Arkansas at Utah. Lean: Utah -12.5. Pick: Under 54.5.

Spread.Utah -12.5 is a modest lean, not the 5.2-point bargain implied by EL’s -17.7. The dossier’s early FanDuel -7.5 and DraftKings -9.5 have already disappeared; Arkansas’s rebuild and Utah’s home advantage are substantially priced. My qualitative assessment still favors Utah’s quarterback continuity and defensive pressure against KJ Jackson, whose opener was only his second career start. Quincy Rhodes’ uncertain availability adds downside to Arkansas’s defensive front, but I would not price him as absent. The strongest counterarguments are Utah’s own coaching transition, concerns about hits on Devon Dampier, and the danger of extrapolating from an FCS blowout. DraftKings’ overwhelming Utah handle confirms this is a popular position, not evidence of an undiscovered edge. I prefer laying less than two touchdowns at the verified major-book number. No complete game-specific availability report or definitive Wednesday injury resolution was verified, keeping conviction below a clear 60-point lean.

Total.Under 55.5 is a small preference despite EL’s 57.76. Arkansas’s inherited offensive rating looks vulnerable with a new staff and inexperienced quarterback facing Utah’s pressure. Saturday-night south winds of 10 to 20 mph provide conditional support, although exact kickoff conditions remain unverified. The over case is credible: Dampier and Wayshawn Parker offer explosive scoring, Rhodes’ uncertainty could weaken Arkansas’s defense, and turnovers can create short fields. Utah’s 66-point opener also warns against assuming a slow game; I found no verified tempo evidence supporting that assumption. Several dossier books have already lowered totals by two points, reducing remaining value. Taking 55.5 rather than 54.5 matters. I would not subtract wind again from EL’s projection, which already includes it.

Where our number looks wrong.Both first_year_head_coach=false flags are incorrect: Utah is beginning Morgan Scalley’s first full season and Arkansas Ryan Silverfield’s first season. EL’s calibration stretches an already uncertain early-season margin, so -17.7 overstates the precision of the Utah case. The totals documentation also contradicts its output: it says no weather is incorporated, yet explicitly subtracts 3.521 points for wind. Arkansas’s 2025 offensive production is a questionable baseline for its rebuilt roster and inexperienced starting quarterback.

Spread.No steam either way, so the tilt has to come from news. Utah's news is good: projected right guard Keith Olson and rotation lineman Isaiah Garcia are back this week per Morgan Scalley (Deseret News 9/8), RB Daniel Bray is tracking through concussion protocol, and Devon Dampier went 15-of-19 for 176 and 3 TD in a 633-yard, 66-point opener (WholeHogSports 9/8). Arkansas is the opposite: an entirely first-year staff under Ryan Silverfield, 14.5% returning production, 3.0 yards per carry and 0-for-2 on throws of 20-plus yards against FCS North Alabama, and 8 penalties for 74 yards including three holding calls on LT Kavion Broussard (WholeHogSports 9/8-9/9). A rebuilt line with a holding problem, at 4,600 feet, at night, in a sold-out Rice-Eccles they are simulating crowd noise to prepare for, is how a two-touchdown cover happens. A thin lean, not a bet.

Total.This is the cleanest signal on the game. Books do not take a number down two-plus points into that much one-way over money unless their own price says lower. Weather agrees: NWS hourly for Salt Lake City has 82F and 14 mph SSW wind at the 8:15pm MT kick, building to 21 mph south by 9pm. And the game script agrees: Arkansas managed 415 yards, 3.0 yards per carry and nothing downfield against an FCS defense behind a first-year staff, and Ron Roberts's defense held North Alabama to 14. Utah's 66 came against Idaho. Take the best number available, 55.5.

Where our number looks wrong.Three concrete problems. (1) The dossier carries Utah as first_year_head_coach: false. That is factually wrong - Morgan Scalley is in his 1st season after Whittingham stepped down (Wikipedia, retrieved 9/9), so Utah's preseason prior of 13.996 never took a coaching-transition haircut, and Arkansas's block does not carry the flag at all even though Ryan Silverfield's entire staff is new. Both sides of the -17.7 are built on a bad coaching input. (2) The in-season refit DROPPED Utah from 13.996 to 11.664 after a 66-14 win, because an Elo-style margin refit reads a 52-point win over an unrated FCS team as underperformance. Then the calibration map stretches the raw 15.2 up to 17.7 anyway. Two errors pushing opposite directions, neither of them information. (3) Arkansas at 14.5% returning production with a fully new staff and +6 portal net is effectively a roster the model has never seen; a prior anchored to 2025's 2-10 team carries almost no signal, so the true error bar on -17.7 is much wider than the stated 12.7 calibrated margin error. I would not use the 5.2-point gap as evidence of anything here, and my Utah lean rests on the availability and Week 1 efficiency news, not on the model.

Spread.What's real: the market moved hard toward Utah, from -7.5 (FanDuel, 8/30) and -9.5 (DraftKings, 9/1) at open to a -12/-13 consensus by 9/9, a full week of steam already pricing in Utah's dominant complementary showing (66-14 over Idaho; QB Devon Dampier 15/19-176-3TD plus 31 rush yards, with backup Byrd Ficklin used in a run package ~15 times) against Arkansas's still-unknown rebuilt roster (14.5% returning production, 2-10 in 2025, first FBS-caliber test under new DC Ron Roberts, who has so far only beaten FCS North Alabama). With that much of the 'Utah is scary good' story already baked into a number that has moved 3-5 points, and a Week 1 blowout raising backdoor/garbage-time cover risk on a lay this large, I'd rather take the points with Arkansas.

Total.The NWS hourly forecast I pulled today for kickoff (~8:15pm MT / 10:15pm ET) shows a much calmer night: clear skies, 73-76F, winds only 3-7 mph, 0% precipitation, meaningfully lighter than what the model assumed, which if anything argues the model's own number is a touch too low, not too high. Despite that, the total has actually moved down 2 points since open at three of five books (BetMGM) while public tickets are reportedly about 97% on the over (VegasInsider), a reverse-line-movement tell that sharp money is on the under. Combined with Arkansas's unproven, rebuilt offense (14.5% returning production) against a Utah team built to control the clock, I'll side under.

Where our number looks wrong.More importantly, the model is still mostly last year's/preseason priors for both teams (Arkansas returns only 14.5% production, Utah 67.9%) after one game each against overmatched non-conference opponents (FCS North Alabama, FCS Idaho), so it hasn't actually seen either team play a real game yet, and it cannot see that the market itself already moved 3-5 points toward Utah since the opener, information that argues whatever value existed in laying additional points with Utah is largely already gone.