On this season's tape, the Steelers and Browns are the same team. Pittsburgh is still favored on the road because of what it did last year, and that leaves the Browns at +126 in a game I'd call a coin flip. That's the bet.
Our simulator never sees a betting line. It builds every player's projection from this season's play-by-play and plays the game 200 times. Cleveland wins 54% of them, by less than a point on average. I shade that down to about 51%, because the sim can't see that Cleveland is starting a backup center tonight against T.J. Watt and Alex Highsmith. My call is Browns 21, Steelers 20. The market makes Cleveland a 43% shot.
Both teams won close ones on Sunday.
Week 3 at a glance
Steelers 30, Bengals 27
- Rodgers: 19 of 34, 292 yards, 3 TD, 1 INT
- Jaylen Warren: 17 carries, 127 yards, plus 3 catches for 49
- Roman Wilson: 38-yard TD on the opening drive
- Derrick Harmon strip-sacked Joe Burrow with 44 seconds left
Browns 21, Panthers 18
- Watson: 16 of 30, 144 yards, 2 TD
- Harold Fannin Jr.: 7 catches, 51 yards, both touchdowns
- Down 18–13, an 8-play, 55-yard drive for the win
- Ronnie Hickman picked off Bryce Young with under 30 seconds left
Why the market likes Pittsburgh
Because of 2025. On last year's play, Pittsburgh would be about a touchdown better than Cleveland. Blend in this season and the gap closes to nothing before home field. On this year's tape alone, Cleveland has been much the better team, though three games of anything is a shaky foundation.
The market's number sits near last season. The sim's sits near even.
Implied margin, by rating
Pittsburgh's offense is wading through mud
Aaron Rodgers is 42, and the problem has a very specific shape. He's 2 of 20 on throws that travel 20 yards or more in the air. An average quarterback would have completed six or seven of those. He's also throwing deep twice as often as he did last year, and holding the ball longer while he waits for it to come open. Every one of those shots has been a wasted down.
Here's the part that surprised me. Underneath, he's sharper than he was in 2025. Throws under 10 yards have been better than last year, and his line isn't the main culprit either. Pittsburgh gives up pressure at about a league-average rate, and Rodgers has been bad even on clean dropbacks. The offense stalls because he keeps reaching for something that isn't there anymore.
Sunday's 292 yards look like a turnaround. Mostly they're three plays, two of them catches by Roman Wilson. Take those away and the Steelers were flat against one of the worst defenses in football.
Jaylen Warren is the part that's working, and tonight he has the backfield to himself with Rico Dowdle out. That matters more than it sounds. Dowdle was averaging 2.5 yards a carry, and without Warren, Pittsburgh's run game is the least successful in the league. Even his 127 yards come with an asterisk. Two runs of 35 and 26 made the day, and his other 15 carries went for 66. That's still better than he ran in the first two weeks, against a Cleveland run defense that's been just as leaky as Pittsburgh's.
If I'm wrong about Pittsburgh, it's because Rodgers stops forcing it. A Steelers offense that lives on short throws and Warren can move the ball on anybody.
Cleveland is a little better than its close wins look
The Browns' record looks lucky. A blowout loss, then two wins by a combined seven points. Look at who they played, though. Week 1 was at Jacksonville, the second-best team on either schedule, and Cleveland's offense has faced tougher defenses than most. The defense is why they're 2–1. Mason Graham has four of the team's seven sacks from the interior, and the few takeaways they've had come from the secondary, including Hickman's interception to finish off Carolina.
Then there's Denzel Boston. At Tampa in Week 2, Cleveland had gone eight drives without a touchdown. Late in the third quarter Boston caught a slant and broke three tackles on a 55-yard run to the end zone that tied the game at 16. Blake Whiteheart scored the go-ahead touchdown in the fourth, and the Browns haven't lost since.
It's a great story. The numbers are less romantic. The offense was less bad after that play, mostly thanks to the rest of that Tampa game, and it moved the chains no more often than before. Boston's own role hasn't grown, either. He drew 13% of Cleveland's targets on Sunday, his lowest share of the season.
Cleveland's two rookie receivers (total EPA, 2026)
So is Boston real? Yes, with a footnote. He's one of the league's most efficient receivers per target, and he does it 15 yards downfield, then adds yards after the catch on top. Take away his two longest plays and he looks more like a solid starter than a star. One of those two came in garbage time at Jacksonville anyway.
KC Concepcion is the more interesting case. Cleveland's first-round pick leads the team in targets and has done almost nothing with them. The reason is how he's used. Cleveland keeps throwing him screens and passes behind the line of scrimmage. He gets more separation than anyone on the roster, and the Browns keep using him like a handoff. His ugly completion numbers mostly trace back to Watson's off night against Carolina. If Cleveland sends him downfield a couple of times, his line looks very different.
| Boston | Concepcion | |
|---|---|---|
| Targets | 15 | 20 |
| Catches | 9 | 12 |
| Yards | 195 | 80 |
| EPA per target | +0.85 | −0.17 |
| Avg. target depth | 15.0 | 5.6 |
| YAC over expected | +4.8 | −2.7 |
| Separation (yds) | 2.6 | 3.5 |
| Target share, wk 1/2/3 | 18/23/13% | 23/20/30% |
| Sim tonight: targets | 4 | 9 |
| Sim tonight: yards | 27 | 57 |
The simulator follows usage, so it hands Concepcion nine targets and Boston four. Given how Cleveland has actually split the ball, that's probably about right.
One thing does tilt toward Boston tonight. Pittsburgh's corners are thin, with Jalen Ramsey trying to play through a broken wrist. If Ramsey is limited, that's the matchup Cleveland should go after.
