SharpEdge: Broncos +3 on Monday Night Is the Best Number on the Week 1 Card
First issue, so the honest part comes first. The model behind this letter does not beat the market. Backtested walk-forward from 2023 to 2025 against closing lines, it made 437 plays and covered 51.7% of them against the spread, the head start the underdog gets: Seahawks -3 means Seattle has to win by more than 3 for that bet to cash. Break-even at standard odds is 52.4%. So the model is a flashlight, not a crystal ball. It shows us which side of a number the football supports. The money comes from 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) and from 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). Every line below is DraftKings via ESPN. Totals are the over/under on combined points from both teams.
The Card
| Game | Play | Line (book) | Tier | One-line why |
|---|---|---|---|---|
| DEN @ KC | Broncos +3 | +3 (DraftKings) | Play | Model makes it a pick'em, Denver has the best defense in the game, and we hold the whole key number 3. |
| BUF @ HOU | Texans +1.5 | +1.5 (DraftKings) | Play | Model has Houston favored by 1.0. Last year's No. 1 defense at home, in a 2.5-point gap, the backtest's sweet spot. |
| GB @ MIN | Under 46.5 | 46.5 (DraftKings) | Play | Model total 41.9. Minnesota scored 19.4 and allowed 18.8 a game last year, and nobody here hits big plays. |
| WAS @ PHI | Eagles -5.5 | -5.5 (DraftKings) | Lean | Model says Eagles by 8. Washington's defense ranked 30th. The number stays under 7, but Philly's offense was ordinary. |
| NYJ @ TEN | Over 38.5 | 38.5 (DraftKings) | Lean | Two bottom-five defenses, a 6.2-point model gap, and both teams went 11-6 to the over. The offenses scare us. |
| BAL @ IND | Colts +3.5 | +3.5 (DraftKings) | Pass | Biggest gap on the board at 6.9 points. Gaps that size hit 47.7% in the backtest. That is the model being wrong, not Vegas. |
Sizing: we 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 get half a unit. That is 4 units of exposure this week. If a Play's number moves through the key number before you get there, say Denver goes to +2.5, it drops to a Lean and half a unit. Never chase a number that left.
Model vs Market
| Game | DraftKings line | Model line | Edge (pts) | Total | Model total | Lean |
|---|---|---|---|---|---|---|
| NE @ SEA | SEA -3 | SEA -1.2 | 1.8 | 44.5 | 45.3 | NE / Over |
| SF @ LA | LA -3.5 | LA -6.7 | 3.2 | 48.5 | 46.9 | LA / Under |
| TB @ CIN | CIN -3.5 | TB -0.7 | 4.2 | 50.5 | 47.1 | TB / Under |
| NO @ DET | DET -7 | DET -5.5 | 1.5 | 49.5 | 45.3 | NO / Under |
| NYJ @ TEN | TEN -1.5 | TEN -2.1 | 0.6 | 38.5 | 44.7 | TEN / Over |
| BAL @ IND | BAL -3.5 | IND -3.4 | 6.9 | 47.5 | 47.2 | IND / Under |
| ATL @ PIT | PIT -3.5 | PIT -2.9 | 0.6 | 42.5 | 44.0 | ATL / Over |
| CHI @ CAR | CHI -3 | CHI -2.4 | 0.6 | 47.5 | 44.5 | CAR / Under |
| CLE @ JAX | JAX -8.5 | JAX -8.7 | 0.2 | 40.5 | 42.9 | JAX / Over |
| BUF @ HOU | BUF -1.5 | HOU -1.0 | 2.5 | 44.5 | 44.3 | HOU / Under |
| MIA @ LV | LV -3 | MIA -2.0 | 5.0 | 40.5 | 42.6 | MIA / Over |
| GB @ MIN | MIN -1.5 | GB -0.3 | 1.8 | 46.5 | 41.9 | GB / Under |
| WAS @ PHI | PHI -5.5 | PHI -8.0 | 2.5 | 44.5 | 43.7 | PHI / Under |
| ARI @ LAC | LAC -9.5 | LAC -6.5 | 3.0 | 47.5 | 44.5 | ARI / Under |
| DAL @ NYG | DAL -3 | NYG -1.6 | 4.6 | 48.5 | 49.3 | NYG / Over |
| DEN @ KC | KC -3 | DEN -0.2 | 3.2 | 43.5 | 41.5 | DEN / Under |
How to read it. Each team gets a 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. Ours uses 1.8 points of home field and builds the rating from net 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) and scoring margin. The profiles are 2025 full-season numbers pulled 35% back toward average for the new year, so a quarterback change like Kyler Murray now listed as Minnesota's starter is not in the math.
