How Data Could Be Making It Rain For Basketball Teams.
We are all familiar with the classic statistics that we have been using for years to understand basketball. These stats, such as points scored and shots made, give us a good idea of how well players perform on the court. However, these traditional stats do not tell the whole story.
Just like the FBI uses DNA, forensic science, computer technology, cyber security, electronic surveillance, biometrics and more to help them in their investigations in retaining details, exploring new avenues, new leads vs eliminating others, sports teams can take advantage of similar technologies evolving rapidly in the field of data analytics, player movements, machine learning algorithms who studies every move, every position, player's skills comparator, and performances to better determine how a player influence the team even when he doesn't have the ball, which is often 90% of the time. Any team that underestimate such valuable and intricate insight will likely be left behind once their superstars retire...
How data transformed the NBA - The Economist
Often the the difference between failure and success in basketball is simply a thin line. See how The Houston Rockets, who boast stars like James Harden, have used data analytics to help them become championship contenders in recent seasons.
Data Analytics on basketball players movements
Calculating Shot Probability
Where are the defenders
Leaner players skilled, more agile over heavy players
Increased Three-Pointers as a percentage of all shots in the NBA has increased
3-D players movements calculated
Could Moneyball's theory work in the NBA?
"Moneyball" was a revolutionary new way of analyzing the game, using statistics and analysis to find valuable players that other teams overlook because they don't use the same analytical process. It offers great insight into some overlooked strategies.
"For decades, the metrics used to measure the performance of individual players have been terrible. Total points, total rebounds, total assists, total steals, etc. are awful metrics, as they focus on action with the ball...at any given time, 90% of the players on the floor don't have the ball in their hands. This means that the most frequently used stats don't account for 90% of action...not to mention that these metrics are deceptive.
Michael Beasley, for example, can score at a high level and fill up the stat sheet, but his other contributions are so poor that he finished his 4th season coming off the bench. Allen Iverson used to log quite a few assists every game, but this was because he played extended minutes and only dished when it was a sure bucket...not because he regularly tried to setup his teammates.
Conversely, Shane Battier rarely fills up the stat sheet, but his contributions to the team as a defender, smart passer and cutter, and leader are top notch." - Jason Lancaster
Image from Wikipedia Commons
We agree with Jason Lancaster that NBA teams who take advantage of the latest stats and analytical tools such as "player tracking" software SportVU will be less likely to make the same old mistakes. Betting on free agents can be a risky move, but if you find the right player and coach them well they might just turn into your lucky charm. The more data that scouts gather from potential players will allow for better drafting decisions in the future which may provide even more of an edge to their team.
An NBA Regular Season Leaders's Stats Looks like this.
It already contains a fair amount of stats ranging from player impact estimate, offensive rating, defensive rating, net rating, usage percentage, assist percentage, all the way to rebound percentage and turnover percentage. But as you can view further below, a software technology such as SportVu can add an abundance of stats that no human could possibly break down and absorb without some structured software.
Photo from RealGM
What is Player Tracking, Machine Learning Algorithm and Probability in Basketball?
Player Tracking is the latest example of how technology and statistics are changing the way we understand the game of basketball. Using six cameras installed in the catwalks of every NBA arena, SportVU software tracks the movements of every player on the court and the basketball 25 times per second. The data collected provides a plethora of innovative statistics based around speed, distance, player separation and ball possession. Some examples include: how fast a player moves, how far he traveled during a game, how many touches of the ball he had, how many passes he threw, how many rebounding chances he had and much more. The information will be available to fans on NBA.com and NBA TV.
"When it comes to the task of classification, machine learning algorithms are using a predicted probability to decide whether or not something belongs to a group. In the case of an NBA positions, a machine learning algorithm is going to predict the probability that the player is either a point guard, shooting guard, small forward, power forward, or center. The algorithm will give us a probability for each position for any given player. It’s these probabilities that could prove to be useful to a front office or discussion among NBA analysts. With each new generation of NBA talent, we have seen more hybridization amongst players when it comes to their skill set. Centers are no longer strictly close-to-the-basket low post players that only come to the 3 point line to set a screen. Modern centers drift out to the three point line to take and make 3’s, and some have the requisite skills to effectively run a fast break. With positional probabilities assigned to a player we can begin to understand the versatility of a player. If we know the positional probabilities of each player on a team, that can be used for a lineup analysis of your own team or the opponent." - Jeremy Lee, B.Sc., Data Science Graduate of Flatiron School & Duke University.
Viziball - Basketball Analytics For Everyone
Viziball is another interesting and fun tool that solves the problem of comparing two basketball players many stats on different aspect of the game by creating an interactive and simple spider chart that allows users to compare any two NBA player in just one click. It also provides detailed information about each aspect of the game, so you can learn more about your favorite players' strengths and weaknesses. - Image from Viziball
Basketball Performance Indicators
Often success in basketball is simply the repeated, periodic achievement of some levels of measured success that impact the "Four Factors" (i.e. scoring the ball, controlling the ball, rebounding the ball, and the free throws), and sometimes success is defined in terms of making progress toward strategic goals.
" What Gets Measures Gets improved" - said Peter Drucker 45+ years ago about "big data".
However, today, too many things are measured. It is important to understand that decisions in sport should be made using only the most relevant data, and it's impractical to try and manage everything. However, managing a few metrics strategically can help better your analysis of trends in random sampling or manual observation.
As Brian Strachman said it eloquently, The first challenge is deciding on what to measure. In the movie "Moneyball, they tell the story of a small group of naysayers who opted to take a step back, and evaluate whether the numbers they relied most heavily on, were appropriate. Take the case of Babe Ruth. He is widely regarded as the best hitter in baseball history. Yet if one were to apply the most commonly used metric to evaluate hitters, batting average, he would rank as tenth in history. The problem lies in the metric. Batting average is number of hits divided by times at bat. Like many business metrics, it is an oversimplification used by humans with limited ability to process large numbers. It gives equal weight to a home run as it does to a single. But when evaluated based on the metric proposed in Moneyball, the “slugging percentage,” The Babe jumps to number one by a large margin. Slugging percentage gives more credit for doubles, triples, and home runs; than to singles, and subsequently changes the result of the analysis. For over 100 years, baseball was using the wrong metric. They needed to include more measurements, for a combined and more accurate view." - Brian Strachman