INSULARITY NFL Forecast
2026 · National Football League

Week 6
Projections.

Model-projected scores, margins, totals and comparisons with the market for every game this week.

Fri 16 Oct – Tue 20 Oct · 14 games
Updated 18 Oct, 23:46 UTC
Model history
Model performanceModel pick vs market lineWeeks 3–6 · published before kickoff
Season · vs market line25–29–446.30%
This week · vs market line3–8–127.27%
Tracking onlyStats-only recordWeeks 3–6 · side of the locked line where the stats-only margin fell; not the official model pick
Vs market line31–23–457.41%
Big gaps (3+ pts)8–466.67%
Biggest gaps (4+ pts)0–30.00%

Week to date

12 graded · 2 pending. The model pick is graded against the line it was locked at (the first forecast with a market line; the rule for games from 1 to 20 October 2026). The margin error uses the last valid forecast before kickoff.

ForecastsGradedModel pick vs market lineAvg. margin error
Publicly verifiable pregame123–8, 1 level6.8 pts

Model pick vs market line: the model pick against the line it was graded on, as described above (level results, no-pick and no-line games listed separately). For analysis and education only.

Friday 16 October
Fri 16 Oct · 00:15 UTCEmpower Field at Mile High
SEA
Seahawks3-2
14
proj 21.7
DEN
Broncos3-2home
17
proj 19.4
SEA 59%Tie 0.4%DEN 40%
Model pickDEN +3.0Tiny edge · about a coin flip (0.26 pts)Locked 13 Oct, 17:00 UTCThen: line SEA −3.0 · model SEA −2.74 · stats-only SEA −0.84Final pre-game: line SEA −2.5 · model SEA −2.24 · stats-only SEA −0.77Model pick: the side of the market line on which the model's projected margin fell at the first forecast with a market line (the rule for games from 1 to 20 October 2026), graded against that line. It is not a probability of finishing on that side of the line. Edge: how far the model's projected margin was from the line (under 0.5 points is about a coin flip). Stats-only: the football-only model, which uses no market line. Big gap: the stats-only number was at least 3 points from the line when the pick was locked (Biggest gap: at least 4 points); a prompt to check why the market disagrees, not a recommendation.
Market line
SEA −2.5
Model line
SEA −2.2
Market total
40.5
Model total
41.1
Stats-only line
SEA −0.8
Stats-only total
42.3
Projected winner
SEA 59% · by 2.2
Result: Model pick correct · margin error 5.2
QBs: Sam Darnold 99.6% / Bo Nix 99.6%
Uses market spread & total
Scored · forecast updated 15 Oct, 23:16 UTCDetails →
Sunday 18 October
Sun 18 Oct · 13:30 UTCDivisionalNeutral · Wembley Stadium
HOU
Texans1-4
34
proj 19.8
JAX
Jaguars4-1
31
proj 22.7
HOU 41%Tie 0.4%JAX 59%
Model pickJAX −2.5Tiny edge · about a coin flip (0.34 pts)Locked 13 Oct, 17:00 UTCAt lock: line JAX −2.5 · model JAX −2.84 · stats-only JAX −2.91Model pick: the side of the market line on which the model's projected margin fell at the first forecast with a market line (the rule for games from 1 to 20 October 2026), graded against that line. It is not a probability of finishing on that side of the line. Edge: how far the model's projected margin was from the line (under 0.5 points is about a coin flip). Stats-only: the football-only model, which uses no market line. Big gap: the stats-only number was at least 3 points from the line when the pick was locked (Biggest gap: at least 4 points); a prompt to check why the market disagrees, not a recommendation.
Market line
JAX −2.5
Model line
JAX −2.8
Market total
42.0
Model total
42.5
Stats-only line
JAX −2.9
Stats-only total
41.6
Projected winner
JAX 59% · by 2.8
Result: Model pick wrong · margin error 5.8
QBs: C.J. Stroud 99.6% / Trevor Lawrence 99.6%
Uses market spread & total
Scored · forecast updated 18 Oct, 12:46 UTCDetails →
Sun 18 Oct · 17:00 UTCDivisionalHuntington Bank Field
BAL
Ravens4-1
10
proj 21.4
CLE
Browns3-2home
13
proj 18.4
BAL 62%Tie 0.4%CLE 38%
