SharpEdge: Carolina +3.5 Is the Best Number on the Board, and the Biggest Model Gap Is a Pass
The Card
| Game | Play | Line (book) | Tier | One-line why |
|---|---|---|---|---|
| DET @ CAR, Sun 8:20 PM ET | Panthers | CAR +3.5 (DraftKings) | Play | The model says Lions by 1.3, and +3.5 gets us past the 3. |
| DAL @ HOU, Sun 1:00 PM ET | Texans | HOU -2.5 (DraftKings) | Play | The model says Texans by 4.7, and a field-goal win cashes. |
| IND @ WAS, Sun 9:30 AM ET | Over | 47.5 (DraftKings) | Lean | The two defenses allow 29.7 and 29.3 a game. Model total 50.0. |
| PIT @ CLE, Thu Oct 1, 8:15 PM ET | Browns | CLE +2.5 (DraftKings) | Lean | The model says Steelers by 0.2, but +2.5 is short of 3. |
| JAX @ CIN, Sun 1:00 PM ET | Jaguars | JAX +2.5 (DraftKings) | Pass | Biggest gap on the board, and gaps that big are usually the model being wrong. |
| NE @ BUF, Sun 1:00 PM ET | Patriots | NE +6.5 (DraftKings) | Pass | The +7 from the open is gone, and the model's case is mostly last year's team. |
Key numbers drive this card: 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. Both Plays sit on the right side of 3. Bet in units: A fixed bet size so results can be compared. One unit is one unit whether that is $10 or $100. Plays get 1 unit. Leans go on the ledger, so bet them at half a unit or skip them. Spread and total prices aren't in today's DraftKings feed, so the ledger grades those at -110.
Model vs Market
The model line is our own number: team strength built from play-by-play efficiency and scoring margin, plus 1.8 points of home field. The spread is the head start the underdog gets. Seahawks -3 means Seattle has to win by more than 3 for that bet to cash. The total is the over/under on combined points from both teams. Edge is how many points apart the model and DraftKings are.
| Game | DraftKings line | Model line | Edge (points) | Total | Model total | Lean |
|---|---|---|---|---|---|---|
| PIT @ CLE | PIT -2.5 | PIT -0.2 | 2.3 | 38.5 | 40.8 | CLE, Over |
| IND @ WAS | IND -3.5 | IND -1.4 | 2.1 | 47.5 | 50.0 | WAS, Over |
| NE @ BUF | BUF -6.5 | BUF -2.2 | 4.3 | 48.5 | 43.6 | NE, Under |
| NYJ @ CHI | CHI -3.5 | CHI -8.6 | 5.1 | 43.5 | 43.5 | CHI |
| JAX @ CIN | CIN -2.5 | JAX -4.1 | 6.6 | 51.5 | 43.3 | JAX, Under |
| ARI @ NYG | ARI -1.5 | NYG -2.7 | 4.2 | 44.5 | 41.2 | NYG, Under |
| LA @ PHI | LAR -3 | LAR -1.4 | 1.6 | 43.5 | 42.1 | PHI, Under |
| GB @ TB | GB -3.5 | TB -0.6 | 4.1 | 39.5 | 45.9 | TB, Over |
| TEN @ BAL | BAL -11.5 | BAL -8.5 | 3.0 | 43.5 | 43.4 | TEN |
| DAL @ HOU | HOU -2.5 | HOU -4.7 | 2.2 | 48.5 | 46.9 | HOU, Under |
| MIA @ MIN | MIN -10.5 | MIN -5.9 | 4.6 | 38.5 | 40.2 | MIA, Over |
| KC @ LV | KC -4.5 | KC -4.2 | 0.3 | 47.5 | 45.0 | LV, Under |
| DEN @ SF | SF -3 | SF -2.6 | 0.4 | 47.5 | 45.0 | DEN, Under |
| LAC @ SEA | SEA -7 | SEA -6.5 | 0.5 | 42.5 | 40.7 | LAC, Under |
| DET @ CAR | DET -3.5 | DET -1.3 | 2.2 | 50.5 | 54.1 | CAR, Over |
| ATL @ NO | NO -3 | NO -1.8 | 1.2 | 48.5 | 46.0 | ATL, Under |
Where the spreads sit on the key numbers: LAR, SF and NO are right on 3, and SEA is on 7. PIT, IND, CHI, CIN, GB, HOU and DET are a half point off 3, and BUF is a half point off 7. ARI, BAL, MIN and KC are off both.
Now the honest part. We tested the model on 2023-2025, using only what was known at the time, and bet only when it disagreed with the final spread by 2 or more points. It went 51.7% on 437 plays. Break-even is 52.4%. The model does not beat the market on its own. Gaps of 2-3 points hit 54.7%, 3-4.5 hit 52.1%, and 4.5+ hit 47.7%. That puts this week's three biggest gaps, Jacksonville 6.6, Chicago 5.1 and Miami 4.6, in the group where the model is usually wrong. The market prices this week's teams and quarterbacks, like Case Keenum in Chicago. Our profiles are 70% last season. The model rates Cincinnati below average, but this year's Bengals are 8th in net EPA per play, offense and defense combined. EPA is 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. Pass all three sides, and pass the 8.2-point Jacksonville under gap too.
