There are lies, damned lies, and then there are statistics. And when it comes to Virginia basketball, traditional statistics horribly misrepresent the team’s actual performance.
Just turn to the Cavaliers’ home page on ESPN. Prominently featured are a selection of four commonly cited stats: points per game, rebounds per game, assists per game and field goal percentage. You’ll notice Virginia ranks no higher than No. 34 nationally in any of the four and doesn’t even break the top 100 in points and assists per game.
Those numbers don’t justify a No. 2 ranking in the AP and Coaches Polls, and the reason is that those stats mean almost nothing. It’s time to set the record straight.
Unless you’ve been living under a rock or really don’t like sports, you’ve probably heard of sabermetrics and the recent proliferation of advanced sports analytics. The basic idea is that traditionally cited statistics in sports are not nearly as useful as we once thought. “Moneyball,” the Michael Lewis book and later Sorkin-scripted movie, launched the concept into popular culture, but it had long before been on the radar of front offices and sports geeks.
To get a sense of why you shouldn’t trust traditional statistics, look no further than the Virginia Cavaliers. The team is averaging 68.3 points per game and 13.7 assists per game, which rank 160th and 104th in Division I, respectively. However, saying our offense is not among the top 100 in the country is ridiculous, given the simple fact that the team plays at one of the slowest paces in all of basketball.
Part of Virginia’s game plan is to milk the shot clock for all its worth, which inherently means the team has far fewer opportunities to score than faster tempo teams. Ostensibly, giving up scoring opportunities may same silly, but taking time on the offensive end means the opposing team’s offensive possessions are in turn more limited. For that reason, a much better indicator of offensive success would be points per possession — commonly referred to as offensive efficiency.
Of course, the converse is equally as true. Virginia ranks first in Division I in scoring defense, allowing a stingy 49.1 points per game. But given that Bennett’s crew likes to play at such a slow pace, teams don’t have the opportunities to score like they would against an average tempo team. As with offensive efficiency, a better metric for scoring defense would be points allowed per defensive possession, meaning that first overall rankings is likely an overstatement of how good our defense actually is.
There are similarly disconcerting issues with almost every statistic in basketball. While field goal percentage may seem like a good indicator of a team’s offensive production, the number is not of nearly as much concern to teams that dominate the offensive boards. Similarly, offensive rebounds don’t matter if a team is making most of its shots or frequently getting to the free throw line. When it comes to winning games, the only thing that counts is putting points on the board before a possession ends and preventing your opponent from doing the same.
With all that in mind, it’s worth looking at how the Cavaliers fare in statistics that actually matter — ones that influence game outcomes.
Ken Pomeroy, a hero among college hoops fanatics, runs an increasingly popular website that specializes in advanced metrics for NCAA basketball. KenPom rankings are among the most respected of these metrics and embrace the mentality that points per possession is much more important than points per game or any other raw statistic.
So despite ranking 160th in points per game, Virginia ranks 6th in KenPom adjusted offensive efficiency — the number of points scored per possession adjusted for opponent quality. Why the huge disparity between the two stats? KenPom has Virginia ranked 350 out of 351 when it comes to adjusted tempo — the number of possessions per game the team would have if it played a team with an average tempo.
Remember when I said Virginia’s slow pace likely overstated their top ranked scoring defense? I was right — kind of. KenPom has Virginia ranked second in adjusted defensive efficiency, not first. That means unlike the Cavaliers’ offensive raw stats, the defensive numbers aren’t nearly as misleading.
Together, Virginia is the only team in the nation with a top 10 offense and a top 10 defense, according to Pomeroy.
KenPom also features a number of highly interesting individual stats. Just as with team stats, individual numbers tend to punish players on teams that play at a slow pace. But on top of that, individual numbers punish players on teams with balanced offenses with lots of ball movement. If every starter is a top-notch player, individual numbers aren’t going to be as flashy as those of an average player on a bad team.
The KenPom offensive rating statistic — based on a formula originally developed by sports statistician Dean Oliver — takes those concerns into account. The statistic captures points generated by a player during possessions in which the player was involved in the action that ended the possession, which could be a made shot, a missed shot, an assist or a turnover. Four Cavaliers rank in the top 200 nationally for offensive rating, including junior guard Justin Anderson who ranks No. 24. Anderson doesn’t even crack the top 100 in points per game.
Perhaps the most exciting thing for a Virginia fan on KenPom is his player of the year standings, which rank the top players in basketball based on offensive and defensive value. Three of the top 10 players are Cavaliers — junior forward Anthony Gill places third, while junior guards Anderson and Malcolm Brogdon rank fifth and seventh, respectively. Those three are in the company Wisconsin’s Frank Kaminsky, Arizona’s Stanley Johnson, Duke’s Jahil Okafor and Kentucky’s Karl-Anthony Towns, among others.
Now for the really good stuff: the national rankings.
KenPom ranks all 351 Division I teams using what he calls pythagorean winning percentage, or what he defines as a “team’s expected win percentage against an average Division I team.” It’s purely a function of adjusted offensive and defensive efficiency. Using the metric, Virginia sits at second in the nation at .9803, just .004 behind No. 1 Kentucky. The Cavaliers have fluctuated between 1 and 2 during the past two weeks, but gave up some ground with the close win against Virginia Tech Sunday.
ESPN calculates a similar statistic called the Basketball Power Index, which was developed by Dean Oliver. The availability of specific information factors that contribute to the ranking is very limited, but like KenPom, BPI adjusts for pace of game and opponent quality. However, BPI also factors in injuries and provides diminishing returns for blowouts, which KenPom does not.
In BPI, Virginia also sits at second, again behind Kentucky. However, Kentucky has a bit more of a lead here than it does in KenPom’s analysis.
This all seems a bit anticlimactic — after all of this explanation, Kentucky and Virginia are still one and two, just like they are in the AP Poll, the Coaches Poll and almost every other ranking you could find online.
There a few notable take-aways, though. First, don’t let anyone tell you Virginia is overrated, as the numbers back up the rankings. Second, the fact that the Cavaliers have been swapping top spots with the Wildcats in KenPom indicates that we’re just as good as they are, and any potential match up could well be the equivalent to a coin flip. And finally, don’t you just feel smarter thinking about rankings in this way, rather than blindly trusting the eyes of sports reporters?
I also hear we have a game coming up against a pretty good basketball team this Saturday, so let’s see if those rankings don’t move us up to No. 1 in less than a week’s time.
Matt Comey is a Sports Columnist for the Cavalier Daily. He can be reached at m.comey@cavalierdaily.com