Delaware State at South Florida. Lean: Delaware State +37.5. Lean: Under 61.5.

Spread.Edge Labs has no spread projection because Delaware State is unrated, leaving no model-versus-market edge to claim. The market’s five-touchdown-plus price already captures the FBS/FCS talent disparity, USF’s 35-1 record against FCS opponents and its eight-game home winning streak. However, -37.5 requires roughly a 50-point USF output at the posted total. The Bulls have scored only 19 and 28 this season, including just 208 yards against FIU, and are implementing Brian Hartline’s system with substantial roster turnover. Delaware State’s unsettled quarterback room and Samari Collier’s shoulder limitation increase blowout risk, but the Hornets produced 514 yards against Stony Brook and held William & Mary to 187 yards despite losing. My rough matchup range is USF by 31-35, making +37.5 playable but not a large position. USF attracting 68% of spread bets and 72% of money suggests the favorite’s obvious advantages are hardly overlooked. I would take the extra point available at FanDuel rather than +36.5 elsewhere.

Total.Edge Labs projects 60.29, only 1.71 points below the available 62 and well inside the model’s historical error, so this is a game-script lean rather than a model edge. The under case is USF’s modest early production, Delaware State’s ability to slow William & Mary, likely favorite substitutions and a thunderstorm risk around kickoff. FanDuel also moved from 62.5 to 61.5. The over deserves respect: Delaware State gained 514 yards against Stony Brook, USF quarterback Michael Van Buren accounted for four touchdowns at Army, and quarterback instability can create short fields; moreover, only 29% of tickets but 54% of money were reported on the over. At 62, however, several plausible USF wins, 45-10, 45-13 or 48-10, stay under, with 62 offering push protection.

Where our number looks wrong.The spread model is effectively unusable here because it cannot rate Delaware State, so it cannot test whether -37.5 is inflated. Its framework also misses the FCS schedule-strength translation, Delaware State’s unresolved quarterback competition, likely second-half substitutions and USF’s almost entirely rebuilt roster. Any inference from USF’s rating alone would be false precision.

Spread.Our rating gives no number here (Delaware State unrated), so this is a market-and-film read. USF is a capable AAC side but its offense has been pedestrian to open 2026: 19 points at home vs FIU and a 28-24 escape at Army, averaging 23.5 ppg versus 40.5 last year with new QB Michael Van Buren still developing (usforacle preview, 9/15). A team that just won by four and has not cracked 30 is being asked to win by 38. Public/preview 'working' numbers sat at -28.5 (youwager 9/13, sportsgrid), yet the live books print -36.5/-37.5 (our DB, odds-api 9/17) - the market has inflated this well past the consensus preview line, which is where I want the points. Dimers' 10k-sim model lands 48-12, a 36-point margin that pushes/loses at -37.5. Rain before 8pm at a 7pm kick can throttle early scoring. Take the points with the dog.

Total.Market 61.5; our ridge total says 60.29 (slight under). USF has not scored 30 in either game (19, 28) and Delaware State's 25.7 ppg is inflated by a D-II opponent (Bowie State) and Stony Brook - it managed 15 against real FCS competition. Dimers' sim lands ~60 combined. A 7pm kickoff carries scattered showers/thunderstorms mainly before 8pm (~82F, light 5mph wind, wtsp/wfla forecasts), which tends to mute first-half scoring, and a lopsided game means USF pulls starters and drains clock late, capping the back end. Everything points modestly below 61.5.

Where our number looks wrong.The rating produces no number because Delaware State is unrated, and the calibration map stretches big favorites larger; there is no priced edge here. The model also cannot see that USF's offense has sharply underperformed last year's output through two games, which is the crux of the cover question.

Spread.Our rating has no number here (Delaware State is unrated as an FCS/MEAC team), so this is a pure market read. FanDuel/DraftKings both sit at 36.5-37.5, one of the largest lines of the week. USF is 2-0 but has not been dominant: 19-9 over a bad FIU team and 28-24 at Army, only 23.5 ppg with a QB (Michael Van Buren Jr., LSU transfer, named QB1 8/25/26) still building chemistry in a new offense under a first-year staff. Covering by 38+ against any FBS-caliber defense, let alone protecting the number against Delaware State's functional MEAC offense (25.7 ppg, wins over Stony Brook and Bowie State), is a big ask given USF hasn't shown that kind of scoring margin yet. Rain is forecast at kickoff, which raises turnover/miscue risk for a young USF passing attack trying to pull away early. I could not verify any USF or Delaware State injuries this week (Covers shows nothing new, no beat-writer reports found); assume health as-is.

Total.The dossier's captured forecast shows rain, 82F, 52% precipitation chance at kickoff (a separate general search corroborates a real shot at scattered storms in Tampa that evening), which typically dents a young, still-gelling USF passing game more than it helps. Neither offense has been especially explosive: USF's two wins came at 19 and 28 points, and Delaware State's own defense has allowed 20+ per game. I found no injury or pace news for either side this week.

Where our number looks wrong.Our model literally has no output here because Delaware State is unrated, so it isn't 'wrong' so much as silent on the one game where a number would matter most. The model's own documented weakness list flags FCS opponents as its least reliable case, and that is exactly this game - there is no calibrated read to check the market against, only the raw fact that USF's actual scoring through two games (19 and 28 points) is well short of what a -37.5 number implies it needs to produce.