Every game read against all three sources. The number in brass is the blend; the meter beside it shows how far apart the sources actually are — which is usually where the edge lives. Every game carries the two-edge test: a statistical edge alone is half an edge. The card is below.
Both columns. Full bet. Jacksonville is not a league-level team and they’re laying 8.5 — the classic public overlay, and what the public is buying is a 30-takeaway season that was the first of its kind in NFL history and will not repeat. Line moved from −7.0 to −8.5, so Cleveland got cheaper. Caveat: this is the biggest number on the card, so it takes the most distortion from our conversions (Commandment 7).
Both columns. Full bet. Every source says Las Vegas shouldn’t lay more than a field goal; the line moved half a point toward Vegas so Miami got cheaper; neither team is a league-level favorite; and we take the hook through 3. The cleanest pick on the card under the filter.
Both columns. Full bet. Dallas is the most publicly-bet team in football and went 2–8 ATS as a favorite, 1–4 as a road favorite — a repeating structural overlay, not a model artifact. We buy low (+2.5 → +3.0), get the key number, and take the smallest spread on the card, which is the least distorted by our own machinery.
Statistical column only. Reduced size. Both independents have Green Bay at pick ’em or better and Sharp’s summer market had them favored in Minneapolis. But the likeliest reason the number moved is the Parsons ACL — information, not bias. I sold this as the strongest play on the board and that was wrong; the market case is inference, not evidence.
Forced, not endorsed. The contest demands five and the slate produced three. Commandment 8 says let the bets find you — a contest that requires a quota is a machine for violating it. Whatever goes in slot five is a tax on entry, and it should be labeled that way rather than dressed up as a read.
Every game page on this site now carries the two-edge test. A model disagreeing with a line is column one only — that is half an edge, and half an edge is a losing bet. Column two asks a different question: why is this number wrong, and who put it there? Public overlay on a famous team, a stale line, an overreaction to last season, a move that already happened. If column two is blank, the verdict is stay away — not a smaller bet.
Stay away is the most common verdict here, and that is the point. Eleven of fifteen this week. A process that finds a bet in most games is a broken process.
Total under and Maye passing under. Maye is #36 of 36 in net Y/A against Cover 6; Seattle ran it on 51% of snaps in the Super Bowl against a 25% baseline — and won.
STAY AWAY — no market case, derivatives onlyTotal over. Neither defense can reach the quarterback — Tampa only pressures by blitzing the league's fastest release, Cincinnati rushes four more than anyone into a top-3 line.
STAY AWAY — sources converge, play the totalSaints team total under and Detroit first half. New Orleans posted a −93 first-quarter margin, the worst by any team this century, and is dead last in explosive plays.
STAY AWAY — Sharp’s line matched Circa exactlyFirst half under. The worst first-half offense in football against the second-worst — Sharp's own data says only the Jets were worse than Tennessee at stalling out early. Both teams score in garbage time.
STAY AWAY — removed. Opposite signs + buying highColts first quarter and Ravens team total under. Lamar was dead last of 36 quarterbacks in red-zone EPA; Indianapolis's one defensive strength is the red zone. Baltimore drives and kicks.
STAY AWAY — 0.28 pts, market agrees with itselfUnder. Two of the slowest, heaviest, least explosive offenses in football — Rodgers threw 10+ yards downfield less than any quarterback in the league, and both teams sit bottom-six in three-receiver sets.
STAY AWAY — thin, and no market storyBears first half and the run game. Chicago is #1 in the NFL in yards before contact; Carolina is #31 in yards before contact allowed. Bryce Young threw 20+ yards on 1.0% of first-quarter attempts — last of 643 QB seasons.
STAY AWAY — number and football cancelSide corrected off the card. An earlier version had the Rams as a value favorite — that applied their home-field grade to a game in Australia. Removing it lands our number on Circa's exactly.
STAY AWAY — neutral site, dead fairCleveland +8.5 — on the contest card. Jacksonville is the first team in NFL history with 30+ takeaways after a season of ten or fewer, went 3–0 in field-goal games and faced seven backup QBs. Their home field grades 0.25, second-lowest in the league.
DOG — CLE +8.5 · both columns clearUnder, and Bills second-half live. Allen took a sack on 32.2% of third-down pressures (#36) into the #2 pressure defense in football — which gets there without blitzing on 37% of dropbacks, above the league average with a blitz.
STAY AWAY — everyone lands on pick ’emCame off the card. Three points have already moved toward Arizona since summer, so we'd be buying the back end of the move. McBride over is the play — the all-time TE reception record holder against a defense that plays dime on 23% of snaps.
STAY AWAY — removed. 3 pts already moved our wayMiami +3.5 — on the contest card. Every source says Las Vegas shouldn't lay more than a field goal. The Raiders had the worst run blocking of any team since 2013 (0.50 yards before contact) and the #32 first-half offense.
DOG — MIA +3.5 · both columns clearNo play — and this is the discipline page. Our sheet said Philadelphia by twelve. 31% of that gap is the Eagles' secondary, which both Hoppen and Sharp independently name as the hole on the roster. We threw our own number out.
STAY AWAY — our model broke, market is rightGreen Bay +1.5 — the strongest signal on the board. Sharp's schedule sheets list Minnesota as a dog to Green Bay at home; Circa gives us a point and a half. Love is #1 in clean-pocket EPA and #2 when blitzed, against the #1 blitz defense — now missing Greenard, Allen and Hargrave.
DOG — GB +1.5 · statistical column onlyGiants +3.0 — on the contest card. New York was 1–7 in one-score games (#31) against the hardest schedule in football. Dallas allowed 511 points, the most in the 66-year history of the franchise, and went 1–4 ATS as a road favorite.
DOG — NYG +3.0 · both columns clearCorrected 09-09, after Circa posted. Sharp's "stated lines" turned out to match Circa exactly on both games where he had one — they're market lookahead numbers, not his own projections. His win totals are often the posted Vegas totals too, and Hoppen's sim takes betting markets as a composite input. That leaves the Master Sheet as the only fully independent valuation here — not the most accurate, just the only one able to disagree with the market for a non-market reason.
So the roles now: the market line is the number to beat, the Master Sheet is the independent read whose divergence is the candidate edge, Hoppen is a semi-independent cross-check, and Sharp's prose is what tells you whether a divergence is real. Sharp's value was never the number — it's the why.
And a standing caution: Circa takes the highest limits in the market. A three-point disagreement is better evidence that our inputs are stale than that we found three points of edge.
The disagreement is the analysis — but read what kind of disagreement it is. On Patriots–Seahawks the three sources are 2.8 points apart while all still favoring Seattle: that's a magnitude split, which usually means one source is mispricing something and a side is worth betting. On Jets–Titans they're 3.7 points apart with opposite signs — they can't agree on which team is better at all. That's not a mispricing, it's genuine uncertainty, and the right response is to stay small or off the side entirely and take the derivative.
When they converge — Bucs–Bengals and Saints–Lions — the side is priced and the money moves to totals, team totals and props.
Prefer a stated line to a derived one. Sharp's chapters list projected lines for many games. Where one exists, use it. Where one doesn't, a win-total gap has to be converted to a spread — and that conversion is the weakest link in this whole system.
Calibrated 09-09: against Sharp's three stated lines, a one-win gap in season totals is worth about 1.6 points of head-to-head spread, not the 2.75 originally used here. The reason is that win totals bake in schedule strength as well as team quality — Baltimore's 11.5 comes with the sixth-easiest slate, Indianapolis's 7.5 with one of the hardest, so their four-win gap wildly overstates the gap between the teams. Sharp's own line on that game is −3.5; his win totals would say −9.75. Every number on this site has been recomputed at the corrected rate.
Each sheet carries the counter-case in red. That section exists so nobody at the table gets caught flat-footed; if a partner raises one of those points, it's a fair objection, not a reach. Full write-ups with result lines for grading live in the vault under Betting/Football/Games/.
Talking points from three sources. Numbers on the left, the argument on the right.
NO Only 0.4 pts — our −2.6 against Circa’s −3.0 is inside the noise band.
NO Sharp’s summer number and Circa agree. No movement, no public-side story, nothing to exploit.
Maye's net Y/A against Cover 6 last season — dead last among every qualifying quarterback at 2.1. Against everything else he's elite: #1 versus non-Cover-6 zone, #7 versus man.
Seattle's Cover 6 rate in the Super Bowl, against a 25% season baseline. They won. Macdonald and Durde are both still in the building — this is a replication question, not a projection.
Games at −0.05 EPA/att or worse: zero in seventeen regular-season games, then four in four playoff games once the league copied Buffalo. Completion rate fell 72% to 58%, Y/A 8.9 to 6.9.
His EPA/att vs Cover 6 — 38% success, 5.9 Y/A. Versus everything else he sits at +0.21 to +0.23. He held the ball 3.15 seconds against it and took a 9.5% sack rate.
New England against playoff teams. They were 12–1 against everyone else. The record is real; the competition wasn't.
Opponent win rate the Patriots faced — the easiest schedule any team has played in 830 team-seasons since 2000.
Seattle against playoff teams — best mark in the NFL — on the eighth-hardest schedule in football.
New England faced the worst schedule of offenses and the worst schedule of quarterbacks in the league. The QB list: Stidham, Ewers, Brady Cook, Huntley, Dart, Flacco, Gabriel, Ward, Rattler. Their #4 scoring defense has not been tested.
Seattle's receiver starters versus their bench — the most top-heavy room in football. Smith-Njigba was roughly 44% of their receiving yards.
Passer rating allowed by Christian Gonzalez, who travels with JSN. New England gave up 6.3 Y/A with him on the field and 7.9 without. The Master Sheet has their secondary tied for the best DB grade in the league.
Shaheed's regular-season touchdowns as a Seahawk. Kupp is at 2.9 catches a game. If Gonzalez erases JSN, there is no Plan B.
Seattle's running back room — the largest single-unit decline in the NFL. Walker to Kansas City, Charbonnet's February ACL surgery, and Price grading out as the lowest-rated first-round back since 2000.
Fleury is the third play-caller in three seasons and a first-time caller. Sharp calls losing Kubiak more significant than losing any player.
Seattle's 12-personnel EPA under Kubiak — that's the scheme that just walked out the door, and Fleury is installing a tight-end-heavy offense through a TE room Hoppen ranks 27th.
