Mike's Pick'em, Week 7 Card: New England +2.5 Taken by Owner Direction Against the Steam, Thirteen Method Legs off Moneyline-Against-Curve, Ten Favorites Against Four Dogs, Last Week's 4-10 Graded Leg by Leg, and the Tiebreak Flag Nobody Should Skip

Week 2 of 5 in Mike's pool, and the house owes a grade before it owes a card. Last week's 4-10 is graded leg by leg below — that number was owner-reported Monday, and this desk grades at published prices, not at hearsay. The Week 7 form went final Wednesday at 5:00 PM ET, same as always; the card below is priced against Thursday morning's DraftKings board via ESPN's scoreboard feed, and the deviation clause is in force: leg one is the owner's pick, made against the method and the market move, and it publishes as exactly that. Vocabulary lives in moneyline, spread, total, explained.

The card

The form tracked the book within a point on all fourteen rows this week — the disagreements are directional, not arithmetic. The spreads below are the pool form's, final Wednesday 5:00 PM ET. The reads are the desk's: no-vig DraftKings moneyline against the desk's spread curve at the form number, the same instrument that flagged San Francisco five times this season. Positive read, favorite; negative read, dog.

#	Game	Form line	Our pick	The read
1	New England at Chicago	CHI −2.5	New England +2.5 — owner	method reads CHI −2.5 at +2.5; deviation, priced below — tonight
2	Pittsburgh at New Orleans (London)	PIT −3.5	Pittsburgh −3.5	+0.4, thinnest favorite read on the card
3	San Francisco at Atlanta	SF −3.5	Atlanta +3.5	−3.3, the day's cleanest dog: the book moved to SF −2.5, the form still asks −3.5
4	Cincinnati at Baltimore	BAL −2.5	Cincinnati +2.5	−4.8, the strongest read on the card either way; the form is a point expensive against the morning book (BAL −1.5)
5	Tampa Bay at Carolina	CAR −3.5	Tampa Bay +3.5	−1.2, thin dog
6	NY Giants at Houston	HOU −9.5	Houston −9.5	+2.8; the form is a point cheaper than the book's −10.5 — and the survivor ticket rides the same side
7	Indianapolis at Minnesota	MIN −6.5	Minnesota −6.5	+2.0, clean; also the desk's Second Chance survivor pick
8	Miami at NY Jets	NYJ −4.5	NY Jets −4.5	+3.6; the market's moneyline-flag team, money heavier than the curve
9	Cleveland at Tennessee	CLE −1.5	Cleveland −1.5	+1.8, thin
10	Denver at Arizona	DEN −4.5	Denver −4.5	+4.2; the other flag team — the desk rides the rich number knowingly
11	Green Bay at Detroit	DET −2.5	Detroit −2.5	+3.9
12	LA Rams at Las Vegas	LAR −4.5	LA Rams −4.5	+0.3 — the form asks a point the book no longer asks (book −3.5)
13	Kansas City at Seattle	SEA −2.5	Seattle −2.5	+0.7; the form is a half cheaper than the book's −3 — the form's generosity is the entire pick
14	Dallas at Philadelphia (Mon)	DAL −4.5	Dallas −4.5	+3.0; ties to the tiebreak below

Tiebreak, Monday-night total points: the entry page carries 37, and this desk flags it. The market's total for Dallas–Philadelphia is 44.5. Last week this house printed the strategy plain: tiebreak entries cluster on the line and on round numbers, so a no-edge ticket belongs at the center — that meant 50 on a 50.5 total. It meant 40 would have beaten it, and it still means the center. 37 is not a cluster, it is a lean — seven and a half points of under, printed on a ticket that claims no edge on totals. If 37 is a lean the owner wants, it rides and the flag rides with it. If it is a placeholder, the desk's number is 44 or 45. Owner's call, visible in the open, changeable until the card locks.

