Steelers 24, Browns 27 — Week 4: Erased 21-10, Lost on a 56-Yarder

24 all, a minute 42 left, and Pittsburgh had just come all the way back. Down 24 to 16, it had faced fourth and 10 at the Cleveland 11. Aaron Rodgers's throw for Darnell Washington fell incomplete, but McNeil-Warren was flagged for pass interference, first and goal at the three. Pat Freiermuth caught the touchdown on the next snap, and Rodgers ran the two-point try in himself. By win probability, the model's read on each team's chance of winning at that moment, the flag was worth 10.7 points and the two-point run 13.7. Then the Browns went 36 yards in a minute 21, and Szmyt kicked a 56-yard field goal with 15 seconds left, 21.8 points of win probability, the biggest swing of the night. Rodgers's last throw, deep right for Roman Wilson, was intercepted. The Steelers lost to the Browns 27 to 24 on Thursday night in Cleveland and are two and two.

The passing game was better than its own September and still made the two biggest mistakes. 45 pass plays, five sacks included, were worth +3.4 expected points. Expected points added is what a snap was worth given the down, distance and field position. At the rate this passing game ran over the first three weeks, those same 45 would have been worth about −10. Rodgers completed 22 of 40 for 299 yards and three touchdowns. DK Metcalf carried the yardage, a hundred 15 on five catches. The throw that restarted the game: down 21 to 10, second and eight at the Pittsburgh 25 early in the fourth, Rodgers went deep right for Metcalf, 43 yards in the air, and Metcalf caught it at the Cleveland 32. +3.3 expected points. Four snaps later Rodgers found Washington deep middle for a 21-yard touchdown. The two-point try, a completion to Jaylen Warren, came up short. Roman Wilson's 12-yard touchdown on second and 12 had opened the scoring in the first quarter, and his 49-yard catch set up a field goal at the end of the half.

The charting is where the cost shows. Rodgers's throws were catchable on 25 of 40, 63 percent, 20-fourth of the 30 charted passing games this week, against 66 percent across the first three weeks. Two were charted interception-worthy, after one in a hundred 12 throws through Week 3, and two were intercepted. The first came on the second snap of the second quarter, up seven to nothing, deep right for Michael Pittman: Ward picked it off at the Cleveland 44, and four snaps later the game was tied. Charters put all five sacks on Rodgers himself, after one in the first three weeks. Cleveland sent five or more rushers on 36 percent of his dropbacks. Play-action stayed in the bag: six fakes on 45 dropbacks, 13 percent, 20-ninth of 30, after 23 percent through Week 3. Four of the six were completed for 64 yards.

The run game opens on its success rate, because it was the steadiest thing Pittsburgh did. 53 percent of 19 rush plays were successful plays, meaning they gained what the down and distance asked for. They were worth +1.2 expected points; at the rate of the first three weeks, about −1.2. Warren carried 17 of the 19 for 93 yards, Travis Homer the other two. Cleveland showed six or fewer defenders in the box on 18 of Pittsburgh's 19 designed runs, 95 percent, the highest rate any of the 30 charted offenses saw this week, after 53 percent through Week 3. The runs gained 4.8 a carry, after 4.6. Warren's longest, 24 yards around left end out of the no-huddle in the second quarter, led to a missed field goal. Boswell's 48-yarder went wide left, down 14 to seven, 8.3 points of win probability.

The pass defense gave the game back through the air. Cleveland's 35 pass plays, two sacks included, were worth +9.2 for the Browns, and on defense a positive number is the bad direction. At the rate this defense allowed over the first three weeks, those 35 would have been worth about −2 and a half. 57 percent were successful, and Deshaun Watson threw for 268. His throws were catchable on 79 percent, the fourth-highest rate of the 30 charted passing games this week, after 74 percent allowed through Week 3. Pittsburgh did make them tight: 10 of 33 throws were contested, a receiver and a defender arriving together, 30 percent, after 10 percent through Week 3. It sent five or more on 12 of 37 dropbacks, 32 percent, thirteenth of 30. The big one beat it anyway: seven all, second quarter, Watson deep left for Denzel Boston, 60 yards to the Pittsburgh 14. Three snaps later Cleveland led. The interception was Spears-Jennings's, deep middle for Boston on the second snap of the fourth quarter, after a pass-interference flag came off on replay review.

The run defense won its matchup. 25 Cleveland rush plays came to −3.2 expected points, and on defense a negative number is the good direction. At the rate of the first three weeks, about plus point one. Only 36 percent were successful. Pittsburgh went light more often than before, six or fewer in the box on 14 of 23 designed runs, 61 percent, after 34 percent through Week 3, and allowed 3.7 a run, after 4.4. On second and goal from the one before halftime, Highsmith and Heyward stopped Quinshon Judkins for a yard loss. And up 21 to 10 in the third, Watt forced Watson's fumble and recovered it himself at midfield, −5.7 expected points for Cleveland. But the runs that scored were the problem: Jaleel McLaughlin's only carry went 28 yards for a touchdown, and Judkins scored from the two.

The turning point was the fumble Pittsburgh could not use. Down 11 with four 13 left in the third, at midfield, the offense got seven yards to Ben Skowronek and 13 from Warren to the Cleveland 30. Then a false start, two incompletions and a 10-yard sack on third and 15, and a punt. 12 drives: three touchdowns, one field goal, two interceptions, and four three-and-outs. Third down was four of 13; the red zone, 12 snaps and two touchdowns. On the last Cleveland drive, a Payton Wilson sack on second and four pushed the kick back to 56 yards. It went through.

Now what repeats. Last week's notebook had three lines. Drops at a high rate: broken. One drop on 25 catchable balls, four percent, after nine percent through Week 3. Contested throws running low: broken the other way, seven of 40, 18 percent, double the nine percent of the first three weeks. And a play-action package that produced on a low rate: the rate went lower still and the four completions came to 64 yards. Motion went up, 41 of 89 charted plays, 46 percent, fifteenth of 30, after 37 percent through Week 3. What goes in the notebook now: five sacks the charters put on the quarterback after one in three weeks, and two interception-worthy throws after one. Is that one opponent, or a quarterback holding the ball?

The stat lines, scored half-PPR. Rodgers led with 22. Roman Wilson, 14.9: three catches on six targets for 74 and a touchdown. Warren, 14.1: 93 rushing yards and three catches for 33. Metcalf, 14. Washington, 10.2 with the touchdown. Freiermuth, 9.2 with the late touchdown. Pittman caught two of four for 15. For Cleveland, Judkins scored 18.6.

That's your Steelers — every game, by the numbers. This is Muffed: we do the same deep dive for all 32 teams. Follow the Steelers show wherever you're listening, so next week's episode lands the moment it's out. No theater. Just the football.

_Data via nflverse. Charting: FTN Data via nflverse._