Ohio at South Alabama. Lean: South Alabama -7. Lean: Under 49.5.

Spread.Edge Labs makes South Alabama -7.5, versus a market centered on -7 with -6.5 available, so the model-to-market value is only about one point. The stronger argument is situational: Ohio produced just 262 yards and was outgained by 185 against Jacksonville State, while starting quarterback Nick Poulos and backup Matt Vezza both exited. Their availability remains unverified; Covers calls Poulos unlikely, but Ohio has not issued a public game-status declaration I could confirm. South Alabama offers substantially more stability with Bishop Davenport, who has 450 passing yards, three touchdowns and no interceptions, and an offense with zero turnovers through two games. The move from -3.5/-4.5 to -7 has already priced most of Ohio’s quarterback uncertainty, making this a modest position rather than a strong bet. I will take the key-number protection at -6.5, but would not lay -7.5. Ohio’s viable rushing game against a Jaguars defense recently gashed by Tulane is the principal backdoor risk.

Total.The over case is legitimate: Edge Labs projects 56.2, South Alabama averages 436.5 yards, Ohio allows 485.5, and light wind should not impede offense. Moreover, South Alabama’s 79 plays per game are not slow. I still prefer under 49.5 narrowly. Ohio’s 29 last week overstates its offense because regulation ended 24-24, seven points came on a kickoff return, and five came in overtime. Quarterback uncertainty encourages a run-heavy, conservative road plan. South Alabama’s nation-leading 38:55 possession average can shorten Ohio’s opportunity count even while the Jaguars sustain drives. The market has already fallen from 53.5 to 49.5, so the best under value is gone; that and the Jaguars’ offensive efficiency keep confidence low. A late thunderstorm is possible, but weather is not the primary case.

Where our number looks wrong.The -7.5 result is plausible but comes from the wrong inputs: final-margin ratings credit Ohio with 29 points without distinguishing five overtime points and a 98-yard kickoff-return touchdown, while ignoring the 447-262 yardage deficit and two quarterback exits. Its calibration also stretches a raw -6.7 margin to -7.5. The model therefore should not be treated as evidence to lay through seven. On the total, 56.2 looks too high because the model cannot see Ohio’s quarterback uncertainty or South Alabama’s possession control.

Spread.Our model calls this a pass (SA -7.5 vs market -6.9, gap 0.6) and is blind to the news that actually moved the number. That is a real subtraction, but the market has now fully priced it, pushing the line through the key number of 7. Ohio is a proven 9-4-in-2025 program, and backup Matt Vezza already came in and won 29-27 over Jax State, so the offense is not empty. I would rather buy Ohio at +7 than lay a full touchdown with a 1-1 Sun Belt home team that just lost at Tulane. This is a low-conviction, key-number value take against a one-directional overcorrection, not a strong position.

Total.The market moved the other way with conviction: totals fell from 52.5-53.5 at open (9/13) to 49-49.5 now (9/16), a 3-4 point drop to the under. That move lines up with Ohio likely starting a backup QB, which caps offensive efficiency, and South Alabama's own offense sputtered in a 24-28 loss at Tulane. Weather is a mild secondary nudge under: hot and humid (~87F) with a 30-44% evening shower/storm chance but light ~5mph wind. I side with the sharp under money and the QB-driven scoring cap over the model's noisy over.

Where our number looks wrong.The model rates South Alabama (-4.638) clearly better than Ohio (-8.88) and lands at -7.5, coincidentally near market, but it reads no injuries and no line movement, so its 'pass' verdict misses the entire story: the number is where it is because of Poulos, and the model would say the same thing whether Ohio's starter played or not.

Spread.The market opened SA -4/-4.5 (9/13) and has steamed to -6.5/-7 across five books by 9/16 (betmgm and draftkings all -7; fanduel -6.5), a clean 2.5-3 point move that lines up with Ohio QB Nick Poulos's injury becoming public. Poulos left the 9/12 win over Jacksonville State early after a hit and was limited to 5 attempts; multiple outlets (RotoWire, Athens Messenger) list him questionable for this game with Matt Vezza the likely starter. Our own number (-7.5) sits almost on top of the now-current market (-7), so the model, which sees none of this news, has converged with an efficient market rather than found daylight, the gap read correctly as a pass. I lean SA anyway: their offense is scoring at 31.5 ppg over two games and Ohio's defense has allowed 38 ppg, while Ohio's O now runs through a backup QB or a limited Poulos on the road in Gulf Coast heat. This is a lean, not a strong edge, the move looks like the market pricing real information, not overreacting.

Total.The total fell from 52.5-53.5 (9/13) to 49-49.5 (9/16), tracking the same QB-driven move as the spread rather than a total-specific signal. Both offenses have shown they can score: SA hung 39 on SE Louisiana and put up 24 at Tulane; Ohio scored 29 on Jacksonville State even with Vezza in relief, and both defenses have been shaky (Ohio allowing 38 ppg, SA allowing 21 but just lost a track meet at Tulane). I lean over the current number at modest confidence, the total's drop looks mechanically tied to the spread move more than to hard evidence scoring will be light, and both offenses are live enough to clear 49.5 in a shootout-capable matchup.

Where our number looks wrong.The model's -7.5 is coincidentally close to the market's -7, but it got there blind, it has no read on Poulos's injury, no in-season weighting beyond two games, and treats the home-favorite math the same at every stadium. It would carry the same number whether Poulos starts, sits, or plays hurt, so its apparent agreement with the market is not evidence it modeled the actual driver of the line move; a full-strength Ohio would likely price closer to its preseason gap than to -7.5.