One of my personal/professional goals for this summer has been to learn how to work with and visualize data. I want to know how to make beautiful data visualizations, and tell interesting stories with the use of data. To that end, I’ve started working through some self-directed reading, and signed up for a couple of open online courses. The first course I completed was Google’s Making Sense of Data course. The course itself is pretty sleek and activity driven (it took me part of two days to complete the whole thing), and it seems primarily designed to introduce users to very basic data management concepts and train them to use their experimental Google Fusion Tables tool. I’ve been working on a couple of data projects using the tool. The first has to do with global university rankings (more on that one later), and the second is a fun project that was intended to answer some questions that have been kicking around in my mind since the San Antonio Spurs won the 2013-2014 NBA championship.
I didn’t follow the most recent NBA season very carefully; in fact, I really only saw parts of a couple of playoff games. Even though I don’t watch the actual games very often anymore, I still read sports journalism fairly avidly, and have been especially intrigued by the recent explosion of interest in and coverage of advanced analytics (witness the meteoric rise of the MIT Sloan Sports conference over the past half-decade, for example). I had a range of questions about the Spurs this year, inspired in part by what seemed to me to be an unusually balanced team: for those who don’t follow basketball closely, it’s worth noting that Gregg Popovich has gained notoriety in recent years for (among other things) judiciously regulating his aging stars’ regular season minutes, and for developing a deep cast of role players which are deployed in novel and ultimately successful ways.
Despite the lack of (or perhaps because of) a single offensive focal point, the Spurs seemed like (from what I saw of them) an absolute offensive juggernaut: one of the few parts of the playoffs I saw was the first half of Game 3 of the Finals, a game in which the Spurs went nearly 11 minutes without missing a field goal, shooting better than 86% in the first quarter and better than 75% in the half. With seven minutes to go in the first half, in fact, they had made 19 of 21 attempted field goals.
It was the most breathtaking display of offensive basketball I have ever seen: fluid, team-oriented, lethal. I couldn’t imagine them losing the series, and after they had defeated the Heat in 5 games and Kawhi Leonard was named Finals MVP, I started racking my brain in search of another comparable team. I didn’t come up with anything, and started wondering whether systematic statistical analysis, albeit of the fairly rudimentary sort of which I am capable, might bear out my hunch.
I decided to gather two fairly simple data sets–I went back and looked at every NBA champion (from 1947) and every team with a winning percentage greater than 75% (teams that won at least 62 games in an 82 game season). I recorded a few pieces of information about each team: their regular season win-loss record (and winning percentage), the team’s average points per game, their three leading scorers (and their PPG), and the number of players on the team who scored more than 8 points per game while appearing in more than 50 games for the team (I later went back and looked at how many players scored more than 7.5 points per game). I also did several calculations using this data, estimating the percentage of the team’s points scored by the leading scorer, the percentage scored by the team’s three leading scorers, and the point differential between the top three scorers on each team, for example. For the purposes of this analysis I will exclude the data from all seasons prior to 1954-1955, the first season in which the 24 second shot clock was in use.
One of the first features that I found unusual about the 2014 Spurs was that their leading scorer during the regular season, Tony Parker, scored just 16.7 points per game (an average which ranked just 41st in a 30 team league). With the possible exception of the 2004 Detroit Pistons, I couldn’t think of an NBA Champion from the recent past with a leading scorer who scored so (relatively) little. A closer look at the data confirmed my suspicion. In the NBA seasons played since 1953-1954, there has never been a league champion whose leading scorer scored fewer points per game than Tony Parker (the Spurs’ leading scorer) in 2013-2014. Here’s a visualization of the data:
If we look at the historical data, we find that the lowest scoring ‘lead scorers’ for NBA champions (measured by points per game, independent of how many games that scorer played) occurred in 1954 and 1955 (the first two years of the shot clock), 1959-1964 (Boston Celtic teams which featured very balanced scoring, often with 6 or more players scoring in double figures), 1976-1979 (just before Larry Bird and Magic Johnson arrived, this stretch is generally considered one of the low points of NBA history, with the league racked by drug addiction and other problems), the notoriously physical 1989-1990 Detroit Pistons (see the recent Bad Boys documentary for a primer), and another famously dull and defensive period in NBA history: 2004-2008 (which featured relatively low scoring lead scorers in every year from this period except for 2006, when Dwyane Wade’s Heat won the title with help from NBA refs).
That’s just looking at points per game, not raw scoring totals. Obviously, the average rate of scoring has varied a great deal in NBA history, with post-1954 championship teams scoring anywhere from 90 (the 2004 Detroit Pistons) to 125 points per game (the 1967 Philadelphia 76ers). And points per game is an average based on the number of games each player appeared in, not a raw measure of total points scored. To help account for both of those potential variants, I’ve also made a graph which shows the percentage of the team’s total points scored by their leading scorer (in raw terms, not per game averages). Using this method, the 2014 Spurs appear even more extraordinary:
If that weren’t enough, the 2014 Spurs look even more like an extreme outlier when you perform the same calculations for top the three scorers. Looking at raw point totals, the top three scorers on the 2014 Spurs (Tony Parker, Tim Duncan, and Marco Belinelli) scored less than 37% of the team’s total points. The team whose three leading scorers composed the next smallest percentage of the total were the 1977 Portland Trail Blazers, with just under 43%. You can explore the historical data below:
As the previous chart indicates, the second notable feature about this most recent Spurs team was their scoring depth. In addition to calculating the percentage of a team’s points scored by their 3 leading scorers, I recorded how many players appeared in at least 50 games and scored at least 8 points per game for each NBA champion. That data can be seen below:
Again, this chart helps visualize just how unique last year’s Spurs were in NBA history. Among all other NBA champions, only the 1966 Celtics had 9 players score more than 8 points per game, and only two other teams in NBA history have even had 8 players hit that benchmark, none since 1977. After the NBA expansion in 1988 and 1989, in fact, only the 1990 Detroit Pistons have even had as many as 6 players score more than 8 points per game in at least 50 games, all of which makes last season’s Spurs even more exceptional within their recent historical context.
This post has already run on longer than I initially intended, so I’ll leave it for now. More to come soon!