Explaining KenPom Ratings and Rankings

Given the next couple articles I’m writing, I figured it would be worth it to put something up explaining the KenPom ratings that so many basketball analysts reference. I’m sure many of you are familiar with them, but for those that aren’t, hopefully this can clear some things up, or even introduce something new.

So KenPom was started by Ken Pomeroy (a Hokie in fact) in 2002. It’s essentially a resource that provides in depth, analytical stats for college basketball. The ranking is based off of offensive rating and defensive rating for each team. A team’s offensive rating is the number of points, on average, they score per 100 possessions. For instance, last season Virginia’s offensive rating was second in the country at 123.4. This season, they’re 120th in the nation with an offensive rating of 102.4. That is, clearly, a significant drop. Defensive rating is, logically, the amount of points a team gives up per 100 defensive possessions. Last season, the Cavaliers’ 89.2 defensive rating was 5th in college basketball. This season, they boast the top rating in the country at 80.5.

The KenPom ranking is based off the differential, or net rating, between a team’s offensive and defensive ratings. For example, last year Virginia was #1 in the KenPom rankings as their net rating was 34.22 (123.4 - 89.2).This year, their 21.84 net rating is 10th (102.4 - 80.5).

As this is essentially an advanced metric for winning margin, it appropriately corresponds very accurately with the actual results. In fact, looking back through the years that Ken Pom has been tracking CBB (since 2002), the National Champion has also been the #1 team on KenPom 10/18 years, and has been in the top 3 another 5 years.

Here’s the breakdown of the National Champ and their KenPom rank:
2002: Maryland 2
2003: Syracuse 8
2004: Uconn 2
2005: North Carolina 1
2006: Florida 1
2007: Florida 2
2008: Kansas 1
2009: UNC 1
2010: Duke 1
2011: Connecticut 10
2012: Kentucky 1
2013: Louisville 1
2014: Uconn 73 (outlier)
2015: Duke 3
2016: Villanova 1
2017: Carolina 3
2018: Villanova 1
2019: Virginia 1

Basically what I’m getting at is that this ranking is an appropriate tell of where teams are and where they may be going, which is why myself and so many others rely on it so heavily. Anyways, hope this helps!

9 Likes

Clear and useful explanation. Thanks!

3 Likes

Very interesting! Thanks for this

4 Likes

You might also mention that the adjusted offensive and defensive efficiency numbers that are most heavily used include adjustments for the number of possessions (pace) in each game and the strength of each opponent.

The National Champ Kenpom ranks you list are the final values, after the champion has been crowned, and so naturally are boosted for the ultimate champion. as far as predictive value, the pre-tournament values (i.e., as of Selection Sunday night) are probably better to review. These are added below in parentheses.

National Champion was #1 team on KenPom on Selection Sunday 3/18 years, and was Top 3 another 7 years

2002: Maryland “2”* (4)
2003: Syracuse 8 (20)
2004: Uconn 2 (5)
2005: North Carolina 1(2)
2006: Florida 1 (6)
2007: Florida 2 (3)
2008: Kansas 1 (1)
2009: UNC 1 (3)
2010: Duke 1 (2)
2011: Connecticut 10 (15)
2012: Kentucky 1 (1)
2013: Louisville 1 (2)
2014: Uconn “73 (outlier)”* (25)
2015: Duke 3 (6)
2016: Villanova 1 (5)
2017: Carolina 3 (3)
2018: Villanova 1 (2)
2019: Virginia 1 (1)

(note: 2002 final ranking for UMd was 3, not 2 - and 2014 UConn was 15, not 73)

2 Likes

Great points! It’s not necessarily a prediction stat but rather a stat that reflects a team’s play so far. Also, thanks for the fact check, don’t know what I was looking at for 2014 UConn :woozy_face:

1 Like