Week 4 ran from Thursday, October 1, when Cleveland beat Pittsburgh 27-24, to Monday night, when Atlanta won 45-24 in New Orleans. Three teams are 4-0: Kansas City, Minnesota and San Francisco. Five are 0-4: Houston, the Chargers, Miami, Tampa Bay and Tennessee. Two of those five, Houston and the Chargers, played in January. Of the three unbeaten teams, only San Francisco did.
Four games is a quarter of a season. It is where the question changes from "was that a fluke?" to "is that who they are?" Last week's issue built a ruler for three games and named what Week 4 would grade: San Francisco's offense, Atlanta's one big game, Minnesota's defense, the Raiders' defense and Houston's 0-3. This one rebuilds the ruler for four, from the same ten seasons, 2016-2025: every team's first four games against the rest of its regular season, and every receiver's first four games against the rest of his.
Here is the ruler, most reliable first. Each number is a correlation between the first four games and the rest of the season; 1.00 would mean the first month tells you the rest exactly, zero that it tells you nothing.
- Receiver target share: 0.70.
- A defense's blitz rate, five or more pass rushers as charted by FTN: 0.68, over the four charted seasons, 2022-2025.
- Team offense, by expected points added (EPA) per play, which measures how many points each snap added to the expected score: 0.54.
- Pass rate over expected (PROE), how much more or less a team throws than down, distance and score predict: 0.51.
- Team defense, by EPA allowed per play: 0.29.
- Receiving touchdowns per game: 0.19.
That ladder is the answer to the title. What a team chooses to do (how often it blitzes, how much it runs, whom it throws to) repeats. How well it does those things repeats about half as much on offense and much less on defense. Touchdowns barely repeat at all.
A fourth game moved the offense number from 0.50 at three games to 0.54, and did not move defense at all: 0.29 at three, 0.29 at four. Last season still counts, at 0.33 for offense and 0.18 for defense, but four games now outweigh it on both sides of the ball.
The bluntest instrument is the record. Over the same ten seasons, 4-0 teams reached the playoffs 20 times in 22. Teams at 3-1 made it 59 times in 84, 2-2 teams 41 times in 103 and 1-3 teams 11 times in 76. Teams at 0-4: none of 23.
No 0-4 team has reached the playoffs in ten seasons. This season has five, and two were playoff teams in January.
Houston lost 34-30 at home to Dallas on Sunday with its best offensive game of the season. The offense ranked third in the league for the week at +0.24 EPA per play. C.J. Stroud went 21 of 31 for 347 yards and two touchdowns, and Nico Collins caught 7 of 8 targets for 118 yards and both scores. It lost anyway because the defense, first in the league across 2025, ranked 25th for the week and ranks 21st over four games.
Last week's issue said the 0-3 base rate ran against Houston and that both exceptions in ten seasons were Texans teams. A fourth loss moves them into a group with no exceptions. Of the 23 teams that started 0-4 from 2016 to 2025, none reached the playoffs. Two of the 23 had been playoff teams the season before, the 2017 Giants and the 2020 Texans, and that did not save either. Only two of the 23 had a top-half offense over those four games, the 2017 Chargers (ninth) and the 2019 Broncos (15th). Houston's offense ranks 19th.
The Chargers are the other January team in the group, and their case is weaker. Their offense ranks 32nd of 32 over four games at -0.16 EPA per play, and Sunday's 30-23 loss in Seattle carried two Justin Herbert interceptions and four sacks. Miami has scored 46 points in four games, the fewest in the league.
None of this says a season is over. It says no team in the last ten has climbed out of this particular hole, and that the two teams in it with the best résumés are a defense that slipped from first to 21st and an offense ranked last.
San Francisco's offense passed its fourth test. Every first-place offense after four games since 2016 stayed in the top ten.
Last week's issue called San Francisco's offense "leaning real" and said the fourth game would grade it. The 49ers ranked tenth in Week 4, their lowest of the season, in a 24-14 win over Denver. They still rank first over four games at +0.29 EPA per play, a tenth of a point per play clear of Buffalo in second at +0.19. San Francisco and Kansas City remain the only two offenses in the top ten in every week, and both are 4-0. Brock Purdy went 19 of 30 for 218 yards and two touchdowns without an interception.
The ruler is kind to this kind of team. In each of the ten seasons, one offense ranked first after four games. Eight of those ten ranked in the top five over the rest of the season, and all ten ranked in the top ten. Widen it to the top five after four games and there are 50 offenses: 23 stayed in the top five and 8 fell to the bottom half.
San Francisco also has last season behind it, sixth in the league across 2025, so the two inputs the ruler weighs agree. A quiet week ranked tenth is what a first-place offense's bad week looks like.
Kansas City is the same story one rung down. The Chiefs rank fifth over four games at +0.17 and won 30-27 in Las Vegas behind Kenneth Walker III, who ran 22 times for 177 yards and two touchdowns. They have a bye in Week 5.
Atlanta's run-first identity is real. Its offense is average, which is a promotion.
