NFL 2026 · Week 4
Kick-off
Mon 5 Oct, 4:00am AEDT
Sun 4 Oct, 1:00pm ET
AU consensus as of 01/10/2026, 12:36:06 pm
Power Comparison
DALRating -1.6
Defence by zone (EPA allowed)
HOURating -1.1
Defence by zone (EPA allowed)
Zone cells are EPA/play vs league mean, left to right in model order. Green = better for that team. Hover a cell for the zone name.
Matchup Edges
DAL offence vs HOU
deep-right passing
+0.52 EPA/play
HOU offence vs DAL
short-right passing
+0.16 EPA/play
HOU offence vs DAL
inside runs
+0.26 EPA/play
Market
Spread (HOU)
Open -2.5 → Now -3·Move -0.5Moved toward model
AU bestHOU -2.5 @ $1.83 PointsBet (AU)
AU best (other side)DAL +3.5 @ $1.78 Dabble AU
Global bestHOU -2.5 @ $1.83 PointsBet (AU)
Total
Open 47.5 → Now 48.0·Move +0.5Moved toward model
AU best overOver 47.5 @ $1.85 SportsBet
AU best underUnder 48.5 @ $1.90 PointsBet (AU)
Global best overOver 47.5 @ $1.87 FanDuel
AU best HOU ML
$1.69 PointsBet (AU)
11 AU books · 33 total · 9 snapshots
The Call
No edge
Model HOU +4.2 · total 48.4Spread · Analysis
HOU -2.5
$1.83 PointsBet (AU)
Edge +1.7PASS
Total · Analysis
Over 47.5
$1.85 SportsBet
Edge +0.9PASS
Houston is favoured at home at -3.0, with the best AU line at -2.5 via PointsBet (1.83), and the model's 4.2-point margin produces just a 1.7-point edge against that line, well short of the 10-point bar required before week 6, so this is a pass. The total shows the same pattern: SportsBet's 47.5 over sits only 0.9 points under the model's 48.4 projection, again too small to act on. Houston's short-right passing game grades at +0.16 EPA per play against Dallas's defence, a modest structural lean but not enough to move either market read. Early season, the rating still blends last year's numbers with the market's post-offseason view of each roster, scaled by continuity, with 5,586 plays of 2026 football only starting to be phased in. The model's season-to-date record sits at 1-0 against the spread through week 3.
On 2025 season charting, Dak Prescott averaged 0.36 EPA per dropback when clean but dropped to -0.28 under pressure, a number worth watching live even though neither market clears the model's threshold this week.
Why · HOU short-right passing (+0.16 EPA/play vs DAL)
Situational
2026 through week 3 (pressure, coverage and personnel: 2025 season) · FTN charting via nflverse · min 40 snaps per cell (25 for box counts)
DAL · 631 dropbacks
Dak Prescott
Clean pocket EPA/db+0.36 (440 snaps)
Under pressure EPA/db-0.28 (191 snaps)
Pressure rate faced30%
Int-worthy rate3%
Drop rate (catchable)6%
QB-fault sack rate3%
League: +0.26 clean · -0.46 pressured
HOU defence: pressure 30% · blitz —
HOU · 446 dropbacks
C.J. Stroud
Clean pocket EPA/db+0.30 (317 snaps)
Under pressure EPA/db-0.43 (129 snaps)
Pressure rate faced29%
Int-worthy rate3%
Drop rate (catchable)3%
QB-fault sack rate2%
League: +0.26 clean · -0.46 pressured
DAL defence: pressure 32% · blitz —
DAL offence vs HOU coverage
| Shell | HOU usage | DAL EPA/db | League |
|---|
| Cover 1 | 29% (178 snaps) | +0.38 (142 snaps) | +0.06 |
| Cover 2 | 23% (143 snaps) | -0.04 (156 snaps) | 0.00 |
| Cover 3 | 22% (138 snaps) | +0.16 (145 snaps) | +0.06 |
| Cover 4 | 12% (76 snaps) | +0.26 (100 snaps) | +0.07 |
| Cover 6 | 8% (48 snaps) | — | -0.03 |
HOU man coverage 32% · DAL vs man +0.33, vs zone +0.07
HOU offence vs DAL coverage
| Shell | DAL usage | HOU EPA/db | League |
|---|
| Cover 3 | 29% (184 snaps) | +0.17 (164 snaps) | +0.06 |
| Cover 2 | 25% (157 snaps) | -0.02 (162 snaps) | 0.00 |
| Cover 1 | 21% (134 snaps) | +0.36 (132 snaps) | +0.06 |
| Cover 4 | 13% (80 snaps) | +0.16 (59 snaps) | +0.07 |
| Cover 6 | 6% (40 snaps) | — | -0.03 |
DAL man coverage 25% · HOU vs man -0.02, vs zone +0.10
DAL tendencies
Play action rate25% · — vs +0.40 EPA (29 snaps)
RPO rate2% (4 snaps)
Screen rate7% (8 snaps)
Personnel11 68%, 12 17%, 21 8%
Def pressure / blitz / man32% / — / 25%
Box vs run, EPA allowedlight +0.13 (36 snaps) · neutral -0.04 (49 snaps)
HOU tendencies
Play action rate19% · — vs -0.07 EPA (25 snaps)
RPO rate3% (5 snaps)
Screen rate3% (4 snaps)
Personnel11 72%, 12 12%, 21 7%, 10 4%
Def pressure / blitz / man30% / — / 32%
Box vs run, EPA allowedneutral -0.07 (38 snaps)
Published Model Results
2026 season · Through Week 6 · Updated 25 Oct 2026, 04:02 pm AEDT
This results panel was subsequently updated after this article was first published.
All rated analysis (W–L–P)| Coverage | Spreads | Totals |
|---|
| Week 6 | 6–3–1 | 7–3–0 |
|---|
| Season to W6 | 37–34–4 | 41–32–2 |
|---|
Includes PASS signals; UNRATED markets excluded.
Week 6 PLAY recommendations: 4–2–1 · +1.50u paper return
One unit risked per PLAY recommendation at its original published decimal price: win = price minus one, loss = minus one, push = zero. Hypothetical paper return, not actual bets. No STRONG recommendations through Week 6.
SEA-DEN: both LEAN analysis calls won; No formal betting selection. Excluded from PLAY returns.
Results source and coverageTracked Bets
2026 season graded betting log through Week 3 · last updated 01 Oct 2026, 12:42 pm
Recorded bets only. Published model selections and historical backtests are tracked separately.
Weeks 1 to 5 are published analysis-first while the current-season play sample is still small. Ratings lean more on prior-season performance, market anchors, roster continuity and supported manual adjustments until the 2026 sample grows. PLAY labels are model recommendations, not confirmation that a bet has been placed. Historical early-season backtest claims are withheld from this display unless independently verified.
Analysis only. Not financial advice. Gamble responsibly. Gambling Help Online 1800 858 858. Read our responsible gambling policy.
savvyplays.comPublished 01/10/2026V3.2