Darnold's turnovers — most in the NFL, costing 100.2 EPA. Seattle was 31st in giveaways and 32nd in fumbles lost.
A.J. Brown's YPRR on 420 career routes with his QB under center, and Maye ranks 10th in under-center rate. Brown pulls safeties out of the box — that is the actual structural antidote to quarters coverage, and it's brand new information the Super Bowl tape doesn't have.
New England's yards per carry over the final nine games — best in the league, on a 41% success rate. They're already 4th in 21 personnel. If Seattle sits in two-high, the counter is to run at light boxes.
Both offenses were elite early: NE #1 in first-half points (17.4/gm, +135 halftime margin), Seattle #1 in first-quarter margin. This is how the under dies before halftime.
Milton Williams played twelve. With him New England was #1 in nearly every run-defense metric — one back carry over 12 yards in 151 attempts. Without him, league average. The whole run defense is one man.
| Market line | Read |
|---|---|
| SEA −1 or better | Best spot. Below every source; Hoppen and Sharp both see 2.5+ points of value. |
| SEA −1.5 to −2.5 | Around the blend. Value on the two deeper sources, roughly fair on the sheet. Solid lean. |
| SEA −3 to −3.5 | At the top of the source range. Thin — Hoppen and Sharp still support it, the sheet doesn't. Total is the better bet here. |
| SEA −4 to −5 | Past all three. Pass the side. |
| SEA −6 or worse | Champion premium paid. New England +points becomes live, with the Master Sheet's QB grade as the supporting argument — but note that's the thinnest of the three sources carrying the case. |
Cleanest expression. Doesn't require picking a winner and survives either team simply being sloppy in Week 1. Vrabel is also 4th-lowest in the league going for it on short 4th downs.
Isolates the Cover 6 edge without needing the spread or total to cooperate.
Seattle's backs go 41% success in quarters 1–3 and 29% in the fourth, against eight-man boxes on 56% of Q4 carries. If they lead late without Walker or Charbonnet, drives stall and the clock stops.
Moneyline note: the Master Sheet's fair price is SEA −108, Hoppen's win gap implies closer to −190. Wide range, so don't treat either as the fair price. Practically: if Seattle's ML sits near −120 or better while the spread is −4 or −5, the ML holds the position more efficiently than laying the points.
The three sources land in a two-point band around Cincinnati −3.8. Little edge on the side, real edge on the total — where the structural read is unusually clean.
NO 0.3 pts. All three sources converge on the side — convergence is the opposite of edge.
NO No line movement story. The market and our sources are looking at the same game the same way.
Tampa Bay blitzes at the 5th-highest rate in the NFL (31%, 185 blitzes) because it's the only way they generate pressure — their non-blitz pressure rate is 29%, 24th. Burrow is blitzed at the lowest rate of any quarterback over the last four years and gets rid of it inside 2.5 seconds more than anyone alive. Tampa's only pressure plan is the one thing Burrow eats.
Tampa's pass defense with Jamel Dean on the field versus off, in EPA/att, over two seasons — 89 passer rating allowed with him, 112 without. Dean is in Pittsburgh now. The secondary is permanently stuck in the "off" column.
Cincinnati's Master Sheet receiver grade — the maximum on the scale, tied with Dallas for the best WR room in the league. Chase and Higgins put up 184 catches, 2,258 yards and 19 TD with Burrow missing most of the year.
Tampa's sack rate when they blitz (24th), producing 11 sacks on 185 blitzes. And opposing QBs posted the 5th-best EPA/dropback in the league when Tampa did pressure them. The pressure doesn't convert even when it arrives.
Cincinnati rushed four on 79% of dropbacks — most in the NFL — and blitzed on 17%, least in the NFL. That four-man rush got pressure inside 2.5 seconds on 15% of dropbacks, 31st. Now it meets Tampa's offensive line, which grades #4 in Sharp's units and tied 3rd-best on the Master Sheet with Wirfs allowing 2 sacks all year.
What Cincinnati allowed when their front seven failed to generate pressure — 29th. Sharp's point is that the corners were fine; the pass rush was the problem. Mayfield should see clean pockets.
Tampa's red-zone TD rate allowed: 69.8%, dead last in football and the 5th-worst mark in five years. Goal-to-go 88%, 28th. Cincinnati's problem is turning drives into field goals — Tampa is the defense that fixes that for them.
Fourth-quarter points Cincinnati allowed last year — 2nd-most in the NFL, with a −31 fourth-quarter margin and zero fourth-quarter comebacks.
Tampa Bay last season — the worst against-the-spread record in Sharp's entire 2025 table. The market has been overpaying for this team for a full year.
Tampa's turnover margin (6th) and penalty margin (2nd), plus +3.6 fumble recoveries over expectation (7th) and the 3rd-best net penalty EPA. Three separate non-repeating gifts stacked into an 8–9 season.
Tampa wins that came by one score — 75%, 4th-highest. Their win list is backup quarterbacks and losing teams: Penix, Tyrod, Rattler, Brissett, Young, Mac Jones.
Cincinnati's EPA lost to turnovers — 3rd-worst in the league — plus 3–5 in one-score games and trailing in 16 of 17 games. Their luck ran the opposite way. Both teams regressing to neutral moves the number toward Cincinnati.
Cincinnati has ranked 27th or worse in pass-block win rate five straight seasons (#30, #30, #27, #32, #28) — the only team in the league that can say that — and this offseason they spent no top-100 pick and signed no free-agent lineman. Same five starters. Master Sheet grades the line 7.4, bottom third.
Cincinnati's record with Burrow versus without. He's missed significant time in three of six seasons, behind that line. This is the single biggest variable in the whole handicap and it's binary.
Cincinnati's defense last year: dead last in yards per play (6.2), 300 points allowed through nine games (2nd-most since the 1970 merger), and eight straight games allowing 27+ points and 350+ yards — the first time that's ever happened in NFL history. Al Golden was retained anyway.
Tampa's offensive health rank last year. Their five most important non-quarterbacks shared the field in three games all season. Mayfield now gets a McVay-tree OC who coached him with the 2022 Rams, in a contract year — Hribar's headline is that this passing game bounces back.
| Market line | Read — blend is CIN −3.8 |
|---|---|
| CIN −2.5 or less | Value on Cincinnati. Below every source. The best price to lay it. |
| CIN −3 to −4 | On the blend. Fair. Thin at best — play the total instead. |
| CIN −4.5 to −5.5 | On the Master Sheet, past Hoppen and Sharp. No edge. Pass. |
| CIN −6 or more | Tampa plus the points. Past every source, and squarely in the overpriced-favorite spot both Hoppen and Sharp warn about. Tampa is also rarely blown out — smallest range of outcomes in the league per Hoppen. |
Both defenses have a structural reason they can't reach the quarterback. Both team notes pre-registered this independently — Hoppen's Bengals read says "elite passing O + bad D = shootouts," and his Bucs read says games versus elite passing offenses lean over.
Max-grade WR room against a secondary that lost the corner whose on/off split was 11th to 32nd. The Tampa note already flags "opponent WR1 props over" as a season-long angle.
Neither team can come back. Cincinnati went 0–9 when trailing entering the fourth with zero comebacks all year; Tampa went 0–8 when trailing at half and their largest deficit overcome all season was seven points.
This game has the two most disruptive nose tackles in football on opposite sidelines. Dexter Lawrence has 108 QB pressures from nose tackle since 2022 — first by a mile. Vita Vea is second, at 32. Cincinnati traded the 10th overall pick for Lawrence, the first top-10 pick the franchise has ever surrendered for a player.
Sharp's head-coach ranks for Zac Taylor and Todd Bowles — Taylor fell from 18th with no panel vote higher than 20th; Bowles dropped from 14th and is openly on the hot seat. Master Sheet grades them 5.0 and 5.5, both bottom-tier. Two desperate staffs raises variance in exactly the late-game decisions that decide one-score games.
Sharp's book projects this exact game, so we have a real line to check against rather than a derived one — and the Master Sheet lands within two-tenths of it.
NO 0.6 pts. Noise.
NO Sharp’s stated line matched Circa exactly at DET −7.0. A market that agrees with itself is not a vulnerable one.
New Orleans' first-quarter point margin last season — the worst by any team this century. 29 first-quarter points scored (#32) against 122 allowed (#31), averaging 1.7 first-quarter points per game.
On the opening drive of games the Saints ranked 32nd in three-and-out rate, 32nd in EPA/play, 31st in yards per drive, 31st in score rate and 30th in points per drive. Kellen Moore's scripting was the single worst in football.
Games in which New Orleans trailed at some point — the most in the NFL. They have not led early against anybody.
Saints' rate of 20+ yard plays — dead last in the league, with 10+ yard plays 23rd and success rate 26th. Now hold that thought for the next section.
Detroit's defensive weakness is specific: Sheppard's man-heavy scheme "sprang explosive leaks" once McNeill and the secondary were compromised, and they allowed 7.3 Y/A (24th). New Orleans is dead last in the NFL at generating explosive plays. The Saints are uniquely badly equipped to attack the one thing Detroit is bad at.
Detroit played base personnel — three linebackers — on 63% of snaps, most in the NFL, with nickel at just 34% (dead last) and zero dime snaps all year. New Orleans runs 11 personnel on 66.7% of plays, 4th-most. If Detroit stays stubborn in base, that's a linebacker on a slot receiver. Watch the first series.
Detroit's pass-block win rate last season, down from 12th — and Goff was still sacked a career-high 38 times (7th most) and hit 117 times, 2nd most in football. That was behind a better line than the one he has now.
New Orleans' sack total (10th) on only the 22nd-ranked pressure rate — one of just three teams with two double-digit sack rushers. Cameron Jordan had 10.5 in his 15th season and Chase Young a career-high 10.0 at an 18.1% pressure rate.
Only Penei Sewell remains from Detroit's 2024 starting five — and he's moving to left tackle for the first time in his career. First-round rookie Blake Miller takes over at right tackle; center Cade Mays has a wrist injury costing significant time. Hoppen names the offensive line as Detroit's X-factor.
Goff's passer rating from a clean pocket — 2nd in the NFL, with 83.7% clean-pocket accuracy (2nd) and 0.17 EPA/att with no pressure (5th). The entire game reduces to one question: can New Orleans get to him?
