Sweet 16 Against the Spread Picks: New Betting System Put to the Test

Over the course of the year, we have tracked the performance of two college basketball prediction systems — KenPom and Sagarin — comparing them both against each other and against the point spreads posted by online betting sites. Our findings indicated two things:

1. when the point spread lay somewhere in between the predictions from KenPom and Sagarin, KenPom was significantly more accurate (66%) than Sagarin.
2. when the point spread lay either above or below both the KenPom and Sagarin predictions, the outcome of the game tended to fall on the opposite side of the point spread as the prediction systems, but not as significantly (55.2%).

Our sample size is admittedly small, but it will grow over time, and we’re ready to start putting our findings to the test.

Applying the trends, above, to each of the Sweet 16 games yields the following eight picks. (Games are listed in chronological order.)

#11 Loyola-Chicago Ramblers vs #7 Nevada Wolfpack

Notes: Bettors should recognize that KenPom and Sagarin do not factor in injuries. Their calculations for Nevada are mostly based on how the team performed with PG Lindsey Drew (29.2 MPG, 8.1 PPG, 4.3 APG), who suffered a season-ending Achilles tear in late February. Sportsbooks can and do take injuries into account when setting the point spread.

No word yet on how KenPom and Sagarin factor in the Sister Jean effect.

#7 Texas A&M Aggies vs #3 Michigan Wolverines

Notes: It’s not as significant as Nevada being without Lindsay Drew, but Texas A&M lost sixth-man Duane Wilson (24.0 MPG, 9.0 PPG, 4.0 APG) in mid February.

#9 Kansas State Wildcats vs #5 Kentucky Wildcats

Notes: While this is technically a neutral site game, Kentucky will have a de facto home-court advantage in Atlanta. That, plus their national popularity, is likely motivating the inflated spread. (More on this below in the Clemson/Kansas section.) Additionally, the status of Kansas State’s leading scorer, Dean Wade, is in question.

#9 Florida State Seminoles vs #4 Gonzaga Bulldogs

Notes: This will be the first time FSU has played in the Pacific Time Zone all season. Despite being in the West region of the bracket, their opening games were in Nashville, TN.

#5 Clemson Tigers vs #1 Kansas Jayhawks

Notes: This game features the biggest gap between the KenPom projection and the point spread. The KenPom model has not been a fan of Kansas this year. Despite being a #1 seed, they are only ninth in overall efficiency on the site. They are fourth on Sagarin, however. They are also a blue blood with a die-hard fan base, which tends to inflate the spread in their favor. Sportsbooks recognize that fans want to bet on their own teams, leading to squads like Kansas, Kentucky, and Duke laying more points than less popular schools, even if they have the exact same resume.

Kansas’ inflated spread is also motivated by the location of the game. Like Kentucky’s game in Atlanta, Kansas will have a de facto home-court advantage playing in Omaha.

#5 West Virginia Mountaineers vs #1 Villanova Wildcats

Notes: West Virginia reserve Wesley Harris (20.5 MPG, 5.5 PPG, 3.6 RPG) could miss the game with a concussion. While he does log over 20 minutes per game, his uncertain status likely has no impact on the point spread.

#11 Syracuse Orange vs #2 Duke Blue Devils

Notes: Duke vs Syracuse is the only Sweet 16 game that is a rematch from the regular-season; Duke beat Syracuse at home, 60-44, on February 24.

Also note that the majority of Duke’s defensive efficiency numbers on KenPom stem from a time when they were playing man-to-man. Duke has since switched to zone, for the most part, rendering their poor early-season defensive performances somewhat inconsequential.

#3 Texas Tech Red Raiders vs #2 Purdue Boilermakers

Notes: This is another game with significant injury concerns. The status and effectiveness of Purdue center Isaac Haas is not clear. KenPom and Sagarin do not account for this. The point spread does.