Cleveland beat Pittsburgh and is 3-1 with a 16 percent favorite record, and our model lifted the Browns to seven wins without believing the record
Updated
Cleveland beat Pittsburgh 27-24 at home on Sunday in a game our model gave the Browns a 16 percent chance of winning, and the Browns are 3-1. That is three straight wins after the opening loss in Jacksonville, and the model's response is the most interesting one of the week. When we ran the rest of the season Cleveland came out at 7-10 with a five percent playoff chance, up from 5.7 wins after Week 3 and nearly double the four wins we projected in July. The model has moved three full wins on the Browns since the summer. It still projects them to lose nine of their last 13, because the roster it sees on paper is a seven-win roster and three one-score wins against teams it had ahead of Cleveland do not change the paper.
The quarterback has been the reason the record is what it is. Deshaun Watson is 80 of 115 for 855 yards with six touchdowns and two interceptions, a 101 rating through four games, and has been sacked eleven times. Our model projects him for 3,485 yards and 16 touchdowns, up from 3,235 and 14 a week ago, and the rating projection has climbed to 92. Denzel Boston has 13 catches for 284 yards and two scores and projects for 77 catches and 1,248 yards as the clear number one. Harold Fannin Jr. has three touchdowns on 17 catches and projects for six. KC Concepcion projects for 63 catches. Quinshon Judkins has 177 rushing yards on 59 carries and projects for just 668, which is a ground game that is not working. The offense projects for 266 points, 30th in the league, and the Browns have scored 81 in four games, so the model is expecting the pace to slow.
The defense is the better half and the model has believed in it since the summer. Cleveland projects to allow 360 points, 19th in the league, but gives up the ninth-fewest yards, which is a bend-but-hold profile. Mason Graham has four sacks in four games. Isaiah McGuire projects for eight, Maliek Collins has jumped from three projected sacks to seven and a half, and Jared Verse's line has dropped from nearly eleven to under seven. Carson Schwesinger projects for 107 tackles. Ronnie Hickman Jr. and Grant Delpit project for two picks each. The team projects for 38 sacks, 21st. The unit held Carolina to 18 and Pittsburgh to 24, and both were games it was supposed to lose by a touchdown.
The division is where the record and the projection diverge most sharply. Cleveland is 3-1 and projects for seven wins. Baltimore is 3-1 and projects for eleven. Cincinnati is 2-2 and projects for ten. Pittsburgh is 2-2 and projects for nine and a half. The Browns are tied for the division lead and project to finish last in it, and when we ran the rest of the season that is exactly what happened in most simulated years. The Browns play the Ravens twice and the Steelers again, and the model gives them 22 percent or less in each.
The next month is where the schedule stops cooperating. A trip to the Jets at 22 percent, Baltimore at home at 18, a trip to Tennessee at 46, a trip to Pittsburgh at 22, and a trip to New Orleans at 40. When we ran that stretch Cleveland went 1-4 more often than anything else, and the Tennessee game is the only one the model treats as a coin flip. A 4-5 record entering November would still be ahead of the seven-win projection, and the Browns would have earned every bit of it.
The player lines are the model giving the Browns credit without giving them the division. Watson's projection climbed from 3,235 yards to 3,485 and from 14 touchdowns to 16, with his rating projection up from 89 to 92 and his rushing line cut from 228 yards to 161. Judkins' rushing projection rose from 597 to 668 and his receiving line from 22 catches to 31. Boston's yardage projection went from 1,153 to 1,248. Concepcion's from 565 to 622. Jeudy's held at 34 catches. On defense, Collins' sack projection went from 3.3 to 7.6, the largest move on the roster, McGuire's from 7.4 to 8.2, and Verse's from 10.8 to 6.7. Delpit's interception projection doubled to over two. Hickman's dropped from three to 2.4. The number to watch is the points scored. The Browns have 81 in four games and project for 266 across 17, which is a 14-point average from here. The model expects the offense to slow down. Watson's last three games say it might not.
The matchup record is the largest overperformance in the league. We gave Cleveland ten percent in Jacksonville and the Browns lost by 24. We gave them 24 in Tampa Bay and they won. We gave them 26 against Carolina and they won. We gave them 16 against Pittsburgh and they won. Three wins from three-quarters of an expected win is a gap no other team comes close to, and it is the reason the projection has climbed three wins in a month while still landing at seven. The market opened Cleveland at five and a half wins. Our model sat a win and a half under it in July and sits a win and a half over it now. The lean is honest about what has happened and skeptical about what comes next. Watson has played well, the defense has held, and three one-score wins are three one-score wins. The next five opponents are a different class.
