Preseason BPI rankings like Michigan State, skeptical about Memphis
Michigan State is a favorite, but not the only favorite.
The Spartans, recently placed atop the preseason AP poll, have a 9% chance to win the national title according to ESPN’s Basketball Power Index, but that’s tied for the most with a somewhat surprising second team: Florida.
The Spartans and Gators rank No. 1 and No. 2 in BPI’s preseason rankings, released Friday. And in terms of chances win it all, they’re the co-favorites.
For the unfamiliar, BPI is our quantitative model that rates every D-I basketball team and projects the entire season, from November straight through the Final Four.
A full breakdown of the machinations of BPI can be found right here, but the abridged version is that in the preseason, ratings are built on four factors:
Quantity of experience on roster (including transfers)
Quality of that experience
Recruiting rankings for incoming freshmen, with an extra emphasis on five-star players
Coach’s past performance
Then we use that rating to simulate the remainder of the season. Once games are played, ratings are updated based on team performance, adjusted for opponent.
Based on that criteria, it’s easy to see why the system has fallen for these two teams.
Let’s take Michigan State first, which brings back perhaps the best player in college basketball: Cassius Winston. The Spartan star finished first in our individual win shares metric among all college basketball players last season — ahead of Zion Williamson and Jarrett Culver — thanks in part due to an opponent-adjusted assist rate that ranked only behind Ja Morant.
And Winston isn’t alone, either. He’s joined by junior forward Xavier Tillman, ranked 12th among players with at least 800 minutes played in our opponent-adjusted defensive rating. One factor BPI is not considering is Joshua Langford’s foot injury, which Tom Izzo said the team would reevaluate in January.
The presence of players like Winston and Tillman, in addition to Izzo’s track record, is enough to offset the lack of a five-star recruit in this year’s class for the Spartans, though they do have a couple of four-stars added to the roster in forwards Rocket Watts Jr. and Malik Hall.
Florida, on the other hand, is all about the newcomers: most notably senior big Kerry Blackshear Jr., who transferred in from Virginia Tech. Last season as a Hokie, Blackshear helped lead a team (along with Nickeil Alexander-Walker) that BPI thought was the 11th-best in the country. He finished 17th in our win shares metric.
In Gainesville, Blackshear should lead a Gator team that also is bringing in a pair of freshman who could make an impact: F Scottie Lewis and G Tre Mann.
So what does being No. 1 and No. 2 in BPI translate to in terms of projections? A 55% and 63% chance to win at least a share of the Big Ten and SEC, respectively. Notably, BPI makes Florida a fairly significant favorite to win the SEC over Kentucky, which has just a 27% chance to earn a share of the conference title.
Both teams are also virtual locks for the NCAA tournament, naturally, and — before a basket has been made — are already better than a 50/50 shot to reach the Sweet 16. The Spartans and Gators’ chances to reach the Final Four are 26% and 24%, respectively, and each has a 9% chance to win the national title (Duke, at 7%, is next-best). Technically Florida has a slightly better chance to win the national title by a few tenths of a percentage point, but the difference can be explained by simulation variance, so the two teams really ought to be thought of as co-favorites.
For what it’s worth: teams in the top four in preseason BPI have won the national championship in 10 of the last 12 years, with both UConn titles as the exceptions (note that this includes seasons that were back-tested in the model’s creation). This season’s top four: Michigan State, Florida, Virginia and Villanova.
It seems only right that we give Virginia some space here, especially because BPI isn’t anticipating as big of a drop-off as most.
While Tony Bennett lost key playmakers such as Kyle Guy, DeAndre Hunter, and Ty Jerome, he has coached the best defense year-in and year-out the past 10 seasons. BPI looks at a coaches’ historical average on offense and defense and historical consistency on both ends of the court. The 2018-19 Cavaliers had their most efficient offense under Tony Bennett, a large reason why they were able to break through and win the title.
With less experience, expect the Cavaliers to lean again on their defense and hope Mamadi Diakite has a breakout season. There has not been a repeat champion since the Florida Gators in 2006-07; BPI gives Virginia a 7 percent shot to pull it off.
In each of the past two seasons, BPI has been down on the Blue Devils relative to the preseason polls. But actually, that is not the case this year.
Duke is 5th in BPI and 4th in the AP Poll. The model has found that elite recruits don’t predict success as much as a high percentage of returning minutes of quality players. Duke brings in three ESPN 100 five-star recruits but this year has the luxury of returning Tre Jones. Duke also brings back role players Javin DeLaurier, Alex O’Connell, and Jack White, who each had over 500 minutes last season.
With more of a mix between experience and highly touted recruits this season, expect Duke to be fighting for a No. 1 seed in March (25% chance per BPI) and making another run at a national title.
There’s a 22.69% chance the national championship winner hails from the Big Ten. And also a 22.69% chance it hails from the ACC. Weird.
That’s probably better news for the Big Ten than the ACC, given the latter’s reputation as the top basketball conference.