Model pickBAL −4.5Tiny edge · about a coin flip (0.07 pts)Locked 13 Oct, 17:00 UTCThen: line BAL −4.5 · model BAL −4.57 · stats-only BAL −3.48Final pre-game: line BAL −3.0 · model BAL −3.06 · stats-only BAL −1.49Model pick: the side of the market line on which the model's projected margin fell at the first forecast with a market line (the rule for games from 1 to 20 October 2026), graded against that line. It is not a probability of finishing on that side of the line. Edge: how far the model's projected margin was from the line (under 0.5 points is about a coin flip). Stats-only: the football-only model, which uses no market line. Big gap: the stats-only number was at least 3 points from the line when the pick was locked (Biggest gap: at least 4 points); a prompt to check why the market disagrees, not a recommendation.
Market line
BAL −3.0
Model line
BAL −3.1
Market total
39.5
Model total
39.8
Stats-only line
BAL −1.5
Stats-only total
41.9
Projected winner
BAL 62% · by 3.1
Result: Model pick wrong · margin error 6.1
QBs: Tyler Huntley 99% / Deshaun Watson 99.6%
Uses market spread & total
Scored · forecast updated 18 Oct, 16:47 UTCDetails →
Sun 18 Oct · 17:00 UTCLincoln Financial Field
CAR
Panthers2-2
25
proj 21.1
PHI
Eagles2-3home
14
proj 18.1
CAR 62%Tie 0.4%PHI 38%
Model pickPHI −1.0Tiny edge · about a coin flip (0.04 pts)Locked 13 Oct, 17:00 UTCThen: line PHI −1.0 · model PHI −1.04 · stats-only CAR −0.41Final pre-game: line CAR −3.0 · model CAR −2.97 · stats-only CAR −3.97Model pick: the side of the market line on which the model's projected margin fell at the first forecast with a market line (the rule for games from 1 to 20 October 2026), graded against that line. It is not a probability of finishing on that side of the line. Edge: how far the model's projected margin was from the line (under 0.5 points is about a coin flip). Stats-only: the football-only model, which uses no market line. Big gap: the stats-only number was at least 3 points from the line when the pick was locked (Biggest gap: at least 4 points); a prompt to check why the market disagrees, not a recommendation.
Market line
CAR −3.0
Model line
CAR −3.0
Market total
39.0
Model total
39.2
Stats-only line
CAR −4.0
Stats-only total
40.4
Projected winner
CAR 62% · by 3.0
Result: Model pick wrong · margin error 8.0
QBs: Bryce Young 99% / Tanner McKee 99.6%
Uses market spread & total
Scored · forecast updated 18 Oct, 16:47 UTCDetails →
Sun 18 Oct · 17:00 UTCMercedes-Benz Stadium
CHI
Bears3-2
31
proj 26.4
ATL
Falcons2-3home
23
proj 25.5
CHI 55%Tie 0.4%ATL 45%
Model pickATL +3.0Tiny edge · about a coin flip (0.02 pts)Locked 13 Oct, 17:00 UTCThen: line CHI −3.0 · model CHI −2.98 · stats-only ATL −1.56Final pre-game: line CHI −1.0 · model CHI −0.99 · stats-only ATL −0.73Biggest gap Stats-only was 4.6 pts from the line, toward ATL. Check why the market disagrees.Model pick: the side of the market line on which the model's projected margin fell at the first forecast with a market line (the rule for games from 1 to 20 October 2026), graded against that line. It is not a probability of finishing on that side of the line. Edge: how far the model's projected margin was from the line (under 0.5 points is about a coin flip). Stats-only: the football-only model, which uses no market line. Big gap: the stats-only number was at least 3 points from the line when the pick was locked (Biggest gap: at least 4 points); a prompt to check why the market disagrees, not a recommendation.
Market line
CHI −1.0
Model line
CHI −1.0
Market total
51.5
Model total
51.9
Stats-only line
ATL −0.7
Stats-only total
46.9
Projected winner
CHI 55% · by 1.0
Result: Model pick wrong · margin error 7.0
QBs: Caleb Williams 82% / Michael Penix Jr. 99%
Uses market spread & totalLineup uncertain
Scored · forecast updated 18 Oct, 16:47 UTCDetails →
Sun 18 Oct · 17:00 UTCMetLife Stadium
NO
Saints1-4
34
proj 23.7
NYG
Giants3-1-1home
27
proj 22.8
NO 54%Tie 0.4%NYG 45%
