Rutgers at Boston College. Lean: Rutgers +3.5. Lean: Under 54.5.

Spread.Rutgers +3.5 is a modest position built around the extra half-point above three. EL makes Boston College -2.4, but that 1.1-point difference is smaller than the model’s demonstrated uncertainty. Rutgers’ four-turnover opener offers some scope for improvement without requiring a wholesale talent upgrade. Surace’s first start adds uncertainty, although a close preseason quarterback competition argues against automatically imposing a large backup penalty. His productive relief appearance came late and proves little. Boston College’s unresolved offensive-line injuries make laying more than a field goal less appealing. The strongest counterargument: BC outgained Cincinnati despite losing comfortably, while Rutgers’ problems extended beyond turnovers. The quarterback announcement is already two days old; I assume it is priced. Most captured books have held -3.5, so FanDuel’s isolated -7.5 opening capture is not evidence of a market-wide four-point move. DraftKings’ Rutgers-heavy handle supports interest in the dog, but cannot establish informed action. Take the hook with limited conviction.

Total.I lean under despite EL's 56.21 projection. Surace's first start and BC's uncertain center/tackle availability create plausible drive-efficiency problems that a historical scoring model misses. BC's conservative opening approach and McKenzie's substantial rushing workload also offer a route to longer possessions, although I could not verify a sustained tempo change. The over case is credible: BC allowed 254 rushing yards, Rutgers surrendered 37 points, Duff supplies explosive-play ability, and benign weather should permit normal offense. BC's 388 yards also suggest better scoring conversion is possible. Nevertheless, I prefer offensive execution risk at 54.5. The dossier's decline from 55.5 at DraftKings and BetMGM means some value has disappeared; under-heavy handle is context, not independent proof. This is a moderate lean, not a strong position.

Where our number looks wrong.The near-10-point Rutgers rating downgrade after one turnover-heavy game looks too aggressive for a margin-only system, especially beside the claim that early ratings remain mostly preseason priors. Conversely, BC's yardage advantage suggests its final margin overstated the performance gap. These effects pull in opposite directions. The model also misses Surace's first start and BC's line uncertainty; -2.4 is not a trustworthy fair-price estimate. The totals estimate of 56.21 likely overstates offensive continuity across these rebuilt units.

Spread.Our number (BC -2.4) and the market (BC -3.5) agree, but both rest on a rating that just overreacted: the in-season refit slammed Rutgers from -0.6 preseason to -10.5 off one game, when the preseason prior (44.4% returning production, 9.3% blue-chip, +2 portal versus BC's 15.6%, 3.9%, -5) had Rutgers roughly a five-point road favorite here. That is an eight-point swing off one Thursday night against a MAC team. The market has moved less than our model did and is quietly on the dog: the number touched -7.5 on 9/4, settled to -3.5 by 9/6, and did NOT tick up when Rutgers lost Lonergan on 9/8. Surace is a genuine downgrade, but Lonergan threw three picks and a fumble and won the camp job by a hair, so it is worth about two points, not four. Take the points through 3.

Total.Weather is a non-factor: NWS has Chestnut Hill clear, upper 60s at kickoff, northwest wind 3-6 mph. The over case is real - BC allowed 34 at Cincinnati with 126 yards after missed tackles, Rutgers allowed 37 to UMass, and both Week 1 games cleared 49 - but the offenses are the worse unit here. Rutgers ran 25 times for 30 yards last week and now hands a first career start to Surace, which argues for a conservative script and stalled drives, and BC managed 15 points with a D2 transfer at quarterback.

Where our number looks wrong.Two ways. First, the in-season v1 refit is reading one game as if it were a season: Rutgers goes from -0.599 preseason to -10.479 after a single loss to a MAC team, which lands it on top of a Boston College team that was 2-10 in 2025 with 15.6% returning production and a -5 portal net. Opponent quality is not being respected - BC's 19-point loss came on the road to a Big 12 team, Rutgers' 16-point loss came at home to UMass - so the refit has, if anything, punished the better roster harder. Second, the rating cannot see that Rutgers is starting a quarterback with 18 career pass attempts, which is the single biggest fact in this game; our -2.4 happens to sit near the market only by accident.

Spread.Our model has BC -2.4 versus a market of -3.5, a 1.1-point gap that sits inside the model's own no-edge zone, so treat this as close to a coin flip rather than a model signal. We lean Rutgers +3.5. Rutgers benched Dylan Lonergan, who has not practiced since his Week 1 injury and threw three picks against UMass, for AJ Surace, who outplayed him in relief (6-of-9, 77 yards, a TD) and now makes his first career start (On3, Daily Targum, 9/8/2026) - a real unknown, not an obvious downgrade. BC is banged up on the offensive line, with center Michael Crounse and LT Veguer Jean-Jumeau questionable per O'Brien (BC Athletics, 9/8/2026), and its defense missed tackles for 126 yards after contact while the offense managed just 15 points against Cincinnati (boston.com, 9/10/2026) - fresh, real weaknesses. Market structure backs the fade too: 64-86% of tickets and money sit on BC, yet the number has been stuck at -3.5 since roughly 9/6-9/7 and never moved further BC's way even after Tuesday's QB news, a soft reverse-line-movement tell (scoresandodds.com, accessed 9/10/2026; dossier line history).

Total.The line tells the story: the total opened as high as 55.5-57.5 and has been bet down to 54-54.5 as of 9/10/2026, despite roughly 93% of tickets and money on the Over (scoresandodds.com; dossier line history, captured 9/6-9/10) - a clean reverse-line-movement signal toward Under. Our totals model likes the Over (56.2-56.9 vs a 54.4-54.5 market), but it is blind to injuries and still mostly running on last year's rosters this early, so we weight the fresher market move higher. AJ Surace's first career start, on the road under Friday-night lights (On3, 9/8), argues for a conservative, lower-variance Rutgers offense, and BC's O-line is still banged up (BC Athletics, 9/8). Kickoff weather is clear with light wind and no precipitation, not a suppressor. Two leaky defenses (BC allowed 34, Rutgers 37 in Week 1) keep this from a lock, but we side with the sharp Under money.

Where our number looks wrong.The model's gap here (-1.1) sits inside its own dead zone, so it isn't 'wrong' so much as uninformative - it has shown no ATS edge at any gap size in two held-out seasons (46-50% in every bucket). More importantly it is structurally blind to the single biggest fact of the week: Rutgers is starting a quarterback with 18 career pass attempts, and it cannot see BC's Week 1 O-line injuries or tackling breakdown either. I'm overriding its slight (and statistically meaningless) BC lean with situational news and market-structure evidence that both point toward Rutgers.