New Orleans' defense in EPA/play — and Hoppen and Sharp independently landed on the exact same figure. The Saints went from 30th to 9th in yards allowed per game (299.8) in Staley's first year, with #7 success rate and #6 yards/play. This defense is legitimately good and the market may not know it yet.
Detroit's 2025 luck stack: 32nd in fumble luck (−7.0 recoveries versus expectation), 30th in field-goal luck, 30th in net penalty EPA, plus 3–5 in one-score games after going 7–2 the year before. They had the best point differential in the NFC North (+68) and finished last.
Detroit's health rank last season — 31st overall and dead last on defense. Sharp lists the injury regression as the strongest over-side argument in either source. It's genuinely hard to be that unlucky twice.
New Orleans' red-zone TD rate — dead last — and their goal-to-go rate was also 44.4% against a 74.4% league average, the worst mark since 2017. Their field-goal percentage was 71.4%, also 32nd. These are the most regressable numbers in football and they only move one direction.
Demario Davis signed with the Jets — 143 tackles, team tackle leader for eight straight seasons, PFF's highest-graded Saint. Sharp centers his departure as the thing most likely to unwind that top-10 defense.
| Market line | Read — blend is DET −6.4 |
|---|---|
| DET −5 or less | Value on Detroit. Below the blend and below Sharp's own projection. |
| DET −6 to −7 | On the blend and on Sharp's stated line. No edge. Play a derivative instead. |
| DET −7.5 or more | New Orleans plus the points. Past every source. The Saints went 9–8 ATS as near-permanent dogs (average line +4.1) and their defense travels better than their offense does. |
The cleanest play. An offense 28th in points per drive and dead last in explosive plays, on the road, against a defense whose only real flaw is giving up explosives. New Orleans went 5–12 to the over last year — unders cashed in 12 of 17.
New Orleans is 32nd in first-quarter points, 32nd in first-drive EPA and posted the worst first-quarter margin of any team this century, and they trailed at some point in all seventeen games.
A line that was 31st in pass-block win rate with four of five starters replaced, against the one Saints unit that's genuinely good. Goff was hit 117 times behind a better group.
This is the easiest schedule in the NFL against the second-easiest. Both teams' win totals are schedule-inflated and both sources say so about both teams — Hoppen calls Detroit's slate "Charmin soft" and says the sim likes them for who they play, not how good they are.
Both kicking games were disasters last year — Detroit 30th in field-goal luck (Bates at 79.4%, 4-of-9 from 50+) and New Orleans dead last at 71.4%. Two teams whose kickers cost them games, in a dome where there's no excuse. Points that went missing last season and shouldn't again.
The sources don't just disagree on the number here — they disagree on which team is better. That's a different problem from a mispriced favorite, and it changes what you bet.
NO The sources have opposite signs — they can’t agree which team is better. That is uncertainty, and Commandment 3 says uncertainty is a stay-away, not a bet.
NO Fails Commandment 9. The number went from Jets +3.0 in summer to +1.5 now. The Jets got a point and a half more expensive — buying this is buying high.
Tennessee had 56% of its first-half drives reach third down — 2nd-worst in the NFL — and Sharp's note names the only team that was worse: the Jets. The two most stalled-out first-half offenses in football, playing each other in Week 1.
The Jets' cumulative halftime deficit last season — no team since 2000 has been worse. They led at half in 2 of 17 games and trailed at some point in all seventeen.
Tennessee's own first-half point margin (#29), trailing by 5.4 a game at the break on 7.5 first-half points per game. They also trailed at some point in every single game.
The Jets' third quarter was just as dead — 28th in EPA/play, 28th in success rate, 30th in points. Only in the fourth, with games already decided, did they post a #4 success rate. Their scoring is garbage-time scoring, which is exactly why the first half is the cleaner bet than the full game.
Geno Smith's rank in EPA/play among qualified quarterbacks last season — worse than Justin Fields and Tyrod Taylor. He also posted a league-high 10.9% sack rate and was sacked on 28% of pressured dropbacks; only Kyler Murray was worse among non-rookies.
Cam Ward's rank in the same metric — dead last. 5.9 Y/A (33rd), 59.8% completions (31st), 55 sacks (tied most in the NFL), sacked on 26% of pressures (36th), and 7 lost fumbles, most in football.
Neither line helps. The Jets allowed pressure inside 2.5 seconds on 24.5% of dropbacks (27th); Tennessee's line graded 30th in Sharp's units and gave up 56 sacks. Two quarterbacks who collapse under pressure, behind two lines that allow it.
The Jets recorded zero interceptions last season — the first team to do that since interceptions became an official statistic in 1933. Four total takeaways, against a prior record low of seven. Turnover margin −19, worst in the league.
They responded by trading for Minkah Fitzpatrick, and added Demario Davis, T'Vondre Sweat, Onyemata, Ossai and Enagbare, plus No. 2 overall pick David Bailey — whose 33.6% college pressure rate was the highest of any player over the last two seasons.
Cam Ward's interceptions and lost fumbles last year. The most turnover-starved defense in NFL history meets the quarterback who gave it away more than anyone. If regression is going to announce itself, this is the spot.
Tennessee's 3–14 came against the hardest schedule in the NFL — .549 opponent win rate, and a league-high ten games each against winning teams and playoff teams. Ward faced nine top-ten pass defenses as a rookie. The Jets' 3–14 came against a normal slate. This is why Hoppen and Sharp favor Tennessee and the Master Sheet doesn't — the sheet grades current talent, they're pricing context.
Sharp's panel ranked Aaron Glenn the 32nd head coach in the NFL — the first time in six years the last-place coach wasn't a rookie. Hoppen calls it "the most pre-fired coaching staff ever." The Master Sheet grades the Jets 4.0 on coaching, tied for lowest in the league. All three sources agree the Jets have the worst staff in football — the sheet just doesn't weight it enough to change its answer.
Tennessee went over in 11 of 17 last season despite that offense — because the defense allowed 8.2 Y/A, dead last. Two bad offenses is not automatically an under when both defenses are also bad. This is the case against a full-game under and the reason to focus on the first half.
Tennessee added No. 4 overall pick Carnell Tate and Wan'Dale Robinson, with Daboll calling plays — and the Jets' pass defense finished last in EPA per attempt allowed at 7.5 Y/A after trading Sauce Gardner. Both team notes independently flag the same angle from opposite ends.
| Market line | Read — blend TEN −1.6, low confidence |
|---|---|
| NYJ favored | Only the Master Sheet agrees. Small Tennessee value, but two of three sources against you is not where to press. |
| Pick'em to TEN −1.5 | Under the blend. Best available side price if you must have one. |
| TEN −2 to −3 | On Sharp's stated line and Hoppen's number, past the blend. Fair to slightly rich. Pass. |
| TEN −4 or more | Jets plus the points. Past every source including both that favor Tennessee. |
The two worst first-half offenses in football, by Sharp's own head-to-head note. Both teams' scoring lives in the fourth quarter after games are decided. Add two new play-callers in Week 1 — Reich for the Jets, Daboll for Tennessee — and early sloppiness is the base case.
Seven interceptions and a league-high seven lost fumbles, against a defense that recorded zero picks all last season and spent the offseason specifically fixing it. Near-mechanical regression meeting the league's most generous quarterback.
When your sources can't agree which team is better, that's not an edge, it's uncertainty. Size accordingly.
These two teams swapped defensive linemen this offseason and now open against each other. T'Vondre Sweat went to the Jets; Jermaine Johnson II went to Tennessee, where Sharp says he "immediately becomes the team's best edge presence." Both face their old team in Week 1.
Tennessee has zero rest-disadvantage games in 2026 — the only team in the NFL — while the Jets have five games with a rest advantage, also league-best. Neither applies in Week 1, but both are worth knowing for the rest of the season.
A referendum on one Achilles tendon. Daniel Jones went down December 7; Week 1 sits at the optimistic end of a standard recovery, and almost everything Indianapolis does well runs through whether defenses respect him.
NO 0.28 pts. Nothing.
NO Sharp’s stated line matched Circa exactly at −3.5. Worth noting for Commandment 5: Baltimore is one of the few teams that would pass the big-favorite test — but there is no edge here to act on.
Lamar Jackson's red-zone EPA per attempt last season: −0.53, dead last among all qualifying quarterbacks — the 35th-ranked QB was at −0.24 and the league average is +0.02. One year after he led the NFL at +0.59.
The cause: Baltimore's red-zone pressure rate on Jackson's 51 red-zone dropbacks, against a 35% league average and a next-closest team at 47%. Only Bryce Young's 2023 has been higher since 2014. Their red-zone offense went #1 in 2024 to #28 in 2025.
And Indianapolis's one genuine defensive strength is the red zone — 11th in TD rate allowed (54%) and 10th in takeaways per drive. Sharp's phrase is that "what saved them was the back end of drives." Baltimore's worst area meets Indy's best one.
Baltimore's offensive line by positional spending, with the receivers 4th-cheapest. They were 26th in pass-blocking efficiency and gave up 45 sacks. Then Linderbaum — the only lineman his backs gained positive EPA behind — left for a record guard contract, replaced by a career backup with 10 starts.
Baltimore's pressure rate on non-blitz dropbacks from Week 3 on — dead last, and no team since at least 2016 has finished a season under 20%. They were 28th in overall pressure rate with 30 sacks and 31st in sack rate.
Indianapolis's offensive line — tied 5th-best on the Master Sheet, 11th in Sharp's units, 7th in run-block win rate. All three sources agree it's the clearest edge either team has, and it's pointed straight at the worst four-man rush in a decade.
But Indianapolis's defense is catastrophic before the red zone: 38% of opponent drives reached it, dead last in football, with punts forced on 27% of drives (31st) and 34 yards allowed per drive (30th) — all against the 6th-easiest schedule of offenses.
Steichen's first quarter is the best in football. With a healthy Jones through Week 10 the Colts scored on 52% of opening-quarter drives and reached the red zone on 60%, both #1 against averages of 34% and 33%. Since 2023 his first-quarter offense ranks top-eight in every category across six different starting quarterbacks. Baltimore's defensive coordinator is calling his first game.
Jones tore it in Week 14. Sharp's timeline: 6-8 months per Schefter against a 9-12 month standard — putting Week 1 at the optimistic end. Hoppen calls betting on Achilles-return efficiency "a pretty bad bet."
