Basketball Analytics
Since the 1950’s, Basketball has revolved around three statistics: points, rebounds and assists. This is the basis for the beloved triple-double, one of the most highly acclaimed achievements in basketball. In 1973-74, the NBA started to track steals, to measure how often a player was able to steal the ball from the person he was guarding. It also added blocks, which was a measure of the number of shots blocked by a specific player. These statistics added a new dimension to measuring the strength of players. During the 1978-79 season, the NBA announced an additional tool for measuring player strength, turnovers (Wikipedia). These six statistics became the staple of basketball knowledge. The are the stat lines in the newspaper, and the way that many measured who should win the MVP at the end of the season. But, as time went on, many started to realize that these did not tell the whole story. Players such as Tim Duncan or Scottie Pippen seemed to be contributing more to their team than these statistics accounted for. These players dominated defensively by forcing poor shots, without racking up steals and blocks. They were excellent at setting screens and moving off ball on offense. It was becoming clear that the world of Basketball needed a new measure of success.
Analytics are defined as the systematic computational analysis of data or statistics. In terms of basketball, this is the process of crunching play by play data from a basketball to give a more comprehensive view of which players contribute to a team’s victory. The main idea behind basketball analytics in specific is the idea that players skill on a per possession basis is much more useful than a player’s skill per game in terms of determining his merit going forward. Consider Nikola Mirotic, a Power Forward for the Chicago Bulls. He scores 10.2 points per game. At an early look, he seems to be an average auxiliary scorer on a decent team. But look a little deeper at his advanced stats. He only plays 20.2 minutes per game. A more flattering and useful statistic says that he scores 26.1 points per 100 possessions (http://www.basketball-reference.com). With the development of these more encompassing statistics, NBA General Managers have an easier time comparing the relative strengths of players, and fans can better understand the skill level of their players.
Another variable that classical statistics fail to reconcile are the impact of the strength of teammates and opponents on the performance of the player. Andrew Wiggins can score 16.9 points per game on the Minnesota Timberwolves, but they are the worst team in the NBA. Now think about Kawhi Leonard, the starting small forward for the San Antonio Spurs. He scores 16.5 points per game. Traditional statistics would tell us that these players are of approximately equal value. This is not true though. Almost everyone would agree that Leonard is a top ten player in the NBA, and that Wiggins has barely developed into a top 50 player (http://www.basketball-reference.com). These two are drastically different because of the workload that their team makes them carry. This is the motivation for the creation of the Real Plus Minus statistic (RPM). RPM is important because it measures the difference when the player is on and off the court, and uses that to attempt to extract the value of the player (ESPN). The also measure field goal percentage and true shooting percentage to decide whether or not a player is efficient as opposed to that player scoring with a high volume. Basketball analytics have provided a more accurate measure of players strength.
The final frontier is defense. Defense is something that is almost entirely qualitatively measured and it is very hard to determine the value of a player’s defensive skills. While RPM does a decent job, the field is still yet to be developed fully. The primary researcher on this topic is Kirk Goldsberry, a writer that has started to do simulations of defense in order to generate defensive shot charts. This allows him to measure how players shoot against a specific defender. Because of this, player’s skills on defense can be measured in a much better way (Goldsberry).
In conclusion, basketball analytics ave progressed to the point where we, as a basketball community, have a much better idea of which players are good and which players are not as good. We can choose players that shoot more efficiently and players that are much better on defense through a numbers based approach. Over the next couple of decades this will approach the point where we fully understand how good each player is at every position. Players are still human though, and there will always be moody things that we cannot measure. Analytics will trudge on, and give it their best shot.
Works Cited
"Basketball Statistics." Wikipedia. Wikimedia Foundation, n.d. Web. 22 May 2015.
"Department of Defense." Grantland. N.p., 24 Feb. 2015. Web. 22 May 2015.
ESPN. ESPN Internet Ventures, n.d. Web. 22 May 2015.
"Nikola Mirotic." Basketball-Reference.com. N.p., n.d. Web. 22 May 2015.