Skip to content
Sharp Six

Against the spread

Week 3 · 16 games · 6 picks · 10 leans

Updated under an hour ago

Week 3

Six games clear 3 stars this week, and only one clears four: Titans getting 3.5 in New York, where Jaxson Dart is out and the number already had Tennessee half a point live before the quarterback news. Take the points there. Behind it, Dolphins getting 11.5 at home against Kansas City is the shape worth your attention, with 93% of the money on the Chiefs and the rating gap still favoring Miami. Broncos, Browns, and Colts are single-digit line calls, fine to note and easy to forget. The live record is 0-0-0, which means nothing yet, and the backtest over the first two weeks sits at 14-17-1 overall with the sharp picks themselves 7-6. Nobody's owed anything for showing up early.

From the Sharp Model, off the week’s picks and the record so far.

The record

0–0–0

nothing decided yet

The Sharp picks, three stars and up, every live one graded against the line it was made on. Pushes are counted and left out of the percentage. Nothing below three stars is in this number.

Week 3 is still being played. Its grades land when the scores do.

By stars

★★★★★
0–0–0 nothing decided
0 picks
★★★★☆
0–0–0 nothing decided
1 pick
★★★☆☆
0–0–0 nothing decided
5 picks

One and two stars are leans. Shown in the table, counted nowhere. 10 leans this season.

By signal

Ratings
0–0–0 nothing decided
fired 16
Our spread against the market line: the prior from win totals blended with capped margins, plus 1.5 for home field.
Rest
0–0–0 nothing decided
fired 2
A gap of 3 days or more between the two teams’ rest. 1.5 points to the rested side.
Public
0–0–0 nothing decided
fired 9
65% or more of the spread money on one side. 1 point the other way.
Weather
0–0–0 nothing decided
not fired yet
Outdoors with wind at 20 mph or a 60% chance of rain. 1 point to the underdog.
Quarterback
0–0–0 nothing decided
fired 4
The starting quarterback out, doubtful, on injured reserve, PUP or suspended. 4 points against his team. Live weeks only.

Each signal is graded on the side it favoured, whichever side the pick took, over every game it fired on, leans included.

Backtest

7–6–054% right

Weeks 1 and 2, run after the fact against the closing line, three stars and up. A demonstration of the system, not a record.

Closing line value

The closing line lands with the first graded week.

The closing line is the sharpest number there is: the market’s last word before kickoff, with every opinion in it. A pick that beats it took a number the market then moved toward, whatever the score did afterwards. A process that beats it consistently is a process, whatever one week’s scores say; one that does not is getting its wins from somewhere else.

Live Sharp picks only. A backtest was made against the close and has nothing to beat; a lean is not a pick. The close is the median spread across books from the last pull before kickoff.

Week 3: the Sharp picks

Three stars and up, which is what a pick means here. Each one carries the case it was published with; the case does not change after kickoff and neither does the line under it.

TEN at NYG

Sun, Sep 27, 1:00 PM ET★★★★☆

TEN +3.5

Market NYG −3.5 over TEN. Our number TEN +3, an edge of +0.5 before the signals. Line now TEN +2.5.

  • Ratings · TEN +0.5
  • Quarterback · TEN +4.0
  • The ratings edge is 0.5 points toward TEN.
  • NYG's top quarterback, Jaxson Dart, is listed Out.

NYG giving 3.5 to TEN, and this desk takes TEN. The rating gap is a modest half point, but the real move is Jaxson Dart sitting Out, worth 4 points on its own by our number. Stack a shaky Giants offense under backup direction against a Titans team that's already covered once this season at 1-1 against the spread, and 3.5 is too many points to be giving away. Four stars, take TEN, and forget that the Titans are winless straight up -- that's not what this number is about.

From the Sharp Model, off the numbers on this card and nothing else.

ATL at GB

Thu, Sep 24, 8:15 PM ET★★★☆☆

GB −6

Market GB −6 over ATL. Our number GB −4.9, an edge of −1.1 before the signals. Line now GB −5.5.

