Final Four should be an all-timer: Stunning metrics make these teams special

ATLANTA – All hail Ken Pomeroy; the king of the calculator, the seer of the stats, the architect of the analytics and the executive of efficiency.

For more than a decade, a website run by a former meteorologist and atmospheric sciences instructor has been the numbers-based Bible of college basketball. Pomeroy’s formula ranks teams by a variety of different metrics including offensive efficiency, defensive efficiency, tempo and luck. But his top-line number assigned to each team, calculated from data over thousands of possessions across the season, is pretty much the gold standard in terms of predicting on a game-to-game basis which team is more likely to win.

And this year?

Pomeroy absolutely nailed it. After nearly 6,271 college basketball games this season, his numbers didn’t just have Duke, Houston, Florida and Auburn as the four best teams in the country; it had those teams ahead of the pack by a wide margin.

In fact, since the beginning of his data in 1997, just 10 teams have had a total Net Rating of +35 or more in his formula. Three of them won the championship, two lost in the championship game, one in the semifinals.

The other four? They’ll be in San Antonio next weekend.

“The four teams that advanced I think are the four best teams in the country,” Auburn coach Bruce Pearl said. “That doesn’t always happen.”

In fact, it almost never happens.

Last year, the Final Four included a team ranked 14th (Alabama) and 45th (NC State). The year before that, 14th (San Diego State), 17th (Florida Atlantic) and 24th (Miami). That’s college basketball. That’s the tournament. Small margins, almost no room for error.

Even KenPom’s computers can’t make perfect predictions about a game played by humans. That’s why, when people really started to pay attention to Pomeroy’s rankings around the early 2010s, he would take some flak on social media if a highly-ranked team in his rankings got upset.

But what the college basketball world has come to understand over the years is that analytics matter in building a narrative about what you’re good at, what your opponent is good at and whether teams are better or worse than what their record or their ranking in the human polls suggest.

And when the numbers are as overwhelming as they were this year, with four teams so clearly separated from the rest, it’s probably a good idea to pay attention. (It’s also worth noting that the teams ranked Nos. 5-8 in KenPom made the Elite Eight and 12 of his top 16 made the Sweet 16.)

So what do the numbers tell us going forward?

They tell us that Duke is the highest-rated KenPom team in 16 years, second only to the 1999 Duke team that went 37-2 and lost to UConn in the national championship game. Again, that means based purely on numbers, Duke is better than the UConn team that won it all last year, better than either of the Florida teams that went back-to-back, better than the 2012 Kentucky team with Anthony Davis, better than either of Kansas’ champions under Bill Self and better than the 2018 Villanova squad that had three players who are all currently starting for NBA teams. ...

Save Story