The four biggest gaps all point at underdogs sitting on or just off 3: Colts +3.5 (6.9), Dolphins +3 (5.0), Giants +3 (4.6), Bucs +3.5 (4.2). Tempting. The backtest says no. Disagreements of 4.5 points or more hit 47.7%, which is losing money. The 3 to 4.5 range did 52.1%, roughly break-even. The only bucket that paid was 2 to 3 points, at 54.7%. That is why the card leans on Houston (2.5) and Philadelphia (2.5), and treats Denver (3.2) as a key-number play more than a model play. Indy is a Pass. Daniel Jones is the listed starter, Baltimore went 6-11 against the spread and 5-9 as a favorite last year, and none of that is why the model likes the Colts by 3.4. The model likes them because of a rating gap the market has already seen and rejected.
New England at Seattle kicks off tonight, Wednesday, 8:20 PM ET. The model has the Patriots 1.8 points better than the number, under our 2-point bar, so no play. If you bet it anyway, +3 is the whole argument: it sits dead on the key number. Read the injury report first. Patriots center Ben Brown and running back TreVeyon Henderson are Out. Seahawks safety Ty Okada is Out, with receiver Tory Horton and safety Nick Emmanwori Questionable. On totals, the model wants the under in 10 of 16 games. We have no totals backtest in the pack, so totals get half the trust and the two on the card are chosen on football, not on the gap.
Line Shop
The pack has one book this week. The only book on the card today is DraftKings, and the multi-book comparison table starts next issue. So here is what you are shopping for. Two of our plays sit on 3. Games land on exactly 3 more than any other margin. If Denver loses by exactly 3, Broncos +3 is a push and Broncos +3.5 is a winner. Same bet, same game, one book paid you and one gave your money back. That half point is worth more than any stat in this letter. The same applies at 7, where Lions -7 in Detroit is a push if they win by a touchdown and an extra point, and -6.5 cashes.
One hint of what shopping catches: the nflverse feed logged Dolphins-Raiders at Raiders -3.5 with a 41.5 total. DraftKings shows -3 and 40.5. Two readings of the same game, a half point apart on the key number. If you like Miami, +3.5 is the number to hunt. If you like Vegas, -3 is the one to keep.
The Lean
DEN @ KC, Monday Sep 14, 8:15 PM ET, Arrowhead, outdoors on grass. No weather data in the pack yet. Both teams are on 7 days rest. Patrick Mahomes against Bo Nix. The ratings say Denver is the better team, 3.4 to 1.4, and after home field the model lands on Broncos by 0.2. DraftKings says Chiefs by 3. The football sides with Denver. Its defense allowed -0.088 EPA per play, sixth best, and a 36.7% success rate, the lowest figure in the whole pack. Success rate is 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). Denver's defense also took the quarterback down on 10.3% of dropbacks while its own line gave up sacks on just 3.6%. Kansas City's offense was 11th by EPA per play at 0.035 and scored 20.6 a game. Last year the Chiefs went 6-11 against the spread, 4-10 as a favorite, and 4-13 to the under. The call: Broncos +3, Play, 1 unit. What makes me wrong: Mahomes at home on a Monday night, and a Denver offense that scored 22.6 a game and needs the defense to carry it. If the number moves to +2.5, it is a half-unit Lean at best.