  • Ratings · ATL +1.1
  • Public · GB +1.0
  • Quarterback · GB +4.0
  • The ratings edge is 1.1 points toward ATL.
  • 67% of the spread money is on ATL. That is a point to GB.
  • ATL's top quarterback, Michael Penix, is listed Out.

The line is GB giving 6 to ATL. Our number is GB giving 4.9 to ATL. ATL's top quarterback, Michael Penix, is listed Out. The side is GB.

Sharp Model case, numbers only. The written case did not land when this pick published.

KC at MIA

Sun, Sep 27, 1:00 PM ET★★★☆☆

MIA +11.5

Market KC −11.5 over MIA. Our number MIA +9.5, an edge of +2.0 before the signals.

  • Ratings · MIA +2.0
  • Public · MIA +1.0
  • The ratings edge is 2 points toward MIA.
  • 93% of the spread money is on KC. That is a point to MIA.

KC giving 11.5 to MIA is 2 points off our number, which has KC giving only 9.5. That gap plus 93% of the spread money piling onto KC — a public number that ugly is its own signal — puts MIA on the right side of this line. MIA is 0-2 straight up and hasn't covered either game, but neither of those numbers moves a rating built on 2 points of edge. Take the points at 3 stars and let the record, currently nothing graded, do the talking.

From the Sharp Model, off the numbers on this card and nothing else.

NE at JAX

Sun, Sep 27, 1:00 PM ET★★★☆☆

NE +3

Market JAX −3 over NE. Our number NE −0.7, an edge of +3.7 before the signals.

  • Ratings · NE +3.7
  • The ratings edge is 3.7 points toward NE.

JAX giving 3 with NE at 0.7 leaves 3.7 points sitting on the table, and that gap is the whole case. Nothing else fired here, just the number, but the number is what moves your record, not the record itself, which sits at 0-0-0 because this is the first entry at 3 stars. Take NE getting the points. If you need a track record to believe a rating, the backtest is right there at 7-6, unspectacular and honest.

From the Sharp Model, off the numbers on this card and nothing else.

BAL at DAL

Sun, Sep 27, 4:25 PM ET★★★☆☆

DAL +3

Market BAL −3 over DAL. Our number DAL +0.4, an edge of +2.6 before the signals. Line now DAL +3.5.

  • Ratings · DAL +2.6
  • Public · DAL +1.0
  • The ratings edge is 2.6 points toward DAL.
  • 78% of the spread money is on BAL. That is a point to DAL.

BAL giving 3, our number has it at 0.4 — a 2.6-point gap that has nothing to do with respect for Baltimore. Add in that 78% of the spread money is stacked on BAL, which is exactly the kind of lopsided ticket count that makes a number soft on the other side. Both teams sit at 1-1 straight up and against the spread, so there's no record doing the deciding here, just the math and the crowd being wrong together. Take DAL and stop worrying about the favorite tag.

From the Sharp Model, off the numbers on this card and nothing else.

PHI at CHI

Mon, Sep 28, 8:15 PM ET★★★☆☆

CHI +3

Market PHI −3 over CHI. Our number CHI −3, an edge of +6.0 before the signals. Line now CHI +4.5.

  • Ratings · CHI +6.0
  • Public · CHI +1.0
  • Quarterback · PHI +4.0
  • The ratings edge is 6 points toward CHI.
  • 80% of the spread money is on PHI. That is a point to CHI.
  • CHI's top quarterback, Caleb Williams, is listed Out.

CHI getting 3 is the side, and the ratings gap alone is 6 points before anything else counts. The public is stacked 80% on PHI, which is its own point to CHI, and that's carrying a real hit against it: Caleb Williams out is worth 4 points back to PHI. Net it out and CHI still clears the number, so this is a 3-star Sharp play with the backtest running 7-6-0. PHI is unbeaten straight up and still can't cover anything at 0-2 against the spread, which tells you where the real value has been sitting.

From the Sharp Model, off the numbers on this card and nothing else.

Week 3: every game

One row per game, one side each. The rows at one and two stars are leans, marked no play: shown so the system’s every call is on the page, counted in no record. The line is the one the side was taken against; the current line appears beside it only when the market has moved.

