What's Changed
Every update is logged here with a timestamp, newest first, so you can see how the picture moved during the week.
Every update is logged here with a timestamp, newest first, so you can see how the picture moved during the week.
Death Valley at night, LSU laying double digits to a Clemson side nobody has seen play (a major book, 1:40 AM ET Wed). Ten points between two preseason-ranked programs is the market saying it believes in one roster's continuity a lot more than the other's — the exact thesis our Week 0 leans went 4-4 testing. What would make us bet it: a verifiable returning-production gap that AGREES with the number's direction plus a settled QB situation on the side we'd take. What would kill it: the line drifting past -10.5 on air, or our post-mortem sizing rule (Angle 5) telling us team leans don't deserve the bullet.
Both teams already played on our watch, which makes this the only Saturday game where we hold verified 2026 tape: UNLV lost 27-21 to Memphis with Jackson Arnold throwing for 291, and Hawaii lost 37-27 at Stanford with Micah Alejado attempting FIFTY-ONE passes. Our Jai'Den Thomas under-14.5-carries read cashed exactly as scripted (12 carries — the committee is real). What would make us bet it: prop lines still priced off 2025 roles that the Week 0 tape just contradicted — that shape went 8-for-2-with-2-voids for us on Saturday. What would kill it: lines that already ate the Week 0 box scores, or our data platform's college windows staying dark past Friday.
Baylor–Auburn in Atlanta at -7.5, UCLA–Cal at -1.5, Western Kentucky–Nevada at -2.5 (all a major book, 1:40 AM ET). Short-number openers are where books hedge and where information moves lines fastest. What would make us bet one: a concrete personnel edge the number hasn't priced — a named starter change, a verified injury, a returning line versus a rebuilt one. What would kill it: nothing but vibes by Friday night; a 1.5-point line with no thesis is a coin flip with commission.
Ohio State -50.5, Texas A&M -40.5, BYU -51.5, Maryland -45.5, Georgia -46.5, Tennessee -46.5. Week 0 said it plainly: our own Look Ahead called these 'unbettable noise' and the two big-spread legs we published anyway (FSU -30.5, SJS +39.5) split — the 43-confidence dog cashed, the 48-confidence favorite didn't cover. Forty-point September spreads are backup-quarterback variance in a costume. What would make us bet one: nothing, and this angle exists so Saturday-morning us remembers that.
Week 0 is graded and the verdict is in: team leans 4-4 (a coin flip, exactly what roster-churn season deserves), role-based player unders 8-2 with two voids — Trader, Thomas, Hall, Hedden, Barfield and the Maiava blowout-script under all cashed; the two misses (Miller, Payne) were both RUSHING volume in blowouts, the one script we said would feed overs. The lesson writes Saturday's plan: props are the flagship, leans get sized like the coin flips they graded as, and rushing-volume unders in projected blowouts get extra scrutiny. What would change this: nothing before Saturday — the sample said what it said, and 10 legs of it is still just 10 legs, which the sizing will also remember.
Thursday: Colorado–Georgia Tech (-6.5), the one real midweek game, plus UMass–Rutgers (-28.5). Friday: Miami -24.5 at Stanford and Fresno State at USC (-22.5), both with fresh Week 0/1 tape to price against. Sunday: Wisconsin–Notre Dame at Lambeau (-20.5) and Louisville–Ole Miss (-7) in Nashville; Monday closes with SMU -3 at Florida State — a ranked road favorite against the team that just handled a 30-point number badly. MLB runs full Saturday boards all weekend. WNBA: dark until ~Sep 17, FIBA World Cup — the league's own break, not our omission. Every line above re-verifies before anything publishes; four-day-old numbers are context, never conviction.
One email a day: every leg, every Edge Gap, and the argument against our own top pick. When a leg misses, that's in there too — same email, same prominence.