BUF @ HOU, Sunday Sep 13, 1:00 PM ET. The pack lists the venue as Reliant Stadium on astroturf and does not list the roof. Josh Allen against C.J. Stroud. Houston's defense ranked first in the league at -0.131 EPA per play with a 38.3% success rate allowed and -0.186 on passes. It got a sack on 8.0% of dropbacks and took the ball away 1.59 times a game while the offense gave it away 0.65. Buffalo's offense is the real thing: third at 0.136 EPA per play, 27.6 points a game, and a rush EPA of 0.11 that should test a Houston front that was fine against the run but not great. Buffalo's pass rate, the share of plays the offense throws instead of runs, was 51%, so expect James Cook's group to work. The market gap is 2.5 points, the bucket that hit 54.7%. Buffalo went 6-9 against the spread as a favorite last year. Houston covered by an average of 4.5 points a game. The call: Texans +1.5, Play, 1 unit. One wrinkle on price: DraftKings has the Texans moneyline at -102 (a bet on who wins, no spread. -166 means risk 166 to win 100; +142 means risk 100 to win 142). The +1.5 only saves you on a one-point loss, so near-even money on the win is a fair swap. What makes me wrong: Houston's offense ranked 21st at -0.013 EPA per play and scored 22.4 a game. Allen does not need many possessions.
GB @ MIN, Sunday Sep 13, 4:25 PM ET, U.S. Bank Stadium, dome, division game. Jordan Love against Kyler Murray. Model total 41.9 against a DraftKings 46.5. Minnesota's offense was 28th at -0.115 EPA per play, scored 19.4 a game, gave up sacks on 11.0% of dropbacks, and turned it over 1.59 times a game. Its defense was fourth at -0.099 and allowed an explosive play rate, the share of plays that gained 20 or more yards, of just 3.7%, the best mark on the slate. Green Bay's offense was fourth at 0.115 with a 0.229 pass EPA, and that is the threat. But the Packers scored 22.3 a game, allowed 20.2, and gave up explosives on only 4.6% of snaps. These two ran 55.5 and 58.4 plays a game. Few snaps, few big plays, one shaky quarterback situation. The call: Under 46.5, Play, 1 unit. What makes me wrong: Murray is not in the numbers, Love shredding a Vikings secondary in a dome, and last year's over-under splits were a coin flip (Minnesota 8-9, Green Bay 9-8).
The Ledger
Record: 0-0-0. Units: 0. Average closing line value: none yet. The ledger starts today, and every Play and Lean in this issue goes on it at the DraftKings number shown. We grade against the spread and track 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. We will never quote a win rate that is not in this table, and we repeat the backtest every time this section runs: the model alone went 51.7% against a 52.4% break-even. The edge, if there is one, is the number, not the prediction.
Ledger Entries
| Game | Market | Side | Line | Odds | Book | Tier |
|---|---|---|---|---|---|---|
| DEN @ KC | spread | DEN | +3 | -110 | DraftKings | Play |
| BUF @ HOU | spread | HOU | +1.5 | -110 | DraftKings | Play |
| GB @ MIN | total | Under | 46.5 | -110 | DraftKings | Play |
| WAS @ PHI | spread | PHI | -5.5 | -110 | DraftKings | Lean |
| NYJ @ TEN | total | Over | 38.5 | -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.
- moneyline: A bet on who wins, no spread. -166 means risk 166 to win 100; +142 means risk 100 to win 142.
- 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.
- 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.
- 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.
- 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.
- 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
Lines, totals, and moneylines: DraftKings via ESPN. Team profiles, power ratings, and the second Dolphins-Raiders reading: nflverse play-by-play. Injuries: official NFL injury reports.
21+. Analysis and opinion, not betting advice. If you or someone you know has a gambling problem, call 1-800-GAMBLER.