GameSideOursEdgeStarsResult
ATL at GB
Thu 8:15 PM
GB −6GB −4.9−1.1
★★★☆☆
–
CAR at CLE
Sun 1:00 PM
CLE +2.5CLE +1.8+0.7
★★☆☆☆no play
–
CIN at PIT
Sun 1:00 PM
PIT +3.5PIT +3.4+0.1
★☆☆☆☆no play
–
HOU at IND
Sun 1:00 PM
IND +2.5IND +0.1+2.4
★★☆☆☆no play
–
KC at MIA
Sun 1:00 PM
MIA +11.5MIA +9.5+2.0
★★★☆☆
–
LAC at BUF
Sun 1:00 PM
BUF −7BUF −7.4+0.4
★★☆☆☆no play
–
NE at JAX
Sun 1:00 PM
NE +3NE −0.7+3.7
★★★☆☆
–
NYJ at DET
Sun 1:00 PM
NYJ +6.5NYJ +3.3+3.2
★★☆☆☆no play
–
SEA at WAS
Sun 1:00 PM
WAS +7WAS +5.5+1.5
★★☆☆☆no play
–
TEN at NYG
Sun 1:00 PM
TEN +3.5TEN +3+0.5
★★★★☆
–
ARI at SF
Sun 4:05 PM
SF −8.5SF −10.8+2.3
★☆☆☆☆no play
–
MIN at TB
Sun 4:05 PM
TB +1.5TB +2.9−1.4
★★☆☆☆no play
–
BAL at DAL
Sun 4:25 PM
DAL +3DAL +0.4+2.6
★★★☆☆
–
LV at NO
Sun 4:25 PM
LV +3LV +1.8+1.2
★☆☆☆☆no play
–
LAR at DEN
Sun 8:20 PM
DEN +2.5DEN +1.5+1.0
★★☆☆☆no play
–
PHI at CHI
Mon 8:15 PM
CHI +3CHI −3+6.0
★★★☆☆
–

Close and CLV land with the grade. A lean carries them for the reader and counts in neither block above.

Past weeks

Week 26–4–016 games · 10 picks · 6 leans · all gradedBacktest

Week 2 was run after the fact against the closing line. That makes it a demonstration of the system, not a record. 16 games, 10 of them at 3 stars or more, and it went 8-8-0.

Sharp Model lead, numbers only.

GB at NYJ

Sun, Sep 20, 1:00 PM ET★★★★★

NYJ +3.5

Right · +0.5

Market GB −3.5 over NYJ. Our number NYJ −2.1, an edge of +5.6 before the signals. Backtest: run after the game, against the closing line.

  • Ratings · NYJ +5.6· right
  • Weather · NYJ +1.0· right
  • The ratings edge is 5.6 points toward NYJ.
  • A 65% chance of rain outdoors is a point to the underdog, NYJ.

The line is GB giving 3.5 to NYJ. Our number is NYJ giving 2.1 to GB. The ratings edge is 5.6 points toward NYJ. The side is NYJ.

Sharp Model case, numbers only. The written case did not land when this pick published.

Final: GB 20 at NYJ 17

CAR at ATL

Sun, Sep 20, 1:00 PM ET★★★★☆

ATL +2.5

Wrong · −28.5

Market CAR −2.5 over ATL. Our number ATL −3.1, an edge of +5.6 before the signals. Backtest: run after the game, against the closing line.

  • Ratings · ATL +5.6· wrong
  • The ratings edge is 5.6 points toward ATL.

The line is CAR giving 2.5 to ATL. Our number is ATL giving 3.1 to CAR. The ratings edge is 5.6 points toward ATL. The side is ATL.

Sharp Model case, numbers only. The written case did not land when this pick published.

Final: CAR 34 at ATL 3

LV at LAC

Sun, Sep 20, 4:05 PM ET★★★★☆

LV +6.5

Right · +18.5

Market LAC −6.5 over LV. Our number LV +1.2, an edge of +5.3 before the signals. Backtest: run after the game, against the closing line.

  • Ratings · LV +5.3· right
  • The ratings edge is 5.3 points toward LV.

The line is LAC giving 6.5 to LV. Our number is LAC giving 1.2 to LV. The ratings edge is 5.3 points toward LV. The side is LV.

