James Madison at San Diego State. Lean: James Madison +2.5. Pick: Under 46.5.

Spread.Our model makes San Diego State -1.8, versus a current market of -2.5, so JMU +2.5 offers only 0.7 points of model value, small, but on the useful side of the key number three. The market moved from SDSU -1.5 to -2.5 while JC Evans’ status became a story, and that adjustment now appears to price a meaningful quarterback downgrade. Evans is genuinely uncertain, but Camden Coleman is not an inexperienced emergency option: he went 13-1 as a Richmond starter and completed 11 of 17 passes for 170 yards in his first JMU start. SDSU’s counterweights are substantial: a seven-game home winning streak, JMU’s 2,546-mile first road trip, and Lucky Sutton’s dependable rushing. However, Jayden Denegal has produced only 182 passing yards through two games, while SDSU’s defense returned just one official starter. That makes laying more than two uncomfortable in a likely close, possession-oriented game. Take JMU +2.5, but only as a modest position because the QB uncertainty and travel disadvantage are real.

Total.The over case is credible: our model projects 50.1, SDSU’s rebuilt defense has only one returning starter, both teams have flashed explosive special teams, and benign weather removes an under catalyst. The under is still preferable. JMU’s 87 points included three punt-return touchdowns, while SDSU’s 53-point opener was aided by two blocked punts, a kickoff-return touchdown and short fields, scoring unlikely to repeat. Against UCLA, Denegal completed only six passes for 45 yards; JMU may again use Coleman instead of dual-threat starter Evans. The total also stayed essentially fixed at 46.5 despite both teams’ headline blowouts, suggesting the market discounted those special-teams scores. With both offenses leaning toward the run and neither passing game proven against FBS resistance, under 46.5 is the stronger side.

Where our number looks wrong.The -1.8 spread is not obviously wrong; it is simply too precise for the available information. The rating cannot price Evans’ uncertain availability, JMU’s coaching and portal rebuild, SDSU returning only one defensive starter, or the cross-country trip. JMU’s 87-3 FCS result also included three punt-return touchdowns, so a margin-driven early-season update can overstate the Dukes’ repeatable strength.

Spread.Our number (SDSU -1.8) and the market (-2.2) essentially agree, so our model reads this as a pass. But the on-field pictures diverge from that consensus. SDSU's offense collapsed to 10 points at UCLA and there is open QB uncertainty (Denegal has struggled; beat writers floated Saunders/Trudeau), while JMU is 2-0 with a defense that generated real havoc and has allowed 8 ppg. JMU is the more proven roster (12-2 in 2025) yet is only a field goal underdog. The line drifting from -1.5 to -2.5 is sharp money on the home side plus a cross-country-trip, 7pm PT body-clock spot that genuinely favors SDSU. That caps this, and JMU's own QB (JC Evans) is banged up with Coleman filling in. Net: I want the points with the better, healthier-at-the-skill-spots defense against a sputtering home offense, but the travel spot and two-sided QB fog keep it a lean, not a bet.

Total.Consensus sits at 46.5 across all five books and hasn't moved. Both teams are run-heavy (SDSU ~61% rush; the preview flags controlled pace for both), which shortens games. SDSU's offense has looked broken outside an FCS blow-up (10 points at UCLA) and has QB uncertainty, while JMU's defense has been the strength of a 2-0 start. JMU also carries its own QB question with Evans dinged and Coleman filling in. Independent projections cluster at or below the number (Dimers 47, the youwager preview 47). Weather is a non-factor: clear, 72F, ~7 mph wind. Our totals model wants ~50, but that is mostly last year's offenses plus venue terms and is blind to SDSU's current offensive collapse and the two-sided QB fog. Grind-it-out script, two solid defenses, uncertain offenses: I lean under.

Where our number looks wrong.The rating leans on SDSU's near-100% returning production and preseason prior and cannot see that the offense managed 10 points against a real defense (UCLA) or the QB carousel. It also credits JMU's rating off an 87-3 win over FCS Wagner, which inflates the in-season number without signal. It sees neither team's actual current form, so its near-coin-flip only accidentally aligns with the market.

Spread.JMU's 2-0 record is inflated by an 87-3 FCS blowout over Wagner; its only real test (Liberty, 20-13) was tight. SDSU's only FBS test (at UCLA, 10-28) exposed a defense returning just one starter, but that was on the road against a Power-conference team, and SDSU gets a full home slate and a rested week to fix issues. The market has moved toward SDSU all week (-1/-1.5 open to -2/-2.5 now across five books), and that steam lines up with real information our rating can't see: JMU starting QB JC Evans (wrist) is day-to-day and backup Camden Coleman has started both 2026 games, with lead RB Seth Cromwell questionable. JMU's underlying profile (12-2, CFP team, strong portal churn) is legitimately good, which caps my conviction, but the combination of market direction plus QB/RB uncertainty on the road for JMU tips this to San Diego State at home.

Total.Looking at actual scoring against comparable competition rather than model output: JMU scored just 20 against a real Group-of-Five opponent (Liberty) and SDSU managed only 10 points on the road against a Power-conference defense (UCLA) - both offenses look ordinary once the FCS blowouts are stripped out. SDSU's defense is rebuilt and has already leaked 28 to UCLA, which argues for more scoring, but JMU is playing on the road with its starting QB questionable and its lead back banged up, which tends to suppress an offense's ceiling. Clear, mild weather (72F, 7mph wind) removes any weather-driven lean. The market itself has barely moved off 46.5 all week, which is its own signal that oddsmakers see this as fairly priced; I lean under on the expectation that QB/RB uncertainty for JMU caps the game below the model's optimistic number.

Where our number looks wrong.The rating is fit on final margins only, so JMU's 87-3 demolition of an FCS team (Wagner) and SDSU's 53-20 win over an FCS team (Portland State) both get folded into the in-season refit with no adjustment for opponent quality within FCS or for the JMU quarterback change. It has no way to reflect that JMU's real offensive quarterback (Evans) is questionable and that the backup's numbers are mostly against overmatched competition, and no way to reflect that SDSU's defensive rebuild already showed real cracks at UCLA. Flat 2.5-point home field also may understate SDSU's specific home cushion (seven-game home win streak) that the market is pricing in via the week-long move toward the Aztecs.