SharpEdge

Line Moves

This table has every game, with the DraftKings open from Oct 5 next to this morning's number. It shows both the spread (The head start the underdog gets. Seahawks -3 means Seattle has to win by more than 3 for that bet to cash.) and the total (The over/under on combined points from both teams.).

The big one is Baltimore at Atlanta. It opened BAL -3.5 and it's ATL -3 now. Lamar Jackson is Out with an ankle and didn't practice, and the pack projects Tyler Huntley at quarterback. The line went through 3 on Baltimore's side, through even, and landed on 3 for Atlanta. That's why it matters so much. Key numbers: NFL games land on 3 and 7 far more than any other margin, so a half point on either side of those numbers is worth real money. Jacksonville went through the other key number, -4.5 to -7.5, after Philadelphia ruled out Saquon Barkley, DeVonta Smith and Marquise Brown. Five other games moved around a key number:

• BUF @ LA went through 3.
• NYG @ WAS moved off 3 to 3.5.
• HOU @ TEN moved onto 7.
• IND @ PIT and SF @ SEA moved onto 3.

Chicago at Green Bay is the odd move. The pack projects Tyson Bagent at quarterback, with Caleb Williams Questionable (hamstring). The side moved a point toward Green Bay and the total still went up 3. Nothing in the pack explains that, so I'm staying away.

As of this morning, no line has moved against our pending plays. If one had, that would be closing line value (CLV: whether the line you bet was better than where the market closed. The best long-term sign a bettor knows what they are doing, because closing lines are the sharpest numbers there are.) going the wrong way, and it would count. Tennessee moved our way: we took TEN +7.5 and DraftKings is at +7, so we hold the half point through the 7. Houston and Washington both took money after the open (6.5 to 7, and 3 to 3.5), but our tickets are at 7.5 and 3.5, so neither move cost us. CLE +2.5 is right where we bet it. None of this is final until the 1:00 PM ET kickoffs.

The Card

First, a word on the model. It runs on power ratings (A number for how good a team is, in points, against an average team on a neutral field. Subtract two ratings, add home field, and you have a model spread.), plus 1.8 points for home field. Tested against 2023-2025 games it hadn't seen, it went 51.7% against the spread on 437 plays. That's short of break-even, measured by hit rate (The percentage of bets that won. At standard -110 odds you need 52.4% just to break even.). Only the 2-to-3-point disagreements cleared that bar, at 54.7%. The 3-4.5 band hit 52.1%, and 4.5-plus hit 47.7%. That's why Miami +7 isn't here even though the model disagrees with the market by 6 points. The model tells me where to look, and the number decides the bet.

TEN +7.5 (Play). The model has Houston by 4.9. Even at today's +7, that leaves the model 2.1 points on Tennessee's side, inside the only band that has beaten the market. Houston is 0-4 against the spread this year, and Tank Dell is Out. The Texans' tackles are a mess on the injury report:

• Blake Fisher and Aireontae Ersery: Questionable, illness.
• Braden Smith: Questionable, foot.
• Trent Brown: no game status after a rest day.

Wrong if: Tennessee's 13.5 points a game can't keep up, even in a 37.5 total.

NYG +3.5 (Play). The model has Washington by 0.4, a 3.1-point gap. Gaps that size hit 52.1% in testing, below break-even, so this bet is about the number: +3.5 still wins on a 3-point loss. Washington is thin. Stefon Diggs and Charles Omenihu are Out, Terry McLaurin is Questionable (hamstring), and three safeties are Out: Nick Cross, Percy Butler and Tyler Owens. Wrong if: the Giants can't attack that secondary. They gain 20 or more yards on 3.3% of plays and give up sacks on 8.9% of dropbacks.

CLE +2.5 (Lean). The model has Cleveland by 0.2, a 2.7-point gap in the band that wins. The Browns also come in on 10 days of rest to the Jets' 7. The pack has Deshaun Watson starting, with Dillon Gabriel Out. It's only a Lean because +2.5 loses if the Jets win by exactly 3. Wrong if: the middle of the defensive line caves with three defensive tackles Out: Mason Graham, Mike Hall Jr. and Kalia Davis.

New: IND moneyline +124 (Lean). The Colts rate 0.3 and the Steelers -0.7. Add home field and the model has Pittsburgh by 0.8, close to a coin flip. DraftKings has PIT -3 and pays +124 on the Indianapolis moneyline (A bet on who wins, no spread. -166 means risk 166 to win 100; +142 means risk 100 to win 142.). That's a 2.2-point disagreement, inside the band that wins, and the price got this good because the line moved from 2.5 onto 3. Pittsburgh is without Michael Pittman, Jamel Dean and DeShon Elliott.