The 2-3 point group is the one that beat break-even, and four games land there. Houston and Carolina are on the right side of 3, so they are the Plays. Cleveland (2.3) only gets +2.5. Its center Elgton Jenkins (concussion) and receiver Tylan Wallace (knee) missed practice and have no game status yet, and Pittsburgh's Rico Dowdle (toe) missed practice too. That makes it a Lean. Washington's 2.1 comes with a catch. The game is at Tottenham Hotspur Stadium in London, but the model still gives Washington 1.8 points of home field. Take that out and the model sits almost on DraftKings' number. The model total doesn't use home field, and it likes the over. Washington's defense is 26th in EPA, Indy's is 31st, and both play fast (66.0 and 62.3 plays a game). Lean over, knowing the backtest covers spreads, not totals.
Line Shop
The only book on the card today is DraftKings, and the multi-book comparison table starts next issue. 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. Because NFL games land on 3 and 7 far more than any other margin, here is what that half point buys. If Detroit wins by exactly 3, Carolina +3.5 wins, +3 pushes (you get your money back) and +2.5 loses. Houston works the other way: at -2.5 a 3-point Texans win cashes, and at -3 it pushes.
Since the open, DraftKings has moved IND -3 to -3.5, CHI -3 to -3.5, LAR -1.5 to -3, NO -2.5 to -3, SEA -6.5 to -7 and BUF -7 to -6.5. That last one took the 7 away from Patriots backers. The Rams move matters most. The model (LAR -1.4) sits near the opening number, but a 1.6-point gap is under our 2-point threshold, and I don't fight a market that moved onto 3.
The Lean
DET @ CAR, Sunday Oct 4, 8:20 PM ET
Bank of America Stadium, outdoors on grass. Both teams have seven days of rest, no weather is listed yet, and nobody is on either injury report. Detroit's offense is 7th in EPA per play and turns it over 0.33 times a game. It has a 50.8% success rate: the share of plays that kept the offense on schedule (roughly 40% of the yards needed on 1st down, 60% on 2nd, all of it on 3rd and 4th). The defense is 29th, and 8.8% of plays against it go for 20+ yards, the worst rate in the league. Carolina throws on 66% of plays and gains 20+ yards on 7.2% of them. It ranks 10th in offensive EPA and 14th on defense, and it takes the ball away 2.0 times a game. Last season Carolina went 10-7 against the spread and beat it by 0.3 points a game. Detroit went 8-9 and beat it by 0.2. This year the Lions are 0-2-1 against it.
The call: Panthers +3.5, Play, 1 unit. The model says Lions by 1.3, and a Lions win by exactly 3 still cashes. What would make me wrong: Detroit sacks the quarterback on 8.7% of dropbacks. If Bryce Young can't stay upright, Jared Goff's offense pulls away.
DAL @ HOU, Sunday Oct 4, 1:00 PM ET
Reliant Stadium on astroturf. The pack doesn't list the roof status. Both teams have seven days of rest, no weather is listed, and nobody is on either injury report. Be clear-eyed here: C.J. Stroud's offense is 24th in EPA with a 40.0% success rate, and Houston is 0-3 against the spread. Part of the model's faith comes from last year's Texans, who went 9-8 against the spread and beat it by 4.5 points a game. Dallas went 7-10 and missed by 2.7. I back Houston anyway because Dallas has the league's worst defense by EPA (0.191 per play) and by success rate allowed (53.7%). Houston's defense allows 37.6%, third-best, and sacks the quarterback on 10.2% of dropbacks. The problem: Dak Prescott's offense is 5th in EPA and gets sacked on only 2.8% of dropbacks. It does gain 20+ yards on just 3.4% of plays, though, so it has to earn every drive.
The call: Texans -2.5, Play, 1 unit. We lay less than 3, and the 2.2-point gap is in the only group that has beaten break-even. What would make me wrong: this year's Houston offense is the real one and Dallas wins a shootout. The total has already climbed from 47.5 to 48.5.
The Ledger
Twelve plays published, ten graded: 2-8-0. We have won 20% of graded bets, when standard -110 odds need 52.4% just to break even, and we are down 6.18 units. The one number that isn't ugly is 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. Ours averages +0.4, so we have been getting decent numbers and losing anyway, on a small sample. Nothing was graded this week, and two entries are still pending with no game attached.
To say it plainly: the model does not beat the market on its own. Its backtest is 51.7% against a 52.4% break-even. What we sell is the number: shopping for it, respecting 3 and 7, and knowing which side of a key number the football supports.
Ledger Entries
| Game | Market | Side | Line | Odds | Book | Tier |
|---|---|---|---|---|---|---|
| DET @ CAR | spread | CAR | +3.5 | -110 | DraftKings | Play |
| DAL @ HOU | spread | HOU | -2.5 | -110 | DraftKings | Play |
| IND @ WAS | total | Over | 47.5 | -110 | DraftKings | Lean |
| PIT @ CLE | spread | CLE | +2.5 | -110 | DraftKings | Lean |
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.
- Units: A fixed bet size so results can be compared. One unit is one unit whether that is $10 or $100.
- 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.
- Success rate: The share of plays that kept the offense on schedule (roughly 40% of the yards needed on 1st down, 60% on 2nd, all of it on 3rd and 4th).
- 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.
- 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.
Sources
DraftKings lines and opening numbers via ESPN. Team profiles, EPA, rest days and against-the-spread records from nflverse play-by-play. Injuries from the official injury reports. Model lines and the 2023-2025 backtest from the SharpEdge model.
21+. Analysis and opinion, not betting advice. If you or someone you know has a gambling problem, call 1-800-GAMBLER.