Buffalo at Penn State. Lean: Under 49.5. No call on the spread.

Spread.Edge Labs makes Penn State -36.9. After adjusting for Buffalo’s unsettled quarterback situation, inexperienced offensive line and Penn State’s improving defensive health, my number is approximately -38.5, still three points below the available -41.5. The market’s move from -38.5/-39.5 to -41/-41.5 already incorporates Buffalo’s loss at FIU and Elijah Holmes’ knee injury. Holmes’ absence would hurt, but backup Jason Wright produced 185 yards and two touchdowns after replacing him, so the downgrade is smaller than for an untested reserve. Penn State’s passing offense can create separation quickly, but its 3.9 yards per rush and two empty red-zone possessions at Temple make laying six touchdowns uncomfortable, particularly once starters rotate out. Betting splits are conflicting across providers and therefore add little conviction. Buffalo’s young roster could unravel against superior depth, but +41.5 leaves room for a 42-7 or 45-7 result. Take the full +41.5; I would not play +40.5.

Total.The total climbed from 47.5-48.5 to 49.5-50.5, while Edge Labs projects 48.35. That small model gap is not independently actionable, but 50.5 creates a playable under number. Penn State’s defense has surrendered nine total points, with its starting unit reportedly allowing only three, and now adds two defensive tackles. Buffalo’s quarterback uncertainty and inexperienced line reduce its scoring expectation. Penn State supplies the danger: Rocco Becht has seven touchdown passes and Buffalo just allowed 33 to FIU. However, the Nittany Lions’ struggling run game, recent red-zone failures and probable second-half rotation can suppress their ceiling in a decided game. Calm, dry weather offers no under assistance, and the upward market move deserves respect. Even so, a likely 42-3, 42-6 or 45-3 range favors under 50.5.

Where our number looks wrong.The model likely understates Penn State by roughly 1.5-2 points because it cannot see Holmes’ knee injury, Buffalo’s 54-player roster turnover, its entirely new starting offensive line, or Penn State’s returning defensive tackles. I still do not think those omissions justify the market being 4.6 points above the model, particularly given the model’s tendency to stretch already-large margins.

Spread.Market has climbed from PSU -38.5/-39.5 at open to -40.5/-41.5, and our model (-36.9) actually sits below the market, quietly leaning the dog. Laying 40-plus asks for a 44-3 type result, but Penn State's own FBS sample argues against it: they managed just 27 points at Temple in Week 2 (27-9), not a track-meet output, under a first-year head coach who may empty the bench once it is out of hand. Garbage-time and pull-the-starters dynamics historically help this size of dog. The counter is real: PSU's defense has allowed only 3 points in two games and Buffalo may be down to backup QB Jason Wright, so a true blowout is live. But at a number this large, with our number and the actual scoring sample both pointing to the dog, I take the points. This is close to a coin flip, hence a modest lean.

Total.Our totals model (48.35) sits just under the 49.5 market, and the fundamentals lean the same way. Penn State's defense has allowed 3 points across two games, and Buffalo's offense is weak (lost 33-20 at FIU) and may be without starting QB Elijah Holmes, with backup Jason Wright struggling in relief. Penn State's own offense scored just 27 at Temple, so its output is not guaranteed to be explosive, and a first-year staff sitting starters in the second half typically drains second-half scoring. The caution is the market itself: the total has risen from 47.5 to 49.5-50.5 all week, money on the over. But with an elite defense, a compromised opposing offense and modest recent PSU FBS scoring, I take the under at a slight edge.

Where our number looks wrong.The rating reads Penn State's 36 PPG and dominant scoring margin off Marshall (45-0) and Temple (27-9), but it cannot see that the offense mustered only 27 on the lone FBS opponent, nor that PSU's suffocating defense (3 points allowed in two games) could keep this game lower-scoring than a 40-point cover implies. It also ignores Buffalo's QB injury on both sides of the ledger.

Spread.Our model's calibrated number (PSU -36.9) sits within the pass zone against the market (-41.1), but the market has moved further Penn State's way since Monday's open (-38.5/-39.5 to -41/-41.5) on real news: Buffalo starting QB Elijah Holmes (knee) missed all practice this week and is questionable, with true freshman Jason Wright the likely starter (RotoWire, 9/17). Wright's lone extended look came in relief of a Buffalo offense that managed only 20 points and lost at FIU in Week 2, a Group of Five defense, not a Big Ten one. Penn State has outscored Marshall and Temple 72-9 combined and lost DE Max Granville for the season plus has DE LaVar Arrington II questionable, but its D-line depth still dwarfs Buffalo's O-line. Most of this is already priced into the four-point line move, so I'm not going max conviction, but a true-freshman first start on the road against this defense argues the market hasn't fully absorbed the downside case for Buffalo's offense staying competitive into the second half.

Total.But the news layers the same way: Buffalo is likely turning to a true freshman QB (Jason Wright) making his first start on the road, coming off a game where the current starter and offense combined for only 20 points against a Group of Five defense. Penn State's defense has allowed 4.5 points per game through two contests. In a Buffalo blowout script, Penn State likely empties its bench once comfortably ahead, capping the market's usual garbage-time scoring, while Buffalo's offense with a freshman under center is a real risk to be shut out for stretches. Weather is mild (65F, 5 mph wind, 2% precip) and not a factor either way. This is a lean, not a strong conviction, since lopsided Big Ten-vs-MAC games can also see backups feast against MAC reserves late.

Where our number looks wrong.The model's -36.9 number is still mostly preseason prior (Buffalo -9.8 composite, Penn State first-year-coach discount) plus two blowout wins over weak competition; it has no way to see that Buffalo's presumptive starting QB is out and being replaced by a true freshman making his first career start against a much better defense, or that Penn State is already missing a projected D-line starter for the year. Both cut the same direction the market has already moved and neither is in the model's raw margin.