Why the moneyline instead of the spread: the pack shows the +124 price but not the price on IND +3. If your book has +3 at a normal price, take that, because a 3-point Steelers win pushes instead of losing. The model was only tested on spreads, not moneylines, which is one more reason this is a Lean. Wrong if: Pittsburgh's 10 days of rest to Indy's 7 shows up on the field, or the Colts' passing game plays like its -0.211 EPA per play (Expected Points Added: how many points a play moved the team toward scoring compared with an average play from the same spot. Positive is good on offense, negative is good on defense.).

Where the card stands. All four kick off Sunday at 1:00 PM ET:

The ledger's pending list also has four empty rows (weeks 3, 4, 4 and 5) with no game, side or line. They aren't bets and there's nothing to grade. They are counted in the 21 published plays, and I'd rather point that out than have you wonder.

Bet every ticket flat at 1 unit, Leans included, because that's how the ledger keeps score. Units: A fixed bet size so results can be compared. One unit is one unit whether that is $10 or $100. That's four tickets and four units at risk. On a 4-10 ledger, nobody should be raising bet sizes.

Line Shop

Line shopping (Checking the same bet at several sportsbooks and taking the best number. Getting +3.5 instead of +3 is free equity, and it costs nothing but a minute.) is where real, repeatable edge comes from. This week the pack only has DraftKings, so I can't tell you which book has the best number. If you have other accounts, here's what to look for:

The Lean

BAL @ ATL, Sunday 8:20 PM ET. The quarterback is the reason for the 6.5-point swing. Lamar Jackson is Out, and so are center Cam Jurgens (concussion) and Trey Hendrickson (finger). Zay Flowers is Questionable with a foot. Baltimore's offense ranks 3rd at 0.186 EPA per play, and the model can't tell how much of that leaves with Jackson. It still has Baltimore by 0.3, because its ratings come from past games, not tonight's depth chart. The backtest says to ignore exactly this kind of disagreement. Atlanta isn't a safe pick either. Its passing game sits at -0.191 EPA per play, it turns the ball over 2.0 times a game, and it went 2-5 against the spread as a favorite last season. The market already did the work here. No bet on the side, and none on the total, which dropped from 46.5 to 43.5.

PHI @ JAX, Sunday 9:30 AM ET, Tottenham Hotspur Stadium. If you're reading this late, it's already underway. Philadelphia is without Barkley, Smith, Brown, center Drew Kendall and safety Marcus Epps, and Jonathan Greenard is Questionable. The Eagles were already struggling: -0.107 EPA per play (29th), sacked on 9.4% of dropbacks, 0-4 against the spread. Jacksonville has the top power rating on the board at 4.6. It allows 13.0 points a game and forces 2.0 turnovers a game. The model says JAX -6.2, which leans Philadelphia by 1.3. That's under our 2-point cutoff, and the ratings don't account for who's out. On the total, the model's 41.8 against the market's 41.5 is no edge at all. The market was right to push this through 7. No bet.

The Ledger

Nothing was graded this week, and there was nothing on the ledger from Thursday's TB @ DAL. A 28.6% hit rate against 52.4% break-even is bad, and -6.36 units says so in real money. The one number pointing the right way is average CLV at +0.18: on average, we've bet a better number than where the line closed. Fourteen graded plays is a small sample for both numbers.

Ledger Entries

Plain English

• Spread: The head start the underdog gets. Seahawks -3 means Seattle has to win by more than 3 for that bet to cash.
• Total: The over/under on combined points from both teams.
• Key numbers: NFL games land on 3 and 7 far more than any other margin, so a half point on either side of those numbers is worth real money.
• Closing line value: CLV: whether the line you bet was better than where the market closed. The best long-term sign a bettor knows what they are doing, because closing lines are the sharpest numbers there are.
• Power rating: A number for how good a team is, in points, against an average team on a neutral field. Subtract two ratings, add home field, and you have a model spread.
• Hit rate: The percentage of bets that won. At standard -110 odds you need 52.4% just to break even.
• Moneyline: A bet on who wins, no spread. -166 means risk 166 to win 100; +142 means risk 100 to win 142.
• EPA per play: Expected Points Added: how many points a play moved the team toward scoring compared with an average play from the same spot. Positive is good on offense, negative is good on defense.
• Units: A fixed bet size so results can be compared. One unit is one unit whether that is $10 or $100.
• Line shopping: Checking the same bet at several sportsbooks and taking the best number. Getting +3.5 instead of +3 is free equity, and it costs nothing but a minute.

Sources

Lines are from DraftKings via ESPN (the Oct 5 open and this morning's numbers). Team profiles come from nflverse play-by-play, 2026 weeks 1-4 blended with last season. Injuries are from the official NFL injury reports. The record is from the SharpEdge ledger.

21+. If you or someone you know has a gambling problem, call 1-800-GAMBLER.