Sharp Model case, numbers only. The written case did not land when this pick published.

Final: LV 26 at LAC 14

DET at BUF

Thu, Sep 17, 8:15 PM ET★★★☆☆

DET +5.5

Wrong · −4.5

Market BUF −5.5 over DET. Our number DET +2.2, an edge of +3.3 before the signals. Backtest: run after the game, against the closing line.

  • Ratings · DET +3.3· wrong
  • The ratings edge is 3.3 points toward DET.

The line is BUF giving 5.5 to DET. Our number is BUF giving 2.2 to DET. The ratings edge is 3.3 points toward DET. The side is DET.

Sharp Model case, numbers only. The written case did not land when this pick published.

Final: DET 31 at BUF 41

CIN at HOU

Sun, Sep 20, 1:00 PM ET★★★☆☆

CIN +3

Right · +17.0

Market HOU −3 over CIN. Our number CIN −0.4, an edge of +3.4 before the signals. Backtest: run after the game, against the closing line.

  • Ratings · CIN +3.4· right
  • The ratings edge is 3.4 points toward CIN.

The line is HOU giving 3 to CIN. Our number is CIN giving 0.4 to HOU. The ratings edge is 3.4 points toward CIN. The side is CIN.

Sharp Model case, numbers only. The written case did not land when this pick published.

Final: CIN 20 at HOU 6

CLE at TB

Sun, Sep 20, 1:00 PM ET★★★☆☆

CLE +8.5

Right · +12.5

Market TB −8.5 over CLE. Our number CLE +5.8, an edge of +2.7 before the signals. Backtest: run after the game, against the closing line.

  • Ratings · CLE +2.7· right
  • Public · CLE +1.0· right
  • The ratings edge is 2.7 points toward CLE.
  • 89% of the spread money is on TB. That is a point to CLE.

The line is TB giving 8.5 to CLE. Our number is TB giving 5.8 to CLE. The ratings edge is 2.7 points toward CLE. The side is CLE.

Sharp Model case, numbers only. The written case did not land when this pick published.

Final: CLE 23 at TB 19

NO at BAL

Sun, Sep 20, 1:00 PM ET★★★☆☆

NO +8.5

Right · +15.5

Market BAL −8.5 over NO. Our number NO +6.5, an edge of +2.0 before the signals. Backtest: run after the game, against the closing line.

  • Ratings · NO +2.0· right
  • Public · NO +1.0· right
  • The ratings edge is 2 points toward NO.
  • 86% of the spread money is on BAL. That is a point to NO.

The line is BAL giving 8.5 to NO. Our number is BAL giving 6.5 to NO. The ratings edge is 2 points toward NO. The side is NO.

Sharp Model case, numbers only. The written case did not land when this pick published.

Final: NO 24 at BAL 17

PHI at TEN

Sun, Sep 20, 1:00 PM ET★★★☆☆

TEN +7

Right · +3.0

Market PHI −7 over TEN. Our number TEN +3.6, an edge of +3.4 before the signals. Backtest: run after the game, against the closing line.

  • Ratings · TEN +3.4· right
  • Public · TEN +1.0· right
  • The ratings edge is 3.4 points toward TEN.
  • 92% of the spread money is on PHI. That is a point to TEN.

The line is PHI giving 7 to TEN. Our number is PHI giving 3.6 to TEN. The ratings edge is 3.4 points toward TEN. The side is TEN.

Sharp Model case, numbers only. The written case did not land when this pick published.

Final: PHI 24 at TEN 20

JAX at DEN

Sun, Sep 20, 4:05 PM ET★★★☆☆

JAX +2.5

Wrong · −4.5

Market DEN −2.5 over JAX. Our number JAX −2.5, an edge of +5.0 before the signals. Backtest: run after the game, against the closing line.

  • Ratings · JAX +5.0· wrong
  • Public · DEN +1.0· right
  • The ratings edge is 5 points toward JAX.
  • 81% of the spread money is on JAX. That is a point to DEN.

The line is DEN giving 2.5 to JAX. Our number is JAX giving 2.5 to DEN. The ratings edge is 5 points toward JAX. The side is JAX.