Model pickNYG +0.5Tiny edge · about a coin flip (0.11 pts)Locked 13 Oct, 17:00 UTCThen: line NO −0.5 · model NO −0.39 · stats-only NYG −3.44Final pre-game: line NO −1.0 · model NO −0.89 · stats-only NYG −3.47Big gap Stats-only was 3.9 pts from the line, toward NYG. Check why the market disagrees.Model pick: the side of the market line on which the model's projected margin fell at the first forecast with a market line (the rule for games from 1 to 20 October 2026), graded against that line. It is not a probability of finishing on that side of the line. Edge: how far the model's projected margin was from the line (under 0.5 points is about a coin flip). Stats-only: the football-only model, which uses no market line. Big gap: the stats-only number was at least 3 points from the line when the pick was locked (Biggest gap: at least 4 points); a prompt to check why the market disagrees, not a recommendation.
Market line
NO −1.0
Model line
NO −0.9
Market total
46.0
Model total
46.5
Stats-only line
NYG −3.5
Stats-only total
43.6
Projected winner
NO 54% · by 0.9
Result: Model pick wrong · margin error 6.1
QBs: Tyler Shough 99% / Jameis Winston 99.6%
Uses market spread & total
Scored · forecast updated 18 Oct, 16:47 UTCDetails →
Sun 18 Oct · 17:00 UTCDivisionalGillette Stadium
NYJ
Jets2-3
19
proj 18.0
NE
Patriots3-2home
26
proj 25.1
NYJ 27%Tie 0.4%NE 73%
Model pickNE −7.0Tiny edge · about a coin flip (0.16 pts)Locked 13 Oct, 17:00 UTCAt lock: line NE −7.0 · model NE −7.16 · stats-only NE −8.42Model pick: the side of the market line on which the model's projected margin fell at the first forecast with a market line (the rule for games from 1 to 20 October 2026), graded against that line. It is not a probability of finishing on that side of the line. Edge: how far the model's projected margin was from the line (under 0.5 points is about a coin flip). Stats-only: the football-only model, which uses no market line. Big gap: the stats-only number was at least 3 points from the line when the pick was locked (Biggest gap: at least 4 points); a prompt to check why the market disagrees, not a recommendation.
Market line
NE −7.0
Model line
NE −7.2
Market total
42.5
Model total
43.1
Stats-only line
NE −8.4
Stats-only total
43.5
Projected winner
NE 73% · by 7.2
Result: Model pick level with the line · margin error 0.2
QBs: Geno Smith 99.6% / Drake Maye 99%
Uses market spread & total
Scored · forecast updated 18 Oct, 16:47 UTCDetails →
Sun 18 Oct · 17:00 UTCRaymond James Stadium
PIT
Steelers3-2
26
proj 20.4
TB
Buccaneers1-4home
19
proj 18.0
PIT 60%Tie 0.4%TB 40%
Model pickTB +3.0Tiny edge · about a coin flip (0.10 pts)Locked 13 Oct, 17:00 UTCThen: line PIT −3.0 · model PIT −2.90 · stats-only TB −1.52Final pre-game: line PIT −2.5 · model PIT −2.40 · stats-only TB −0.79Biggest gap Stats-only was 4.5 pts from the line, toward TB. Check why the market disagrees.Model pick: the side of the market line on which the model's projected margin fell at the first forecast with a market line (the rule for games from 1 to 20 October 2026), graded against that line. It is not a probability of finishing on that side of the line. Edge: how far the model's projected margin was from the line (under 0.5 points is about a coin flip). Stats-only: the football-only model, which uses no market line. Big gap: the stats-only number was at least 3 points from the line when the pick was locked (Biggest gap: at least 4 points); a prompt to check why the market disagrees, not a recommendation.
Market line
PIT −2.5
Model line
PIT −2.4
Market total
38.0
Model total
38.4
Stats-only line
TB −0.8
Stats-only total
41.3
Projected winner
PIT 60% · by 2.4
Result: Model pick wrong · margin error 4.6
QBs: Aaron Rodgers 99.6% / Jalon Daniels 99%
Uses market spread & total
Scored · forecast updated 18 Oct, 16:47 UTCDetails →
Sun 18 Oct · 17:00 UTCDivisionalLucas Oil Stadium
TEN
Titans0-5
26
proj 18.7
IND
Colts2-3home
20
proj 25.3
TEN 28%Tie 0.4%IND 71%
