% | 43.0% |
| 2019 | 45.6 | 51.8% | 43.1% | 52.3% |
| 2024 | 45.8 | 53.3% | 51.1% | 55.9% |
| 2025 | 46.0 | 53.9% | 49.8% | 53.3% |
Overtime games have held between 3.5% and 8.2% of the schedule, the high in 2015. A third of games are now played indoors (33.8% in 2025, 22.6% in 1999) as domes and retractable roofs have replaced open stadiums. Totals have followed scoring: the average closing total line was 40.1 in 1999 and 44.9 in 2025, and the over has hit 52% of the time in each of the last two seasons.
Are NFL quarterbacks getting older?
No. Weighted by dropbacks, the average age of a quarterback with 200 dropbacks was 29.4 in 1999, peaked at 29.5 in 2001, fell to 27.1 in 2023 and was 29.0 in 2025. What has changed is the top and the bottom: eight quarterbacks aged 35 or more qualified in 2020 against one in 2011, and seven did in 2025. Running backs have got younger and more numerous: the lead back carried 70% of his team’s rushes in 2002 and 54% in 2021, and 1,000-yard rushers fell from 23 in 2000 to seven in 2015 before climbing back to 17 in 2025.
| Season | QB age | QBs aged 35+ | 1,000-yard rushers | 4,000-yard passers |
|---|---|---|---|---|
| 1999 | 29.4 | 5 | 14 | 5 |
| 2004 | 28.5 | 2 | 18 | 5 |
| 2009 | 28.7 | 3 | 15 | 10 |
| 2014 | 29.2 | 5 | 13 | 11 |
| 2016 | 29.1 | 5 | 12 | 13 |
| 2020 | 29.3 | 8 | 9 | 12 |
| 2023 | 27.1 | 5 | 12 | 10 |
| 2025 | 29.0 | 7 | 17 | 6 |
Thirteen passers threw for 4,000 yards in 2016 and six in 2025, even with a seventeenth game, as offenses lean back toward the run. Ages are season minus birth year.
How was this measured?
Play-by-play, schedules, lines and player seasons come from nflverse, 1999 to the current season, refreshed weekly. Expected points added, success and explosive plays are nflfastR model outputs. Air yards, and so depth of target, exist from 2006; the shotgun and no-huddle flags are not reliable before 2002, and the report starts those panels where the data does. Regular season only, and a partial season is drawn dotted.
Per-game rates count both teams. A fourth-down attempt is a pass or run on fourth down, out of every fourth-down play including punts and field goals. A home cover is a game where the home margin exceeded the closing home spread; version 3 of the report corrected that formula, which had been adding the spread instead of subtracting it and made home teams look like they covered 60% of the time. The rule-change markers in the report are context, not causation.
The data lives in a Power BI semantic model on Microsoft Fabric. Every number on this page was produced by an AI assistant running the report’s own queries against that model, and the same model powers the live public report.
Want answers like this from your own data?
This started as public data. We added location and conditions, built a model worth trusting, and pointed an AI at it. We do the same thing with sales, operations and finance data that nobody outside your company will ever see.