Pro Football Week 1 Board — 2026

Model lines for every game, published before kickoff. Generated 2026-08-17 by v1 (teams + starting QB, recency-weighted ridge). Lines use each team's most recent starting quarterback — early in the week that assumption can be stale, and it is shown for every game so you can judge it yourself.

Backtest vs the market (816 games, 2023-2025)

Model error 10.28 pts · closing line 9.74 pts · ATS 49.9% (breakeven 52.4%)

The market is still ahead of this model, and this site will always say so. How the model works

Kickoff	Matchup	Model line	Home win prob	Assumed QBs (away / home)
Wednesday 09-09	NE @ SEA	SEA -7.7	72.1%	Drake Maye / Sam Darnold
Thursday 09-10	SF @ LA	LA -2.2	56.5%	Brock Purdy / Matthew Stafford
Sunday 09-13	DAL @ NYG	NYG -0.6	51.8%	Dak Prescott / Jaxson Dart
Sunday 09-13	WAS @ PHI	PHI -8.3	73.6%	Josh Johnson / Jalen Hurts
Sunday 09-13	GB @ MIN	MIN -0.8	52.3%	Jordan Love / J.J. McCarthy
Sunday 09-13	ARI @ LAC	LAC -7.7	72.0%	Jacoby Brissett / Justin Herbert
Sunday 09-13	NYJ @ TEN	TEN -2	55.9%	Brady Cook / Cam Ward
Sunday 09-13	ATL @ PIT	PIT -5	64.7%	Kirk Cousins / Aaron Rodgers
Sunday 09-13	MIA @ LV	LV +0.2	49.5%	Quinn Ewers / Kenny Pickett
Sunday 09-13	BAL @ IND	IND +2	44.1%	Lamar Jackson / Riley Leonard
Sunday 09-13	BUF @ HOU	HOU -0.2	50.5%	Josh Allen / C.J. Stroud
Sunday 09-13	NO @ DET	DET -9.4	76.2%	Tyler Shough / Jared Goff
Sunday 09-13	TB @ CIN	CIN -1.4	54.4%	Baker Mayfield / Joe Burrow
Sunday 09-13	CHI @ CAR	CAR +3	41.0%	Caleb Williams / Bryce Young
Sunday 09-13	CLE @ JAX	JAX -8.6	74.4%	Shedeur Sanders / Trevor Lawrence
Monday 09-14	DEN @ KC	KC -1.5	54.6%	Jarrett Stidham / Chris Oladokun

Power ratings

Points versus a league-average team on a neutral field. Home-field advantage: +2.25 pts.

1	SEA	+5.16
2	BUF	+4.53
3	DET	+3.95
4	LA	+3.74
5	BAL	+3.32
6	PHI	+2.85
7	HOU	+2.75
8	GB	+1.91
9	DEN	+1.73
10	SF	+1.72
11	JAX	+1.45
12	MIN	+1.28
13	KC	+1.13
14	NE	+1.02
15	IND	+0.75
16	TB	+0.64
17	CHI	+0.48
18	LAC	+0.47
19	PIT	+0.16
20	DAL	-1.29
21	CIN	-1.33
22	MIA	-1.73
23	ATL	-1.79
24	WAS	-2.06
25	NO	-2.15
26	ARI	-2.44
27	NYG	-3.40
28	CLE	-3.66
29	CAR	-3.84
30	LV	-4.64
31	TEN	-5.19
32	NYJ	-5.52

Quarterback values

Points above an average starter, estimated jointly with team ratings. Argue with the model, not with us.

Sam Darnold	+2.72
Patrick Mahomes	+1.67
Lamar Jackson	+1.56
Drake Maye	+1.36
Matthew Stafford	+1.36
Joe Burrow	+1.36
Jalen Hurts	+1.29
Jared Goff	+1.26
Brock Purdy	+1.23
Derek Carr	+1.23
Josh Allen	+1.20
Jordan Love	+1.17
Trevor Lawrence	+1.13
Bo Nix	+1.06
Kyler Murray	+0.95
C.J. Stroud	+0.90
Justin Herbert	+0.74
Tua Tagovailoa	+0.65
Jaxson Dart	+0.60
Tyler Huntley	+0.56
Aidan O'Connell	+0.53
Jayden Daniels	+0.52
Aaron Rodgers	+0.45
Daniel Jones	+0.44
J.J. McCarthy	+0.33
Caleb Williams	+0.24
Ryan Tannehill	+0.22
Tyler Shough	+0.20
Baker Mayfield	+0.20
Deshaun Watson	+0.16
Davis Mills	+0.15
Dak Prescott	+0.15
Anthony Richardson	-0.15
Shedeur Sanders	-0.16
Case Keenum	-0.17
Tim Boyle	-0.18
Mason Rudolph	-0.19
Desmond Ridder	-0.19
Jarrett Stidham	-0.20
Malik Willis	-0.29
Brandon Allen	-0.29
Easton Stick	-0.30
Kirk Cousins	-0.32
Chris Oladokun	-0.34
Bailey Zappe	-0.36
Trey Lance	-0.37
Tanner McKee	-0.37
Clayton Tune	-0.42
Carson Wentz	-0.46
Joshua Dobbs	-0.46
Philip Rivers	-0.48
Marcus Mariota	-0.50
Andy Dalton	-0.52
Kenny Pickett	-0.54
Jake Haener	-0.60
Jameis Winston	-0.63
Tommy DeVito	-0.63
Bryce Young	-0.71
Brady Cook	-0.78
Russell Wilson	-0.79
Sam Howell	-0.82
Tyrod Taylor	-0.92
Gardner Minshew	-0.99
Quinn Ewers	-1.02
Jake Browning	-1.10
Joe Flacco	-1.10
Cam Ward	-1.40
Mac Jones	-1.43
Cooper Rush	-1.53
Spencer Rattler	-1.53
Jacoby Brissett	-1.79
Geno Smith	-2.14

Model version history

Version	Verdict	Gap to closing line
v0 teams-only	baseline	+0.69
v1 + starting QB	adopted	+0.54
v2 EPA models	rejected out-of-sample	no change
v3 rest/bye/divisional	rejected in tuning	no change

Walk-forward backtest, 2023-2025 regular seasons, no lookahead. The market is the benchmark and it is still ahead.

Model outputs are published for research and entertainment. Nothing here is betting advice, and the honest record above should make clear why: this model does not currently beat the closing line. If someone tells you their model does, ask for their version history.