Oregon State vs Houston Prediction, Picks & Odds — Sep 5, 2026

Two books have this posted, Bovada and DraftKings, between 20 and 20.5, consensus 20.2. We project Houston by 24.5. The 4.3 points in between rest on no 2026 game: our in-season form weight for both teams is exactly zero. Last season's regressed power rating is worth 11.81 points to Houston and home field a flat 4.5 — 16.31 together. The remaining 8.19 is offseason bookkeeping: portal 3.19, quarterback 2.37, returning production 2.35, roster talent 0.25, skill transfers 0.03. Strip those out and our projection posts Houston -16.3, 3.9 under the market; keep them and it lands 4.3 over. The market is pricing about half the offseason adjustment we apply in full.

The biggest of those terms is barely Houston's doing. Of the 3.19-point portal component, Houston contributes 0.65: 18 in, 16 out, 51 incoming star-points against 42 outgoing, net weighted +9.3. The other 2.53 is Oregon State's — 20 in against 35 out, 57 star-points in against 95 out, net weighted -58.4. Nothing Oregon State added is rated above three stars, and its four top additions are two tackles, Broderick Shull from Auburn and Kwan Johnson from Eastern Illinois, plus two long snappers. The departures include two four-stars, quarterback Gabarri Johnson and defensive lineman Kelze Howard. That is star-count arithmetic, not snap counts, carrying 3.19 points of the projection.

The quarterback term adds 2.37 more, and most of it is a penalty rather than a rating. Conner Weigman — four FBS seasons, 0.255 career EPA, 0.89 usage, credibility 0.73 — is flagged settled and grades +2.70; Willie Fritz confirmed him at Big 12 Media Days in July. Maalik Murphy grades +1.44 on his own record (three FBS seasons, 0.192 EPA, credibility 0.62) and is the projected starter in the Aug. 21 Portland Tribune depth chart. But Oregon State's job is modeled open and three deep — Brady Jones (Western Michigan) grades -1.46, Braden Atkinson (Mercer) +1.14 — so the team-level term is -0.80, a full 2.24 points below Murphy alone. More than half our projection-market disagreement is a discount for not knowing, and it would largely evaporate the day JaMarcus Shephard names a starter.

What the record says at this number

We price Houston to cover -20.2 in 57.4% of runs. That is the share of a 40,000-game simulation built on the same 24.5 projection — the margin disagreement restated, not independent evidence for it. The realized record is flat. Favorites laying 16.7 to 23.7 have covered 49.7% since 2013 (n=1,238). Favorites of 21 or more, 49.2% (n=1,473). A nine-win favorite laying 15 to 25 on a team that won three or fewer, 49.0% (n=257). Favorites whose own run game ranked bottom-20 the prior year, 44.3% (n=70) — Houston's ranked 128th of 136.

Two well-populated cohorts point opposite ways. Oregon State went 2-10, and underdogs in this band off a two-win-or-worse season kept it within one score 29.9% of the time against 23.8% band-wide (n=234, control n=1,227), the favorite averaging 18.2 points of margin — under the number — and covering 45.3%. Against that: these two met last year, Houston 27-24, and when a rematch follows a game decided by three or fewer and the market has re-rated it to 17 or more, the favorite wins 95.3% of the time by an average of 22.0 (n=127, cover rate 48.8%). That 22.0 sits between the market's number and ours, nearer the market.

The live question is Oregon State's team total

The cohort that fits Oregon State most precisely says the roster damage shows up on the underdog's side of the scoreboard, not in the margin. Underdogs of 14 or more returning 32% or less of their prior-season receiving production, against a favorite returning 45% or more: 251 games since 2014. They averaged 15.0 points, median 14, were held to 17 or fewer 67.7% of the time and to 13 or fewer 46.6%, and reached 24 in 19.9%. The baseline for all 2,716 big underdogs is 17.1, 58.2%, 39.1%, 26.1%. The margin barely moved: -22.6 against -22.2. Oregon State returns 31% of its receiving production (18% by roster-verified yards) to Houston's 49%. We project it for 12.5 points, modal score 38-10.

Trent Walker, the largest pass-catching share at 22% usage and 0.467 EPA, transferred to Houston and projects as its starting Z; David Wells Jr. (17.1%) went to Tulsa, and Tastean Reddicks (13.0%) and tight end Bryce Caufield (12.0%) are out of eligibility. The three pass-catchers brought in to replace them — Xayvion Noland (Texas-Rio Grande Valley), Adonis McDaniel (Mercer), Eric Olsen (Utah Tech) — have no FBS snaps between them; all four of Houston's incoming skill transfers came from FBS programs. Two returners carry usage: running back Cornell Hatcher Jr. (17.4%) and tight end Gabe Milbourn (9.7%).

The band turns on whether the underdog reaches 24. In 1,125 games with a closing spread of 16.7 to 23.7 since 2014, the favorite wins 90.5% by an average of 19.4. In the 338 where the dog scored 24 or more (30.0%), the favorite's margin collapses to 8.8, the finish is within one score 45.9% of the time, and the dog wins outright 23.1%. In the 391 where it was held to 13 or fewer, the favorite wins by 28.7 and the dog wins 0.5%. Oregon State's cohort clears 24 less than one time in five.

0%

-38

0%

-32

0%

-24

0%

-18

1%

-10

3%

-4

7%

+4

12%

+10

16%

+18

18%

+24

17%

+32

12%

+38

7%

+46

Oregon State winsHouston wins

Market: Houston −20Projected: Houston by 24 (95%)

The range of outcomes for this game. Each bar is how often the final margin lands in that range — the more it leans toward Houston, the more likely Houston wins.

GRIDPEX.COM

The defense inside Houston's rating has turned over

Houston's 2025 defense is a real part of that 11.81-point carry term: 40th of 136 in opponent-adjusted defensive efficiency, 8th of 136 in explosive plays allowed, graded a strength at 0.94. The personnel behind those grades is mostly gone. The five biggest defensive departures, all graduated or gone to the NFL, are edge Nadame Tucker (14.5 sacks, 55 tackles), end Eddie Walls III (5.5, 60), safeties A.J. Haulcy (89 tackles, three interceptions) and Juwon Gaston, and defensive back Wrook Brown. Tucker and Walls alone account for 20 of the team's 24 sacks in 12 games; the best returning total is Brandon Mack II's 5.0. Pass-defense continuity is 0.20 and rush-defense 0.23, against 1.00 for the run offense — which is where the 81.2% returning production, and its +3.97 credit, actually lives.

The pass rush being replaced was not much of one: 71st nationally in sacks at 2.0 a game, 113th in tackles for loss. If the eighth-place explosive ranking came from coverage, as 48 passes defended (22nd) suggests, cornerback Will James returning with three interceptions matters more than the sack ledger implies.

So what does the number mean. A projection carrying zero in-season information disagrees with two posted books by 4.3 points, and the portal and quarterback terms alone — 3.19 and 2.37 — exceed that disagreement between them. Week 1 is not itself a reason to widen the range: openers in this band (n=94) produced outright dog wins 8.5% of the time and one-score finishes 23.4%, against 9.5% and 24.7% later in the year (n=1,046); only the cover rate moved, 42.6% against a 49.3% control. The thing to watch is Oregon State's team total: reach 24 and the band's history says the game changes shape; fail to, and the 12.5 we project sits just under the 15.0 those cohort offenses averaged.