Jonathan Taylor's EPA per rush through Week 10 versus after — from best in the NFL to 43rd of 46. Same back, same line, same scheme. Only the quarterback changed. Success rate 48% to 36%, 6.0 YPC to 3.3, explosive rate 12% to 3%.
Loaded boxes on Taylor's first-half runs before and after. Sharp's phrase: the run game "lost its oxygen" when defenses stopped respecting the quarterback. Hoppen reaches the same place from formation data — Taylor is far better under center, and a recovering Achilles pushes Jones toward shotgun.
Indianapolis's record in Jones's starts versus everyone else's last season. They also went 7-0 against bottom-16 efficiency teams and 1-9 against the top 16 — and Baltimore is a top-16 team.
Sharp calls this the single best in-season betting trigger in either source: Jackson's career record when he gets 7+ designed runs is 35-9 (80%), versus 44-27 (62%) with six or fewer. Last three years it's 7-1 against 26-16. In the playoffs, 3-1 versus 0-4. His 2025 rushing fell to a career-low 26.9 yards a game.
Jackson's EPA per dropback before and after the Week 4 hamstring. Through four weeks he was #1 in YPA, #3 in EPA, #1 in TD rate and #1 in explosive-gain rate. He was on the injury report in 12 of his last 13 games. Both sources treat 2025 as injury noise, not decline.
Baltimore's first new head coach since 2008, an offensive coordinator who has never called a play in an NFL game, and a defensive coordinator whose last unit finished 24th in points. Hoppen's pre-registered read: "the market may over-trust the brand early." Week 1 is exactly that spot.
One-score records for Baltimore and Indianapolis. Baltimore also went 0-2 in field-goal games and 3-6 at home after winning 23 of their previous 27 there. Both are positive-regression candidates in close games — it cancels out, but neither record reflects the roster.
| Market line | Read — blend BAL −3.1 |
|---|---|
| BAL −2 or less | Value on Baltimore. Below every source including Sharp's stated line. |
| BAL −3 to −3.5 | On the blend and on Sharp's number. Fair. Play a derivative. |
| BAL −5 or more | Colts plus the points. Past all three, and this is a road favorite with a first-time play-caller in Week 1. |
The best scripted first quarter in football — 52% score rate, 60% red-zone rate, both #1, and top-eight across six different quarterbacks since 2023 — against a defensive coordinator calling his first NFL game.
Worst red-zone offense in the league by a distance, behind the third-cheapest line, against the one thing Indianapolis's defense actually does. This profile produces field goals, not touchdowns.
Only if the injury report says Jones is limited or the offense opens shotgun-heavy. Both sources reach this independently, and Taylor's collapse from #1 to 43rd was entirely quarterback-driven. Check Friday.
Live trigger: count Lamar's designed runs in the first half. Seven or more and his career win rate is 80%; six or fewer and it's 62%. No market prices this, and it's visible by the second quarter.
A collision of two opposite second-half collapses: Atlanta's offense dies at halftime, Pittsburgh's defense dies in the fourth quarter. One of those trends has to break, and the total is where that gets paid.
NO 1.1 pts — real but thin, and it rests on the model rather than the football.
NO Circa sits inside Sharp’s summer band. No move, no overlay, no story about who is on the other side.
Pittsburgh's fourth-quarter score rate allowed per drive was 60.4% — 32nd in the league, and 826th out of 830 team-seasons since 2000. They were also #32 in fourth-quarter drives allowed to reach the red zone (40%), #31 in yards per drive and #30 in punts forced. One of the worst closing defenses of the century.
Atlanta's second-half point margin (#28), including being outscored by 54 in the third quarter alone while averaging 2.88 third-quarter points. Their first half was +32 (#12). Same team, two different seasons, every week.
Bijan Robinson's EPA per rush, first half to second half, out of 49 qualifying backs. He led the NFL in first-half rushing yards (976) at 6.1 YPC. After the break: 4.0 YPC, 5.6% explosive, #47 of 49. Sharp: "It was the largest shift in the NFL" — and the concepts, personnel, down usage and game state all stayed the same.
Games Atlanta led at some point — 82% — on the way to 8–9. Meanwhile Pittsburgh was 3rd in the league in games led at halftime (11) but 31st in losses while leading at half. Neither of these teams can finish.
Share of Rodgers's throws that travelled 10+ yards downfield — the lowest in the NFL against a 33.1% average. He also had the league's quickest release (2.52 sec), its shortest depth of target (6.0 air yards) and a league-high 31.7% of throws at or behind the line. Career-low 3,322 yards and no 300-yard game all season, playoffs included.
Both teams' 11-personnel usage rank — Pittsburgh 38.7%, Atlanta 45.1%. Atlanta runs 12 personnel on 38.2% of snaps, 2nd-most in the league; Pittsburgh runs 13 personnel at the 2nd-highest rate. Two heavy, run-first, clock-killing offenses.
Points per drive: Atlanta 25th, Pittsburgh 16th. Atlanta's third-down conversion rate ranked 30th; Pittsburgh's drives after a kickoff or punt reached the red zone just 25% of the time (#24).
The vault pre-registered this shape from both sides. Pittsburgh's note: "elite defensive front + veteran QB = lots of one-score games — look at unders and small spreads." Atlanta's: "road dogs have some cover equity in ugly, low-scoring games."
Master Sheet defensive-line grades — a 6.8-point gap, the largest single-unit gap in any game we've done. Pittsburgh's front grades 2nd in the league; Atlanta's grades dead last. Corroborated on both sides: Hoppen has Pittsburgh's interior D-line 2nd and edge 3rd in his depth index, and Sharp ranks Atlanta's front seven 31st.
Atlanta got 57 sacks (2nd, franchise record) on only the 15th-ranked pressure rate — Sharp's point being it wasn't repeatable. Then Floyd, Ebiketie, Onyemata and Elliss all left, Jalon Walker tore his ACL, and James Pearce (10.5 rookie sacks) faces a suspension with availability "at best uncertain."
Pittsburgh allowed pressure on 25% of dropbacks — best in the NFL — with the 3rd-best pass-block win rate. The league's worst pass rush against the league's best protection.
Hoppen's stated way to beat Pittsburgh is pressure — he flags their tackle depth at 27th and says pressure-heavy opponents are the matchup. Atlanta is the one team in football that cannot do that. And it matters less than usual anyway: Rodgers was pressured on 26% of dropbacks, lowest of 36 quarterbacks, largely because of his own release.
Sharp's own word for Pittsburgh's 2025. Turnover margin +12 (#4) worth +83.05 net EPA (#2), 12 opponent fumbles recovered (#2), +5.1 net fumble recoveries over expectation (#2), and 7–3 in one-score games with 70% of their wins by one score. They were 9–2 against non-playoff teams and 1–5 against playoff teams. And Tomlin, the man associated with sustaining that, is gone.
Atlanta's fumble luck ran the exact opposite way — 19 fumbles with 10 lost against opponents' 23 with only 7 lost — plus 0–2 in overtime and 3–4 in field-goal-margin games. The 2nd-luckiest fumble team in the league is hosting the 31st. Both regress toward each other.
Pittsburgh fielded the most expensive defense in the NFL ($175M against $85M on offense) and it finished 14th in points per drive. They allowed 259 passing yards a game, 2nd-most in football, and a 54% completion rate on throws of 10+ yards, third-worst. Watt is off a down year with Heyward and Ramsey both past 30.
Sharp calls this the strongest over-case in either source: Atlanta collapsed after halftime in every measurable way, and Stefanski's entire philosophy is second-half control. Two-time Coach of the Year, brought Bill Callahan in to coach the line. He also fed Nick Chubb 25 touches a game in Cleveland — and Bijan's share was only 68.8% of backfield opportunities, with Allgeier now gone.
| Market line | Read — blend PIT −3.7 |
|---|---|
| PIT −2.5 or less | Value on Pittsburgh. Below every source. The trench gap is real and it's at home. |
| PIT −3 to −4.5 | On the blend and inside Sharp's own road-line band for Atlanta. Fair — play the total. |
| PIT −5.5 or more | Atlanta plus the points. Past all three, against a team whose entire 2025 was fumble and one-score luck, and Atlanta is a volume dog-cover team — no road number above +7.5 all year. |
Two heavy, run-first offenses in the bottom six of three-receiver usage, one of them quarterbacked by the least downfield passer in the NFL who did not throw for 300 yards in a single game last season. Atlanta is 25th in points per drive and 30th on third down.
The regression case is one-sided: Pittsburgh was 2nd in fumble luck, 2nd in turnover EPA and 7–3 in one-score games; Atlanta was 31st in fumble luck and 0–2 in overtime. Take the number wherever it's above a field goal.
The genuinely chaotic part. If Atlanta leads at half — they led in 14 of 17 games — they meet the worst closing defense of the last 25 years. But their own offense is 28th in second-half margin. Whichever collapse arrives first decides this, and the live market won't price either.
Where the Master Sheet dissents is instructive: it has Atlanta better on offense (29.55 to 27.7) and better on coaching, with Pittsburgh winning entirely on defense (+8.1) off that one enormous D-line gap. That's an unusually clean statement of what this game actually is — Atlanta has the better skill talent, Pittsburgh has the better front.
Atlanta's quarterback is unresolved. Penix ranked 25th of 33 in EPA/play, Tua 27th, and among 36 qualifiers Tua had the tied-2nd-highest turnover-worthy play rate. Hoppen's read is blunt: start Penix if healthy, and "Tua starting = fade harder." That interacts badly with Pittsburgh's fumble-recovery history.
The two luckiest teams in football from 2025, playing each other. Both records are fiction — the question is only whose fiction unwinds faster, and whether the trenches settle it first.
NO 0.78 pts on Carolina — and the football runs hard the other way. #1 in yards before contact against #31 in yards before contact allowed.
NO The half-point drift toward Chicago is too small to read. Nothing identifies a wrong side.
Chicago's turnover EPA — #1 in the league and the second-highest by any team since 2013. A +22 turnover margin, 33 takeaways and 11 giveaways (all #1), plus +5.5 fumble recoveries over expectation, also #1. Sharp's phrase: "the NFL's version of Houdini."
