In the last decade, the National Basketball Association (NBA) has been undergoing a major (r)evolution by using data analytics. This change process may not have been visible to the average Joe, but has been going on since the early 2010’s. By then, multiple teams started to deploy an analytics department. Nowadays, these departments exist out of up to 20 data analysts, analyzing every facet of the game. And this strategy clearly paid off. One of the frontrunners in data analytics, the Houston Rockets, have transformed from a mediocre team into a title contender since the introduction of its data analytics team. How did this technology transform one of the biggest sport competitions in the world?
The NBA has been one of the frontrunners in the collection of data in sports. Every match, six high speed camera’s capture the player movements and other data points. These extensive data tracking systems result in up to a million data points per match. These data then gets analyzed by advanced machine learning algorithms, resulting in interesting insights that have changed the league.
First of all, data analytics have changed the overall strategy of the game of basketball. Prior to the data analytics age, many long range 2 point shots were taken by the players. After data analysis, these shots were found to be inefficient in terms of their pay off (measured in points * chance a shot hits the basket). Instead, data analysis showed that 3 point shots were a far more efficient way to score points. These shots had a lower percentage of being scored, but the higher points scored made the trade-off worth it. As a result of this changing strategy, elite three point shooters like Golden State Warriors’ Stephen Curry and Houston Rocket’s James Harden have become some of the most valuable players of the league.
Additionally, teams have been collecting sophisticated data about their players performance and fatigue. Analysis of this data by prediction models has resulted in advanced resting schedules for their players, in order to prevent injuries. Nowadays, teams might bench their key players to rest, even though they may not actually be injured. These schedules have proven to be the key to preventing injuries, in a league where a team can play up to 110 games per year.
Lastly, the last field where data analytics has proven its success, is in the field of scouting. Coaches and general managers now heavily rely on predictive analytics when acquiring new players. Advanced statistical models are created to predict which college & high school players will become the new LeBron James and Michael Jordan.
Data analytics in the NBA surely are not a golden ticket to winning championships. But in a highly competitive league such as the NBA, creating an edge (sometimes as little as 1%) over your opponent by using data analytics might actually make a difference that will win you the game. One thing is for sure, data analytics in the NBA is here to stay!
Very interesting article Jesse! It sounds like magic for a coach. Just run a data analysis on the potential player you want to buy and you know if he can really add value to your team. Or what you mentioned, focus more on practicing the 3 point shot instead of the 2 points shots. In my opinion, sport is something you have to practice because you like it and you get energy from. I know that at the same time there is a huge business model behind it. Look for example at the NBA final, one of the most prestigious event to perform as an artist.
These developments on AI and data in sports scares me. These insights really turned out to be really successful in order to improve the quality of the team. It is very likely that in the near future there will be even more focus on the data. But what scares me, is that coaches and teams tend to forget the real person. In the end, a Basketball player is way more than some statistics. I am really afraid that focussing too much on those parameters, it will in the end damage the whole industry as its not about the sport itself anymore, but about collecting as many data as possible.
Off course, I do see the positive sides of these whole evolution, but at the same time is also see the challenges. How do you think this will evolve?
Hi Divia! Glad to hear you liked my post! I think you don’t have to be afraid that in the future the focus will be on data too much. At the end, it is the athlete that makes the difference. Data analysis only will provide an edge, but will not decide the outcome of the game to such a great extent.
A very interesting article, thank you for it! I was also aware that Houston Rockets was the first team to utilize data analytics in the NBA. As I am mostly a football fan myself I can add here that such methods are also well-known in the English Premier League. Nowadays interesting statistics are coming to the surface such as Goal/minutes, km covered/games etc. WhoScored is a very famous website amongst others that ranks players’ performance based on such statistics. Overall, managers nowadays do not only focus on their tactics but are also interested in the player’s statistics before purchasing new players. As the average fees in the top leagues have increased significantly over the past decade, a more detailed review of the players before purchase is required. I do believe that still a “moment of magic” as many commentators call it, cannot be measured by statistics easily, but it can be predicted to a certain extent if the right indicators are followed. I am interested to see statistics taking a more major role in more sports and see how they can influence the performance of smaller clubs as well.
