Why No. 3 Duke -- fresh off a loss to Clemson -- is still underrated

This may be hard for you to stomach. It's a sentiment not often conveyed to the masses: The Duke Blue Devils are underrated.

Yes, Duke just lost at Clemson, and fell to Stephen F. Austin at home in November. We're aware of the No. 3 AP ranking for a team whose annual expectation is a No. 1 ranking. We're also fully aware of a historically weak ACC. And still, Duke is the best team in the country, and BPI's data and other metrics say it's not that close.

Even after the Clemson loss, Duke remains No. 1 in these widely used predictive metrics: BPI, Sagarin, TeamRankings and Massey. (Duke fell slightly behind Kansas for No. 1 in KenPom on Thursday morning). Why? Most predictive metrics, at their core, use an adjustment to a team's game-by-game scoring margin -- commonly referred to as adjusted net efficiency or adjusted scoring margin. Adjusted net efficiency tells us the average scoring margin of a team, adjusted for opponent and location. Duke is first in the nation with an adjusted net efficiency of 33.9, and Kansas is second at 32.5. Of Baylor, Butler and Gonzaga -- the three other No. 1 seeds in Joe Lunardi's latest version of Bracketology -- Gonzaga is the closest at 7.9 points per game behind Duke.

While that stat would be impressive in most years, this year it is especially so. This season has the smallest standard deviation of adjusted net efficiency at this point in the campaign, going back to 2008. This means that college basketball teams are harder to separate at this point in the season than usual. It is interesting to note that this smaller spread of team efficiency comes from a smaller variance in team offenses, while the spread in defensive ability across teams is largely consistent with where it has been in past seasons. Translation: Duke's lead in adjusted scoring is more impressive considering the parity in what has been a wild year for college basketball.

The eye test may say this Zion Williamson-less Duke team is not as strong as last year's group. But the 2018-19 Blue Devils were more fallible than memories and narratives might suggest. That team lost regular season games with Williamson, including its second loss this season on January 14th (at home to an unranked Syracuse team). Duke's second loss in 2019 also came on Jan. 14. You might also remember scares against UCF and Virginia Tech in the NCAA tournament, prior to a loss to Michigan State in an Elite Eight classic.

But didn't last year's Duke squad have three NBA lottery picks, including the No. 1 overall pick, on its roster? Yes, and while the 2019-20 team does not have a lottery pick at this stage including according to ESPN's latest NBA mock draft, the numbers say this group is every bit as dominant as the one Mike Krzyzewski fielded a year ago.

BPI spins adjusted net efficiency into a predictive metric by considering other factors such as rest, travel, recruiting, coaching, and returning production. With those factors in mind, BPI ranks Duke as the best team by 2.1 points. The only team in BPI history (since 2008) that has had a bigger lead in BPI at this point in the season was the 2015 Kentucky team that started 38-0 and featured nine future NBA players.

Despite featuring Williamson, RJ Barrett and Cam Reddish, Duke struggled from the outside a year ago. Duke was 327th in the country last year, shooting 31% from 3.

This season, Matthew Hurt, Joey Bakerand Cassius Stanley are all shooting over 40% from 3-point land. In addition, Tre Jones has improved his 3-point shooting from 26% last year to 36% this year. As a team, Duke is shooting 37% from 3, which is 40th in the nation. That improvement is all the more impressive considering that on average teams are down 1-2% from 3 after the 3-point line was moved back to 22 feet, 1 inches.

Here's the scary part. Despite improving from the outside, Duke is not in the top 300 in terms of the percentage of shots it attempts from 3. This team should be shooting even more from outside, and when it does, watch out.