NFL Sunday, Week 6: The Machine and the Market Argue About Philadelphia — 65 Cents Against 40, the Card Rides Three Home Teams the Book Discounts, the Teaser Needs Kansas City and a Banged-Up San Francisco, and the Survivor Desk Holds Two Tickets on the Chiefs

One Sunday, twelve games, and one argument worth twenty-five points — plus a Monday-nighter holding the second leg of the teaser. The machine and the market agree on almost everything this week, and where they disagree, they disagree about the same team from opposite directions. Here is the whole board, every ticket the house is holding, and every number that grades tonight.

The card

Quarter-Kelly stakes, capped at two units. All three legs are home teams the model likes better than the book does.

Leg one carries the argument, so the argument gets printed. The market made Philadelphia a 42.5% team on Thursday night after Jalen Hurts and Saquon Barkley were ruled out. The loop's injury digest measured that damage at −2.81 — the deepest number on the Sunday board — and the model still makes the Eagles a 65.0% winner. That gap is not a secret the market missed; it is a disagreement about how much a backup quarterback costs. The backtest's injury weights measure the average starter-out effect across hundreds of games; the market prices this backup, today. Both instruments are doing their jobs, and they are twenty-five points apart. The market moved overnight — 3.05 on Saturday morning, 2.85 this morning — four points toward the machine before a ball was kicked. The desk has not declined the leg; the deviation clause stays open until kickoff, and this paragraph is the disclosure.

Fourth and off the card: Atlanta +8.7% (55.5% vs 46.8%), the beneficiary of Chicago's −1.61 digest. The MAX_BETS=3 cut keeps it in the notebook.

The full board

*Forecast only — PRO•LINE had not posted at publish. Washington at San Francisco is Monday night (the second teaser leg); it is not on today's board and the loop will carry it tomorrow. The pass/fake edges on the three clamped rows (Titans, Jets, Chargers at +8.9 to +13.5 on the other side) are floor-manufactured and named as such; the house rule passes them.

Every number in the P(ours) column already carries this morning's injury digest — Philadelphia's −0.110, Green Bay's −0.035, Kansas City's +0.059 — so the edges price the news, not the schedule.

Where the venues disagree

Pittsburgh at Tampa Bay is the stale price of the morning. Mayfield was ruled out Saturday; PRO•LINE's fair makes Pittsburgh a 60.0% team, and Hyperliquid's outcome token still sits at 56.1% — a 3.9-point gap in the same market. The model, with the Mayfield damage priced in, says 65.0%. Nobody here cards a leg the sheet passed, but the read belongs on the record: the cheapest Pittsburgh ticket in the market is the HL token, and the book of record agrees with the model's direction.

Philadelphia is the opposite shape: both venues agree with each other (39.1 and 40.0) and both disagree with the machine. If the desk rides the leg, the ticket is the PRO•LINE moneyline at 2.85 — Hyperliquid's 0.391 token implies a worse price than the board offers.

The teaser card

Card 1: Kansas City −1.5 (from −7.5, crosses 7 and 3) + San Francisco −1.5 (from −7.5, crosses 7 and 3), priced at an assumed 1.91 — the real PRO•LINE two-team payout remains unconfirmed, and that confirmation is still the single most valuable piece of information this engine is missing. Breakeven is 72.4% per leg; the market's own numbers clear it on Kansas City (75.6%) and on San Francisco (75.6%) — with a flag. Nick Bosa is out (SF digest −1.28), which trips the engine's injury flag, and the next engine run may re-roll the card without the 49er leg. The Tampa Bay leg that opened this card on Thursday came off when Mayfield was ruled out; if that version was already placed at a window, it grades as placed, slip price logged when it surfaces.

The survivor card

Three tickets, one deviation, all published before kickoff (the Week 6 brief):

The desk's published call was the Rams, the season's biggest favorite at 86.3% carrying 42.51% of America; the owner took the other seat — he did not want to be on the bus with everyone — and the deviation published with its costs in the open: eleven and a half points of win probability against the desk's call, and the portfolio's all-three survival drops from 54.7% to 47.4% in exchange for a free fade. Both tripwires were tested Saturday at 6:00 PM and both held — Kansas City moved to −7.5 without crossing the wire, Indianapolis never re-priced — so the card stands as printed. One game can now kill two tickets: if the Chargers win at Arrowhead, the ESPN and Second Chance entries both die on a 20.4% outcome, and tonight's grading opens by saying exactly that, in advance, on the published numbers.

What we owe

Every carded leg grades against the prices on this page — the Eagles at 2.85, the Giants at 2.45, the Packers at 2.20, the survivor tickets against the market's own percentages, and the teaser leg by leg, with the 49er settling Monday night at the slip's number. The London game grades untracked: no line, no ticket, no asterisk needed. The engine line through this morning's sheet: 48-52 on positive-edge bets, +21.61u, mean Brier 0.2373 across 216 graded predictions, every one written down before first pitch (the ledger).

Twelve games Sunday, one Monday, four tickets, one argument about Philadelphia, and a scoreboard that settles most of it by midnight. The card is on the record. The grades land tonight.

Bring your own pencil.