Michigan Wolverines vs Rutgers Scarlet Knights prediction: Without Lonergan, Rutgers likely hands Michigan's stingy defense to AJ Surace

A backup against the best

Rutgers loses its quarterback the week it draws the nation's stingiest defense per play. That one change shapes everything about this Halloween afternoon in Piscataway, where Michigan is hunting its second conference win and has beaten Rutgers nine straight times.

Dylan Lonergan is out after a head injury knocked him from the Northwestern game early in the first quarter. Rutgers hasn't named a starter, but AJ Surace is the likely choice, with Sean Ashenfelder behind him. Surace went 7-of-18 in relief at Northwestern, a game Rutgers lost 42-14.

Lonergan was why the Rutgers passing game worked: 8.4 yards per attempt and 10 touchdowns in six games. His favorite target, KJ Duff, is doubtful with an undisclosed injury after 774 receiving yards and nine touchdowns.

The defense waiting for them allows 4.1 yards per play, best in FBS. It ranks ninth in passing EPA allowed (expected points added per play, a measure of how much damage opponents do through the air). It also ranks second in points allowed per scoring opportunity, at 2.08, so even the drives that cross midfield tend to end in field goals or nothing.

Michigan has its own health questions. Safeties Rod Moore and Jordan Young are questionable, as are edge rushers John Henry Daley and Dom Nichols. Rutgers gives up 2.57 sacks per game, 114th nationally, and a new quarterback rarely fixes protection.

Michigan's path runs on the ground

Rutgers allows 199.1 rushing yards per game, 129th of 138, and ranks 133rd in rushing EPA allowed. Recent form is worse. Over the last three games, opponents posted a 52.9% success rate (the share of plays that keep an offense on schedule) against the Scarlet Knights, who allowed 39.7 points per game in that stretch.

Our projection has Michigan rushing for 227 yards, a run-first, slow plan aimed squarely at that weakness. Michigan's own ground game is ordinary. Its 2.92 line yards per carry (yardage credited to the blockers) ranks 95th, and Jordan Marshall averages 3.9 yards a carry. Offensive linemen Andrew Sprague and Ace Hamilton are out, and tackle Blake Frazier is questionable.

The matchup still tilts toward Michigan because Rutgers rarely disrupts anything. Its havoc rate (share of plays with a tackle for loss, sack, takeaway or pass breakup) ranks 119th. That shields a Michigan offense that ranks 118th in havoc allowed.

Rutgers' best answer without Lonergan is its backfield. Clay Thevenin has 272 yards over the last three games at 6.3 a carry. Michigan ranks 10th in rushing EPA allowed, though, and that is a harder wall than anything Northwestern or Maryland offered.

Michigan's finishing problem

The worry for anyone laying the points sits at the other end. Michigan scores 20.3 points per game, 119th nationally. It scores 2.79 points per scoring opportunity, 125th.

The Indiana game captured it. Michigan reached the red zone four times and came away with six points, losing 13-6 in overtime, and Tommy Carr turned the ball over twice in the first half. Carr is still expected to make his second straight start after Bryce Underwood was benched, though Kyle Whittingham called the decision "to be determined" on Monday. Underwood threw four interceptions over his last three games.

The market keeps expecting margin that doesn't come. Michigan was favored against Iowa, Minnesota and Penn State and failed to cover all three, losing two outright. It averaged 13.7 points over its last three games.

Both teams also give the ball away. Michigan commits 1.86 turnovers per game, 130th, and Rutgers commits 2.0, 132nd. A short field from a Surace mistake could be the cleanest way for this Michigan offense to score.

What the number says

Michigan is a 14.8-point favorite, up from 13.8 at the open, with the total at 45. Our model lands almost on top of that: Michigan by 15.05, with a fair total of 45.18. It projects a 30.1-15.1 final and gives Rutgers a 17% chance to win.

For the first half, our number is Michigan by 7.34 with a total of 23.54.

If Michigan covers, the ground game is probably how. In our tested games, favorites who covered also cleared their rushing line 33.9% of the time, against 25.3% if the two were unrelated. With Rutgers this soft against the run and its passing game depleted, that link fits this game better than most.