Of the top-seven turnover-EPA teams since 2013, every one cratered the following year — the 2024 Bills went +112.9 to +6.0, and over half went outright negative. And Chicago's takeaway producers left: Byard (team-high 7 INTs), Gardner-Johnson and Edmunds are all gone.
Carolina's net point margin entering the fourth quarter across the season — and they won eight games anyway. One of only three teams in 471 this century to manage that. Add 7–3 in one-score games, 6–2 in field-goal games, and seven of their eight wins by one score. Expected record: 6.5–10.5.
Carolina's defensive penalties all season — fewest since 2016 — good for #1 in defensive penalty EPA (+14.57) and #4 in net penalty EPA. Another input that does not repeat. Both of these teams are worse than their records, and both regress. For the side, that largely cancels.
Chicago leads the NFL in yards before contact per rush (2.01). Carolina ranks 31st in yards before contact allowed. That is about as direct a collision as matchup analysis produces. Bears backs also went 46.1% success (#2) with 43% of runs gaining 5+ yards, first in football.
The same story in protection: Chicago was #1 in pass-block win rate (sacks 68 → 24, sack rate 10.8% → 4.1%), and Carolina was #31 in pressure rate with the NFL's worst QB-hit total (54). Carolina paid Jaelan Phillips $120M to fix it, but Nic Scourton tore his ACL.
Carolina's Master Sheet offensive-line grade — already regraded down from 8.2 on 09-08 for Ekwonu's patellar tendon and Moton's NFI, both out 4+ weeks. Your sheet is the only source carrying that injury; it's baked into the −2.05 above.
Master Sheet coaching grades. Ben Johnson is tied for second-highest in the league, and Sharp documents the fingerprints — early-down play action 24% → 44% (#30 → #2), motion 47% → 61%, and #4 in fourth-down decision EPA.
Bryce Young threw 20+ air yards on 1.0% of his first-quarter attempts — last among every qualifying QB season over twenty years. His first-quarter average air yards was 4.8, 44th of 45. Sharp's read is that the staff "neutered" him deliberately: "control him like a robot to prevent errors."
Carolina's first-half points per game (#30), with 4.24 in the second quarter (#29) and a first-half point margin of −90 on the season. They held a halftime lead in only five games, average margin −5.0.
Chicago's first-down EPA per play — 2nd in the NFL, up from 25th — with the league's highest rate of 10+ air-yard throws on early downs. Ben Johnson scripts; Carolina does not.
Young's overall standing: 32nd in EPA per attempt, 30th in success rate, 35th in YPA — and those were career highs. He failed to clear 200 passing yards in ten games.
Hoppen's verdict on Chicago's defense without the turnovers: "if turnover luck doesn't repeat, this is a bottom-5 defense." Sharp's numbers agree — #26 in total defensive efficiency, #27 in early-down success, #29 in pressure rate, #31 pass-rush win rate, and 4.8 YPC allowed to backs (#29).
Caleb Williams' off-target rate — worst among 33 qualifiers — with a 58.1% completion rate (40th of 45) and bottom-seven accuracy two years running. And Carolina's secondary is legitimately good: Horn with 5 INTs, Mike Jackson with an NFL-high 19 pass breakups, Moehrig at 5.9 yards per target, 6th-fewest explosive plays allowed.
Your own sheet has Carolina's defense rated higher than Chicago's — 23.4 to 20.6, with edges at both defensive line (8.0 to 6.6) and secondary (8.4 to 7.0). That's the entire reason it prices this at −2.05 instead of −3, and it lines up with what both other sources say about the Bears' defense.
Sharp's rest data looks like a Bears argument — 42-35-5 ATS since 2020 without a rest deficit, and 11-1-1 ATS in 2025 — and Week 1 gives nobody a deficit. But that 11-1-1 was produced by the same historic turnover luck that's regressing. Citing it here double-counts the luck.
| Market line | Read — ours CHI −2.2, Circa −3.0 |
|---|---|
| CHI −3.0 (current) | Pass. 0.8 points of model value on Carolina, offset by a clean two-way trench mismatch running the other way. Genuinely close to fair. |
| CHI −4.5 or more | Carolina plus the points. Past every independent source, against a team whose defense is bottom-5 once the turnovers normalize. |
| CHI −1.5 or less | Chicago. Below our own number, and the run-game matchup is the most lopsided thing on this card. |
Chicago is 2nd in first-down EPA with the league's best scripted early-down passing; Carolina is 30th in first-half scoring with a −90 first-half margin and a quarterback who is statistically the most conservative first-quarter passer of the last twenty years.
First in the NFL in yards before contact against 31st in yards before contact allowed, behind the league's best run-blocking line, against a front missing Scourton.
The number says Carolina by eight-tenths of a point. The football says Chicago. When those cancel, there's no bet.
The Rams are surrendering a home game to play this in Australia — which is the whole story of the number. Strip their home field out and Circa's −4.0 is almost exactly right.
NO 0.03 pts once the neutral site is handled. The tightest number on the board.
NO Not applicable — there is no disagreement to explain.
The field-goal rate San Francisco's opponents converted last season — the worst opponent FG rate any team has faced since 2002, with 6.54 kicks missed over expectation against them, by far the most in football. Add +8.58 net field goals over expected (#2) and 5–1 in one-score games, and you have a schedule of gifts that does not repeat.
San Francisco's penalty margin was the third-best in the league on the second-fewest flags — after finishing +32 the year before. A swing that large in one season is noise wearing a jersey.
The Rams' own kicking ran the opposite way — #32 in field-goal EPA (−10.92), #28 in FG rate, 1–4 in field-goal games and 0–2 in overtime. Mevis then went 12-of-13 after Week 10 and 6-of-6 in the playoffs. Two teams whose kicking luck is about to trade places.
The one real matchup edge: the Rams were #1 in EPA per rush, #1 in back success rate and #1 in first-down rate on the ground; San Francisco was #30 in rush success rate allowed at 4.5 YPC to backs.
San Francisco was #30 in pressure rate and #32 in sacks (20 — third-fewest in franchise history). The Rams allowed 23 sacks, tied for fewest in the league, with all five starters back. The pass-rush collision that would decide this game doesn't exist — SF is the wrong team to make Stafford uncomfortable, and Stafford grades 23rd–25th under pressure.
Hoppen calls the Rams' offensive line "the weakness" and the X-factor; Sharp grades it #5 in the league — fewest sacks allowed, #4 pressure rate. That's the flattest disagreement between the two sources on this slate, and it's about the unit the game turns on.
Stafford threw 28 consecutive touchdown passes without an interception, an NFL record, and became the first quarterback ever with four TDs and no picks in three straight games — at 37. His 48.1% end-zone conversion against a 37.4% league rate is the regression flag: 49 of 55 comparable QB seasons since 1990 threw fewer touchdowns the year after.
Our number and Circa's agree to within three-hundredths of a point once the neutral site is handled. There is nothing here.
Two teams at opposite extremes of field-goal luck. If either kicker's early misses move a live number, that's mean reversion being mispriced as form.
#1 EPA/rush against #30 rush-success-allowed is the only genuine mismatch, and it's the way a −4 becomes a −10.
The biggest number-edge on the slate, in the game with the biggest unknown. Both of those things are true at once, and they mostly cancel.
YES 4.1 pts, the largest on the slate. Both independent sources land four points below the market.
YES Commandment 5, run in reverse. Jacksonville is nowhere near a league-level power and they are laying 8.5 — the textbook public overlay. What the public is buying is last year’s 30-takeaway surge, the most non-repeating input in football. And the line moved −7.0 → −8.5, so Cleveland got cheaper (Commandment 9).
Jacksonville's takeaways. They are the first team in NFL history to record 30+ takeaways after a season of ten or fewer, and points off turnovers went from 25 to 110 — first in the league.
Their record in field-goal games (the league's only perfect mark) and overtime (best in football), plus 6–3 in one-score games, +3.5 fumble recoveries over expected, #4 in field-goal luck, four return touchdowns and 9–1 against losing teams. Health ranked 7th.
Quarterbacks Jacksonville faced who were backups or rookies — all seven in the second half of the season — on the #5 easiest schedule of opposing offenses. The 2025 record was assembled out of exactly the inputs that don't repeat.
ESPN's schedule pain index rates Jacksonville #32 in 2026 — the worst mark since 2002: 35 days between home games, no back-to-back home games all season, seven of the final ten on the road. It starts now.
Cleveland's backs averaged 0.38 yards before contact — dead last, and the worst of 448 team-seasons since 2012, with a huge gap to second-worst. They run into a Jacksonville defense that was #1 in rushing yards allowed and the third team since 2000 to give up no individual 75-yard rushing game.
On first down Cleveland was 32nd in EPA per pass, 32nd in pass success, 32nd in yards per attempt, 32nd in first downs per attempt and 32nd in passer rating — which cascaded into 32nd in average distance to go on third down. Their receivers scored four touchdowns all season.
Nobody in either source names a Cleveland Week 1 starter. The room is Watson, Sanders and Gabriel. Watson is #59 of 59 over five years in EPA per play and #59 in sacks per pressure; Sanders drew an NFL-high 51% pressure rate and caused 29% of his own pressures (#44 of 45). This is very likely what moved the line a point and a half.
Cleveland's line is a rookie left tackle plus three imports, with only one returning starter — after eight different combinations played 50+ snaps last year. Jim Schwartz also resigned after being passed over for the head job; Sharp calls that a bigger loss than Myles Garrett.
Cleveland's defense was #1 in football on a per-drive basis — #2 in total efficiency, #2 against the pass, #4 against the run, #1 in yards allowed per drive (24.0), #2 in EPA per play — on below-average spending, with 53 sacks (3rd, a franchise record) and 117 tackles for loss (#1). The entire starting secondary returns.
Cleveland's game totals are the lowest band in Sharp's entire book, clustered at 39.5–41.5. In a 40-point game, 8.5 points is an enormous share of the expected margin — big numbers are structurally harder to cover in low-total games.
And Jacksonville's offense isn't the counterweight the number implies — 18th in yards per play with Sharp's #29-ranked running back room, and Lawrence lost 50.8 EPA to drops, the most in the NFL (that part regresses in his favor).
| Market line | Read — ours JAX −4.4, Sharp's look −7.0 |
|---|---|
| JAX −8.5 (current) | Cleveland plus the points — on the card, but see the caveat. Four points clear of our number, 1.5 clear of Sharp's, in the lowest-total game in the book. |
| JAX −6.5 or less | At Sharp's number and closing on ours. Edge mostly gone. |
Biggest raw edge on the board, and Sharp's own line agrees the market is too high. But this is the pick most exposed to our conversion bias — both our numbers use compressive scales that systematically produce dog value on large spreads, and the quarterback is genuinely unknown.
