Thirty-Five Carries, and the Number Doesn't Hold

With eight minutes and twenty-seven seconds left in the third quarter, Justice Hill took a handoff up the middle and ran forty-seven yards before Avieon Terrell pushed him out of bounds at the Atlanta 18.

The yardage is not the most useful part. The tackler is. On an interior run, forty-seven yards from where it started, the player who finally arrived was a cornerback — someone whose assignment puts him off the line of scrimmage by design. The front had been displaced, the linebackers had been passed, and the safeties had been passed, in that order, before anyone laid a hand on him.

Atlanta came into Sunday night having allowed a longest run of seventeen yards across the entire season, second-fewest in the league. Hill nearly tripled it on one carry. He finished the night with four.

That drive went five plays and seventy-four yards and ended in a Baltimore touchdown. One carry by the backup running back was forty-seven of the seventy-four.

Yesterday I printed the objection against my own number. Atlanta's league-best 48.3 rushing yards allowed per game had been assembled against four of the seven lowest rush-rate offenses in football — Pittsburgh at 33.9 percent, Carolina at 35.4, Green Bay at 31.1, New Orleans at 34.0 — while Baltimore runs on 53.1 percent of its plays. I set the test in advance and I set it as a number rather than a feeling. Twenty-five carries. If the 48.3 survived twenty-five carries, the number described a defense. If it dissolved, then the objection I had raised against myself was the entire story.

Baltimore ran thirty-five times for two hundred and eight yards.

That is not a narrow miss on the threshold. It cleared it by ten carries, and the yardage deserves to be stated the way it actually lands: Baltimore gained more on the ground on Sunday night than Atlanta had allowed in its previous four games put together. The four-game total was 193. One night was 208. The season figure moved from 48.3 to 80.2, a jump of nearly thirty-two yards a game delivered by a single opponent, and the rate went from 2.7 a carry to 3.8.

So the number did not hold and the objection was correct. On Friday I wrote that Atlanta owned the league's best run defense and that nobody had run at it yet. The first half of that sentence was a fact about arithmetic. The second half was the part that carried the risk, and Sunday night is the thing it had been waiting for.

Now the part I had wrong.

Henry was contained, and Atlanta still lost on the ground

My stated assumption was that volume would be the mechanism — that Derrick Henry, who had taken eighty-nine carries in four games while every Atlanta opponent combined had managed seventy-one, would be handed the ball until the front wore through. The box score does not support that.

Henry finished with twenty-four carries for a hundred and two yards. That is 4.3 a carry, within a tenth of the season rate he brought into the building, and he did not score a rushing touchdown. Against the defense his workload was supposed to expose, Henry was held to roughly what he had been averaging against everybody else. On a per-carry basis, at the point of attack, Atlanta's front did the job it was built to do.

The 5.9 a carry Baltimore finished with is not Henry's number. Remove his twenty-four carries and the other eleven produced a hundred and six yards. Hill's four produced seventy-three of those.

This is a different failure from the one I described, and the distinction is the whole diagnosis. A defense ground down by a power back has a problem with strength at the line. A defense that holds that back to his average and then concedes forty-seven yards to his backup has a problem with gap integrity behind the line — a second level arriving late, or arriving in the wrong gap, against run concepts it has not had to defend because none of the first four opponents bothered to make it.

Atlanta's front seven did not lose a test of strength on Sunday. It lost one of geometry, and then it lost two of leverage.

Both plays that decided the game were runs

Atlanta led 23-15 with two minutes and thirty-one seconds to play, at home, having just banked a forty-eight-yard Nick Folk field goal.

Baltimore scored with forty seconds left, Tyler Huntley to Devontez Walker for seven yards, and went for two. It was the second time they had gone for two. The first came in the third quarter after Henry's eighteen-yard touchdown catch, and they threw it: Huntley to Mark Andrews, incomplete, attempt failed.

The second time, they handed it to Henry around left end. It succeeded. Tie game.

In overtime, Huntley ran it in himself from a yard out.

Both of the plays that actually decided this game were runs, and the second two-point call is the one to sit with. Baltimore had already tried to convert through the air once and failed. With the game on the line at the end of regulation, in a hostile building, against the defense leading the league in rushing yards allowed per game, Jesse Minter chose to run at it — and then chose to run at it again from the one in overtime. That is a judgment about what Baltimore believed it had learned over the preceding fifty-nine minutes. It was rewarded twice.

The asterisk, in both directions

Baltimore did this without Lamar Jackson, who took zero snaps. That cuts two ways and both deserve saying.

In Atlanta's favor: a Jackson offense is a harder run-defense problem than a Huntley offense, because the quarterback becomes a carrier the structure has to account for on every snap, and the 48.3 had never faced that either. Sunday was not the maximum version of this examination.

Against Atlanta: Jackson's absence is also why Baltimore's intent was so legible. Huntley completed twenty-four of thirty-two for two hundred and one yards, which is efficient and almost entirely unthreatening in the way that should let a front commit downhill without fear. Atlanta knew what was coming, at home, with the crowd, and still conceded two hundred and eight yards and both of the short-yardage conversions that mattered. A defense that is genuinely elite against the run holds up better than that against a backup quarterback handing off to a backup running back.

Where this leaves five games

Atlanta is 2-3 and 0-2 at home. Both home games this season are losses, and a point differential that was plus-four on Sunday morning is now minus-three. Carolina, at 2-2, holds the division on its record inside it, and Atlanta sits a game back of a team that beat it by thirty-one points in this same building three weeks ago.

The two road wins by twenty-one apiece are still on the ledger and they still happened. But the shape of this team after five games is a club that wins in Green Bay and New Orleans and has not yet won in front of its own people, and the defensive identity that was supposed to be the floor underneath all of it drew its first real examination on Sunday night and did not pass.

I am keeping the 48.3 in the record, because it was true on the day it was printed. It simply turns out to have been a description of four opponents rather than a property of this defense.