Model pickIND −7.0Tiny edge · about a coin flip (0.14 pts)Locked 13 Oct, 17:00 UTCThen: line IND −7.0 · model IND −7.14 · stats-only IND −7.65Final pre-game: line IND −6.5 · model IND −6.64 · stats-only IND −7.77Model pick: the side of the market line on which the model's projected margin fell at the first forecast with a market line (the rule for games from 1 to 20 October 2026), graded against that line. It is not a probability of finishing on that side of the line. Edge: how far the model's projected margin was from the line (under 0.5 points is about a coin flip). Stats-only: the football-only model, which uses no market line. Big gap: the stats-only number was at least 3 points from the line when the pick was locked (Biggest gap: at least 4 points); a prompt to check why the market disagrees, not a recommendation.
Market line
IND −6.5
Model line
IND −6.6
Market total
43.5
Model total
44.0
Stats-only line
IND −7.8
Stats-only total
44.4
Projected winner
IND 71% · by 6.6
Result: Model pick wrong · margin error 12.6
QBs: Cam Ward 99.6% / Daniel Jones 99%
Uses market spread & total
Scored · forecast updated 18 Oct, 16:47 UTCDetails →
Sun 18 Oct · 20:05 UTCDivisionalSoFi Stadium
ARI
Cardinals1-4
18
proj 19.2
LA
Rams2-3home
16
proj 32.2
ARI 14%Tie 0.4%LA 86%
Model pickARI +13.5Tiny edge · about a coin flip (0.03 pts)Locked 13 Oct, 17:00 UTCThen: line LA −13.5 · model LA −13.47 · stats-only LA −10.20Final pre-game: line LA −13.0 · model LA −12.98 · stats-only LA −10.32Big gap Stats-only was 3.3 pts from the line, toward ARI. Check why the market disagrees.Model pick: the side of the market line on which the model's projected margin fell at the first forecast with a market line (the rule for games from 1 to 20 October 2026), graded against that line. It is not a probability of finishing on that side of the line. Edge: how far the model's projected margin was from the line (under 0.5 points is about a coin flip). Stats-only: the football-only model, which uses no market line. Big gap: the stats-only number was at least 3 points from the line when the pick was locked (Biggest gap: at least 4 points); a prompt to check why the market disagrees, not a recommendation.
Market line
LA −13.0
Model line
LA −13.0
Market total
51.0
Model total
51.4
Stats-only line
LA −10.3
Stats-only total
52.1
Projected winner
LA 86% · by 13.0
Result: Model pick correct · margin error 15.0
QBs: Jacoby Brissett 99.6% / Matthew Stafford 99.6%
Uses market spread & total
Scored · forecast updated 18 Oct, 19:46 UTCDetails →
Sun 18 Oct · 20:25 UTCAllegiant Stadium
BUF
Bills4-1
33
proj 29.8
LV
Raiders3-2home
30
proj 24.5
BUF 70%Tie 0.4%LV 30%
Model pickBUF −6.0Tiny edge · about a coin flip (0.32 pts)Locked 13 Oct, 17:00 UTCThen: line BUF −6.0 · model BUF −6.32 · stats-only BUF −5.36Final pre-game: line BUF −5.0 · model BUF −5.32 · stats-only BUF −5.44Model pick: the side of the market line on which the model's projected margin fell at the first forecast with a market line (the rule for games from 1 to 20 October 2026), graded against that line. It is not a probability of finishing on that side of the line. Edge: how far the model's projected margin was from the line (under 0.5 points is about a coin flip). Stats-only: the football-only model, which uses no market line. Big gap: the stats-only number was at least 3 points from the line when the pick was locked (Biggest gap: at least 4 points); a prompt to check why the market disagrees, not a recommendation.
Market line
BUF −5.0
Model line
BUF −5.3
Market total
54.0
Model total
54.2
Stats-only line
BUF −5.4
Stats-only total
56.7
Projected winner
BUF 70% · by 5.3
Result: Model pick wrong · margin error 2.3
QBs: Josh Allen 99.6% / Kirk Cousins 99.6%
Uses market spread & total
Scored · forecast updated 18 Oct, 19:46 UTCDetails →
Sun 18 Oct · 20:25 UTCDivisionalGEHA Field at Arrowhead Stadium
LAC
Chargers0-5
6
proj 17.2
KC
Chiefs4-0home
22
proj 25.1
LAC 25%Tie 0.4%KC 75%