Lowest totals in Sharp's book for a reason: the worst first-down passing offense in football against a defense that's #1 per drive, plus the #1 run defense against the worst run-blocking team since 2012.
The best offense in football against the best defense in football, and every number on the board says pick’em. This is the most honestly-priced game on the slate — and the one where the structural argument is loudest.
NO 0.46 pts. Every source lands within half a point of pick ’em.
NO The one-point drift toward Buffalo since summer is ordinary. Nothing identifies a mispriced side.
Josh Allen’s rate of sacks taken per third-down pressure, by year: 14.7, 23.1, 14.5, 9.8 — then 32.2% last season, #36 in the league. His third-down sack rate was 15.1% (#39) on 19 third-down sacks (#41). Even unblitzed on third down, 43% of his dropbacks ended in a hit or a sack, against 25% and 19% the two years before.
Houston’s defense pressures at 42.2%, #2 in football — and gets there on 37% of dropbacks without blitzing (#4), a rate higher than the league average WITH a blitz. When they do blitz they arrive in 2.27 seconds, 4th-fastest. This is the exact defense that breaks the exact thing Allen broke on last year.
Khalil Shakir’s average target depth — the lowest of any receiver with 90+ targets in 752 such seasons since 2007. Only 38.9% of his catches gained a first down, second-worst of any receiver in twenty years. Allen’s own aDOT has fallen every season since 2022 (9.2 → 8.7 → 8.3 → 7.3, #32, worst in the NFL), with 26% of throws at or behind the line.
And that dink-and-dunk offense runs into Sharp’s #1 secondary, fronted by Derek Stingley Jr. — back-to-back first-team All-Pro and PFF’s highest-graded coverage corner since 2023. Buffalo’s receivers were 18th in catches, 22nd in yards, 23rd in touchdowns and 27th in target rate per route.
The counterweight, and it is enormous. Buffalo’s run defense allowed 136.2 rush yards/gm (#28), 5.1 YPC (#30), #26 yards before contact, #32 yards after contact — and added nothing inside all offseason. Houston’s schedule of run defenses eases from #3 to a projected #21, the single largest rushing-schedule improvement in the league, and it starts here.
David Montgomery on man/gap runs: #1 in EPA per rush, #1 in first-down rate (44%), #2 in YPC (6.6). Houston ran man/gap on 38.8% of carries, 2nd-most in football. The one thing Houston does well is aimed straight at the one thing Buffalo cannot do.
Houston’s net field goals made over expectation — #1 in football. They went 48-of-52; their opponents went 18-of-27, the worst rate any defense faced. Add 7–5 in one-score games and 19 interceptions forced (#3) and the 2025 record has a large luck component that does not carry.
Stroud has fallen ten spots in Sharp’s rankings. Career-low 217.2 yards/game, and a playoff run of 87 dropbacks, 51.9% completions, zero touchdowns, five interceptions and six sacks across Pittsburgh and New England. Sharp’s word for it: “an abject disaster.” The Master Sheet is harsher still — QB 12.0 of 24, near-replacement.
Houston’s red-zone touchdown rate, 30th, eleven points below league average. Inside the twenty Stroud was #40 in success, #41 in yards per attempt (3.0), #28 in EPA. The cause is self-inflicted: 78% of early-down red-zone passes came from 3+ WR sets (#4) while 63% of red-zone runs used fewer (#13). They tip themselves.
Buffalo led at halftime in only 7 of 17 games and was outscored by 11 points at the break on the season — then scored 7.1 in the third quarter (#3) and 8.9 in the fourth (#1). Allen is #1 of 41 quarterbacks in second-half EPA over three years. Houston, meanwhile, was outscored by 40 after halftime and scored 3.41 third-quarter points a game (#26). Whatever the first half says, the second half has been Buffalo’s.
Under half a point of edge. Every source is inside a field goal of pick’em and they disagree about why. There is no bet here at this number.
#1 defense, #28 offense, a 30th-ranked red zone and the worst third-quarter team in football, against a passing game with the shortest target depth in the league. Both teams’ scoring is structurally capped.
Buffalo is #1 in fourth-quarter scoring and #1 in three-year second-half EPA; Houston is minus-40 after halftime. If Houston leads at the break, the live number is the play, not the pregame one.
#1 man-gap back, #2 man-gap rate, against #30 YPC allowed and #32 yards after contact, with the biggest run-schedule easing in football.
Our sources say Arizona is worth almost two points here. Sharp’s summer number says the market has already paid three of them out. Those two facts are why this came off the card.
YES 1.8 pts, and the largest health-regression case on the board.
NO Fails Commandment 9 — this is buying high. Sharp’s summer number was LAC −11.5 and Circa is at −8.5. Three points have already moved our way. We would be buying the back end of a move that already happened, which is a different bet than the one the model thinks it is making.
The Chargers’ record in one-score games (.750, #3) and field-goal games (#4), plus 1–0 in overtime. Six of their eleven wins came by one score, Dicker hit 92.7% (#5) for +20.46 field-goal EPA (#7), and they drew the #8 easiest schedule with 11 games against non-playoff teams (#3 most). Sharp’s own verdict: “an 11-win team that probably didn’t deserve to be an 11-win team.”
Los Angeles’s net rest for 2026: #32 in the NFL, the 4th-largest deficit since 1990, #2 since 2000, #1 since 2013. Seven rest-disadvantage games is a franchise record and only one other team since 1998 has had that many. From Week 3 on they play the league’s #2 hardest schedule, with a rest disparity in eight straight games, Weeks 6 through 13. None of that helps Arizona today — but it is why the season-long price is wrong.
The 2025 Chargers offensive line: #32 pass-block win rate, #31 run-block win rate, #30 PFF pass block, #32 PFF run block, #3 most-injured line in football. Their most-used five-man combination played 181 snaps — 16.5% of offensive plays, the lowest rate in the NFL. By January they were on the fifth combination.
Herbert was hit more than any quarterback in football — T-2nd most by any QB in twenty seasons — and took 54 sacks (#3). That is the third-most sacks ever absorbed by a quarterback on an 11-win team in the Super Bowl era; the average for the 200 eleven-win teams since 2000 is 31.6.
Arizona’s side of the ledger. They finished #1 worst in player health in the entire NFL — 32nd on offense, 31st on defense, 32nd overall — plus #2 hardest schedule, #2 worst field-goal luck, #5 worst fumble luck, #9 worst turnover EPA, 2–8 in one-score games (#3 worst) and 0–5 in field-goal games (worst in football). Murray played five games. Paris Johnson Jr. and Marvin Harrison Jr. missed five apiece. Their starting line played together on 23.7% of snaps.
Trey McBride set the all-time single-season record for a tight end, breaking Ertz’s 116 by ten, and became the first tight end ever with back-to-back hundred-catch seasons. Arizona’s 36% tight-end target share is #1 in the NFL. The Chargers used 12 personnel on 5.8% of snaps (#32) and played dime-or-bigger on 23% (#3 most), with Perryman available ten games and Colson at 218 career snaps. That is the matchup.
The Chargers’ record in games Joe Alt started and finished. With Alt on 184 dropbacks Herbert posted 67.5% completions (10th), 9.1 aDOT (2nd), 8.3 YPA (4th), a 7.1% TD rate (2nd). On 431 dropbacks without him: 7.3 aDOT (25th), 6.9 YPA (22nd), 4.2% TD (26th), 10.1% sack rate (31st). Sharp’s note is that his EPA per dropback with Alt on the field would have led all qualified quarterbacks over a full season. If Alt and Slater are both back, this line is not wrong.
Fewest allowed in the NFL. Jesse Minter’s 2025 defense: 1.80 points per drive (#5), #5 total yards, #5 pass, #8 run, 20.0 PPG (#9), 35.2% on third down (#5), 46.9% red-zone TD rate (#4), and opponents reaching the red zone on 28% of drives — #1 in football.
Brissett is not the problem the record suggests. Among 45 qualifiers he was #18 in EPA, #19 in success rate, #20 in completions, #23 in yards per attempt — ahead of Murray, Rodgers, Hurts, Mayfield and Lamar Jackson. Clean-pocket accuracy 81.7% (8th), 97.0 rating clean vs 62.3 pressured (32nd). Sharp: his play “was certainly NOT the sole reason” for 1–11.
Harbaugh’s go-rate on 4th-and-2 or less from his own 40 or beyond — the lowest in the league by a mile against a 78% average. Over two seasons he is 30th in go rate and 31st in fourth-down decision EPA. That is a structural cap on blowout margins in a game he is favored by eight and a half.
1.8 points of model value and the largest health-regression case on the board. But we are buying after a three-point move, into a defense that led the NFL in touchdowns allowed. Removed from the contest card. At +9.5 or better it is a real bet.
The all-time record holder at his position, on a team with the league’s #1 tight-end target share, against a defense that plays dime on 23% of snaps with a linebacker room that has 218 career snaps behind an injured starter.
#29 red-zone touchdown rate, an NFL-worst 5.2% red-zone interception rate from Herbert, 47% fourth-down aggression, and only 14% of receiver catches gaining 20+ (#25). This offense does not produce blowouts even when it wins.
#32 net rest, the #2 hardest schedule from Week 3 on, eight consecutive rest-disparity games, and a defense losing its coordinator. The price on their season number is the real edge, not this spread.
Two of the four worst rosters in football, and the market is charging a full field goal plus the hook for the home one. Our number says three-and-a-half is half a point too many — and Sharp’s summer line agrees.
YES 1.4 pts, with both independents and Sharp’s summer number all saying Las Vegas should not lay more than a field goal.
YES Line moved LV −3.0 → −3.5, so Miami got cheaper (Commandment 9). Neither team is remotely a league-level favorite, so Commandment 5 clears it. And we take the hook through the key number of 3.
Las Vegas’s yards before contact per rush. Sharp’s framing: “not just #32 in the NFL last year, but the worst for any team in the NFL since 2013.” Add 3.6 YPC (#32), −0.23 EPA per rush (#32) and a 29.6% success rate (#32). A clean sweep of last place.
