Baltimore Ravens at Indianapolis Colts. Pick: Indianapolis Colts +3.5. No call on the total.

Spread.I take Indianapolis +3.5, with modest conviction. Our Indianapolis +0.6 number differs from the market by 2.9 points, but its poor ATS record makes that a starting point, not an established edge. Baltimore deserves more credit than its injury-affected 2025 performance: Jackson reports being healthy, while Jones must demonstrate his post-Achilles form. Those considerations plausibly explain much of the market premium. Nevertheless, getting more than a field goal is attractive against a new Baltimore offensive system with Gwyn making his first NFL start at center. Flowers' limited participation adds uncertainty to Baltimore's passing ceiling, while Madubuike's uncertain readiness matters against Taylor. Indianapolis' established coaching structure and returning starting quarterback offer a credible competitive-game script. FanDuel still shows +3.5 despite its Baltimore moneyline shortening in favoritism from -205 to -180. That supports price sensitivity, not a claim of hidden sharp action. Incomplete Indianapolis practice designations and Jones' recovery prevent a stronger position.

Total.I lean under 47.5, narrowly. The strongest over argument is substantial: correcting our model's upward bias still leaves 50.82, Jackson is healthy, and Jones' return could restore Indianapolis' offensive efficiency. Madubuike's uncertain readiness also helps the over case. Against that, Flowers is limited with a hamstring injury, Baltimore has a new coordinator and first-time starting center, and Jones' familiar name does not eliminate post-Achilles uncertainty. A run-heavy approach could reduce possessions, although verified 2026 pace data are unavailable. FanDuel displays a decline from 48.5 to 47.5, so some offensive risk is already priced and the best under number is gone. I make only a small additional downward adjustment for those execution risks. Roof status is unconfirmed; weather supplies no meaningful under argument.

Where our number looks wrong.Indianapolis +0.6 likely understates Baltimore's advantage by treating injury-affected 2025 performance as too representative and quarterback continuity as equivalent to health. Jones returning from an Achilles tear and Jackson reporting full health are materially different situations. Conversely, Baltimore's center replacement and current Flowers limitation are missing risks. The totals model's 52.02 should first become 50.82 after the supplied 1.2-point bias correction; it still insufficiently reflects offensive transition, recovery uncertainty and current injuries. Its closed-roof assumption is unverified.

Spread.Our v3 number makes Baltimore only a 0.6-point favorite while the market prices 3.5, a 2.9-point gap toward Indianapolis, and ESPN's FPI (9/9/26) lands at Ravens -2.4, also short of the price. Three independent estimates all sit under the number. The line has not moved off 3.5 since our first capture on 8/30 even though DraftKings showed 70% of spread tickets and 66% of spread money on Baltimore as of 9/8; books absorbing the public side without moving is the shape of a slightly rich favorite. The news tilts the same way. Baltimore opens on the road under a first-year head coach (Jesse Minter, hired 1/22/26) who is also calling the defense, with a new OC in Declan Doyle, and gives Jovaughn Gwyn the first start of his career at center replacing Tyler Linderbaum; Zay Flowers (hamstring) and Nnamdi Madubuike (neck) were limited on 9/9. Indianapolis had all 53 men practice 9/9 and Daniel Jones took first-team reps from day one of camp. Take the healthy home team and the hook through 3.

Total.This is the strongest market signal on the board. The total opened around 49.5 and has been ground down to 47.5, with DraftKings alone dropping from 48.5 to 47.5, and roughly 87% of the handle sits on the under, the heaviest under-side money of the sixteen Week 1 games. Our totals model reads 52.0, but it is no reason to fade that: its holdout MAE (10.46) is worse than the market's, it runs about 1.2 points high, and it shrinks toward the mean, manufacturing over leans on lower totals. The fundamentals agree with the money: Baltimore is installing a new offense under a first-year staff with a first-time starting center, Zay Flowers is limited with a hamstring, Indianapolis' quarterback takes his first live contact since December Achilles surgery, and both defenses are talented. Weather is neutral indoors. Under 47.5.

Where our number looks wrong.Two ways. First, the v3 rating spread (BAL 2.08 vs IND 0.62) is a frozen preseason fit that cannot see the single biggest structural change in this game: Baltimore has a brand-new head coach, a new offensive coordinator, and a center making his first career start, while the flags only record 'new_head_coach: true' with no penalty attached. If anything the model's Baltimore rating is too generous relative to what actually takes the field Sunday. I am siding with Indianapolis because the market price, the FPI number, the unmoved line against heavy public Baltimore money, and the news all point the same direction, the v3 number merely fails to contradict that. Treat the model as one weak vote, not a 2.9-point edge.

Spread.What tips it toward Indianapolis for us is what the model can't see: Baltimore fired John Harbaugh in January and enters Week 1 under first-time HC Jesse Minter, a defensive-minded hire with no Ravens play-calling track record (baltimoreravens.com, 1/22 and 9/9). Two projected defensive starters, DT Nnamdi Madubuike (neck) and LB Teddye Buchanan (knee), are day-to-day off major 2025 injuries rather than clean gos (baltimoreravens.com, 9/8-9/9). Indianapolis, by contrast, gets a fully healthy Daniel Jones ('100%,' colts.com 9/7) and Jonathan Taylor at home in a building that's tough on road offenses. The market hasn't moved off -3.5 despite 70-84% public money on Baltimore (Complex.com, 9/8), which reads as an efficiently-set number rather than one we should chase further.

Total.Both quarterbacks are healthy (Daniel Jones full-go post-Achilles, colts.com 9/7; Lamar Jackson reported in 'peak condition,' baltimoreravens.com), and both offenses were efficient in 2025 (Colts +0.08 EPA/play, Ravens +0.05, per our own season data). Lucas Oil Stadium's retractable roof and a mild, no-wind Indianapolis forecast (AccuWeather, checked 2026-09-09) mean weather isn't a scoring suppressant either way. Meanwhile roughly 87% of total handle is already on the under (OddsShopper/prediction-market recaps, week of 9/8), a lopsided public lean the market hasn't pushed the number down hard against, itself a mild reason to consider the other side. We'll take the over.

Where our number looks wrong.The model is structurally blind to injuries (zero NFL injury rows) and frozen as of 8/19, so beyond the correct new-HC flag it cannot see that Baltimore's staff turnover brings real play-calling uncertainty, or that two Ravens defensive starters are not clean Week 1 gos.