Three teams tied for the best record in football last season. New England, Denver and Seattle all finished 14-3 — and if you only watched the highlights, they probably blur together into "the good teams." They shouldn't. The data says they were three completely different machines: one was carried by the league's best offense, one by the league's most relentless pass rush and an almost unbelievable record in close games, and one by a suffocating defense and a single receiver who swallowed half his team's passing game.
That's what this issue is: the 2025 season retold by what actually happened on the field — the stuff that decided games, some of which never made a highlight reel. It's also the background you'll want loaded before 2026 kicks off, because this season's story only makes sense as a sequel.
The champion looked old-fashioned on purpose
Seattle won the Super Bowl, and the shape of the team would have made sense to your grandfather: defense, run commitment, and one great receiver.
The defense was the real thing — second-best in football by expected points allowed per play. (EPA — expected points added — is the measuring stick we'll use all season: it values every play by how much it actually moved the scoring math, instead of counting yards that don't matter.) The offense, as a whole, was 13th — fine, not fearsome. And the identity was deliberate: only two teams in the league ran the ball more often, relative to what game situations called for, than Seattle did.
What made the passing game work was concentration. Jaxon Smith-Njigba didn't just lead the league in receiving yards with 1,793 — he accounted for half of Seattle's downfield passing game, a 50.1% share of the team's air yards, the highest in football, on the league's highest target share. And Sam Darnold — a punchline three years ago — quietly completed passes at a rate 4.3 points above what his throws' difficulty predicted, fourth-best among all qualifying quarterbacks. He didn't carry the team. He didn't have to. He hit what was there, seventeen games in a row.
Was it a fluke run? The data says mostly no — and we'll come back to that below.
The best offense in football lived in Foxborough
New England's 14-3 was built the opposite way: the league's #1 offense by EPA per play, and it wasn't particularly close on passing downs — the Patriots' pass offense also ranked #1.
The engine was Drake Maye, and his season deserves to be remembered precisely. He completed 71.9% of his passes against an expected rate — based on the difficulty of every throw he attempted — of 62.8%. That gap, +9.1 points over expectation, wasn't just the best in football; it was more than four full points clear of second place. Whatever you think the ceiling of that team is, the 2025 floor was real: the best throwing season in the league, on the league's best offense.
The team that won on the ball
Chicago won the NFC North at 11-6, and all season the story was "the defense is back." Here's the precise version, because precision is the whole point of this newsletter.
The Bears' defense forced 33 takeaways — most in the NFL, on 23 interceptions and 10 recovered fumbles — and the team's +22 turnover margin also led the league. But by expected points allowed per play, that same defense ranked 19th. Ordinary, snap to snap. What Chicago had wasn't a great defense; it was a historic ball-hawking season attached to an average one.
And they needed every takeaway, because the passing offense had a problem the wins concealed: Caleb Williams completed 58.1% of his passes against an expected rate of 65.0% — a gap of −6.9, ranking 36th of the 36 qualifying quarterbacks in football. Chicago won a division with the least accurate qualifying passer in the league. That is a remarkable, strange, true sentence about the 2025 season, and it happened mostly because of the ball. Remember it when we talk about what tends to repeat.
Denver's season came down to a possession, thirteen times
Denver's 14-3 had real structure under it: the league's most productive pass rush (68 sacks, most in football), a top-six defense by EPA per play, and an offensive line that allowed the fewest sacks in the league — first place at both ends of the line of scrimmage.
But here's the number that defined their season: Denver went 11-2 in games decided by one score. Thirteen of their seventeen Sundays came down to a possession, and they won eleven. Nobody in the league lived closer to the edge, and nobody walked it better. Whether that's a skill a team can keep is one of the oldest questions in football analytics — and it's the entire subject of Issue 3, because a decade of data has a strong opinion about it.
The collapse that wasn't quite a collapse
Kansas City went 6-11 and missed the playoffs, and the easy story was decline. The data's story is stranger. By points scored and allowed, the Chiefs played like a nine-or-ten-win team — their record ran three and a half wins below their underlying performance. The sharpest number: Kansas City went 1-9 in one-score games. Denver's season, inverted. The same coin, landing tails ten times.
Patrick Mahomes' own season splits in two directions. His accuracy over expectation, −2.9, was the lowest of his eight-year career and ranked 31st among qualifiers — that part is real and worth watching. The other direction is what that number was measured through: he played fourteen games before a torn ACL ended his season three weeks early. He's been fully cleared for camp. What Kansas City actually was in 2025 — a broken team or an unlucky one with a diminished superstar — is a genuinely open question, and it's exactly the kind the weekly Read exists to track.
Real or Noise — 2025 edition
The verdicts, in the calibrated language we'll use all season. These grade the story, not the team.
- "Seattle was a fluke champion": noise. The profile is an earned one — a top-two defense, a points-scored-and-allowed record that supports 13 wins, and a merely good 6-3 in one-score games. The championship wasn't luck wearing a uniform.
- "Chicago's defense is elite": noise — with a real thing inside it. The ball-hawking was genuinely historic (#1 in takeaways and margin). The down-to-down defense was 19th. Those are different claims, and only one of them tends to carry into the next season — turnover margin barely repeats year to year.
- "Drake Maye is now the best QB in football": scope it correctly and it's real. Best accuracy in football, by daylight, on the best offense — that's verified. "Best QB" is a bigger sentence than one season of data should be asked to carry. We'll say the part we can prove.
The Receipts
Verified numbers worth keeping in your pocket.
- Denver finished first at both ends of the line of scrimmage — most sacks produced (68) and fewest allowed (23, a league-low 3.4% of dropbacks). Line play doesn't trend on social media; it decided a #1 seed.
- Matthew Stafford threw 46 touchdowns against 8 interceptions — the most touchdown passes in football, across all 17 games. The league's best deep passing season belonged to a man in year seventeen of his career… and we'll leave 2026 expectations for Issue 3.
- The biggest single lever in football, in one line: on charted dropbacks last season, quarterbacks produced +0.257 EPA per play from a clean pocket and −0.464 under pressure — a swing bigger than the gap between the best and worst offenses in the league. Next issue is entirely about this hidden layer of the sport.
Where to hear it
Muffed is an AI that reads every play, every box score, and every advanced stat from all 32 teams — and tells you the real story of the game. Every number above was checked against the source data before it shipped. That's the house rule, and the receipts are always in the footer.
The audio version of any team's season is one click away: Seattle, New England, Chicago, Denver, Kansas City. During the season, Muffed builds a weekly show for your favorite team or your fantasy roster — muffed.ai.
Next week: The sport under the sport — what 19,800 charted dropbacks reveal about how games are actually won.
Methodology: team and player figures from nflverse play-by-play, 2025 regular season. Accuracy over expectation (CPOE) via NFL Next Gen Stats via nflverse. Pressure split: FTN Data via nflverse, 2025 season. "Played like a nine-or-ten-win team" = pythagorean expectation — the record a team's points scored and allowed would normally produce. Data via nflverse.