Sharp Model case, numbers only. The written case did not land when this pick published.

Final: JAX 13 at DEN 20

NYG at LAR

Mon, Sep 21, 8:15 PM ET★★★☆☆

NYG +6.5

Wrong · −15.5

Market LAR −6.5 over NYG. Our number NYG +1, an edge of +5.5 before the signals. Backtest: run after the game, against the closing line.

  • Ratings · NYG +5.5· wrong
  • Rest · LAR +1.5· right
  • The ratings edge is 5.5 points toward NYG.
  • LAR come in on 11 days' rest to NYG's 8.

The line is LAR giving 6.5 to NYG. Our number is LAR giving 1 to NYG. The ratings edge is 5.5 points toward NYG. The side is NYG.

Sharp Model case, numbers only. The written case did not land when this pick published.

Final: NYG 6 at LAR 28

GameSideOursEdgeStarsResult
DET at BUF
Thu 8:15 PM
DET +5.5DET +2.2+3.3
★★★☆☆
Wrong · −4.5DET 31 at BUF 41
CAR at ATL
Sun 1:00 PM
ATL +2.5ATL −3.1+5.6
★★★★☆
Wrong · −28.5CAR 34 at ATL 3
CIN at HOU
Sun 1:00 PM
CIN +3CIN −0.4+3.4
★★★☆☆
Right · +17.0CIN 20 at HOU 6
CLE at TB
Sun 1:00 PM
CLE +8.5CLE +5.8+2.7
★★★☆☆
Right · +12.5CLE 23 at TB 19
GB at NYJ
Sun 1:00 PM
NYJ +3.5NYJ −2.1+5.6
★★★★★
Right · +0.5GB 20 at NYJ 17
MIN at CHI
Sun 1:00 PM
MIN +4.5MIN +2.5+2.0
★★☆☆☆no play
Right · +10.5MIN 9 at CHI 3
NO at BAL
Sun 1:00 PM
NO +8.5NO +6.5+2.0
★★★☆☆
Right · +15.5NO 24 at BAL 17
PHI at TEN
Sun 1:00 PM
TEN +7TEN +3.6+3.4
★★★☆☆
Right · +3.0PHI 24 at TEN 20
PIT at NE
Sun 1:00 PM
PIT +4.5PIT +2+2.5
★☆☆☆☆no play
Wrong · −12.5PIT 3 at NE 20
JAX at DEN
Sun 4:05 PM
JAX +2.5JAX −2.5+5.0
★★★☆☆
Wrong · −4.5JAX 13 at DEN 20
LV at LAC
Sun 4:05 PM
LV +6.5LV +1.2+5.3
★★★★☆
Right · +18.5LV 26 at LAC 14
MIA at SF
Sun 4:25 PM
MIA +12.5MIA +10.8+1.7
★☆☆☆☆no play
Wrong · −9.5MIA 13 at SF 35
SEA at ARI
Sun 4:25 PM
SEA −3.5SEA −2.7−0.8
★☆☆☆☆no play
Right · +20.5SEA 31 at ARI 7
WAS at DAL
Sun 4:25 PM
WAS +3.5WAS +1.9+1.6
★☆☆☆☆no play
Wrong · −13.5WAS 20 at DAL 37
IND at KC
Sun 8:20 PM
KC −6KC −8.4+2.4
★★☆☆☆no play
Wrong · −3.0IND 30 at KC 33
NYG at LAR
Mon 8:15 PM
NYG +6.5NYG +1+5.5
★★★☆☆
Wrong · −15.5NYG 6 at LAR 28

Close and CLV are blank on a backtest week: the line at pick already was the close, so there was nothing to beat.

Week 11–2–016 games · 3 picks · 13 leans · all gradedBacktest

Week 1 was run after the fact against the closing line. That makes it a demonstration of the system, not a record. 16 games, 3 of them at 3 stars or more, and it went 6-9-1.

Sharp Model lead, numbers only.

CLE at JAX

Sun, Sep 13, 1:00 PM ET★★★★☆

CLE +8.5

Wrong · −15.5

Market JAX −8.5 over CLE. Our number CLE +4, an edge of +4.5 before the signals. Backtest: run after the game, against the closing line.