Deshaun Watson is the swing piece. When he scrambles, Cleveland picks up big chunks. When they call his number on designed runs it's the opposite: they lose ground, including three short-yardage stops last week. Pittsburgh rushes four and rarely blitzes. If that keeps him in the pocket, it takes away his best way of making plays.
How this goes wrong
Pittsburgh's pass rush against Cleveland's patched-up interior. This is the matchup I'm most worried about, and it's the reason I shade the sim. Cleveland is missing center Elgton Jenkins and guard Teven Jenkins, so Luke Wypler starts at center next to a backup. That line was already among the worst in football at turning pressure into sacks. Watt and Highsmith have 6.5 of Pittsburgh's 8 sacks. Watson is a good quarterback when he's kept clean and a bad one when he isn't. If Pittsburgh lives in the backfield, the Browns lose this game.
The pick
Browns moneyline +126
- Our win %
- 51%
- Break-even
- 44%
- Sim win %
- 54%
This is the one I like. I think these teams are even, and the price says Pittsburgh is clearly better because of a season that's over. Keep it small: the line problem is real, and so is the gap between my number and the sim's.
How often it hits, against what the price needs
On the side
Browns +2.5. If you'd rather have the points, the sim covers 61% of the time, and a coin-flip game decided by a field goal lands on your side of the number. I prefer the moneyline price, but this is the safer way to say the same thing.
Watson over 29.5 rushing yards. I'd take a small bite. The scrambles carry this one, and a collapsing pocket gives him more reasons to run. My worry is the same rush getting home before he can escape, which turns would-be scrambles into sacks.
The total. I'm staying away. My number is about 41 and the sim's is higher, so the over is the lean. But rain is likely by the fourth quarter, and taking the over means betting on a Pittsburgh offense I just spent a whole section knocking.
Warren's 67.5 rushing yards is right where we have him, Dowdle or no Dowdle, so I'm leaving it alone.
Reference: who's out
Steelers
- RB Rico Dowdle: out (toe)
- CB Jalen Ramsey: questionable (broken wrist)
- DL Yahya Black: questionable (personal)
- CB Joey Porter Jr.: traded to Dallas, hadn't played in 2026
Browns
- C Elgton Jenkins: out (concussion); Luke Wypler starts
- G Teven Jenkins: out (back), 4th straight game
- WR Tylan Wallace: out (knee)
- CB Tyson Campbell and S Grant Delpit: cleared
Reference: projected stat lines
Median lines from the 200 sims, assuming everyone plays. PPR is the average, with the 10th–90th percentile range next to it. The sim doesn't know about tonight's offensive line or cornerback absences.
| Player | Median line | Any TD | PPR | 80% range |
|---|---|---|---|---|
| Deshaun Watson QB · CLE | 229 pass yds, 1.7 TD, 0.8 INT, 31 rush yds | 31% | 19.8 | 12–29 |
| Aaron Rodgers QB · PIT | 248 pass yds, 1.5 TD, 0.6 INT | 19% | 16.0 | 8–24 |
| KC Concepcion WR · CLE | 9 tgt, 5 rec, 57 yds | 48% | 15.2 | 7–27 |
| Jaylen Warren RB · PIT | 17 car, 65 rush yds, 3 rec, 12 yds | 46% | 15.0 | 7–25 |
| Quinshon Judkins RB · CLE | 14 car, 53 rush yds, 2 rec, 15 yds | 35% | 12.3 | 6–21 |
| DK Metcalf WR · PIT | 6 tgt, 4 rec, 46 yds | 34% | 12.2 | 4–22 |
| Harold Fannin Jr. TE · CLE | 6 tgt, 3.5 rec, 40 yds | 24% | 10.3 | 2–22 |
| Roman Wilson WR · PIT | 4.5 tgt, 3 rec, 36 yds | 23% | 9.0 | 2–19 |
| Denzel Boston WR · CLE | 4 tgt, 2 rec, 27 yds | 29% | 8.5 | 2–17 |
| Pat Freiermuth TE · PIT | 4 tgt, 3 rec, 25 yds | 9% | 6.7 | 1–13 |
| Michael Pittman WR · PIT | 3 tgt, 2 rec, 18 yds | 16% | 6.2 | 0–13 |
| Jerry Jeudy WR · CLE | 3 tgt, 2 rec, 18 yds | 16% | 5.8 | 0–13 |
Reference: the tape and the model
2026 unit ranks (longer bar = better)
Rank of 32. The better unit's number is in bold. "Competitive" means win probability between 10% and 90%.
| Team | Week 1 | Week 2 | Week 3 |
|---|---|---|---|
| PIT | W 20–13 vs ATL 31st | L 3–20 @ NE 16th | W 30–27 vs CIN 17th |
| CLE | L 10–34 @ JAX 2nd | W 23–19 @ TB 24th | W 21–18 vs CAR 12th |
The small number next to each opponent is its 2026 rank in opponent-adjusted net EPA/play.
The simulator projects each player first and builds the score from their box scores. That conversion is a formula fit on 4,222 real team-games from 2018–25: 2.9 + 5.72 per offensive touchdown + 0.017 per yard, R² 0.80, with σ 4.4 points of noise for field goals and everything else. No Vegas numbers go in, and 2-point tries are left out. Here's where we already know it leans:
- Usage over efficiency. It projects receivers from how often they're targeted, which is why it likes Concepcion more than Boston.
- Track records for veterans. It gives Rodgers and Metcalf more credit than their 2026 tape has earned, so it's probably generous to Pittsburgh's passing game.
- High totals without a Vegas line. Every game starts from a neutral 44.5 baseline, which is part of why my 41 sits under the sim's 45.
- Blind to injury news. It assumes everyone plays, so it can't see Cleveland's missing linemen or Pittsburgh's thin cornerback room.
Where the total lands
Built on three games of data and written Thursday afternoon, before inactives.