Tulane at Kansas State. Lean: Tulane +20.5. Lean: Under 49.5.

Spread.Edge Labs makes Kansas State -17.3, versus a current market of -20.5, so the numerical value lies with Tulane; however, the model’s documented 12.7-point margin error makes the 3.2-point disagreement modest. The dossier’s Sept. 17 history shows several books moving K-State from -19.5 to -20.5, meaning the Wildcats’ dominant start is increasingly priced. K-State’s Sept. 14 team notes report a 105-10 scoring margin and no offensive touchdowns allowed, but those results came against Nicholls and Washington State. Tulane’s Sept. 12 team recap showed improvement after installing Kadin Semonza, with 212 passing yards and no interceptions, while Jaylin Lucas ran for 214 yards. That gives the underdog enough explosiveness for a backdoor or competitive middle quarters. Action Network’s Sept. 17 split showed K-State attracting 60% of bets but only 40% of money, though the underlying sample was not disclosed. My fair line is approximately K-State -19, making +20.5 a small, price-sensitive play rather than a strong position.

Total.The model’s 50.45 is effectively identical to Under 50, so this is matchup-based rather than model-driven. K-State’s Sept. 14 notes show 607 rushing yards through two games and a defense allowing only 144 yards per game; that run-first profile can produce long possessions while suppressing Tulane’s scoring. Tulane had only 23 first downs across its first two games, according to Action Network’s Sept. 17 matchup data. The over case is legitimate: Semonza stabilized the offense, Lucas provides breakaway speed, and the forecast is not prohibitive. Still, Tulane’s uncertain RB depth and difficulty sustaining drives create a plausible 34-10 or 35-10 game. Taking 50 rather than 49.5 also adds valuable push protection.

Where our number looks wrong.The precision is more suspect than the direction. It likely understates Kansas State somewhat because final-margin ratings cannot fully capture Tulane’s massive roster turnover, first-year staff and in-season quarterback change. Conversely, two K-State blowouts against limited opposition can inflate an early Elo-style refit. I land nearer K-State -19 than the model’s -17.3, so most of the apparent 3.2-point edge is not actionable.

Spread.Market sits KSU -20.5 (up from -19.5 at open); our number is -17.3, so our model and I both see the home price as a touch rich. KSU's 2-0 resume (71-3 vs FCS Nicholls, 34-7 vs a rebuilding Washington State) inflated its in-season rating to 9.329, but neither opponent tells us it can bury a Power-conference underdog by three touchdowns. Tulane's offense was inept at Duke (3 pts) yet functional vs South Alabama with QB Kadin Semonza (named wk-2 starter, Yahoo 9/6), and the Green Wave have won 2 of the last 3 meetings. Add a hot, windy, wet forecast (15-26 mph, ~80% rain) and an 11am-local kick that muddies scoring, and this profiles as a game where a competent dog stays within a big number. I lean Tulane +20.5. This is a lean, not a hammer: KSU's defense (10 points, zero TDs allowed in two games) could smother Tulane's shaky offense and win comfortably.

Total.The kickoff forecast is windier than the captured 12.3 mph our model used: 15-26 mph with ~80% rain and 90-degree heat (NWS/AccuWeather, 9/17), which suppresses passing and kicking. KSU's defense has been suffocating (10 points and zero TDs allowed across two games) and Tulane's offense was shut down at Duke, so a shootout is unlikely; Tulane's own recent under trend (Under 6-1 last 7) fits. KSU can score, but against a stout-enough Tulane defense in wind, a 31-13 / 34-10 type line lands under the number. The public is 62% to the over, adding contrarian value to the under. Lean Under 49.5.

Where our number looks wrong.The dossier's Covers injury rows list 'B. Barnett' and 'D. Duff' as questionable KSU quarterbacks; those are not KSU's starter. Avery Johnson is the healthy starter (3 TDs on 9/12), so the QB-injury flags look like a bad roster match and should be ignored. Separately, the in-season rating jumped KSU from a 5.2 preseason to 9.3 purely on two blowouts of weak opponents, likely overstating true strength; the rating reads final margins only and cannot see that the schedule was soft.

Spread.But the model is blind to injuries, and the injury picture here cuts toward Kansas State by more than 3 points. Tulane HC Jon Sumrall said LB Chris Rodgers and TE Ty Thompson are still awaiting a diagnosis or second opinion with no timeline (SI Tulane, 9/15), a starting linebacker and receiving tight end both in doubt for a defense that already allowed 17 points to Duke and an offense scoring just 15.5 PPG through two games. Kansas State's own questionable tags (Barnett, Duff, Koch, Lovett, Morrell) mostly hit backups and depth, not starting QB Avery Johnson, who is 67.3% passing with zero sacks taken in two starts (KStateSports, 9/14). Kansas State is 2-0 by a combined 105-12; Tulane lost 3-17 at Duke. I side with the market over the model and take Kansas State -20.5.

Total.Kansas State's defense has allowed just 5 PPG through two games and Tulane's offense is averaging only 15.5 PPG with TE Ty Thompson hurt and unresolved; Tulane's own defense, already down LB Chris Rodgers, has to face a Kansas State offense scoring 52.5 PPG. In a game with a 20.5-point spread, backups typically see snaps once the outcome is decided, capping the ceiling even for a hot Kansas State offense. Wind is a moderate 12.3 mph at kickoff, which the model itself scores as a 2.4-point drag on scoring. I lean Under 49.5.

Where our number looks wrong.The model can't see that Tulane's defense and pass-catching TE are both clouded by unresolved injuries (Rodgers, Thompson, per Tulane HC 9/15), which likely explains why the market sits 3.2 points above our number; the gap reads more like missing information the model can't ingest than mispriced market value.