  • Ratings · CLE +4.5· wrong
  • The ratings edge is 4.5 points toward CLE.

The line is JAX giving 8.5 to CLE. Our number is JAX giving 4 to CLE. The ratings edge is 4.5 points toward CLE. The side is CLE.

Sharp Model case, numbers only. The written case did not land when this pick published.

Final: CLE 10 at JAX 34

ATL at PIT

Sun, Sep 13, 1:00 PM ET★★★☆☆

ATL +6.5

Wrong · −0.5

Market PIT −6.5 over ATL. Our number ATL +2.4, an edge of +4.1 before the signals. Backtest: run after the game, against the closing line.

  • Ratings · ATL +4.1· wrong
  • The ratings edge is 4.1 points toward ATL.

The line is PIT giving 6.5 to ATL. Our number is PIT giving 2.4 to ATL. The ratings edge is 4.1 points toward ATL. The side is ATL.

Sharp Model case, numbers only. The written case did not land when this pick published.

Final: ATL 13 at PIT 20

NO at DET

Sun, Sep 13, 1:00 PM ET★★★☆☆

NO +7

Right · +6.0

Market DET −7 over NO. Our number NO +4, an edge of +3.0 before the signals. Backtest: run after the game, against the closing line.

  • Ratings · NO +3.0· right
  • The ratings edge is 3 points toward NO.

The line is DET giving 7 to NO. Our number is DET giving 4 to NO. The ratings edge is 3 points toward NO. The side is NO.

Sharp Model case, numbers only. The written case did not land when this pick published.

Final: NO 30 at DET 31

GameSideOursEdgeStarsResult
NE at SEA
Wed 8:20 PM
NE +3NE +1.5+1.5
★★☆☆☆no play
PushNE 10 at SEA 13
SF at LAR
Thu 8:35 PM
SF +3.5SF +3.2+0.3
★☆☆☆☆no play
Right · +23.5SF 27 at LAR 7
ATL at PIT
Sun 1:00 PM
ATL +6.5ATL +2.4+4.1
★★★☆☆
Wrong · −0.5ATL 13 at PIT 20
BAL at IND
Sun 1:00 PM
IND +3IND +1.9+1.1
★☆☆☆☆no play
Wrong · −15.0BAL 41 at IND 23
BUF at HOU
Sun 1:00 PM
HOU +1.5HOU −0.7+2.2
★★☆☆☆no play
Wrong · −3.5BUF 36 at HOU 31
CHI at CAR
Sun 1:00 PM
CAR +3CAR +0.2+2.8
★★☆☆☆no play
Wrong · −19.0CHI 59 at CAR 37
CLE at JAX
Sun 1:00 PM
CLE +8.5CLE +4+4.5
★★★★☆
Wrong · −15.5CLE 10 at JAX 34
NO at DET
Sun 1:00 PM
NO +7NO +4+3.0
★★★☆☆
Right · +6.0NO 30 at DET 31
NYJ at TEN
Sun 1:00 PM
TEN −1.5TEN −2.3+0.8
★☆☆☆☆no play
Wrong · −14.5NYJ 23 at TEN 10
TB at CIN
Sun 1:00 PM
TB +3.5TB +3.2+0.3
★☆☆☆☆no play
Wrong · −2.5TB 27 at CIN 33
ARI at LAC
Sun 4:25 PM
ARI +8.5ARI +6.6+1.9
★★☆☆☆no play
Right · +20.5ARI 26 at LAC 14
GB at MIN
Sun 4:25 PM
GB +1.5GB +0.7+0.8
★☆☆☆☆no play
Wrong · −15.5GB 22 at MIN 39
MIA at LV
Sun 4:25 PM
LV −3LV −3.2+0.2
★☆☆☆☆no play
Right · +11.0MIA 13 at LV 27
WAS at PHI
Sun 4:25 PM
WAS +6WAS +4+2.0
★★☆☆☆no play
Right · +4.0WAS 22 at PHI 24
DAL at NYG
Sun 8:20 PM
NYG +3NYG +0.2+2.8
★★☆☆☆no play
Right · +11.0DAL 20 at NYG 28
DEN at KC
Mon 8:15 PM
DEN +2.5DEN +2.4+0.1
★☆☆☆☆no play
Wrong · −18.5DEN 10 at KC 31

Close and CLV are blank on a backtest week: the line at pick already was the close, so there was nothing to beat.

