New Orleans Saints at Detroit Lions. Lean: Under 49.5. No call on the spread.

Spread.Take New Orleans +7, verified at BetMGM. Edge Labs makes Detroit -10.7, but its raw margin is only 8.4; the calibration manufactures 2.3 additional points without demonstrated ATS value. My judgment places this closer to Detroit -6.5 than double digits. Detroit's missing starting safeties and center matter, while a new offensive coordinator and unresolved left-guard competition add execution uncertainty. Shough's starting role is confirmed, and New Orleans retains credible receiving options against the depleted secondary. Those absences are public information, not discoveries the market necessarily missed. Nevertheless, seven provides useful push protection. VSiN reported Detroit remaining at seven despite heavy DraftKings ticket support, consistent with resistance to a higher favorite price; conflicting FOX splits prevent treating that as definitive sharp evidence. The counterargument is substantial: Detroit's offensive weapons are largely available and its talent advantage can overwhelm this number. I could not verify a current Saints Wednesday injury report, limiting conviction.

Total.Under 49.5 is a small lean. Applying the dossier's stated 1.2-point historical upward bias numerically lowers 49.72 to 48.52, only 0.98 below market. Detroit's center absence, unsettled left guard and first game under Petzing offer plausible drive-stalling mechanisms. The over case is strong: indoor conditions, Detroit's available skill players, and missing starting safeties create scoring opportunities on both sides. The captured DraftKings increase from 48.5 to 49.5 also argues against assuming suppressed scoring, although other captured books stayed unchanged. I give the offensive-transition risks slightly more weight at this price, without inventing a slower pace: no reliable current pace projection was verified. Missing Saints practice information further limits confidence.

Where our number looks wrong.Detroit -10.7 is too aggressive. Calibration stretches an already substantial raw advantage, while the injury-blind ratings miss Detroit's safety and center absences. Returning QB/head-coach flags also obscure the offensive-coordinator change. These shortcomings support shrinking toward the market, not claiming a precise 3.8-point edge.

Spread.Our number (Detroit -10.7) sits nearly four points past the market's -6.9, but the model's own holdout says that gap carries no ATS information (47.6%), the ratings are frozen at Aug 19, and it sees zero injuries, so I will not buy Detroit off it. Market structure argues the other way: the spread has held -6.5/-7 since the Aug 30 open, FanDuel is back down to -6.5, and aggregated public splits show roughly 87-90% of tickets on Detroit (cleatz.com, retrieved 9/9). Public money that cannot move a number is money books are happy to take. The news cuts both ways. New Orleans is without WR Jordyn Tyson (IR), DE Cameron Jordan (ruled out 9/9), DT Bryan Bresee, and probably Alvin Kamara (limited Wednesday). Detroit is without both starting safeties, Brian Branch and Kerby Joseph, for at least four games. A depleted but healthy-at-quarterback dog getting the key number +7, against a rebuilt Detroit offensive line that still has to block Chase Young, is worth a lean rather than a bet.

Total.Dome, so no weather variable to price. The market is 49.5 at every book we captured; our model says 49.72, but the same model documents a +1.2-point holdout bias against closing totals, which puts its honest fair value near 48.5 - a point of under value, not over. Personnel points the same way: New Orleans has lost its explosive skill players (Kamara limited at best, Tyson on IR, Chandler on IR), leaving Etienne and a rookie-heavy receiver room behind a quarterback with nine career starts, and Saints games stayed under in 12 of 17 in 2025. The genuine over risk is Detroit's missing safety duo plus a Lions offense whose games went over 10 of 17. I will take the under at 49.5 and would pass below 48.5.

Where our number looks wrong.Three ways. First, -10.7 is a calibration artifact as much as a football opinion: the map stretches large raw margins, so an 8.4-point raw edge reads as double digits, and the documented 47.6% holdout ATS means the 3.8-point disagreement with the market is noise, not signal. Second, the ratings are frozen at Aug 19 and injury-blind, so they miss the two facts that actually matter this week: Detroit playing without both starting safeties (Branch and Joseph, PUP through at least Week 4) and New Orleans's post-freeze additions (Etienne, Fant, Elliss, Tyree Wilson) alongside its subtractions. Third, the qb_returning flag gives Shough full continuity credit for what is really a nine-start sample; the model is not distinguishing a proven returning starter from a second-year unknown. Direction of the error: our Detroit number is too big, and the market at -6.5/-7 is closer to right than we are.

Spread.Our v3 model rates Detroit -10.7 at home, well past the -6.9 market consensus and the live -7 (-6.5 at FanDuel) line, a 3.8-point gap our own calibration flags as a 'pass' since the model has zero proven ATS edge (47.6% in the 2022-2025 holdout) and is blind to injuries. That caveat noted, this week's real injury news independently supports Detroit: HC Kellen Moore confirmed DE Cameron Jordan (hamstring) is OUT for the opener, and Alvin Kamara (MCL sprain) was only a limited practice participant Wednesday with no guaranteed Sunday status, leaving free-agent signee Travis Etienne Jr. as New Orleans' clear lead back behind an offense that averaged under 20 points a game across Tyler Shough's nine 2025 starts. Detroit's own report is clean, Amon-Ra St. Brown, Jared Goff and the rest of the starting unit are full participants. The number hasn't moved off -7 in ten days despite this news, but sharp money (75% of dollars per scoresandodds) is stacked on Detroit against a near-even 59/41 ticket split. I'm backing Detroit to cover on the real personnel gap, not our model's inflated number.

Total.Both models sit on top of the market here, our totals model's 49.72 is just 0.2 above the 49.5 consensus, essentially no signal, and the model's own holdout bias runs about 1.2 points high, which pulls its 'true' read closer to 48.5, toward the under. New Orleans is missing its top pass rusher (Jordan, OUT) and rolling with a hobbled/limited Kamara behind an offense that scored under 20 points a game across Shough's 2025 starts; a shorthanded, ball-control-limited Saints attack caps total possessions and scoring more than it inflates garbage-time volume against a Detroit team that plays aggressively but doesn't need to force extra plays. Ford Field is a fixed dome, so weather is not a factor either way. Public and sharp money both lean under (82% of bets, 65% of dollars per scoresandodds), a crowded trade but one consistent with the injury picture on the Saints' side of the ball.

Where our number looks wrong.Our -10.7 number is a preseason-frozen rating stretched further by a calibration map that inflates large favorites, and it has zero demonstrated ATS edge (47.6% in the 2022-2025 holdout), it should be read as noise, not a real point value. It also cannot see that Jordan is out or that Kamara is hobbled; the model would output the same -10.7 whether or not either player suited up, so any overlap between its number and this week's actual injury news is coincidence, not model skill. If anything the model is more likely wrong on magnitude (too big a favorite) than on direction.