Model pickKC −8.0Tiny edge · about a coin flip (0.34 pts)Locked 13 Oct, 17:00 UTCThen: line KC −8.0 · model KC −8.34 · stats-only KC −9.19Final pre-game: line KC −7.5 · model KC −7.84 · stats-only KC −9.20Model pick: the side of the market line on which the model's projected margin fell at the first forecast with a market line (the rule for games from 1 to 20 October 2026), graded against that line. It is not a probability of finishing on that side of the line. Edge: how far the model's projected margin was from the line (under 0.5 points is about a coin flip). Stats-only: the football-only model, which uses no market line. Big gap: the stats-only number was at least 3 points from the line when the pick was locked (Biggest gap: at least 4 points); a prompt to check why the market disagrees, not a recommendation.
Market line
KC −7.5
Model line
KC −7.8
Market total
41.5
Model total
42.3
Stats-only line
KC −9.2
Stats-only total
42.0
Projected winner
KC 75% · by 7.8
Result: Model pick correct · margin error 8.2
QBs: Justin Herbert 99.6% / Patrick Mahomes 99%
Uses market spread & total
Scored · forecast updated 18 Oct, 19:46 UTCDetails →
Monday 19 October
Mon 19 Oct · 00:20 UTCLambeau Field
DAL
Cowboys2-3
27.7
GB
Packers3-2home
26.6
DAL 55%Tie 0.4%GB 44%
Model pickDAL +1.0Tiny edge · about a coin flip (0.16 pts)Locked 13 Oct, 17:00 UTCThen: line GB −1.0 · model GB −0.84 · stats-only GB −3.86Now: line DAL −1.0 · model DAL −1.17 · stats-only GB −3.79Model pick: the side of the market line on which the model's projected margin fell at the first forecast with a market line (the rule for games from 1 to 20 October 2026), graded against that line. It is not a probability of finishing on that side of the line. Edge: how far the model's projected margin was from the line (under 0.5 points is about a coin flip). Stats-only: the football-only model, which uses no market line. Big gap: the stats-only number was at least 3 points from the line when the pick was locked (Biggest gap: at least 4 points); a prompt to check why the market disagrees, not a recommendation.
Market line
DAL −1.0
Model line
DAL −1.2
Market total
54.0
Model total
54.3
Stats-only line
GB −3.8
Stats-only total
52.9
Projected winner
DAL 55% · by 1.2
QBs: Dak Prescott 99.6% / Jordan Love 99.6%
Uses market spread & total
Latest pregame forecast (may still update) · forecast updated 18 Oct, 23:46 UTCDetails →
Tuesday 20 October
Tue 20 Oct · 00:15 UTCLevi's Stadium
WAS
Commanders1-3-1
22.5
SF
49ers5-0home
30.0
WAS 26%Tie 0.4%SF 74%
Model pickSF −7.5Tiny edge · about a coin flip (0.05 pts)Locked 13 Oct, 17:00 UTCAt lock: line SF −7.5 · model SF −7.55 · stats-only SF −9.24Model pick: the side of the market line on which the model's projected margin fell at the first forecast with a market line (the rule for games from 1 to 20 October 2026), graded against that line. It is not a probability of finishing on that side of the line. Edge: how far the model's projected margin was from the line (under 0.5 points is about a coin flip). Stats-only: the football-only model, which uses no market line. Big gap: the stats-only number was at least 3 points from the line when the pick was locked (Biggest gap: at least 4 points); a prompt to check why the market disagrees, not a recommendation.
Market line
SF −7.5
Model line
SF −7.5
Market total
52.0
Model total
52.5
Stats-only line
SF −9.3
Stats-only total
51.0
Projected winner
SF 74% · by 7.5
QBs: Jayden Daniels 99% / Brock Purdy 99.6%
Uses market spread & total
Latest pregame forecast (may still update) · forecast updated 18 Oct, 23:46 UTCDetails →

Scores are expected values (means), not predicted final scores. Win probabilities include a small regular-season tie probability. Market spread and total are shown for context and, for the combined model, are model inputs. Every forecast is frozen with a timestamp before kickoff; results are scored against that frozen version.