How it works

The market is not the input. Every team starts the season with a prior: its win total against the average team’s, at 0.85 points a win. As games are played, a second rating is fitted to the margins actually posted, with every margin capped at 21 points, because a blowout says less about next week than the score suggests. The two are blended by how much has been played: results carry a weight of played divided by played plus 6, so week 3 is a quarter results and three quarters prior. Home field is 1.5 points, constant. Our spread is the difference in ratings plus home field, and the edge is our spread against the market line.

Four signals move the edge by fixed amounts, and each keeps its own record above. Rest: a gap of 3 days or more between the two teams’ rest is 1.5 points to the rested side. Public: when 65% or more of the spread money is on one side, 1 point goes the other way. Weather: an outdoor game with wind at 20 mph or more, or a 60% or higher chance of rain, is 1 point to the underdog on the market line, and nothing on a pick’em. Quarterback: a team whose starting quarterback is out is docked 4 points, on live weeks only, because an injury status is a fact about now and not about a week that has already been played.

The side is whichever team the adjusted number favours, and every game gets one; a number of exactly zero goes to the home side. Stars are bands on the size of that number: under 1.5 is one star, under 3 is two, under 4.5 is three, under 6 is four, and anything more is five. Three stars and up are the Sharp picks and get a written case. One and two stars are leans, not picks: the side the number tilts to without getting far enough to call, so they sit in the week table marked no play and are counted nowhere, not in the headline, not by stars, not in the backtest.

A pick is frozen the first time it is published with a line before kickoff: the side, the stars, the signals, the case and the line it was made against never change after that. A later run can update the current line, the score and the grade, and nothing else. A game that reaches kickoff without a line gets no row. When a live pick grades, the closing line is recorded beside it: the median spread across books from the last pull before kickoff, and how far the market moved toward our side between the pick and the close. Beating the close is the measure that survives one week’s scores, and it is kept in its own block above. Weeks 1 and 2 were run after the fact against the closing line and are labelled a backtest: a demonstration of the system, not a record.

Nothing here is fitted. The constants are stated so they can be argued with, and they stay where they are until there are enough weeks to fit them honestly. The case on a pick and the lead on a week come from the Sharp Model, off the numbers on this page and nothing else; when the written case does not land, a numbers-only case stands in and says so. This page is a record of a system, kept in public. It is not a recommendation to do anything with it.

Week 3 ratings32 teams, best first. Points better than the average team.
TeamRatingPriorMarginsWeight
SF+5.2+0.8+5.20.25
SEA+4.2+1.7+4.20.25
KC+4.0+1.7+4.00.25
CIN+3.5+1.7+3.50.25
NE+3.4+1.7+3.40.25
BUF+3.1+1.7+3.10.25
MIN+3.10.0+3.10.25
BAL+3.0+2.5+3.00.25
CHI+2.8+0.8+2.80.25
LAR+2.6+2.5+2.60.25
PHI+1.3+1.7+1.30.25
DAL+1.2+0.8+1.20.25
JAX+1.20.0+1.20.25
DET+0.6+1.7+0.60.25
NO+0.5−0.8+0.50.25
LV+0.1−2.5+0.10.25
DEN−0.5+0.8−0.50.25
HOU−0.8+0.8−0.80.25
CAR−0.8−0.8−0.80.25
GB−1.0+0.8−1.00.25
NYJ−1.2−2.5−1.20.25
TB−1.30.0−1.30.25
PIT−1.40.0−1.40.25
NYG−1.8−0.8−1.80.25
IND−2.4−0.8−2.40.25
LAC−2.8+0.8−2.80.25
WAS−2.8−0.8−2.80.25
TEN−3.3−1.7−3.30.25
ARI−4.1−4.3−4.10.25
CLE−4.2−2.5−4.20.25
ATL−4.4−0.8−4.40.25
MIA−7.0−4.3−7.00.25