Can ‘Press Virginia’ Slow Down the Wildcat Attack?

Eric Schmidt
Analyzing NCAA Basketball with GCP
4 min readMar 23, 2018

Authored by: alokpattani@google.com

Four more down.

As the second week of the NCAA Tournament progresses, the teams that survive inch ever closer to San Antonio. Given the wild ride this year so, some of the basketball faithful among us may be clinging to time-honored traditions — and time-honored metrics. And while we’ve been flexing BigQuery muscle around some new stats and analyses that have never been attempted before with public data, we’ve also been putting it to the test on foundational concepts, too.

With the NCAA and SportRadar data set on hand, let’s see what tonight’s Villanova-West Virginia matchup might have in store.

Background

A team’s offensive and defensive efficiencies are typically measured by looking at points per 100 possessions (scored and allowed). Evaluating teams on a per-possession basis accounts for the 351 Division-I teams playing at different paces, and is a way to capture how well a team is playing, rather than the volume of points alone.

Going deeper, we can look at four core facets of the game that contribute to offensive and defensive efficiency, known as the Four Factors. Initially developed by basketball analytics pioneer Dean Oliver and brought to the college game by basketball analyst Ken Pomeroy, the Four Factors are:

  • Shooting (measured by effective FG percentage — FG percentage, adjusted for 3s)
  • Turnovers (measured by turnovers per possession)
  • Rebounding (measured by rebounds per available rebounds)
  • Getting to the Foul Line (measured by free throws made per field goal attempt)

There are really eight Factors — four on offense, four on defense — and studies have shown these eight metrics to be fairly independent and cumulatively comprehensive. That is, they represent different skills at the team level, and they combine to explain almost all the variation in offensive and defensive efficiency.

Many NBA and college teams use the Four Factors as key performance indicators for how they are playing on both ends of the court. But with a few tools and the NCAA dataset, we can do it, too.

Analysis Process

We read data from BigQuery into R using the bigrquery package. We then adjusted for schedule strength (most importantly opponent strength, but also home/road) by using ridge regression, which we implemented in R using the glmnet package. Each statistic is adjusted for opponents’ strength on the other side of the ball, e.g. a team’s offensive rebounding is adjusted for how good their opponents are at defensive rebounding. This type of adjustment is quite important in college basketball stats, since the 351 Division I teams play varying levels of competition throughout the season.

Finally, we used Hadley Wickham’s ubiquitous ggplot2 package, the standard for data visualization in R, to generate a few graphics.

With these tools in place, we calculated every Division I team’s efficiency and Four Factors for the 2017–18 season, ranked them, and visualized the results. Here’s how Villanova and West Virginia stack up.

Offense, Meet Defense

Villanova is one of the title favorites among the teams still alive, and the plot below will help you see why. The graph has a bar for each Division I team, in each statistic, with the given team’s bar highlighted and labeled with its statistic and ranking among all 351 teams. From this, we can easily see that Villanova is excellent on offense, ranking best in Division I in offensive efficiency. Moreover, their offense is great because they shoot efficiently (again, top in the country) and take good care of the ball (very low turnover %).

Villanova’s defensive efficiency is also very good (#16 in NCAA), based on solid shot defense, good defensive rebounding, and not sending opponents to the FT line frequently. Their (relative) weaknesses are scoring from the FT line and forcing opponent turnovers.

Let’s compare to the same plot for West Virginia:

Though less efficient than Villanova (particularly on offense), West Virginia still ranks highly on both sides of the ball. The Mountaineers’ offensive efficiency comes from great offensive rebounding and low turnovers. On defense, they do a good job defending shooters and they are great at forcing turnovers — the “Press Virginia” nickname is well earned.

Taking it one step further, we can compare the two teams’ profiles to find some keys to Friday’s game:

  • Villanova’s defense ranks much better at defending shots than West Virginia’s offense does in shot making. Will the Mountaineers knock down enough shots to keep up with the high-powered Wildcats attack?
  • Villanova rarely turns it over, but West Virginia creates a lot of turnovers with its pressure defense. Who wins this battle of strength vs strength?
  • Both teams rank much better in offensive rebounding than their opponent does on defensive rebounding. In a potentially close game, a couple extra opportunities on the offensive end could prove vital.

Both Villanova and West Virginia advanced through the first two rounds relatively easily, so let’s hope for a more competitive game Friday night in Boston.

--

--