Futures bettors take note: the NCAA Selection Committee is providing a preview of the top 16 teams this coming Saturday. Before you plan on how to take advantage of the rankings, here's what to expect.
This coming Saturday (Feb. 11), the NCAA Selection Committee will present its first ever in-season preview of the Division 1 Men's Basketball Championship bracket. It will unveil the top-four teams in each region at this point of the season. The seedings are not etched in stone, but the announcement is intended to “give fans a glimpse to what the men's basketball committee is thinking at this point of the season,” according to Michigan State AD and chair Mark Hollis, “creating a buzz as we look towards Selection Sunday.” What can bettors take away from the broadcast? Plenty.
Firstly, there is no telling what 16 teams will make rank in the top four 29 days after the announcement when the final brackets are determined. The last two weeks has seen dozens of upsets; the blue bloods are vulnerable. Yesterday (Feb. 4), in fact, tied a record for the most top-10 AP teams to lose on a single day with six. They include No. 2 Baylor Bears, No. 3 Kansas Jayhawks, No. 5 Arizona Wildcats, No. 7 West Virginia Mountaineers, No. 8 Kentucky Wildcats, and No. 9 Virginia Cavaliers. The Jayhawks witnessed a 51-game home winning streak snapped, and John Calipari's Wildcats have dropped three of four for just the second time in five years. It is probably not cliché this year to argue the tournament is wide open.
Secondly, we know the committee is going to lean on the Rating Percentage Index (RPI). This has proven an advantage for bettors and modelers for years. Created in 1981, the RPI is a measure of strength of schedule and how a team does against that schedule in D-1 contests only. Its biggest flaw is that it does not take margin of victory into account, making its predictive accuracy flimsy. It's a results-oriented system. A 2013 study by the Harvard Sports Analysis Collective showed ratings that incorporated point differential were much stronger in this regard. Ken Pomeroy's rankings, for example, proved 3 percent better picking winners on a game-by-game basis in the tournament between 2007-12. Where models that employ margin of victory shined in comparison to RPI is determining the Final Four and tournament winner. The RPI correctly predicted 9 of the semi-final contenders in the five-year period and zero winners based on it's ratings entering the tourney; KenPom notched 10 in the Final Four and three champions. Other models proved even more predictive.
The NCAA held a meeting in Indianapolis two weeks ago to discuss entering new metrics into the selection process to make it more accurate and fair. Predictive models like KenPom, Sagarin, and BPI are likely to enter the fray, as the committee creates some sort of composite index combining them with the results-oriented RPI. But any change in protocol will have to wait until next year as the NCAA conjures up and implements a new model.
Baylor, Villanova, Louisville, and Kansas sit atop the RPI as of publication. Look for at least three of these four as No. 1 seeds in Saturday's announcement. If any hold the top spots in mid March is anyone's guess. Futures bettors should likely avoid wagering the committee's top team from winning it all, both Saturday and when the final seedings are announced.