Washington vs USC Prediction — College Football Week 5, Saturday, October 3, 2026

Every Washington vs USC prediction this week starts with the same split screen: a Huskies team that just throttled Washington State and a Trojans offense that has been lighting up everyone else. The market has settled on Washington +8.5 at -108, which on its face reads like a double-digit mismatch. The model doesn't see it that way.

Run the numbers and USC still comes out ahead on a neutral card. But the margin is nowhere near 8.5, and that gap between what the market is charging and what the game projects to is the entire reason this is a play.

Why the model is on Washington +8.5

Start with the model's own math. It opens USC by 3.3 points on a straight projection, then shaves half a point off for the newly announced suspension of linebacker Desman Stephens II, landing at Washington +2.8. That's the full gap: USC a favorite, but by less than three points, not the 8.5 the board is asking bettors to lay. The desk isn't betting Washington to win. It's betting that the Huskies won't lose by more than the number says they will, and that's a much easier case to make.

The Stephens suspension matters here because of who's on the other side of it. Demond Williams Jr.'s mobility is the swing factor, and a thinner linebacker corps is exactly the unit that's supposed to contain a quarterback who can extend plays with his legs. Take away depth there and Washington's quickest path to staying inside the number gets wider.

The case against

USC's argument for winning outright is straightforward: sustained passing production. The Trojans scored 42 against Rutgers and 27 against Oregon, numbers that dwarf Washington's 21.3 points per game across its three FBS opponents. The 48 the Huskies hung on FCS Eastern Washington doesn't move that comparison; it's a different level of competition and the model treats it that way.

USC's path runs through a Washington secondary that's dealing with its own issues, with Karnley probable rather than fully cleared and CJ Christian still out. If USC's passing game travels the way it has the last two weeks, that's where it happens.

Washington's counter is built on defense, not offense. The Huskies' 24-10 win over Washington State is the blueprint the desk wants repeated: control the game, limit possessions, keep it close. That plan leans on Stewart being back in the rotation even if still working his way to full speed and VanSumeren getting cleared, while Elinneus Davis's return to practice and Rashid Williams getting the green light offer real but partial relief up front. Hatchett remains unavailable, and Jordan Washington, Carr and Limar are all still out, with McMahan's season-ending injury cutting into rushing depth on top of it. None of that is a clean bill of health, but the defensive line doesn't need to be whole to replicate a low-scoring script against a USC offense that still has its own uncertainties at linebacker.

The number that matters

USC projects to win this game. That part isn't in dispute. The dispute is by how much, and 8.5 points is simply more cushion than a 2.8-point model edge justifies. That 5.7-point gap between the market line and the projected margin is the play, not a belief that Washington walks out of Los Angeles with the upset.

USC still carries its own injury questions beyond Stephens, with Mosley, Bowman, Graham, Johnson and Treter all listed with uncertain statuses and a reported season-ending timetable for Simms that hasn't been confirmed by the school. None of those individually swing the number, but stacked together they're part of why the model isn't willing to hand USC a bigger cushion than 2.8 points in the first place.

The verdict

This is a play. The desk is on Washington +8.5 at -108, betting the gap between the market's number and a projected 2.8-point USC win, not betting on a Huskies upset. If the linebacker thin spot opposite Williams holds up the way the model expects, that gap is where the value lives.