Raiders first-half scoring per game — #32. Also #32 in first-quarter EPA per pass (−0.32), #31 in EPA per play, #31 in yards per play, and outscored by 111 points at halftime (#30). Sixteen of their 64 sacks allowed came in the first quarter alone.
Most in the NFL, an 11.1% sack rate, and 26% of all pressures turned into sacks — #32. Which is the one place Miami can’t capitalize: the Dolphins were 26th in pressure rate at 31.8%, and Chop Robinson fell to an 8.4% pressure rate, 149th of 219 qualifiers. The mismatch exists; neither team can exploit it.
Malik Willis’s record in games with double-digit dropbacks where the opponent scored 15 or more. Sharp, flatly: “in his four NFL seasons, he has never won a game in which the opponent scored more than 14 points.” His three career wins: 16–10, 17–10, 30–14.
The share of 2025 offensive and defensive snaps Miami is replacing. Sharp’s benchmark: “turnover above 35% on one side is considered a major rebuild… clearing 50% on both sides is nearly unprecedented.” They carry ~$180M in dead money — 60% of a $301.2M cap paid to players not on the roster, destroying the old record of $127.7M.
De’Von Achane’s PFF rushing grade — #1 among all 55 qualifying backs. He was #1 in the rate of runs gaining 10+ yards (16.8%), 2nd in yards per touch, at 5.7 YPC with 4.1 yards after contact and 1,838 scrimmage yards. He is the best player who will be on this field.
Miami’s depth index has five separate units ranked dead last — receiver, tight end, edge, corner and safety. Opponents completed 52% of throws of 10+ air yards against them, third-worst in football. If Las Vegas can protect at all, the field is open.
The Willis paradox. On the samples that exist he is first in EPA per play among 60 quarterbacks with 40+ dropbacks last year, and first among 62 with 100+ dropbacks since 2024. His 89 attempts in Green Bay produced 10.9 yards per attempt, +0.37 EPA and 79% completions. Over two years he leads all quarterbacks in fantasy points per snap (0.73), ahead of Lamar Jackson and Josh Allen, on a league-high 14.5% scramble rate.
Sharp grades Las Vegas #1 in the NFL in net rest edge, #1 in prep and #1 in rest. Miami is #24 in net rest, #31 in prep, #30 in rest, with exactly one game all season in which they have more rest than their opponent — #32. Not a Week 1 factor, but it is the season-long shape.
The Mendoza caution, if the rookie plays instead. His college sack rate against pressure was 21%, 60th of 78, and 73% of his college throws came from a clean pocket — 6th-highest in the class. Behind the worst run-blocking line since 2013, that is a bad combination.
Miami’s giveaway profile should improve regardless: 29th in interceptions thrown (18) and 27th in total turnovers (24) — while Willis has thrown zero interceptions and lost two fumbles across two seasons. The one place Miami’s teardown might quietly help them.
1.4 points of model value, both independent sources on the same side, and Sharp’s summer number half a point better for us than Circa’s. Getting the hook through 3 in a game between two bottom-five rosters is the whole bet.
The #32 first-half offense against a defense that is bad but not fast, with a quarterback who has never won a game where the opponent scored 15. Totals are banded 39.5–47.5 on the Raiders all year for a reason.
The #1 graded back in football is the one clean talent edge on the field, on an offense with nothing else. Volume is close to guaranteed.
#32 in first-quarter EPA per pass, with a quarter of their sacks allowed coming in the opening frame. If they start slow again the live number overreacts to a structural fact.
Our Master Sheet spits out Philadelphia by twelve. The market says four and a half. When a model disagrees with a market by that much, the model is usually the thing that’s broken — and here it is. This page shows the work.
NO Discarded. The sheet said −12.2 on units both other sources call Philadelphia’s weakness. Our honest number is Hoppen’s −4.7 — which is the market.
NO There is no case to make. When your own number lands on the market’s number, there is no vulnerability to find.
Overgraded, and it is 31% of the whole gap. Sharp, verbatim: “Safety is the hole — Blankenship gone, Sydney Brown traded, leaving Mukuba, Michael Carter and DeJean in base.” Hoppen, independently: the safeties are “the exploitable spot.” Both notes’ own betting reads say to attack the deep middle. A room two sources call the roster’s weakness is graded at 80% of scale.
Overgraded on reputation — 28% of the gap. Barkley’s yards before contact fell 2.64 to 1.36; stuff rate went 16% to 21%; the team’s first-down yards per rush fell #9 to #28 and EPA per rush #8 to #18; third-down distance #13 to #30. Sharp explicitly rules out the injury excuse: with Lane Johnson on the field they averaged 4.2 YPC, without him 4.0. And there is a new play-caller installing a new blocking scheme.
The sharpest single contradiction in the sheet. Barkley in 2025: 4.1 YPC (33rd of 49), 45% success (39th), 24.3% of runs gaining nothing (47th) — 2,005 yards down to 1,140, eleven hundred-yard games down to three, at 29. Washington’s backs, graded 1.0 of 3, actually out-produced him: 4.5 YPC (11th), 42.6% success (4th), 3.53 yards after contact (4th). The sheet grades the worse backfield two and a half times higher.
Washington is undergraded here — the same 45% number the Raiders get. Washington had 42 sacks (#12) and spent heavily: Oweh at 4yr/$100M (7.5 sacks, 15.9% pressure rate in twelve games), Chaisson (7.5 sacks, 14.5%), Omenihu, Chenal, Settle, with Armstrong back. Pressure rate 35.6%, #20 — not bottom-three. Seven projected new defensive starters means the grade is marking last year’s unit.
And here the sheet argues against the spread — it has Daniels 2.5 points ahead of Hurts. Sharp’s data cuts the other way: Daniels fell to −0.05 EPA per pass, 61% completions, 6.7 Y/A, QBR down more than 20 points to 44.7, while Hurts posted a career-high 25 touchdown passes at a career-low 1.3% interception rate and Philadelphia is 29–8 when he carries ten or more times. Correcting the QB grade makes the sheet worse, not better — which means the entire twelve has to be defended on the units above. It can’t be.
Washington’s defense is a genuine historical outlier — the only defense since 2021 to allow 35%+ of drives into the red zone AND 67%+ touchdowns once there (37% / 68%). Plus 38.2 yards per drive allowed (#32), 78% down-set conversion (#32), 23% three-and-outs forced (#32) and 384 yards a game, most in the NFL.
The average opponent opening drive against Washington — worst in the NFL, worst by any Washington defense this century, 13th-worst by anyone in 26 years. First quarter: 101 points allowed (#28), a −64 first-quarter margin (#31), and seven punts forced in 39 first-quarter drives — the worst rate by any defense this century.
Philadelphia’s rest differential across Weeks 1–9 — easily the worst in the NFL, four disadvantage games and zero advantages, inside a season-long −15 (#2 hardest, their worst since 2012). History: since 2015 at a 3+ day rest disadvantage the Eagles are 7–14–1 straight up and against the spread (−36.4% ROI); since 2020, 2–6–1 ATS at −52.3% ROI, third-worst in football.
The Eagles’ regression flags. 8–4 in one-score games, second-most such wins in the league, with 72.7% of all their wins coming by one score (5th-highest); a #1 red-zone touchdown rate of 70.5% against a 57.3% league average and their own 57.4% the year before; and the #2 healthiest roster two years running, with Hurts, Barkley and Smith missing zero games.
Philadelphia was #32 in penalty margin (26 more flags than opponents), #32 in defensive penalty EPA and #31 in net penalty EPA (−23.6) — and 61% of all second-half penalties in Eagles games went against Philadelphia, most in the league. Structural, not variance.
Fangio’s pass defense allowed fourteen passing touchdowns — fewest in the NFL and the franchise’s lowest since 2001 — at 6.4 yards per attempt (#5), with the league’s best third-down and red-zone pass defense and 14.9 points allowed per game from Week 10 on, best in football. Quinyon Mitchell allowed a 42.4% completion rate, best of any full-season corner; DeJean allowed zero slot touchdowns. Both first-team All-Pro.
Washington’s offense was not the problem — 33.0 yards per drive (#10), a 24% three-and-out rate (#2), #10 in early-down success, #11 rush efficiency, and a 65.2% red-zone touchdown rate (4th). Daniels played 38.4% of snaps, appeared in seven games and finished four; he and McLaurin were healthy together for five days all season.
McLaurin is aimed exactly at the graded hole: 14th among 100+ route runners, a 30.9% target share, 45.7% of the team’s air yards when on the field, 14.0 air yards per target, and 55% of his targets producing a first down or touchdown — 3rd in the NFL. Vertical, between the numbers, against three safeties nobody trusts.
Washington’s own regression case is real too: #31 turnover margin, #32 in opponent fumbles lost, #30 in roster health with nine starters missing 7+ games (against one the year before), 2–5 in one-score games after 8–4, and 0–3 in field-goal games (#32) after 4–1 (#5). Both teams regress. That is why the market is at 4.5 and not wider.
Our honest number is Circa’s number. The only version of this game with an edge is the one built on unit grades that two independent sources contradict in writing. Not on the card.
The one place where the broken grade points at something real. A 14-air-yard, 31%-target-share receiver against the safety room both books call the exploitable spot. Attack the middle, not the perimeter.
Philadelphia’s offense was the worst of the Sirianni era — 22.3 PPG (#19), 311 yards (#24), #28 in motion rate with Super Bowl personnel — against the #4 easiest schedule of defenses, and it went 5–5 in its last ten. Now there is a new play-caller, the fifth in five years.
A −22 rest differential over nine weeks, on a team that is 2–6–1 ATS at a 3+ day deficit since 2020. That is where the Eagles number is actually mispriced, not here.
The cleanest signal on the board. Both independent sources make this a pick’em or better for Green Bay, Sharp’s own schedule sheets have the Packers as the road favorite, and the market is laying a point and a half the other way.
YES 1.45 pts. Both independents sit at pick ’em or better for Green Bay.
NO Honest downgrade. Sharp’s summer market had Green Bay favored here and Circa now gives them 1.5 — but the most likely cause is the Micah Parsons ACL, which is information the market priced correctly, not public bias. This is a bet on the football with the market column unproven. Size it accordingly.
