SharpEdge

The Card

Every bet here is a spread: the head start the underdog gets. Seahawks -3 means Seattle has to win by more than 3 for that bet to cash. Both Plays sit a half point past a key number. 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. A Play (I would bet it) is 1 unit. A Lean (I like the side, not the number) is half a unit. A Pass (stay away) is nothing. That's 2.5 units on the week. If either Play loses its half point before you bet, it drops to a Lean.

Model vs Market

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. Our model uses 1.8 points for home field. Total: The over/under on combined points from both teams. Edge is the gap between DraftKings and the model, in points, toward the team named.

Where the lines sit against the key numbers: four sit right on one (JAX -7, CIN -7, CHI -3, LAR -3), eight are a half point off, and DAL -8.5, MIN -1.5 and DET -5.5 are clear of both.

Here's the honest part. We ran the model walk-forward over 2023-2025 against closing spreads. It bet 437 games where it disagreed with the market by 2 or more points and won 51.7%. Break-even is 52.4%, so on its own the model does not beat the market. Edges of 2-3 points won 54.7%, edges of 3-4.5 won 52.1%, and edges of 4.5 or more won 47.7%. When the model and the market are far apart, the model is usually the one that's wrong.

That's why I'm passing on the two biggest gaps, Tampa and Miami. The ratings are built from scoring margin and EPA per play. 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. Nothing in that recipe knows Baker Mayfield missed practice. Then there's Miami. The profiles rank the Dolphins 32nd of 32 in net EPA (offense and defense combined) and Cincinnati 12th. The ratings get pulled toward average because four weeks is a small sample, and with a gap this size I trust the profiles and the market more than the rating.

The other passes, quickly. Arizona, the Chargers and Baltimore all sit in the band that won 52.1%. Green Bay's 2.0 is in the good band, but the line is right on 3, and Chicago ranks 5th in net EPA to Green Bay's 26th. For my New Hampshire regulars: Las Vegas ranks 8th in net EPA to New England's 18th, and the Pats turn it over 2.5 times a game. I'm not laying 3.5 with that.

Totals need a warning label. The model pulls every offense 35% toward the league average of 22.5 points, so it calls under on the biggest totals (BUF @ LA: 54.5, model 47.9) and over on the smallest (HOU @ TEN: 37.5, model 42.2). That's the design talking, not the football. The backtest also covers spreads only, so no totals on the card.

Line Shop

DraftKings is the only book on the card today. 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. If Washington wins by exactly 3, Giants +3 is a push (money back) and Giants +3.5 cashes. If Houston wins by exactly 7, Titans +7 pushes and Titans +7.5 cashes.

For now, here's what I can show: where DraftKings opened on Oct 5, where it sits now, and the line in the nflverse game data. You can't bet the nflverse line. It's a second read on where the market is.

Both Plays have moved our way since the open. If your book has Titans +7 or Giants +3, that's a worse bet, not the same one.

The Lean

HOU @ TEN: Titans +7.5 (DraftKings), Play

Sunday Oct 11, 1:00 PM ET, Nissan Stadium, outdoors on grass. No weather in the pack yet. It's a division game, both teams are on 7 days of rest, and neither team has anyone on the injury report. Cam Ward hosts C.J. Stroud.

EPA per play: Houston -0.026 (19th), Tennessee -0.085 (25th). The defenses rank 21st and 24th. 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). Houston is at 41.6%, Tennessee 41.2%. That's a coin flip, not a touchdown. Pass rate: The share of plays the offense throws instead of runs. It tells you which teams feed receivers and which feed backs. Houston throws 64% of the time into a Tennessee pass defense giving up 0.203 EPA per throw. Tennessee throws 58%. Explosive play rate: The share of plays that gained 20 or more yards. Explosives are how bad teams win and good teams cover. Houston is at 7.6%, Tennessee 5.1%. Houston's defense sacks the quarterback on 9.0% of pass plays, and Tennessee allows sacks on 4.8%. Neither team gives it away much: 0.5 turnovers a game for Houston, 0.75 for Tennessee.

Houston is 0-4 against the spread (ATS) this year. Tennessee went 7-10 ATS last year and missed the number by an average of 3.8 points a game.

