The correlation between scoring margins and player offensive efficiency

Way back on December 1, 2010 Loyola was 7-0 and Butler was coming to play the Ramblers at the Gentile Center. There was a lot of hope in Rogers Park that Loyola would shake off its status as a lower division member of the Horizon League and start challenging the big boys.
Well, for 40 minutes the Ramblers did it. They traded punches with Butler until the final possessions and ended up falling 65-63. In just a 57 possession game it was quite the performance. Afterwards though there were a lot of questions about how Brad Stevens contained the Loyola offense. The bright, young coach talked about how stopping Geoff McCammon was the key to Loyola's offense and winning the basketball game.
That a struck a nerve with me that night. Throughout the remainder of the season I wondered, "Just how important is one player to a team's offense and to winning?" In the series of posts here and the ones that will follow over the next few days I hope delve into that issue a little deeper.
First a little bit about the method. All of the data I'm going to use in these posts is simple correlations between offensive efficiency and point margin. Why scoring margin and not straight wins and losses? Because margin gives a clearer picture of how well you played. Yes, ultimately the result is what matters, but how much you win by puts a meaningful spin on that result column.
Thus Loyola's first 10 games look like this for example: 16, 12, 22, 11, 18, 28, 1, -2, -10, 28.
But to start off let's get a baseline for these numbers. For instance, how does the offensive rating of a superstar playing on a quality team correlate with their scoring margins?
Glad you asked. Here's the correlations for some "notable" college basketball players this season:
  • Derrick Williams, Arizona: 0.561
  • Kemba Walker, Connectic: 0.547
  • Jimmer Fredette, BYU: 0.546
  • JaJuan Johnson, Purdue: 0.267
  • Jared Sullinger, Ohio State: 0.249
  • Nolan Smith, Duke: 0.136
So the more "balanced" the team the lower these numbers are going to be. It's possible for a "star" players' contributions to get lost in the mix because they're not the only players on the court. 
Still, I think the numbers above are pretty intuitive and will help us moving forward. Williams, Fredette and Walker were expected to carry their respective teams. Kemba managed to drag a young Connecticut team all the way to the National Championship and in order to do it he had to put up some very impressive offensive statistics.
Moving forward I'm not suggesting anyone in Chicago is on the same level as Walker, who will probably be a Lottery Pick in the NBA Draft this June, but there are certainly some players that are just as important to their teams. Though they might not be the ones you'd expect.
Tomorrow we'll answer the question, "Was Stevens right?" when we look at the correlations from players on Loyola. Friday I'll do a short post on two key players from UIC (I think you can guess who.) Finally, early next week I'll dive into Northwestern and DePaul's numbers. The Wildcats' correlation might be the most surprising of all.

Leave a comment