Week 9 review: best calls, bad beats and the trends that moved

Week 9 is in the books. The model's Strong picks - the model is 6+ points off the market on the spread or total - went 6-8-1, and all of the model's total picks went 2-7. Spread picks went 10-14-1, reported apart.

Best calls

• Colorado State at Utah State: Under 60.5. Final: Colorado State 14, Utah State 31.
• Georgia Tech at Pittsburgh: Pittsburgh -8.5. Final: Georgia Tech 11, Pittsburgh 34.
• Oklahoma State at Iowa State: Iowa State +9.5. Final: Oklahoma State 49, Iowa State 51.
• UConn at Air Force: Over 52.5. Final: UConn 26, Air Force 35. Backed by Model 6-10 pts above the total (18-12).
• Northwestern at Oregon: Oregon -3.5. Final: Northwestern 26, Oregon 34.

Bad beats

• Illinois at Maryland: Maryland +2.5 missed by 0.5. Final: Illinois 20, Maryland 17.
• Tulane at Charlotte: Tulane -8.5 missed by 1.5. Final: Tulane 38, Charlotte 31.
• Ohio State at USC: Over 50.5 missed by 2.5. Final: Ohio State 18, USC 30.
• Oregon State at Fresno State: Under 57.5 missed by 2.5. Final: Oregon State 22, Fresno State 38.

Trends on the move

• New model struggle: Model picks the underdog to win (spread), the model now 25-40.
• New model struggle: 14-21 pt spread (spread), the model now 24-36.
• New model struggle: Clear offensive edge (total), the model now 13-21.
• New model struggle: Model 6-10 pts above the total (spread), the model now 12-18.
• Dropped out: Average-paced game (spread) cooled off and no longer qualifies.
• Dropped out: Spread hasn't moved (total) cooled off and no longer qualifies.
• Dropped out: Model 6-10 pts above the total (total) cooled off and no longer qualifies.

What we're working on

• Boston College at Duke: the model had Duke by 9.8; Duke won by 45. The biggest input was offensive efficiency edge.
• Baylor at UCF: the model had UCF by 0.8; Baylor won by 32. The biggest input was defensive efficiency edge.
• Wisconsin at Iowa: the model had Iowa by 1.2; Iowa won by 30. The biggest input was defensive efficiency edge.
• Early-season calibration: the model has run low on the home margin in weeks 1-2, and we are testing whether its early-season adjustment is sized right.
• Line movement: how far a line moves between open and close is now a tracked situation, and we are watching whether it earns a place in the trends.

The Thursday preview has this week's Strong picks. See this week's picks.

Trends describe what has happened; they are not guarantees. Not betting advice.