Twenty-Fourth to First, and the Number Holds

Atlanta allowed 126.2 rushing yards a game in 2025 and finished 24th in the league. Through four games of 2026 it allows 48.3, which is first, and those two numbers belong to the same franchise running most of the same front seven.

That is the largest single change in this team's statistical profile. The figure circulating this week is the 48.3, and in the five pieces of this week's coverage I read, not one of them set it beside the 126.2. That is the figure that makes it mean anything.

The line, whole

193 rushing yards allowed in four games. 71 rushing attempts faced, the fewest in the league. 2.7 yards a carry, first. The longest run anybody has gotten on this defense all season is 17 yards, which is second. Two rushing touchdowns allowed, fifth.

Baltimore, over the same two seasons, did not move: 106.6 a game in 2025, 10th; 100.0 now, 10th. Atlanta shed 77.9 rushing yards a game and climbed 23 places.

The improvement was also announced before it arrived. A cached March club-site note on the defensive staff's offseason priorities names run defense as the 2025 weakness being targeted, with rotational depth added to address it — cached, not live, and labeled that way on purpose. A stated intention in March plus a league-best number in October is a different class of evidence than a four-game curio.

Why I will vouch for this one

Yesterday I wrote that ESPN's team statistics were returning 0 for Atlanta's yards allowed and 0 for its points allowed while ranking the Falcons Tied-1st in both, and that "an unpopulated field wearing a first-place rank is the most dangerous shape a bad number takes." Twenty-four hours later I am handing you another ESPN defensive first place. So here is the work.

The 48.3 does not live on the surface that was broken. It comes off the by-team statistics endpoint, Opponent Rushing split, and it is populated: 193 yards, 71 attempts, 2.7 a carry. I did not take the rank on trust either. I pulled opponent rushing for all 32 teams and sorted it myself, ascending. Atlanta is first at 48.3. Tampa Bay is second at 72.8, which means Atlanta allows 24.5 fewer rushing yards a game than the second-best run defense in the league. Baltimore is 10th.

Then the step that settles it. The 193 reconciles to four box scores with nothing left over: 58 at Pittsburgh, 68 against Carolina, 17 at Green Bay, 50 at New Orleans. The 71 attempts do the same — 21, 22, 9, 19. An unpopulated field cannot add up to four separate afternoons. That is the whole difference between the rank I refused yesterday and the one I am printing today, and it is the only reason I am printing it.

The rate traveled, the volume has not arrived

Now the objection, which is mine rather than somebody else's. Facing 71 attempts is fewer than 18 a game, and some of that is scoreboard rather than run-stopping. Atlanta led Green Bay by three scores and New Orleans by three scores, and both opponents put the ball away: 9 carries and 19 carries. A per-game average is a volume statistic, and part of this volume was donated.

The control for that is the rate, and the rate survives it: 2.7 yards a carry, also first in the league, across all 71 of those attempts. Take the volume out of the argument and the achievement is still standing.

Three days ago I ran the same separation on Atlanta's run offense and found that 59 of 205 yards at New Orleans came on one carry — a single snap holding up a season number. Different unit, opposite result. No game holds this one up: the worst is 68, the best is 17, and all four are under 70. The symmetry is exact enough to be uncomfortable. Atlanta's longest run this season is 59 yards. Atlanta's longest run allowed is 17.

Which brings the limit, and it is a large one. Derrick Henry has 89 carries in four games. That is 22.25 a game by himself, against the 17.8 rushing attempts a game that every Atlanta opponent has managed combined. One back's workload is bigger than the entire run volume this defense has been asked to handle. Whatever the 48.3 measures, it has not measured that.

The counterweight is his efficiency: 374 yards at 4.2 a carry, which ties the lowest full-season figure of his career and sits well under the 5.9 he averaged in 2024 and the 5.2 in 2025. He is also finishing more often than he ever has, 7 rushing touchdowns in four games. A back running at a career-low rate and scoring at a career-high one is not a solved problem — he is a volume problem, and volume is the one input Atlanta's number has never been handed.

The front seven the injury report does not describe

Atlanta plays a base 3-4. The active roster lists five defensive ends: Samson Ebukam, Za'Darius Smith, Cameron Thomas, LaCale London and Zach Harrison. ESPN's Week 5 game report carries two of them as Questionable — Za'Darius Smith with a concussion, Ebukam with a hamstring — and on the depth chart those two are the outside linebackers, Smith on the weak side and Ebukam on the strong side. Atlanta's own Thursday report has Smith not practicing at all and Ebukam limited, which is the paperwork doing real work on the edge. It does none at defensive end. The players the depth chart actually starts there, London and Brandon Dorlus, appear on neither document. Thomas is on the Thursday report at full participation with a shoulder, and not on the game report. Ten front-seven players on that depth chart are on neither one.

So the framing this week wants — the league's most physical back arriving against an Atlanta front missing its edge — is not available from the documents. Three of five defensive ends appear somewhere, two of those three are the starting outside linebackers, and the two the depth chart starts at defensive end appear nowhere. Neither club has posted a game status for Sunday yet — every line on both official reports reads unspecified, so nobody is Out and nobody is Doubtful. Atlanta has seven names on its report, three of them hamstrings, Ebukam, Divine Deablo and Yasir Abdullah, which I will count and not explain. Baltimore has thirteen, Lamar Jackson's left ankle sprain among them, and I wrote that one yesterday.

Where the number got lost

Atlanta's own viewing guide for Sunday carries the sentence: "They currently boast the league's best run defense, allowing 48.3 yards per game, but that will be tested by Derrick Henry." No noun phrase anywhere near that pronoun owns it. The antecedent is Atlanta. At least one aggregator read it the other way and handed the number to Baltimore, along with a rushing-offense figure Baltimore does not have. The Ravens allow 100.0 a game, 10th, and run for 153.5, fourth.

The national short-form preview frame has Baltimore as the immovable object here. Dave Choate at The Falcoholic, locally, has the better-evidenced read: the Ravens "just don't have quite the same caliber of run game or run defense as the Falcons." That is checkable, and it checks out. Atlanta is second in the league in rushing offense at 181.0 a game and first in rushing defense; Baltimore is fourth and 10th.

None of which decides Sunday night. Four games is four games, 193 yards is a small sample, nothing on either injury report is settled yet, and at home this season Atlanta is 0-1 — the one game at Mercedes-Benz Stadium was Carolina 34, Atlanta 3, which is also why Carolina and not Atlanta leads this division. Atlanta has one win in its last five against Baltimore, and it came in 2010.

What I will defend is the number and who owns it. Twenty-fourth to first is not a hot streak inside a box score. It is the single thing about this football team that genuinely changed, and it changed in the one category the staff said in March it was going after. Sunday at 8:20 is the first night anybody actually runs at it.