Two weeks ago the Falcons' offense ranked 32nd of 32. It has now ranked second and first in consecutive weeks, scoring 35 in Green Bay and 45 in New Orleans after 13 and 3 in the first two games. On Monday it posted +0.42 EPA per play, the best mark of Week 4, and scored touchdowns on 6 of its 10 drives without a turnover.
The game was decided early and on the ground. On the third snap, second-and-1 from the Atlanta 41 with 13:43 left in the first quarter, Bijan Robinson ran around left end for a 59-yard touchdown. He finished with 19 carries for 145 yards and two scores. Brian Robinson added three touchdowns on 14 carries, and Atlanta ran for 207 yards and five touchdowns in all. Michael Penix Jr. went 15 of 20 for 223 yards and a touchdown, 96 of those yards to Drake London.
Last week's issue said Atlanta's Week 3 was "one of three." It is now two of four, and the ruler has a direct answer for how much the second big game counts: no more than the two bad ones. Across ten seasons, an offense's EPA in its third and fourth games predicts the rest of its season at 0.42; its first and second games predict it at 0.44. Recent games carry no extra weight. The honest read is the four-game figure, and that puts Atlanta 18th at -0.02 per play. That is the middle of the league, reached from the bottom of it.
Ten offenses since 2016 made the same turn, bottom ten over their first two games and top five over the next two. Four of the ten finished the rest of the season in the top half and three in the top ten; the median finished 17th. The schedule tempers it further: Green Bay's defense ranks eighth once its Atlanta game is set aside, but New Orleans's ranks 24th.
What is real is how Atlanta plays. The Falcons have the league's lowest PROE through Week 4, -10.3, meaning they throw 10.3 percentage points less often than the situation predicts. Of the ten offenses that ranked most run-leaning after four games from 2016 to 2025, nine stayed among the seven most run-leaning over the rest of the season; the exception was the 2024 Chargers. The plan is the part to believe. Whether it keeps producing +0.37 EPA per rush, Monday's figure, is the 0.54 question, and two games do not answer it.
Defense is still mostly noise at four games. Minnesota's blitz is the exception.
Minnesota is 4-0 with an offense that ranks 27th over four games. The defense is why. It ranks first at -0.18 EPA allowed per play (on defense, a negative number is good), and it is the only defense in the top ten in every week: tenth, ninth, fourth and fourth. On Sunday it held Miami to 10 points in a 15-10 win.
How it does it has not changed in four weeks. Minnesota sent five or more pass rushers on 111 of 153 charted dropbacks, 72.5%, and led the league in each of the four weeks, including 14 of 20 against Miami. Washington is second at 47.4%; the league rate is 31.2%. No defense in the four charted seasons before this one blitzed that often over its first four games. The highest was 58.6%, by Minnesota in 2023.
That is the most reliable number in this issue. Blitz rate carries from the first four games to the rest of the season at 0.68, and the first-month blitz leader in each of 2022-2025 finished the rest of its season first, second, sixth and third. Minnesota blitzing is the identity to believe. Minnesota's blitzing working is the 0.29 question. Of the ten defenses that ranked first after four games from 2016 to 2025, five were top five the rest of the way and seven top ten. Minnesota also ranked fourth across 2025, so last season and this month agree.
The Raiders are the cautionary case. Last week's verdict on them was "the record, yes; the defense, wait." The defense ranked seventh, fifth and 13th in the first three weeks, then 24th in Sunday's 30-27 loss to Kansas City, when Walker ran for 177. It still ranks fourth over four games at -0.10, after 20th across 2025. The ruler on that kind of defense is cold: of the 50 that ranked top five after four games, 14 stayed top five and 15 finished in the bottom half. Las Vegas is 3-1, and 3-1 teams have made the playoffs 59 times in 84.
Seattle is the defense with both inputs behind it. It ranks second over four games at -0.14 EPA allowed per play and ranked second across 2025, and it is 3-1 after Sunday's win over the Chargers.
Keep the target shares. Discount the touchdowns.
The receiver half of the ruler splits along the same line as the team half: the usage numbers repeat and the outcome numbers do not. Four games of target share carry into the rest of the season at 0.70, yards per game at 0.60, yards per target at 0.28 and touchdowns per game at 0.19. Of the receivers who ranked in the top 12 in target share after four games over the last ten seasons, 121 counting ties, 62 were top 12 the rest of the way and 95 were top 24.
Jaxon Smith-Njigba is the share to believe. He has 42 of Seattle's 115 targets, 36.5%, first among the 83 wide receivers and tight ends with at least 16 targets, with 481 yards and six touchdowns. Week 4 was his quietest game, 5 catches on 6 targets for 76 yards against the Chargers, and the share still leads the league. The six touchdowns are the part the ruler discounts.
George Kittle is the other side of the line. He has four touchdowns on 22 targets, a 20.0% share that ranks 39th of the 83, and he caught one of them against Denver. Christian Watson also has four touchdowns, on a 21.3% share that ranks 32nd. Touchdown totals built on middling shares are the 0.19 kind of number.
