Prediction
No Picks Against the Favorite: The Model's Big Ten Board for Week Five
The model has no pick against the betting favorite on the Week 5 Big Ten board, the first time that has happened since Week 0. Its picks against the favorite are 1-7 this season, so the open question is the margin: the model and DraftKings are more than five points apart in two of the eight games.
Everything below is the model’s output, run Friday morning with DraftKings lines… these are not my picks… and it covers the eight games on the Big Ten slate, including Nebraska at home against Maryland.
Last week, in two numbers
Week 4 went 5-4 straight up and 4-5 against the spread. Season to date: 39-16 and 25-28-2. The full breakdown is in Week Four graded: 5-4, and the model had Nebraska’s 31 exactly.
Carry one number from that grade into this board: picks against the favorite are 1-7. None of the eight below goes against the favorite, so that record cannot move this week.
Nebraska vs. Maryland
The model has Nebraska 33-20, a solid label at 78.5%. DraftKings has Nebraska by 14.5, so the model projects a 12.7 point margin and gives Maryland 1.8 more points than the line does. Its largest factor for Nebraska is current power rating, and it raises no flags on the game.
Kickoff is 3 p.m. CT Saturday at Memorial Stadium on FS1, for Homecoming. Nebraska is 4-0, its best start since 2016, and won its Big Ten opener for the first time since 2019.
The history is close. Nebraska leads the series 3-1, and last October’s 34-31 win at Maryland came after trailing 31-24. The visiting team has won each of the last three meetings, and the last two were decided by a total of six points.
Maryland is 2-2 and 0-1 in the Big Ten, coming off a 54-3 loss to UCLA, the game the Week Four board had Maryland winning 28-24. The number to watch is its passing offense, 283.0 yards a game, 28th in the country and sixth in the Big Ten, inside a 431.3 yard, 32.3 point average. Nebraska’s pass defense allows 143.0 yards a game, seventh nationally, and its scoring defense allows 12.0, tenth.
Head coach Matt Rhule was looking ahead the moment the Michigan State game ended. “I’m going to watch tape one more time and I will be on to Maryland,” he said afterward. His read on the week: “It’s just straight tactical.”
The closest call on the board
Penn State at Northwestern. The model has Penn State 24-23 at 50.2%, and flags it as a coin flip where it has no real conviction. DraftKings has Penn State by 2.5, so the model and the line are on the same side and the model sees the game as 2.4 points closer.
The biggest gaps with the line
In each of these, the model and the line pick the same winner and disagree on the margin:
- Ohio State at Iowa: the line has Ohio State by 14.5, the model by 7.7. A 6.8-point gap, the widest on the board.
- Purdue at Illinois: the line has Illinois by 10, the model by 15.5. 5.5 points.
- Washington at USC: the line has USC by 8.5, the model by 4.4. 4.1 points, and the model flags it as a thin pick driven by pedigree, where last season’s ratings disagree.
The board
| Game | Model | Line |
|---|---|---|
| Penn State at Northwestern | Penn State 24-23 | Penn State by 2.5 |
| Michigan at Minnesota | Michigan 24-21 | Michigan by 5.5 |
| Michigan State at Wisconsin | Wisconsin 30-17 | Wisconsin by 10 |
| Ohio State at Iowa | Ohio State 28-20 | Ohio State by 14.5 |
| Maryland at Nebraska | Nebraska 33-20 | Nebraska by 14.5 |
| Purdue at Illinois | Illinois 35-20 | Illinois by 10 |
| Washington at USC | USC 28-24 | USC by 8.5 |
| Indiana at Rutgers | Indiana 38-14 | Indiana by 24.5 |
Nothing is bold this week because no pick goes against the favorite. Lines are DraftKings, pulled Friday morning, and they will move before kickoff.
Confidence labels
Two picks carry the strong label (Indiana at Rutgers at 93.3%, Purdue at Illinois at 83.1%), two solid (Michigan State at Wisconsin and Nebraska), two lean (Ohio State at Iowa and Washington at USC), and two toss-up (Penn State at Northwestern and Michigan at Minnesota).
Season to date, strong picks are 29-2, solid 5-3, lean 0-4 and toss-ups 5-7. The model’s own season check still reports the labels out of order, by a gap it flags as larger than chance, though the samples are small.
What the model says about itself
Backtested on 2022 through 2025, it picks the winner about 70% of the time and does not beat the closing spread. Straight up this season it has been better than that at 39-16. Against the spread it is 25-28-2, which is the part the backtest said to expect.
Husker Nation, the question this board leaves open is the margin, and the grade on it, early next week, tells us whether the model’s gaps with the line were signal or noise.