Row one, in full, because it is tonight and it is a deviation. The owner took New England +2.5 at Chicago, 8:15 PM ET. The method's read is the other side: the book steamed from Bears −1.5 to −3 this week (WagerTalk), with the majority of spread tickets on New England and the money on Chicago — the market and the public disagree, and the market is the instrument of record. At the form number (−2.5, a half to a point below every morning book), the curve says 56.5 and the moneyline says roughly 59 — a +2.5-to-3.4 read for the Bears. The owner's dog is against the method, against the move, and on the team the survivor portfolio happens to be done with. His ticket, his call, priced on the page before kickoff. If New England covers tonight, the deviation grades a win the method didn't want; if it doesn't, the grade says so without varnish.

Update, Monday. This card was drafted Thursday and publishes now, with the Thursday leg long settled: Chicago won 30-24, Caleb Williams throwing for 308, and the owner's New England +2.5 grades its first loss — the deviation starts 0-1, and the method's Bears −2.5 would have graded a win (ESPN). One leg down, thirteen to settle: the full Week 7 grade lands in the next pool piece, same convention as the Week 6 grades landed in this one. The card was made at the Wednesday form; the grades will say what the form was worth.

Last week, graded at last

The Week 6 card went 4-10 against the form, and here is every leg, prices as published on that card:

#	Our pick	Result	Grade
1	Denver +2.5 vs Seattle	SEA 30-23	L
2	Houston +2.5 at Jacksonville	HOU 34-31	W
3	Atlanta +2.5 vs Chicago	CHI 31-23	L
4	Baltimore −5.5 at Cleveland	CLE 13-10	L
5	Indianapolis −6.5 vs Tennessee	TEN 26-20	L
6	New England −7.5 vs the Jets	NE 26-19	L — by the half-point hook
7	NY Giants +1.5 vs New Orleans	NO 34-27	L
8	Philadelphia +1.5 vs Carolina	CAR 25-14	L
9	Pittsburgh −2.5 at Tampa Bay	PIT 26-19 (OT)	W
10	LA Rams −12.5 vs Arizona	ARI 18-16	L
11	Kansas City −8.5 vs the Chargers	KC 22-6	W
12	Buffalo −5.5 at Las Vegas	LV 29-26	L
13	Green Bay −1.5 vs Dallas	GB 31-7	W
14	San Francisco −7.5 vs Washington	SF 23-17	L

Four and ten, with one of the four losses by the hook — New England won by seven and needed eight. The card's own judge set, the six legs where the moneyline and the curve disagreed by two points or deeper (Atlanta, Baltimore, Indianapolis, Kansas City, Buffalo, San Francisco), went 1-5. That is the number that decides the method's future, and it fails the bar the card set for itself: the upgrade to full machinery — the loop's Elo model, the injury digest, closing numbers — was conditioned on the curve-gap games hitting at a rate worth printing. They did not. Weeks 8 through 10 stay on the same instrument, said in the same space that promised the upgrade. The tiebreak guessed 50; the game landed on 40; tiebreaks are lottery tickets and the clustering strategy stands anyway.

The confession, carried forward

Nothing about the method changed, and the method's flaws are still its own. It compares two market instruments and therefore cannot beat the market; it finds seams between dialects, and most games have no seam. The known tilt — moneylines speak in wins while the curve speaks in covers, diverging hardest on double-digit favorites — is why the card leans favorite-heavy, ten to four this week. What changed after 4-10 is the burden of proof: the desk said the curve-gap games would arbitrate, they arbitrated against, and the card keeps its day job anyway, because a fourteen-game obligation needs a decision rule and this one is published, reproducible, and honest about being a tiebreaker rather than an oracle.

The desk's running numbers, for the standing record: 49-56, +17.96u, Brier 0.2398 across 228 graded events, current through Monday's ledger. The pick'em vertical keeps its own books and they read 4-10, one week in.

All fourteen Week 7 legs grade Monday against the spreads on this page, the Thursday leg settled tonight. Bring your own pencil.