Western Kentucky vs Nevada Prediction, Picks & Odds — Sep 5, 2026

Our projection makes Western Kentucky a 6.5-point favorite at Nevada on Sept. 5. Bovada and DraftKings both post 2.5, and the quoted range across the two books is 2.5 to 2.5. Where our four-point disagreement comes from matters more than its size: the projection contains zero games of 2026 football. The in-season form component is weighted 0.0 for both teams and contributes 0.0 points. Every 2026 opponent-adjusted field our projection normally leans on — offensive PPA, success rate, explosiveness, defensive havoc, points per scoring opportunity — is empty for both. What is left is last season's rating and this season's roster accounting.

The projection decomposes with no residual. Last season's body of work — 2025's power rating regressed toward the mean — is worth 9.78 points toward Western Kentucky, the largest term. Returning production adds 6.25 more, the transfer portal 0.65. Against those, roster talent credits Nevada 5.52, home field at Mackay Stadium 4.5, and the quarterback comparison 0.17. Net: Western Kentucky by 6.5.

◄NevadafavorsWestern Kentucky►

Last season's body of worklast year's power rating, regressed toward the averageWestern Kentucky +10

Returning productionhow much of last year's production is back this seasonWestern Kentucky +6

Roster talentoverall talent level of the roster — recruiting pedigree plus an independent preseason roster gradeNevada +6

Each ingredient of our projection and the side it favors — bars toward Western Kentucky are why it's the pick (home field is counted separately).

GRIDPEX.COM

The two carry-forward terms — last season's results, and how much of last season's production is back — total 16.03 points toward Western Kentucky, both applied at full strength by a model with no current-season evidence to update them with. Reconciling to 2.5 does not require discarding either one. Shave the pair from 16.03 to 12.03, about a quarter of its weight, leave every other term alone, and our number lands on the market's.

One term appears to run backwards

The returning-production credit needs the same scrutiny, because the headline figure and the detail beneath it are not measuring the same thing. Western Kentucky is credited with 83.7% of last season's production back, against 35.5% for Nevada. Its pass-offense continuity reads 90% — and the largest departure charged against that unit is quarterback Maverick McIvor, at a 34% share. The 90% is receiving production; it carries no discount for the passer who is gone. Nevada is the team not replacing a starter: it loses a 13% passing share in Chubba Purdy but returns Carter Jones, who held a 67.3% usage share in 2025.

The rushing half is cleaner and less flattering. Four running backs take four of the five slots on Western Kentucky's departures list — Marvis Parrish (28% of carries), La'Vell Wright (22%, to Pittsburgh), George Hart III (16%) and Jalen Hampton (15%) — with receiver Matthew Henry fifth. That backfield produced the 84th-ranked rushing offense of 136. The replacements are Oklahoma State transfer Kalib Hicks, whom we rate 53 on our 0–99 player scale, non-FBS transfer Sincere Baines from Maine, and three-star depth-chart starter Ajay Allen. One skill leader returns: slot receiver K.D. Hutchinson, at 13.9% usage.

Nevada's problem is larger and better documented. It returns none of last season's rushing production and 21% of its passing production, off an offense that ranked 128th of 136 overall and 135th through the air. Across 154 FBS team-seasons since 2015 that returned 30% or less of their receiving PPA, 61% saw offensive efficiency decline the following year, at a median of minus 0.042 PPA per play. The comparison groups give the shape: the 663 seasons returning 30–65% declined 52% of the time at a median of minus 0.004, and the 513 seasons above 65% improved, median plus 0.024. Nevada starts from 128th.

The defense inverts the usual reading of Nevada's ranks. Its pass defense finished 36th, which reads like a foundation until you check what left. Four of the five listed departed defenders are defensive backs, worth 61.8 impact points between them, including Murvin Kenion III (59 tackles, five interceptions) and cornerbacks AJ Odums and Michael Coats Jr. The two defensive backs who return total 14.1. That rank was coverage, not pressure: 19.0 sacks, 103rd nationally, and Jonathan Maldonado took 5.0 of them to Ole Miss. What Nevada returns at the top is Dylan Labarbera on the line, attached to the run defense that finished 126th.

Both quarterback jobs are open

Neither team has named a starter. Nevada lists three candidates across a 10-point rating spread: Carter Jones at 76, who started seven games as a true freshman in 2025 (64.3% completions, 1,019 yards, six touchdowns against eight interceptions); UCLA transfer Luke Duncan at 69, with one 2025 start; and fifth-year AJ Bianco at 66, the only one of the three with three FBS seasons behind him. Jeff Choate pushed the call to the weekend of Aug. 22-23, and no winner has been announced. Western Kentucky's field is two and flatter: Rodney Tisdale Jr. at 75 over Florida State transfer Brock Glenn at 73, who has three FBS seasons to Tisdale's one and the better career-best EPA, 0.382 against 0.101. Our projection prices the whole question at 0.17 points toward Nevada, against 9.78 for last season's results.

What does the number mean for a bettor? Less than a four-point argument implies. Across 3,469 FBS games since 2013 with a closing favorite of zero to six points, the favorite won 58.1% of the time and covered 48.8% — a coin flip before the vig and a loser after it. Our simulation puts Western Kentucky at 66.4% to win against a market-implied 55.3%, which is the same disagreement restated as a probability. And our projection publishes a standard deviation of 5.8 points on each team's rating: the uncertainty on a single input is larger than the whole gap.

Games priced like this one are also not the tight finishes the number advertises. In 1,412 games since 2013 with a spread of zero to six and a total between 44 and 54, 48.9% finished inside one score and 21.6% were decided by three or fewer — but the average final margin was 12.2 points, and the underdog won outright 42.1% of the time. Our simulation reproduces the same barbell: 38.8% one-score, 25.5% decided by 20 or more, with an 80% band running from Western Kentucky by 27 to Nevada by 14.

Turnovers are the largest single lever, and history treats them more emphatically than our own paths do. In 2,493 FBS games since 2014 with a spread of zero to six and an uneven turnover count, the team that won the turnover battle won 76.7% of the time. Our sensitivity table gives the factor a 31-point swing here, but it lifts Nevada only to 50.7%, because Nevada starts at 33.6%. Winning turnovers usually wins the game; here it returns the home team to a coin flip.

The total is the wider disagreement: 48.6 projected against 53.2 posted, 4.6 points against the spread's four. We publish no lean on it, because the totals projection has no graded track record yet. And the lookalike cohort's 49.8 average total is not support for the under — that sample was selected on totals between 44 and 54, so its average is an artifact of the filter, not evidence about a 53.2 number.

One angle worth closing off. Western Kentucky travels from Bowling Green, Ky., to Reno, and across 88 games since 2014 pairing a Mountain West host with an Eastern- or Central-time visitor, the visitor won 36.4% and lost by an average of 7.0 points. Against the spread it is nothing: in the 57 of those with a consensus line, the home team covered 50.9% against an average number of minus 4.2. The travel effect is already priced, and here it points the wrong way — the traveling team is the favorite.

Our number is on the record: Western Kentucky by 6.5, covering 2.5 in 58% of simulations. It is carried by two terms our projection cannot yet update, it contains a talent component that runs against our own talent index, and its returning-production credit rewards the team that is replacing its quarterback. Favorites of this size have covered 48.8% of 3,469 games. Track the disagreement; don't bet it.