Duke vs. Virginia Prediction, Odds & Pick
Prediction, odds and the model's read
Market vs our model
we are 15.2 toward 
UVA
| Market | Our model | |
|---|---|---|
| Spread | UVA −6.5 | UVA −21.7 |
| UVA win | 71% | 92% |
Calibrated against results, not against the closing line. Read this as our confidence, not as an edge over the market.
Scouting report
2026 projection · last season, roster and staff
Where Virginia gets at Duke: 36th at hitting chunk plays, against a unit 123rd at keeping everything in front of them. Duke projects no clear way in against Virginia. Virginia is also 67th at surviving second and twelve, against a unit 119th.
HowUVAattacks
DUKE
- 1
Take shots
hitting chunk plays · projected
UVA#36 of 138DUKE#123 of 138 - 2
Trust the offense behind schedule
surviving second and twelve · projected
UVA#67 of 138DUKE#119 of 138 - 3
Run it
running it · projected
UVA#47 of 138DUKE#93 of 138
The two teams
Duke
Duke throws it: Manny Diaz's offenses have been pass-leaning across 5 seasons. The strengths are defensive, led by 24th at stuffing runs at the line. The way in is 123rd at keeping everything in front of them, and they are 119th at closing out passing downs.
On fourth down he goes for it about as often as his peers do from the same spots (+2.5%). Against what the expected-points model recommends he is −22.8% — the profession as a whole is negative on this.
Duke has played 2 games against an average slate, adjusted for here. That football carries about 0% of what is below; the rest is still the preseason projection.
Projected strengths
- Defensegetting stopped at the line#24projected
Projected soft spots
- Defensehitting chunk plays#123projected
- Defensesurviving second and twelve#119projected
- Offensegetting stopped at the line#117projected
Virginia
Virginia is in year 5 under Tony Elliott, whose offenses have been balanced. The strengths are defensive, led by 26th at closing out passing downs.
On fourth down he goes for it about as often as his peers do from the same spots (−0.3%). Against what the expected-points model recommends he is −27.4% — the profession as a whole is negative on this.
Virginia has played 1 game against an average slate, adjusted for here. That football carries about 0% of what is below; the rest is still the preseason projection.
Projected strengths
- Defensesurviving second and twelve#26projected
Ranks are national. Nothing here is last season's number — carrying one forward is measurably worse than saying nothing — and in-season figures are adjusted for who was actually played. This describes the matchup; it is not a prediction and not a pick.
Head coaches only — there is no free feed for coordinators, so on a program where the head coach does not call the offense this describes the staff he hired. Play-calling identity is a style, not a grade: throwing more than the situations warrant is not better than throwing less than they warrant.
Model prediction
how the number is builtOur model simulates the game drive by drive from each side's opponent-adjusted efficiency (UVA Elo 1655, DUKE Elo 1615 for reference), with home-field advantage. That projects UVA -21.7 (92% to win) — 15.2 points clear of UVA's market line of -6.5. That is a disagreement, not a betting edge — sides we favour have not covered at better than breakeven.
Why that percentage is worth reading: across 604 graded in-season games, the calls this model put near 58% came in at 58.2%. The full calibration table is published, bin by bin. Early in the season the cold-start path runs instead and calibrates less well (Brier 0.2028 against 0.1889).
Model as of Sep 1 · through week 1 · drive simulation
Where the game is won
2026 national rank · right is better
offensethe defense it faces
Moving the ball
DUKE90th6thUVAUVA41st74thDUKEThrough the air
DUKE113th17thUVAUVA100th103rdDUKEOn the ground
DUKE37th13thUVAUVA19th31stDUKEStaying on the field
DUKE90th27thUVAUVA116th44thDUKETaking care of it
DUKE21st9thUVAUVA21st75thDUKE
Each bar runs from last in the country on the left to 1st on the right. The filled span is the gap between the offense and the defense it faces.
Season form — 2026
nationally rankedThe matchup, in context
series history + adjusted profilesVirginia leads 40–34 · 27% have been one-possession games
| DUKE | UVA | |
|---|---|---|
| +4.6 (#53) | CORE overall | +13.1 (#27) |
| -1 / -5 | offense / defense | +2 / -11 |
| 1.42 | points / drive | 3.09 |
| 0.25 | allowed / drive | 0.64 |
| 33% | three-and-outs | 18% |
| -15 | pass over expected | -23 |
Bold is the side ahead. CORE strips the situation from every play and solves the schedule out across the league, so these compare two teams that never met.
Key matchups
The analyst read: each team's offense splits crossed against the opponent's defense, with the edges called out.
Create a free account →You get Pro free for 6 months. No card required.
FAQ
Who will win Duke vs. Virginia?
Gridpex's model favors Virginia with a 92% win probability.
What's the predicted spread for Duke vs. Virginia?
The model projects Virginia by 21.7.
What time is Duke vs. Virginia and what channel is it on?
Fri, Oct 23, 7:00 PM ET, at Scott Stadium.
Where this number comes from. Every projection on this page is produced by the same model, run before kickoff on opponent-adjusted efficiency, power ratings and situational data — walk-forward validated on eleven seasons and never adjusted after the fact. How the model works →