Weekly Model Roundup, Nov. 3–9, 2026
Across five leagues from Nov. 3–9, Stats Orb's models picked the winner in 371 of 475 decided games (78%). The NFL had the best week against the betting line at 9-6 (60%); men's…
Read the recap// ratings · stats · projections
Stats Orb turns every final score into team ratings, strength of schedule, game-by-game odds and full-season projections — for the NFL, college football, men's college basketball, the NBA, the NHL and Major League Baseball.
// new · college football featured picks
Only the games where our model and the opening line disagree most. Top picks have hit about 60% against the opening line in back-tests since 2013; nothing that tested below 55% is ever featured.
CHOOSE A PASS
Week 12This week’s picks$49 Rest of the regular seasonWeek 12–13, through championship week$98 Get the picks →Information for entertainment only, not gambling advice. Past performance is not indicative of future results.
// from the articles
Across five leagues from Nov. 3–9, Stats Orb's models picked the winner in 371 of 475 decided games (78%). The NFL had the best week against the betting line at 9-6 (60%); men's…
Read the recapWhy even the best ratings miss by double digits: the typical miss in the NFL, college football, NBA, college basketball, NHL and MLB, and what it means for win probabilities.
Read the guidePoint spreads are sharp forecasts, but they don't explain themselves, cover every game or show a season's outlook. How independent, transparent ratings complement the betting market.
Read the guide01 / what we do
“Who's actually good?”
Every team gets a rating built from who it played and how it did — with an all-margin version that counts every point and a no-margin version that only cares whether you won.
“Who did they beat?”
Strength of schedule, records against the top tier, week-over-week rank movement. The context a win–loss record leaves out.
“What's likely next?”
Win probabilities for every upcoming game, rolled up into the full range of final records — and how likely each one is.
02 / probability, not prophecy
That's not a bad prediction — that's what 57% means. Play the same game a hundred times and the underdog takes its share. Upsets aren't proof the numbers are wrong; they're the numbers working.
That's why Stats Orb shows odds, not guarantees: the chance a team wins a game, the chance it finishes within one game of its projected record, the chance it wins out.
03 / context is everything
Two teams can share a record and have almost nothing else in common. Strength of schedule (SOS) measures how good a team's opponents are, so a 5–0 run through contenders counts for more than a 5–0 run through the bottom of the league.
Our ratings bake that in — and you can sort any league by SOS rank to see who's had the hardest road.
04 / two ratings, one argument settled
Fans have argued about it forever. We publish both answers side by side.
Counts every point. A 42-point win says more than a 3-point win, so dominant teams rise faster. Built for forecasting what comes next.
A win is a win. Only who you beat matters, not by how much — the measure of a résumé, and immune to running up the score.
05 / how a projection is born
Every final score, home or away or neutral site, regular season and playoffs.
Each team's rating is recalculated from everyone it has played, so one upset ripples through the whole league.
Ratings become an expected margin and a win probability for each game still on the schedule.
Those odds combine into the likelihood of every possible final record, plus projected standings.
06 / pick your league
We'd love to hear it.
We use strictly necessary cookies to remember this choice and protect our forms, and — only if you allow it — an analytics cookie that helps us count returning visitors. We never sell or share your personal information. Read our cookie details.
Your browser sent a Global Privacy Control signal, so optional cookies are off.