The call: Titans +7.5, 1 unit. What would make me wrong: Tennessee scores 13.5 a game, so a couple of Houston big plays can settle a game with a 37.5 total. Houston also beat the spread by an average of 4.5 points a game last year.

NYG @ WAS: Giants +3.5 (DraftKings), Play

Sunday Oct 11, 1:00 PM ET, Northwest Stadium, outdoors on grass. No weather in the pack yet. It's a division game, both teams are on 7 days of rest, and neither team has anyone on the injury report. Jameis Winston visits Jayden Daniels.

On paper, this is the same team twice. EPA per play: Giants -0.047 (22nd), Washington -0.045 (21st). The defenses rank 16th and 20th. Success rate: 40.0% Giants, 38.6% Washington. Pass rate: 52% to 55%. Washington gives up 29.8 points a game, the Giants 19.8. Washington's real edge is up front: the Giants allow sacks on 8.9% of pass plays and get them on just 2.2%, while Washington allows 3.5%. The Giants' explosive play rate is 3.3% to Washington's 4.2%. Washington turns it over 0.5 times a game to the Giants' 1.25, but the Giants take it away more, 1.75 to 1.25.

Washington went 7-10 ATS last year and missed the number by an average of 2.9 points a game. The Giants beat it by 1.4 and are 3-1 ATS this year. DraftKings opened Washington -3, and the model says Washington by 0.4.

The call: Giants +3.5, 1 unit. What would make me wrong: the Giants never get to Daniels and never hit a big play. If it goes back to +3, it's a Lean.

TB @ DAL: Bucs +8.5 (DraftKings), Pass

Thursday Oct 8, 8:15 PM ET, AT&T Stadium. The pack lists no roof status or weather. Both teams are on 4 days of rest. Baker Mayfield (thumb) did not practice and has no game status, and the pack projects Jalon Daniels to start. Tampa's kicker Chase McLaughlin and defenders Antoine Winfield Jr., SirVocea Dennis and Benjamin Morrison also missed practice. For Dallas, Cobie Durant, DeMarvion Overshown and Drew Shelton all missed practice with hamstring injuries. Nobody on either list has a game status yet.

Dallas has the 6th offense by EPA per play (0.166) and the 32nd defense (0.201). That defense gives up 0.427 EPA per pass and an 8.3% explosive play rate. Tampa has the 31st offense (-0.154) and the 6th defense (-0.059). But Tampa's passing game is at -0.261 EPA per play, its success rate is 38.7%, it takes sacks on 11.3% of pass plays, and it turns it over 2.25 times a game. Pass rates are 63% for Dallas and 60% for Tampa. Dallas went 2-8 ATS as a favorite last year. Tampa is 0-3-1 this year. DraftKings opened Dallas -10.

The call: Pass. Gaps this big lost money in the backtest, and the rating can't see the quarterback situation. At +8.5 Tampa is already past the 7, so there's no key number to grab. What would make me wrong: that Dallas pass defense hands Tampa the cover.

The Ledger

Record: 4-10-0 on 14 graded plays, out of 18 published. Units: -6.36. 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. Our average CLV is +0.18.

At -110 you need to win 52.4% to break even, and we're at 28.6%. Fourteen bets is a small sample, and I'm not going to hide behind that. The positive CLV is the one number pointing the right way, and it's the only thing this letter is selling, because the model alone does not beat the market. Nothing was graded this week. Four entries in the ledger file are marked pending but have no game details, so they aren't in the record and I won't guess what they were. This week's two Plays and one Lean go on the ledger below.

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.
• 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.
• 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.
• Total: The over/under on combined points from both teams.
• 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).
• Pass rate: The share of plays the offense throws instead of runs. It tells you which teams feed receivers and which feed backs.
• Explosive play rate: The share of plays that gained 20 or more yards. Explosives are how bad teams win and good teams cover.
• 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 via ESPN (lines, openers, totals); nflverse play-by-play (team profiles, ATS and over/under records) and nflverse game lines; official NFL injury reports; this week's depth charts (projected starting quarterbacks); DBTech45 power ratings, backtest and ledger.

21+. Analysis and opinion, not betting advice. If you or someone you know has a gambling problem, call 1-800-GAMBLER.