Making Sense of the Madness: A Look at KenPom and Tempo

A general view of the March Madness logo on center court. Maddie Meyer/Getty Images/AFP.

We are a few weeks away from Selection Sunday and the start of March Madness.

Prior to that, the regular season winds down with rivalry games and teams desperately needing one more high-quality win. Then there’s conference tournament season, with automatic bids at stake and one last attempt to post wins to impress the selection committee.

As teams try to get into position on the NCAA Tournament bubble, you’ll start to hear an alphabet soup of analytics and rankings, as cases are made. Here’s the first in a series of posts looking at which models can help as you try to bet basketball’s crazy season.

KenPom

Ken Pomeroy is the godfather of college basketball analytics. He created the website named … or abbreviated … after him: KenPom.com. His rankings of all 358 teams in Division I are generally considered the gospel for analytic-minded people. He also provides a predicted point spread for each game, which are among the most accurate around.

Everyone uses KenPom, so his ratings are more a requirement to get on everyone’s level more than providing an edge at this point. But his method of producing the rankings can help provide value as you try to pick winners.

Tempo Free Statistics

Pomeroy breaks down basketball at a possession level, which helps point out the value of looking at tempo in college basketball. A team may allow 60 points per game, but that doesn’t necessarily mean it’s good at defense. It could simply play at a snail’s pace, limiting the number of possessions—and scoring opportunities—for opposing teams.

Similarly, teams may give up a higher number of points in a game, but their strong defense is masked by the run and gun style they play, which produces a large number of possessions in a game.

KenPom looks at offensive and defensive efficiency, which essentially measures points per possession, to see how good a team is at each end, regardless of how fast or slow they play.

For example, Arizona gives up 65 points per game, while Virginia gives up 60. It would seem that the Cavaliers are significantly better on defense, but UVA plays at the second slowest tempo in the country, while Arizona plays ninth fastest. A Wildcats game typically has 13 more possessions than a Virginia game, so allowing 65 points at that pace is tougher than allowing 60 at Virginia’s.

The per possession numbers show that. Arizona has a defensive efficiency of 90.2, meaning it allows 0.902 points per possession. Virginia is at 98.6, or close to 10 percent worse on a given possession. So KenPom has Arizona as the country’s eighth-best defense, while Virginia is at No. 79.

So if a team draws an unfamiliar tournament opponent who has a stingy points per game, it’s worth looking at the tempo that opponent plays. Is it an Arizona or a Virginia.

On the flip side, Nebraska scores 72.6 points per game, while Villanova scores 73.7. So it seems that both teams have relatively even offenses. But Nebraska plays at the No. 12 tempo in the country, getting 10 possessions more per game than Villanova, at No. 347. Scoring 73.7 points at Villanova’s pace is a significantly bigger accomplishment, and the efficiency numbers show that. Nova is at 118.8, while Nebraska is 102.4.

So, is a high-scoring opponent a Nebraska or a Villanova? The difference can help point toward a potential winner.

Reducing the Noise

Real-world statistics always have noise that may mask the true meaning, and one of the biggest sources of noise in college basketball is tempo. So, while everyone is looking at KenPom team rankings and point spreads, taking the extra step to break down a team’s tempo and efficiency on both ends of the court can help provide an edge, even on a site everyone stops at before making their NCAAB picks.