Browns 21, Panthers 18 — Week 3: Won It on Third and Ten

Third and 10 at the Carolina 14, one 55 left in the game, Cleveland down 18 to 13. Deshaun Watson had been hurt on the play that put the ball there, sat out one snap, and came back for the next two. On the second of them he threw short left to Fannin, and Fannin scored from 14 yards. 23 points of win probability on one snap — win probability being the model's read on each team's chance of winning at that moment — the biggest swing of the game. By expected points added, what a snap was worth given the down, the distance and the field position, measured against what that situation was already worth, it was +3.4. Watson hit Boston for the two-point conversion, 21 to 18. The Browns beat the Panthers 21 to 18 in Cleveland on Sunday, trailed for about seven minutes, all of them in the fourth quarter, and the model had them as low as 22 percent after Carolina's touchdown with four 29 left.

Start with the defense, because it is why there was a last drive to win. Carolina's 50 pass plays were worth −5.8 expected points — on defense a negative number is the good direction. At the rate this pass defense allowed over the first two weeks, those same 50 would have been worth about +4 and a half. At last season's rate, when Cleveland ranked in the top five in the league against the pass, about −6. So Sunday looked like last season, and not like Week 1, when this unit ranked last in the league at Jacksonville. With the run defense added in, it ranks fifth of the 28 teams that have played this week by total expected points allowed, despite 262 net passing yards allowed, twentieth so far this week. The stand to close the first half: up 10 to nothing, a 30-yard catch by Hubbard, two roughing-the-passer penalties and a holding call had carried Carolina to first and goal at the Cleveland four with 47 seconds left. Two incompletions, then McGuire's 13-yard sack on third and goal, and Carolina kicked a 35-yard field goal. And the finish: fourth and 16 at the Carolina 16 with one 27 left, stopped eight yards short, and then, after the offense gave the ball back, third and 10 with 16 seconds left, Bryce Young short left for Waller, and Hickman intercepted at the Cleveland 46.

Third down was the other half of it. Carolina converted four of 15, and Cleveland's 11 stops, 73 percent, rank fourth among the 28 teams so far this week. The run defense did its work upstream of that: only 30 percent of 23 Carolina rush plays were successful plays, meaning they gained what the situation asked for. 98 yards and −2.6 expected points, ninth of 28 per play. At the rate this run defense allowed over the first two weeks — it ranked 30-first in Week 2 at Tampa Bay — those plays would have been worth about +1 and a half. It bent on Carolina's go-ahead drive, when three straight runs gained three, 11 16 yards; two plays later Young hit Metchie from eight yards to make it 18 to 13.

The passing game has an odd line. 127 net passing yards is the second-fewest of the 28 teams that have played this week; only the Giants had fewer. But 33 pass plays — sacks included, and there were two — were worth +1.1 expected points, and Cleveland's best passing game of all last season was worth plus point eight. At the rate the Browns passed last season, when they ranked last in the league, these 33 would have been worth about −9 and a half. At the rate of the first two weeks, about zero. The value came from no interceptions that counted and two short touchdowns on the downs that mattered most. Besides the winner, there was fourth and goal from the two with one 43 left in the first half, Watson short right to Fannin: +4.3 expected points, more than any other snap in the game. 36 percent of the pass plays were successful, 20-third of the 28 teams that have played this week.

The run game opens on its baseline, because it is the unit that moved most from Week 2. 28 rush plays, counting Watson's scrambles, for 134 yards and −3 and a half expected points. At the rate Cleveland ran over the first two weeks — thirtieth in the league in Week 2 — those plays would have been worth about −6. Still below average: 20-first of 28 this week per play, and at last season's rate, when the run game ranked 20-fifth, about −1. Two runs set up the first-half touchdown, 11 yards on the first snap after Carolina's failed fourth down and 17 to the Carolina 10 on second and eight. The biggest came from the quarterback: third and three at the Carolina 45, two 55 left, down five, Watson scrambled 21 yards to the 24, 17 and a half points of win probability. And the one that nearly undid it: up 21 to 18, one oh five left, fourth and one at the Carolina 15, Judkins up the middle for a two-yard loss, 17 points of win probability handed back.

The last drive started at the Cleveland 35 with four 25 left. After the scramble and an incompletion, an illegal shift wiped out a five-yard completion, and on second and 15 from the Carolina 29 Watson's deep throw was intercepted in the end zone by Evans — erased by a roughing-the-passer call on Okereke that moved the ball to the 14, 17 points of win probability on a play that goes in the book as no play. Shedeur Sanders took one snap while Watson was out, a pass to Judkins for no gain; Watson threw incomplete on second down, and then came third and 10.

The rest of the drive chart: 10 drives, two touchdowns, two field goals from Szmyt from 48 45 yards, five punts, the fourth-down stop, and no offensive turnovers. Three were three-and-outs. Cleveland's one giveaway came on special teams: Corley fumbled the second-half kickoff and Carolina recovered at the Cleveland 30, −5.4 expected points, which became a Carolina field goal. Third down was five of 15, tied for twentieth so far this week, two of the conversions Watson scrambles. The longest drive of the day by the clock went nowhere. It spanned the first and second quarters, nearly nine minutes from the Cleveland nine to the Carolina 43, with back-to-back false starts turning third and one at the Carolina 36 into third and 11, a converted fourth and three, and a holding penalty that wiped out a 23-yard completion to the Carolina 11. It ended in a punt, and the game stayed scoreless for its first 24 minutes.

For anyone new to this show, here is where that sits. Last season the Browns went five and 12 with the league's last-ranked passing game by expected points per play, a run game 20-fifth and a third-down offense in the bottom five — and a defense in the top five against the pass, the top 10 against the run and eighth on third down. Week 1 was a 34 to 10 loss at Jacksonville in which that pass defense ranked last in the league and the offense converted one third down in eight. Week 2 was a 23 to 19 win at Tampa Bay with a passing game seventh in the league and a run game and run defense thirtieth and 30-first. Sunday was the first of the three games in which both halves of the defense played at or near last season's rate, and the passing game again beat anything Cleveland produced last season, this time on a hundred 27 net yards. Three games do not settle much, and a three-point win that turned on a penalty is not a verdict. But the defense this team was built on showed up against the pass and the run in the same game for the first time this season.

Cleveland hosts the Steelers on Thursday night, October first.

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

And if you play fantasy, there's a Muffed show built around your exact roster — free at muffed.ai.

_Data via nflverse._