Jordan Love’s EPA per dropback from a clean pocket — #1 in the NFL. Overall he was #2 in EPA per dropback (0.22), 6th in rating (101.2), 6th in interception rate (1.4%), 8th in touchdown-to-interception ratio. Sharp’s flat statement: “Jordan Love just might be the NFL’s most underrated quarterback.”
Where Minnesota’s three 2025 quarterbacks finished in EPA per play out of 57 qualifiers. McCarthy specifically: #36 of 36 in interception rate, #35 in EPA, #35 in completion rate, #35 in rating, #35 in accuracy, with a 14.4% catastrophic-play rate — worst in the NFL and the first non-rookie since JaMarcus Russell to pair sub-58% completions with a 14%+ catastrophic rate.
And Kyler Murray is not the fix. Since 2022 on throws of 10+ air yards he ranks #50 of 50 in EPA per attempt; on 25+ air yards he is #36 of 36 in EPA, yards per attempt and completion rate; between the hashes over four years, #47 of 50. He has been healthy for one of the last five seasons. Sharp: “Minnesota upgraded from disaster to uncertainty at quarterback.”
Love is #2 in the NFL in EPA when blitzed (+0.31) — and Flores’s defense led the league in blitz rate (48%), pressure rate (44.5%), and the rates of rushing five and six-plus, using four-man rushes on only 50% of snaps (#32). Green Bay’s quarterback is specifically excellent at the specific thing Minnesota specifically does.
Third down. Green Bay converted at #2 in football against a 39.5% average; Minnesota’s offense converted at #32. Sharp’s caveat is honest — “the Packers ran pure on third downs, which could be difficult to replicate” — but the gap is eighteen points wide.
The Minnesota defense that carried a 9–8 team is dismantled. It was #3 in yards per play, #2 in passing yards allowed, #2 in red-zone touchdown rate, #4 on third down, #1 in pressure rate, #3 in points allowed, 49 sacks (#4). Greenard traded to Philadelphia, Allen and Hargrave cut, roughly $50M over the cap. The front seven fell from #1 to #20.
Minnesota’s 2025 quarterback play on the first fifteen scripted plays: #6 in EPA, #5 in success, #5 in yards per attempt, 69.2% completions. After the script ran out: #35 in EPA, #34 in success, #33 in yards per attempt, 53.4%. Sharp: “When the answers were handed to him, he was a top-five quarterback. When he had to find them himself, he was the worst quarterback in the NFL.”
The rate at which Love was pressured last year — a career high — and his EPA under pressure was −0.37, #21. Green Bay’s line fell from #10 to #27, was 29th in pressure rate allowed, and its starting five played together on 18% of snaps (#30). Walker left for Carolina and Jenkins for Cleveland; the plan is Morgan to left tackle and Rhyan to center with a fifth-rounder as the only addition. Against the #1 blitz-and-pressure defense in football, that is the way Green Bay loses this game. The Master Sheet agrees: Minnesota’s line grades 9.2 to Green Bay’s 7.0.
Parsons and Gary produced twenty of Green Bay’s thirty-six sacks. Parsons is rehabbing a December ACL and may open on PUP; Gary was traded to Dallas; Enagbare is gone. Van Ness has 8.5 sacks in three seasons. The rest of the roster generated sixteen sacks all year. Sharp’s own note: “the most expensive pass-rush acquisition in NFL history did not translate into elite team-level disruption.”
Green Bay’s interceptions, year over year — from T-#3 in the league to #28, with takeaways per drive falling from #3 to #25. The coverage split is the tell: 6.4 yards per attempt allowed (#6) but #24 in EPA allowed per pass. Nixon allowed 6 touchdowns and a 105.1 rating; Valentine 7 and a 121.2.
Fumble luck flatters both sides and cancels: Green Bay was #5 (+4.3 net recoveries over expectation), Minnesota #4 (+4.5, with opponents losing a league-high 13 of 22). Kicking runs the other way, hard — Minnesota was #3 in field-goal rate, #1 in extra points, #1 in field-goal EPA (+32.7), #3 in field-goal luck, all fade candidates; Green Bay was #23, #27, and #27 in net field goals over expected, with McManus at a league-worst 42.9% from 40–49 before being replaced.
The strongest signal on the slate: both independent sources at or better than pick’em, and Sharp’s summer market had Green Bay laying points in this building. We are getting a number two to three points better than the one the market printed in June.
If the contest side is taken, the outright is the same opinion at a better price. The 1.5 is nearly meaningless as a cushion; the disagreement is about who is better, not by how much.
#32 on third down, a quarterback who is #50 of 50 downfield since 2022, and a passing game that was top-five only on scripted plays.
Green Bay scored 3.0 first-quarter points a game (#29) but allowed 1.5 (#1). A scoreless or trailing first quarter is their normal shape, not a signal.
A small edge on the biggest positive-regression team in football, against a defense that just allowed more points than any team in the sixty-six-year history of the franchise.
YES 0.6 pts — the smallest number on the card, and deliberately so. Small spreads take the least distortion from our conversions (Commandment 7).
YES Dallas is the most heavily public team in football, and the price shows it: 2–8 against the spread as a favorite, 1–4 as a road favorite. That is a repeating structural overlay, not a model output. Line moved +2.5 → +3.0, so we buy low and get the key number of 3.
The Giants’ record in one-score games — .125, #31 in football, tied for last in one-score wins and 30th in one-score losses. Sharp: “Flip their one-score losses, and the Giants were a 10–7 team.” Seven of thirteen losses were one possession; five more were blown double-digit leads. They were outscored by 36 in the fourth quarter (#26) despite scoring the 12th-most fourth-quarter points — the collapses were defensive.
Points allowed by Dallas — the most in the sixty-six-year history of the franchise. 30.1 per game (#32), 2.88 per drive (#32), 252 passing yards a game (#32), and #32 in early-down success, total efficiency, red zone and explosive rate, #31 in pass defense, #30 against the run and on third down.
Dallas’s touchdown-to-interception ratio allowed through the air — the 5th-worst by any defense in the last decade. The takeaways have not come back since Dan Quinn left.
Jaxson Dart’s rookie split is the most extreme in either conference. Passing EPA −15.09 (#28 of 36 quarterbacks) against rushing EPA +46.34 — 3rd among all rookies since 2000, behind only Jayden Daniels and Cam Newton. His 17.5 EPA on designed runs was #2 in the entire NFL behind Josh Allen, and his nine rushing touchdowns are the third-most ever by a rookie quarterback. On third down he was +0.13 EPA per attempt, #6 in the league — the best rookie third-down mark of 33 qualifiers since 2020, against a group average of −0.18.
Dallas drew 133 accepted penalties, most in football, for 1,136 yards (#2) — 31 offensive holds (tied worst) and 23 false starts, with Smith, Guyton and Pickens at seven apiece. (The honest caveat: the penalty EPA equation actually favored them, 3rd on offense and 11th on defense.)
The Giants’ offense was quietly one of the league’s most improved and the record hid it: #10 in EPA per play, #8 in yards per drive (33.9), a turnover on only 9% of drives (#13). First-quarter points went to #11 after ranking #32, #32 and #30 in the three prior years — and #7 excluding Russell Wilson’s three starts. Excluding those: yards per drive #8 (from #28), red-zone drive rate #10 (from #26), touchdowns per drive #11 (from #30).
Dexter Lawrence is in Cincinnati, and his on/off split is the biggest downside variable in the Giants’ book. With him on the field New York was #13 in yards per attempt, #15 in pressure rate, #8 in pass success, #10 in EPA per pass, #8 in sack rate. Without him: #32, #31, #31, #30, #30. The longer-window version shows the same shape. The 2026 Giants are the “without” column.
Dallas’s run game is real and New York’s run defense is not. Cowboys rush EPA +0.09 (#5, up from #27), 44.1% success (#7), 4.6 YPC (#9); Javonte Williams went 1,201 and 11 at 4.8. The Giants allowed 5.5 yards per carry to running backs — three-tenths worse than the 31st-place Bengals — on 1.85 yards before contact. Parker’s Dallas scheme is built on light boxes, which is exactly what Dallas got on 32.5% of Williams’s carries (#3).
Prescott’s play-action EPA per attempt last year, on 173 dropbacks worth +51.9 total — against +35.7 total across 568 play-action dropbacks from 2020 to 2024. His 11-personnel EPA per play went #30 to #2, EPA per pass #27 to #3, early-down passing #23 to #4. His CPOE was +4.4 (#3) on an expected completion rate of 62.9% — the hardest throw diet in the league.
New York’s interior defensive line is now Reader, Fotu and a 32-year-old Shelby Harris after Roy Robertson-Harris tore an Achilles in offseason workouts, on top of the Lawrence trade. And Sharp calls the No. 5 overall pick playing off-ball linebacker value-diminishing with Burns and Carter already owning the edges.
Dallas played the #2 easiest schedule in football last year — the #7 easiest slate of defenses, the #3 easiest slate of offenses, the #2 easiest slate of run defenses — and went 7–9–1, 1–5 against playoff teams, trailing in sixteen of seventeen games. 2026 is the #4 hardest schedule with the largest year-over-year jump in difficulty in the league “by a mile.”
Dallas’s against-the-spread record as a favorite last season, including 1–4 as a road favorite — against 5–2 as a dog, on a 7–10 overall ATS mark and 13–4 to the over. The Giants went 9–7 ATS and 11–6 over as a team that was an underdog in sixteen of seventeen games.
And the Giants’ 1–7 one-score record came against the hardest schedule in the NFL, with ten games against winning teams — the most of any team — going 2–8 in them. Positive regression on both the record and the slate.
A small, clean edge on the league’s biggest one-score-regression candidate, at home, against a favorite that went 1–4 as a road chalk and just allowed the most points in franchise history. Getting the key number of 3 is most of the value.
First-quarter scoring jumped from #32/#32/#30 to #11 — #7 without Wilson’s starts — while the fourth-quarter collapses were defensive. Bet the halves, not the game, against the #32 scoring defense in football.
#2 in the NFL in designed-run EPA behind Josh Allen, third-most rookie rushing touchdowns ever, against a defense that finished #32 in red zone and #30 against the run.
Dallas was 13–4 to the over, the Giants 11–6, and Dallas’s defense set a franchise record for points allowed. Both books point the same way here.