Troy at Missouri. Lean: Troy +26.5. Lean: Over 50.5.

Spread.The refreshed page shows +26.5 at DraftKings, FanDuel and BetMGM; the extra point matters, although direct sportsbook execution was not verified. Our Missouri -25.5 supplies only a two-point disagreement, nowhere near an established model edge. The stronger argument is Missouri's willingness to shorten games when ahead, combined with Hardy's expected continued absence. Neither is necessarily overlooked: Hardy's recovery is public, and the dossier already records movement from Missouri -27.5 to -26.5. I would not label that sharp action without reliable splits. Troy's rushing production and turnover-free quarterback performance last Saturday offer a plausible route to enough possession and scoring to stay inside this number. The opposing case is substantial: Missouri's passing attack produced against Kansas, while Troy's results came against much weaker opposition. Green's return also argues against indiscriminately downgrading Missouri's line. This is a modest price-sensitive dog lean, not a claim that Troy approaches Missouri's quality.

Total.Under 50.5 is a small lean against our model's 51.13. Missouri ran 15 consecutive fourth-quarter plays against Kansas, providing actual evidence for a clock-consuming lead script. Hardy's expected absence reduces explosive rushing potential, although Roberts demonstrated capable replacement production. Troy's 14-of-29 passing against Alabama State raises questions about sustained drives against Missouri. The over deserves respect: Simmons threw for 292 yards and three touchdowns, Troy produced explosive plays, and warm conditions offer no meaningful weather suppression. Short fields or late reserve-defense scoring could easily overturn the under case. DraftKings' captured move from 49.5 to 50.5 also opposes this selection.

Where our number looks wrong.The calibration stretches a 21.7-point raw margin into 25.5 without demonstrated ATS value. It cannot distinguish healthy Missouri from a team missing Hardy, or competitive scoring from late-game clock management. Conversely, treating every Covers flag as a fresh absence would overcorrect: Green returned against Kansas. The totals model's 51.13 is only 0.63 above 50.5, and its historical weighting misses current personnel and game-management choices. Neither model supports precise confidence here.

Spread.Our calibrated number is Missouri -25.5 against a -26.7 consensus, and my own projection lands in the same place: Troy's offense is bad enough (81st in scoring, 104th passing, 31% on third down) that it should be held near 13-17, but Missouri needs roughly 41 to clear 26.5 and its ground game is the unit under stress. Ahmad Hardy has still not practiced (CBS 8/23, Covers 9/12 OUT) and Malae Fonoti tore an ACL in week one (ABC17 9/8), leaving Drinkwitz calling the room 'probably the biggest point of concern we have right now.' Troy's best unit is exactly that matchup: 24th nationally against the run, 77 rush yards allowed per game, with preseason Sun Belt DPOY Donnie Smith up front. Parker's group also led Clemson 16-3 at halftime as a 30-point road dog in 2025.

Total.My central projection is roughly Missouri 39-41, Troy 13-16, or 53-56, a couple points clear of 50.5. The cleanest edge in the game is Missouri's passing offense (324.5 yards/game, 17th; Simmons at 75% with no interceptions) against Troy's 88th-ranked pass defense, and Troy's 31% third-down rate (119th) hands Missouri extra possessions rather than draining the clock. Missouri's red zone defense has allowed scores on all opposing trips so far, so Troy's few chances should become touchdowns rather than field goals. Both Missouri games have gone over this number. The under case is real - Troy's front is 24th against the run, Missouri's backfield is depleted, and a 100th-anniversary blowout gets the starters pulled early - which is why this is a lean rather than a position.

Where our number looks wrong.Two things. First, the number is manufactured by the calibration stretch, not by the rating: the raw margin is 21.7 and the 2026-09-07 map inflates it to 25.5. The model's own belief about this game is 21.7, four points friendlier to Troy than the number it reports, and the stretch is what drags it up next to the market. Second, the rating is blind to the single most relevant fact here - Missouri is playing a 26.5-point favorite without the SEC's leading rusher and without its RB2, with Drinkwitz on record that the room is his biggest depth concern. That should move the number toward Troy and the model cannot see it. Offsetting honestly: the rating also cannot see that Troy's -7.79 was earned against Sam Houston and an FCS team, so it is not obviously too low on Troy either. Net, the gap reading of 'pass' is the right posture; the model just gets there for the wrong reasons.

Spread.This line has been remarkably stable, it opened Missouri -27.5 and has only drifted a point to -26.5, versus our model's -25.5 and the current market consensus around -26.5/-27, all clustered together with no real disagreement to exploit. The case for Missouri leans on schedule quality: Troy's 2-0 start came against Sam Houston (FBS, but a weak one) and Alabama State (FCS), so its resume tells us little about how it holds up against a real SEC roster. Missouri, by contrast, has scored 46 ppg and allowed 17.5 through two games (including a real road win at Kansas) despite missing preseason All-American RB Ahmad Hardy the entire time, meaning this offense's production isn't dependent on a player who's still out. New starting QB Austin Simmons (Ole Miss transfer) has played well enough to keep that scoring pace. I don't see a market inefficiency here so much as a well-priced mismatch; my confidence reflects agreement with the number rather than a discovered edge, tempered by the usual backdoor-cover risk once a 26+ point margin triggers backups in the second half.

Total.Our totals model (51.1) and the market (50.1-50.5) are essentially in agreement, so there's no real model-vs-market case here, this is a close read either way. The lean toward the over comes from the matchup: Missouri is averaging 46 points a game even missing its best skill player, against a Troy defense that has only had to deal with a Group-of-5-or-lower level of competition so far. In a game this lopsided, the total's fate mostly depends on whether Missouri keeps its offense on the field once the lead is safe, which cuts both ways, a running clock caps points, but backups scoring against a clearly overmatched Troy defense would still add them. Modest rain risk (16% precipitation) and wind (~6 mph) are not strong enough on their own to force an under. Given the genuine uncertainty in how this game's script plays out, I'll take a light lean to the over rather than a strong conviction.

Where our number looks wrong.The model's rating for Troy is still largely preseason-anchored (portal net of -13, zero blue-chip recruits in the last four cycles) and its in-season refit after two wins over sub-par competition hasn't meaningfully moved that number, which is probably correct, since neither opponent tells us much. The bigger risk for our number is on the Missouri side: it has no way to credit that Missouri has produced its current per-game output without its best running back, which could mean the team's true offensive ceiling is being undersold rather than oversold.