Drake London sits between the two. His share is 26.9%, 13th of the 83, and his 14.8 yards per target ranks second, behind Zay Flowers' 16.0. The share is the part that travels. The efficiency is a 0.28 number. And on the league's most run-leaning offense, the pie is small: Atlanta's 93 targets through four games are the fewest of any team, so the plan above caps his volume. Chris Olave has the most targets in the league, 48, on a 28.1% share that ranks 11th.
Real or Noise
The verdicts, calibrated on four games. These grade the story, not the team.
- "San Francisco has the best offense in football": real. First over four games at +0.29 EPA per play, top ten every week, sixth across 2025. Every one of the ten offenses that ranked first after four games since 2016 stayed in the top ten.
- "Atlanta's offense is fixed": the plan is real; the efficiency is noise for now. The league's lowest PROE, -10.3, and the most run-leaning offense after four games has stayed among the seven most run-leaning nine times in ten. The offense itself ranks 18th over four games, and games 3 and 4 predict no better than games 1 and 2.
- "Minnesota's defense is elite": the blitz is real; the results lean real. Five or more rushers on 72.5% of dropbacks, a rate no defense reached over its first four games in 2022-2025, on a measure that carries at 0.68. First by EPA allowed, top ten every week and fourth across 2025.
- "Houston can still dig out": the ruler says no team has. No 0-4 team in 23 since 2016 reached the playoffs, including two that had been there the year before. Houston's offense ranks 19th and its defense 21st.
- "Jaxon Smith-Njigba is the league's No. 1 target": real. The six touchdowns: noise. A 36.5% share leads the league on a measure that carries at 0.70; touchdowns carry at 0.19.
The Receipts
Verified numbers worth keeping in your pocket.
- Four games predict the rest of the season at 0.70 for receiver target share, 0.68 for defensive blitz rate, 0.54 for offensive EPA per play, 0.51 for PROE, 0.29 for defensive EPA and 0.19 for receiving touchdowns per game. Ten seasons, 2016-2025; blitz rate over 2022-2025.
- Teams that started 0-4 from 2016 to 2025 reached the playoffs 0 times in 23; 4-0 teams, 20 times in 22. This season has five 0-4 teams and three 4-0 teams.
- Minnesota has sent five or more pass rushers on 72.5% of opponents' dropbacks through Week 4. The highest first-four-games rate in 2022-2025 was 58.6%.
Read more
- San Francisco's Week 4 against Denver and Kansas City's in Las Vegas
- Houston's 34-30 loss to Dallas and the Chargers in Seattle
- Atlanta, Bijan Robinson and Drake London
- Minnesota's 15-10 win over Miami and the Raiders against Kansas City
- Jaxon Smith-Njigba's Week 4, Nico Collins' and George Kittle's
- The boards behind this issue: team EPA, pass rate over expected, defense blitz rate, target share and stat stickiness
- Last week's issue: Which three-game patterns are worth believing?
Follow the players you care about on Muffed, across all your leagues, and every Wednesday their weeks and the Read land in one email.
Next week: San Francisco's first-ranked offense visits Seattle's second-ranked defense, Houston and Tennessee meet at 0-4, and Atlanta's run game gets a fifth game to tell us whether it is a plan or a streak.
Methodology: Week 4 of 2026 is the 16 games from Thursday, October 1 through Monday, October 5; all 32 teams played. Team offense and defense are EPA per play on regular-season plays whose type is a pass or a run (scrambles included, penalty no-plays excluded), offense as the possession team and defense as EPA allowed, ranked among all 32 teams; "over four games" is the same measure across Weeks 1-4 together, and a weekly rank is that week alone. Green Bay's and New Orleans's defensive ranks with the Atlanta game set aside drop every play Atlanta ran from the whole league's table. PROE is the mean of nflverse's pass_oe on plays that carry it, as on the pass rate over expected board. Blitz rate is the share of charted dropbacks on which FTN counted five or more pass rushers, from the charting week snapshot. The four-game ruler is 2016-2025: for teams, each team's first four regular-season games (320 team-seasons; 288 with a prior season, for the last-season comparison) against its fifth game onward, as Pearson correlations; rank-based rates rank the first four games and the rest separately within each season; playoff rates use each team's record after four games and whether it played a postseason game that season, with ties counted as half a win and 0-4 meaning no wins or ties. The three-game figures repeat Issue 8's method (first three games against the fourth onward). Blitz-rate stickiness uses the same windows over 2022-2025 (128 team-seasons). For receivers: wide receivers and tight ends with at least 16 targets in at least three of their team's first four games and at least six games after, on one team in each window (822 player-seasons); target share is the player's targets over his team's targets in the games he appeared in; top-12 ranks are within each season among that group. This season's receiver figures use the same definition through Week 4 (83 qualifiers). Records are from nflverse's schedule file; player lines are nflverse's weekly player stats. Data via nflverse; FTN Data via nflverse.