BPI is high on the Big Ten though, and it actually rates slightly ahead of the ACC in average BPI rating (+6.8 vs. +6.7) this preseason. Both trail the Big 12 in this metric, though the Big 12 typically tops this list because it has fewer teams and lacks cellar-dwellers like Northwestern or Wake Forest that drag down the conference average — whether those teams ought to be involved in a conversation about conference strength is debatable. The Big Ten also leads the way in terms of most top-25 teams, with 7, besting the ACC (5).
We’ve already covered Michigan State, but BPI also features Wisconsin at No. 6 (more on that in a bit), Purdue at No. 11 and Maryland, Iowa, Penn State and Ohio State between 19-25.
In terms of average BPI, after the Big 12, Big Ten and ACC comes the Big East, SEC, AAC and Pac-12, in that order. But all is not lost for the Pac-12, despite a low average rating. Last season we noted that Arizona State was the highest-ranked team from the conference in the preseason … at No. 45 overall. But this season the conference boasts a top-10 squad and a title contender in Oregon, which has the best chance to win at least a share of its conference title among major conference teams. The Ducks return a top contributor in point guard Payton Pritchard but have also added two significant recruits in C.J. Walker and N’Faly Dante — though Dante is currently ineligible and that is not accounted for by BPI.
While the Ducks give the Pac-12 hope, it’s not all great news for the conference: it’s still projected to put fewer teams into the NCAA tournament (2.7) than the Atlantic-10 (3.1), for example.
Perhaps BPI is making up for our preseason Football Power Index (FPI) that had the Badgers 38th in the preseason despite being 19th in the AP Poll.
So what is BPI seeing that the voters (and Wisconsin fans themselves) are not? Experience. The Badgers finished 2018 No. 17 in BPI and have the highest possible experience grade in the preseason. The 23-11 record in 2018 was against the fourth-hardest schedule. Despite losing a star in Ethan Happ, Wisconsin returns a total of 4640 minutes from last year’s squad plus will get transfer Micah Potter from Ohio State later in the season.
Former Wisconsin prep star Trevor Anderson will also have his first full season with the Badgers after transferring from Green Bay and suffering a season-ending knee injury after eight games last year. Greg Gard has consistently produced efficient offenses and that consistency plus player experience reduces BPI’s variance around Wisconsin. This is one of the bigger outliers in the preseason rankings and BPI could be dead wrong, or we may be saying “told you so” in 5 months. Either way, we’ll find out a lot about the Badgers between now and Feb. 5, when they wrap up the toughest 12-game stretch of any team in the country with a game at Minnesota.
The Cyclones lost one of our NBA draft model’s darlings in Talen Horton-Tucker, along with Nick Weiler-Babb, Marial Shayok and Lindell Wigginton from its 2018-19 squad that earned a No. 6 seed in the tournament but flopped in the first round.
BPI is banking on Iowa State’s returning talent and incoming transfers to make up the gap. Sophomore guard Tyrese Haliburton — who BPI is particularly bullish on in part due to efficient scoring despite low usage — and senior forward Martin Jacobson headline the former group. Rasir Bolton is the key incoming transfer out of Penn State, where he averaged 12 points per game as a freshman.
BPI’s love for Iowa State goes so far that it actually makes the Cyclones the Big 12 favorite (barely, but still), a fairly shocking take over traditional powerhouse Kansas and threatening contender Baylor. As often is the case in the Big 12, however, there are a whole cluster of teams at the top: and after Iowa State Big 12 teams make up three of the next six teams in BPI.
Penny Hardaway has the No. 1 recruiting class and a lot of energy around his program. BPI is not buying it.
The model is built predictively on what we knew about each D-I team heading into the season and how each team fared that season. BPI typically knocks teams that rely solely on highly touted recruits, as they have not fared as well as experienced teams with good players. .
So how is it that Duke teams largely comprised of talented freshman crack the preseason BPI top 15 and Memphis is nowhere close?
Duke has Coach K, who has a great track record that’s built into the model. In comparison, we have only one season of information on Hardaway, and Memphis finished 64th in that season. It also does not help that its few returning minutes have only an average offensive efficiency of 99.5. That said, Memphis is in an unusual situation. Typically, coaches with one season of D-I coaching experience don’t land No. 1 classes. Should Penny Hardaway be the first of his kind, BPI will adapt quickly.
Myles Powell racked up over 23 points per game last season in a year in which Seton Hall reached the NCAA tournament. Now Powell is a preseason All-American alongside names like Cassius Winston and Markus Howard and Seton Hall was picked by Big East coaches to win the conference. Despite all that, BPI remains skeptical.
Though most of the Pirates — Powell included — are back for another run and BPI has a general affinity for experience, the model just isn’t overly impressed by the overall quality of what’s coming back. Seton Hall ranked 55th in opponent-adjusted net efficiency last season. BPI thinks they’ll improve from that, but not into the top 25.