New: Be A GM College lets you run your NIL budget, fight the portal, and play the season. Try it.

Weekly Lines · 2018

Week 9.

13 games and 8 plays, replayed walk-forward on angles chosen from the seasons before 2018. Every number is open.

This week's plays: 5-3, +2.7u flat at -110. Season through week 9: 45-46, -6.0u flat at -110.

GameMarketTotalFinal
OAK @ SFThu, Nov 1 20:20 ETSF -1.544OAK 3 at SF 34SF -1.5 closed · SF covered
ATL @ WASSun, Nov 4 13:00 ETWAS +1.546.5ATL 38 at WAS 14WAS +1.5 closed · ATL covered
CHI @ BUFSun, Nov 4 13:00 ETBUF -10.538CHI 41 at BUF 9BUF -10.5 closed · CHI covered
DET @ MINSun, Nov 4 13:00 ETMIN +4.547DET 9 at MIN 24MIN +4.5 closed · MIN covered
GB @ NESun, Nov 4 20:20 ETNE +556GB 17 at NE 31NE +5 closed · NE covered
HOU @ DENSun, Nov 4 16:05 ETDEN +146.5HOU 19 at DEN 17DEN +1 closed · HOU covered
KC @ CLESun, Nov 4 13:00 ETCLE -7.551KC 37 at CLE 21CLE -7.5 closed · KC covered
LAC @ SEASun, Nov 4 16:05 ETSEA +148.5LAC 25 at SEA 17SEA +1 closed · LAC covered
LA @ NOSun, Nov 4 16:25 ETNO -2.557.5LA 35 at NO 45NO -2.5 closed · NO covered
NYJ @ MIASun, Nov 4 13:00 ETMIA +343.5NYJ 6 at MIA 13MIA +3 closed · MIA covered
PIT @ BALSun, Nov 4 13:00 ETBAL +1.547PIT 23 at BAL 16BAL +1.5 closed · PIT covered
TB @ CARSun, Nov 4 13:00 ETCAR +655.5TB 28 at CAR 42CAR +6 closed · CAR covered
TEN @ DALMon, Nov 5 20:15 ETDAL +4.540TEN 28 at DAL 14DAL +4.5 closed · TEN covered
The market line as posted, and how each game finished.

The plays

A play is an angle with a prior record, staked by that record. The model's number alone is never a play.

The angle board mined, walk-forward record behind each

DET @ MIN
WIN DET +4.5 2.0u
  • model fades them by 1.0+, got blown out last week, worse record, in a dome 35-17 (67%)
  • model fades them by 1.0+, got blown out last week, in a dome, on turf 50-31 (62%)
  • model fades them by 1.0+, got blown out last week, division game, in a dome 32-20 (62%)
  • and 3 more angles on this side
LV @ SF
WIN LV -1.5 1.0u
  • model fades them by 3.0+, got blown out last week, worse QB, better coach 16-10 (62%)
  • model fades them by 3.0+, lost last week, worse QB, better coach 51-34 (60%)
  • model fades them by 3.0+, favorite, got blown out last week, worse QB 30-20 (60%)
PIT @ BAL
WIN PIT +1.5 1.0u
  • model likes them by 1.0+, on the road, elite QB, division game 73-41 (64%)
  • model likes them by 1.0+, on the road, better QB, division game 122-77 (61%)
  • model likes them by 0.5+, on the road, elite QB, division game 77-50 (61%)
TB @ CAR
WIN TB +6 1.0u
  • model fades them by 0.5+, underdog, better QB, high total (48+) 22-14 (61%)
TEN @ DAL
LOSS TEN +4.5 1.0u
  • model fades them by 2.0+, on the road, better QB, on turf 88-53 (62%)
GB @ NE TOTAL
LOSS Over 56.0 1.0u
  • over when high total (48+), home team off Monday night 37-16 (70%)
  • over when on turf, home team off Monday night 71-50 (59%)
  • over when home team off Monday night 152-111 (58%)
KC @ CLE TOTAL
LOSS Under 51.0 1.0u
  • under when windy, home team a big underdog 17-7 (71%)
  • under when high total (48+), home team a big underdog 29-13 (69%)
TEN @ DAL TOTAL
WIN Over 40.0 1.0u
  • over when in a dome, low total (41 or under) 89-54 (62%)
  • over when on turf, low total (41 or under) 179-140 (56%)

Read it straight. Five rules frozen 2026-08-28 before Week 1, git-timestamped, append-only. Flat 1u. Historical: the frozen card went 91-83, +6.8u over 2015-2025, 7 of 11 winning seasons; the weather family 362-334-11, +19.1u over 2010-2025. This season was replayed walk-forward: the angles were chosen on the seasons before it and fired on its games, so nothing here was scored on the games that found it. It is a replay, not a published record. 2026 is the first season we posted before kickoff.