Hi Vasileios! Nice to hear that you enjoyed this read. I’m a big football fan myself, so I’m glad that data analytics is making it’s entrance into football. One of my favorite data analysis related stories in football is how Leicester City FC used data analysis to pick up bargains like Riyad Mahrez & N’golo Kante from the French 2nd league. Only to sell them for major profits a couple of seasons later, after they won Leicester the Premier League title as well! I could only hope to hear more of such stories for the upcoming years!
Hi Jesse, thank you for the interesting blog post. It’s quite amazing to see how data analytics have impacted many sports. The two-point shots versus three-point shots is a very clear example of data contradicting something that many people believed to be true. Three-point shots were probably considered too risky with a high rate of failure compared to two-point shots. Data analytics showed that this is indeed true, but the three-point shots actually still have a higher expected value. Personally, I’m a big fan of seeing team and player related statistics, they can provide very interesting storylines and show the strengths and weaknesses of teams and players. I also believe that it’s quite hard for many viewers to analyze sports at a higher level. However, statistics can provide some of this high-level analysis to regular viewers, for example a stat like Pass% in football, which indicates the percentage of successful passes. I hope that the teams and sports broadcasts keep utilizing data to improve the quality of the games and the view experience.
Hi Jens, i’m glad to hear that you enjoyed the post! I also thoroughly enjoy the use of data and statistics in football. One of my favorite shows to do so is Match of the Day. They take the analysis of football and its statistic to another level! Really improves the viewing experience i think!
Hi Jesse, really nice post! I was scrolling through all the different posts and this one definitely caught my attention. I am aware that data analytics is becoming a bigger part of several sports, take the formula one for instance, where an individual F1 team has several dozens of analysts. However, I had no clue that it was also such a big part of the NBA. Especially the part about the less efficient long range two point throws is interesting. Although it might seem as something that is somewhat self-evident, I believe it is really valuable for those teams to learn it from the hard data. I am specifically wondering what your opinion is on the usage of data analysis within this sport? Personally, I think that AI and machine learning can be a powerful and helpful tool which most of the times improves a sport. However, one could also argue that it rules out the ‘gut feeling’ which a lot of people love about their sport. Again, really nice post!
Hi Tristan, glad to hear you enjoyed it! I’m a big fan of F1 as well! Personally I believe that data analysis plays an even bigger role in F1 than in the NBA. I really liked the following video about data strategy in F1. You should give it a try! link: https://www.youtube.com/watch?v=4CFkltWIc8o
Very nice blog post, I’m a huge NBA fan so this topic was very interesing to me! I was aware of the use of these analytics in the NBA and it has truly amazed me. The NBA website alone shows so many specific details that it seems a bit over the top from time to time. However, it also seems like it could be very effective as you explained because anysized edge, is indeed an edge over the opponent. What I think about often regarding the use of analytics is its viability in other sports. To me, data analytics seems ideal for the NBA as it involves quick and intensive play on a relatively small court. Compare this to football (not american football) for example, it seems to me that data analytics might be less important. There are multiple important players in certain football teams that just do not appear on the stat sheet by means of assists/interceptions/goals/passes. These players are for example used due to their experience to position and coach other players on the pitch. I think it would be very interesing to see how data analytics would portray these type of players, as the older, experienced players often come of the bench in the NBA and will only have limited gametime. Football games are ofcourse completely different as the substitutions work differently. Similarly, data analysis has proved the 3pt shot to be more valuable than the 2pt shot, which also led to players taking more and more 3pters each NBA season. However, I cant imagine similar analysis being performed in other sports that lack definite distinctions between for example a 2pointer or a 3pointer. In football it is difficult to decide whether a player should shoot from 20 metres out or 22 metres, as both will lead to the same result when succesful, a goal. I know for a fact that footballteams have started experiencing with data analytics, however I can’t imagine it being as succesful and valuable as it is right now in the NBA.
Hi Nasreddine! Nice to hear you are a fellow NBA fan! I agree that data analysis will probably prove to be more useful in the NBA than in football. Due to the nature of the game, as it is an intensive game with lots of different outcomes (e.g. differences in points per basket).
One example i really liked regarding how statistics proved the impact of experienced players: Data analysis actually showed that during the GSW’s championship runs, the experienced 6th man Shaun Livingstone actually played a very significant role in the outcome of the championships. His experience and playing style were found to be crucial factors in the playoffs!