Arizona at BYU. Pick: Arizona +7.5. Lean: Under 47.5.

Spread.Arizona +7.5 offers the valuable touchdown hook against an EL number of BYU -6.1, although that 1.4-point difference is not a demonstrated edge. The stronger argument is matchup-specific: Fifita remains Arizona’s starter, while BYU loses safety Raider Damuni. Arizona’s current depth chart lists its established offensive starters; the two questionable Arizona players in the dossier are not listed in its offensive rotation, so treating those flags as major offensive losses would be excessive. BYU deserves favoritism, and expected linebacker returns partially offset its secondary loss. Its home advantage could also exceed EL’s flat allowance. Market history is mixed: DraftKings moved toward BYU, while FanDuel moved toward Arizona, making a unified sharp-money narrative unjustified. Monday’s injury news is likely substantially priced by Wednesday. I prefer the experienced visiting quarterback plus more than a touchdown, but neither the news nor the unproven model supports a strong position.

Total.The under deserves respect: both defenses have strong prior-season results, BYU expects linebacker reinforcements, and the captured market has already fallen three to six points at several books. Possible running-back absences and the dossier’s wind forecast also favor fewer points. However, 47.5 now demands a relatively restrained game from two established quarterbacks. Damuni’s absence creates a plausible passing vulnerability, while Arizona’s depth chart does not establish major losses among its primary offensive options. EL’s 48.19 offers almost no independent support, but its 51.458 ridge baseline shows how heavily subsequent adjustments suppress scoring. I could not corroborate the 14.8-mph kickoff wind, so I would not mechanically accept that deduction. Take a modest over position at the reduced number; the decline alone does not prove an overreaction.

Where our number looks wrong.The spread model misses Damuni’s absence and BYU’s coordinator transition, but also misses returning linebackers and potentially understates Provo home advantage. Its precise -6.1 is unjustified after FCS openers. The totals documentation contradicts itself: it says weather is excluded, yet explicitly subtracts 2.911 points for wind. That unverified wind input materially drives the forecast. BYU’s official game-week page also lists No. 14, versus No. 15 in the dossier.

Spread.Our rating makes this BYU -6.1 into a market that opened -7.5 and still sits -7.5 at seven books, priced about -110 both ways with no movement all week (Sportsbook Review, 9/9). A Pima County judge granted a TRO on 9/9 restoring RB Ismail Mahdi (859 rushing yards, 4 rushing TDs, 17 catches), LB Max Harris (78 tackles, third on the team) and P Sullivan Moon; all three sat out week 1 and all are eligible Saturday (SI, 9/9). Arizona has not confirmed it will activate them, but the line had not moved as of my last odds read. BYU meanwhile loses starting safety Raider Damuni for weeks with Tommy Prassas starting, and RB2 Sione Moa is day-to-day (Deseret 9/7, BYU depth chart via Yahoo 9/9), and is breaking in DC Kelly Poppinga after Jay Hill left for Michigan. These two needed double overtime last year, BYU 33-27. +7.5 buys the hook past the key number with a fifth-year QB in Noah Fifita. A lean, not a full-size bet.

Total.The move is the story. A 3 to 6 point unanimous drop plus pricing that says fair is still below 47.5 is the strongest signal on this game, and it has a football reason: both defenses allowed 7 points in week 1, both allowed about 19 per game in 2025, and BYU's veteran line is built to shorten games. Against it: weather is benign, so there is no wind suppression, and both teams averaged 31-plus points in 2025. I side with the move while admitting I am buying at the bottom of it, hence the modest number.

Where our number looks wrong.Two places. First, the spread model is still essentially a preseason prior with one FCS blowout folded in: BYU 63-7 over Utah Tech and Arizona 35-7 over Northern Arizona tell it nothing, and it reads no roster news at all, so its -6.1 is not independent support for Arizona. I am taking the dog on this week's news, not on the 1.4-point gap, which the documented 46-50% held-out ATS record says is worthless on its own. Second, the totals model's wind layer subtracts 2.911 points off a captured 14.8 mph reading, but the NWS forecast for Provo on Saturday (api.weather.gov, retrieved 9/9) is SSE 7 to 15 mph, so that penalty is likely roughly double what it should be and the 48.19 is understated, which argues against the total side I took. Flat 2.5 home field also looks light for LaVell Edwards Stadium at roughly 4,500 feet in September heat.

Spread.Neither side carries a confirmed impact injury: Arizona's Fifita (262 yds, 2 TD, 91.0 QBR Wk1) and BYU's Bachmeier (16/20, 198 yds, 3 TD Wk1) are both healthy and playing well; the questionable/out tags on both rosters are depth pieces, unconfirmed beyond our own Covers scan (ESPN/Rotowire shows no data for either team). Both Wk1 wins came against outmatched FCS/G5 opponents and tell us little about a true road Big 12 environment. I lean the points with Arizona, but lightly - the market found this number early and hasn't moved off it since.

Total.That kind of uniform, multi-book decline usually means the market overpriced Week 1's 63-7 and 35-7 scorelines (both against clearly overmatched non-conference opponents) and has since corrected toward a more realistic Big 12 road-game number. Weather does not push this under: the NWS forecast for Provo on Saturday calls for a 93F high, dry and mostly sunny by kickoff, warmer and calmer than the dossier's captured 82F/14.8mph snapshot, so I am not leaning on weather here - I'm on the under mainly because the market itself already did the work.

Where our number looks wrong.The model's own held-out testing also shows zero proven ATS edge at any gap size, so its 'pass' here should not be read as evidence against betting this game either way - it is simply outside the model's demonstrated competence.