Given the one-and-done culture in college basketball, rating teams in the preseason is one of the most difficult exercises of any sport. Preseason polls tend to overvalue programs’ reputations, and the historically successful programs often end up on top. With the proper inputs, however, it’s possible to create a predictive rating system that can objectively identify which teams will be strongest for the coming season.
With that in mind, ESPN Stats & Information has created a preseason version of its Basketball Power Index, or BPI, for the 2016-17 season. This automated set of rankings is designed to predict how strong each team will be going forward and then use that information to predict single-game and season outcomes.
Preseason BPI methodology
When discussing which college basketball teams will be best next season, you are likely to hear about Duke’s coach, Kentucky’s recruiting and Villanova’s returning talent. Not coincidentally, these are all important factors when projecting a team’s performance for the coming season.
Preseason BPI has four main inputs: percentage of roster returning (including transfers), past performance of those returning players, composite recruiting rankings (from 247 Sports) and coaches’ past performance. All of these inputs are adjusted for competition faced and, when applicable, pace. For a complete description read our explainer here or scroll to the bottom.
These four factors interact to produce a team’s predicted offensive and defensive rating for the coming season. Those ratings are then combined to form preseason BPI, which represents a team’s projected point differential (per 70 possessions) against an average Division I team on a neutral court (all else equal).
Although these ratings are not perfect, they have been accurate when predicting team strength. Seven of the last nine NCAA champions have ranked in the top six in preseason BPI, with the 2011 and 2014 Connecticut championships representing the outliers.
Let’s run through the 1-through-351 rankings:
The Duke Blue Devils sit atop preseason BPI with by far the highest average offensive projection in the country. Led by AP Preseason All-American Grayson Allen and a superb recruiting class, the Blue Devils should have a perfect mix of experience and young talent this season. With four five-star recruits (Harry Giles, Jayson Tatum, Frank Jackson and Marques Bolden) and our system’s highest-rated group of returning offensive players in the country, a deeper or stronger team heading into the 2016-17 season won’t be found.
Villanova ranks right behind Duke with a projected top-three offense and defense. The Wildcats will have to replace two of their most important players -– Ryan Arcidiacono and Daniel Ochefu -– but the rest of the roster is intact with Josh Hart, Kris Jenkins and Jalen Brunson all back. The Wildcats return 87 percent of available minutes (including transfer Eric Paschall), and those returning players are among the most efficient in the country.
Oregon surprised many by winning the Pac-12 and claiming a No. 1 seed in the NCAA Tournament last season, but the Ducks aren’t going to sneak up on anyone this year. Despite losing two of their top five scorers, the Ducks return nearly everyone else to a team that only got stronger as the season progressed. Assuming Dillon Brooks is healthy, Oregon should be even better than last season.
Two teams that might sneak up on some are NC State and Ohio State. Both teams are unranked in The Associated Press poll but rank in the top 12 of preseason BPI. NC State added two five-star recruits, point guard Dennis Smith Jr. and center Omer Yurtseven, and returns a bunch of talented big men such as Malik-Abdul Abu and BeeJay Anya. Though both five-star recruits have obstacles to overcome -– Smith Jr. is coming back from a torn ACL and Yurtseven is suspended for the first nine games -– they should have an impact later on. Like NC State, the Buckeyes return some efficient big men, but their success under Thad Matta, who has a top-five adjusted net efficiency rating among coaches since 2007-08, is driving their BPI rating.
The Big Ten might be the most crowded conference in the country with its top six teams separated by 3.2 points on a neutral court. BPI favors Purdue (45 percent chance to win a share of the conference title) with the duo of Isaac Haas and Caleb Swanigan clogging the middle, but there are a number of teams that could challenge for the title. Three of those teams are ranked between 16th and 20th in BPI.
The honor of being the top projected mid-major team belongs to Gonzaga. Despite losing All-Americans Kyle Wiltjer and Domantas Sabonis, the Bulldogs should not miss a beat with the addition of transfers Nigel Williams-Goss (from Washington), Jordan Mathews (from Cal) and Johnathan Williams III (from Missouri). Saint Mary’s returns nearly everyone from a team that finished in the top 40 in BPI last season, so the Gaels should challenge Gonzaga for the WCC title.
This section is filled with mid-majors that could make some noise come March. The A-10 race should be interesting with Dayton (25th), VCU (32nd), Davidson (39th) and Rhode Island (44th) in the top 50. Princeton (40th) returns just about everyone and should warrant attention in the Ivy League, while Wichita State should not take too much of a step back (despite losing Ron Baker and Fred VanVleet) as Gregg Marshall is among the most impactful coaches in the nation.
Looking for a team that could take a step forward? Minnesota finished 8-23 last season, but Richard Pitino’s squad returns nearly everyone and added the 32nd-ranked recruit in the ESPN 300, small forward Amir Coffey. The Golden Gophers likely aren’t going to win the Big Ten, but they have a good chance to at least double their win total from last season.
On the flip side, BPI expects Texas A&M and Oklahoma, who met in the Sweet 16 as top-three seeds last season, to take a step back. Both teams have huge holes to fill and neither has spectacular recruiting classes or returning players to fill those lost minutes. Similarly, Maryland is much lower in BPI (65th) than in the AP poll (25th) after losing four of its top five scorers from last season. Although its top scorer, Melo Trimble, returns, he was not particularly efficient last season.
Preseason BPI methodology
Preseason BPI is built on the player and coach level and is meant to predict where a team’s BPI rating will be at the end of the season. In the simplest sense, BPI looks at the percentage of minutes a team is returning, how good those returning minutes have been, how a team will fill its non-returning minutes and the strength of the coach on the sideline. Of course, preseason BPI isn’t that simple, so we’ll take you through the four components in depth.
The first component is the percentage of the roster returning. This is based on the minutes played in each previous season for returning players, including transfers. Each player’s highest percentage of minutes played in a season is used for this calculation to account for injuries and other setbacks. Then BPI looks at if those returning minutes were any good. This is captured with a weighted average of a player’s opponent-adjusted offensive and defensive rating (points per 100 possessions) in their careers.
To fill in the remaining minutes that have graduated or left the program, BPI factors in the average recruiting grade of the incoming class (compiled by 247 Sports) from four recruiting services (ESPN, Scouts, Rivals and 247 Sports). It also gives extra weight to five-star players. Finally, BPI captures a coach’s past performance in terms of his average adjusted offensive and defensive rating since 2007-08 (including games at other schools). A first-time head coach gets a replacement-level rating.
These four components hold different weights depending on how many players are returning, but they are combined in a Bayesian model to produce a team’s preseason BPI.
Obviously there will be teams that preseason BPI “gets right” and “gets wrong,” but over the years it has proved accurate.
