Implementation of Artificial Intelligence in the National Basketball Association

Original article was published by Gregory Walfish on Artificial Intelligence on Medium


Implementation of Artificial Intelligence in the National Basketball Association

The National Basketball Association is a forward-thinking organization and is always looking for the best new technological advances. For example, data analytics has become a huge part of basketball. The use of statistical analysis, wearable technologies and computer vision are a vital part of a team’s strategies. A current use of data analysis enables teams to optimize their gameplay by picking the smartest play with the optimal players. Data use has become a part of scouting, calculating the best shot positions and scenarios and used to benefit the health and safety of players.

The shift towards incorporating technology into the game accelerated in 2013 when the league installed cameras in all 30 arenas. This system captures more than 16,000 data points per minute for teams to analyze. Coaches now have access to insights on offensive and defensive pain points and then base strategies off them, a tactic that has improved teams game play [1]. The league is constantly seeking new ways to improve the game through technology. Just last year, in 2019, the NBA hosted a hackathon to build more tools and accelerate innovation.

As more tools and predictive strategies become available, some fans and players have become widely critical of this data focused approach. The greatest critic is that of the 3-point shot, which thanks to analytics, has seen a huge increase because of its risk-reward ratio [2]. We are now seeing a game that is dominantly 3-pointers and in turn losing other aspects of the game [3].

Artificial Intelligence (AI) is a technology so powerful it can perform cognitive tasks. By using neural networks, it is intended that AI can mimic the way the human brain works [4]. Such cognitive tasks could come to include recommending different plays and strategies in real-time [5]. This could be used as an assistant to help coaches decide when to rest a star player or what play to use in crucial game moments. The use of AI will surely propel the current use of data and elevate it to a much higher level of predictability as much more data can be analyzed: posing a threat to the game as we know it. The NBA must take a stance on how it intends to proceed with AI; by either imposing guidelines and rules or allowing for freedom of exploration and use. Both approaches may impact the thrill and culture of basketball, provide insights on the health and well-being of the players and impact fans and the community.

Background:

As computer vision technology continues to become better and arenas continue to be outfitted with more cameras. There is increasingly more data to be analyzed. Historically, data analysis is done by team’s analytics departments which are composed of many human analysts. However, with this influx of footage and data, it is no longer possible for humans to analyze all of it. AI is the solution that can watch and interpret billions of data points. Some NBA teams have already started to use AI in their strategies. In 2016, The Toronto Raptors unveiled their “Digital War Room” developed by IBM Canada [6].

Figure 1: Toronto Raptors IBM Digital War Room (Watson)

Source: https://www.itworldcanada.com/article/toronto-raptors-unveil-a-digital-war-room/380686

This high-tech looking room is showing data from advanced cognitive technology from its AI called Watson. Masai Ujiri, president and general manager of the Toronto Raptors, uses Watson and its data analytics capabilities for many different tasks [7]. For example, Watson analyzes large pools of data that can help identify team deficiencies and ineffective players. It then matches these deficiencies and develops a criterion for possible roster combinations and identifies players with the highest chances of improving the teams in those areas. IBM Watson Trade-off Analytics uses mathematical filtering to narrow results to the most optimal and uses various analytical and visual approaches to help you explore the trade-offs between these options [8].

Watson is just one example of AI playing into scouting and recruitment strategies. Sportlogiq, a Montreal based sports analytics company uses AI to find talent. Sportlogiq’s computer vision software uses videos of games to track players’ movements and the orientation of their bodies during play. Then machine-learning algorithms crunch the numbers to evaluate players’ skills and overall potential. Sportlogiq software picked up that Sean Durzi, a 19-year-old defenseman in the Ontario Hockey Leagues should have been considered a top recruit but didn’t even make it in the top 40 of the NHL draft [9].

In addition to scouting and recruitment strategies, the possibilities of AI for health and safety of players can have a huge impact on reducing the number of injuries. This can be done as AI can predict potential outcomes. The National Football League (NFL) is leading by example by partnering with Amazon, a leader in AI and one of the largest companies in the world, to combat injuries. The NFL and Amazon is working to develop new tools to become better informed on player injuries and how different aspects of the game impact them [10]. Whether it be current game rules or equipment, they hope to combat the issue using AI and ultimately prevent injuries in the league. An approach like this would be very beneficial for the NBA too.

Even though AI has its benefits, the NBA has already encountered some issues with the rise of data analytics. As mentioned, the rise of the 3-point shot has significantly affected the game. Teams now realize that they can score more points by simply changing their shot selection, and they have done just that.

Figure 2: The rise of the 3 in the NBA

Source: https://www.espn.com/nba/story/_/id/26633540/the-nba-obsessed-3s-let-fix-thing

Fans are not happy as they are now seeing less mid-range shots and primarily only free throws, lay-ups and 3 pointers [3].AI has the power to find even more optimal strategies which might turn the game more repetitive and thus less enjoyable to watch.

AI is also capable of predicting match results. A number of experiments have been conducted and the results pinpointed that machine learning models can offer up to 67% predictive accuracy in classifying NBA game results [11]. This predictability will surely affect sports betting, an activity that many fans engage in.

Analysis:

There are possible positive and negative effects of AI in the NBA. The league must decide how to proceed. Whether it be imposing guidelines and rules or allowing for freedom of exploration and use of AI, three factors must be considered by the league: fairness, health/safety of the players and the effects on fans and the community.

1. Fairness:

Ensuring fairness is very important and something that the NBA takes very seriously. If the league decided to implement guidelines and rules, strict measures would have to be put in place. AI technology is so discrete and would be difficult to monitor as very few people understand it. In reality, it would be very difficult for the league to enforce such measures if they do impose rules on how AI can be used. Following the rules and guidelines would be based mostly on trust, which may not be viable as there have been cases of cheating in the NBA in the past. For example in 1999 Joe Smith and the Minnesota Timberwolves were caught in a salary cap sneak that promised Joe more money on future contracts. In addition, this framework of guidelines and rules would restrict access to knowledge and halt innovation of further technological advancements.

On the other hand, allowing for free exploration of use may ensure more equality than regulation ever could. Each team is working independently and freely to get the most benefit from the technology. They can each have their own unique capabilities, further adding onto this new interesting competitive advantage and angle to the game. Contrary to rules and guidelines, this solution will promote further adoption and innovation.

2. Health and Safety:

Health and safety is a very important criteria to consider in terms of AI and the prosperity of the league. As seen in the NFL, viewership in the 2016 and 2017 season saw a dip of 7% and 9% respectively. This poor performance can be accredited to a number of things, including conflicting views on concussions [12]. The NFL has turned to AI to help with reducing injuries. Health and safety of the players is vital and imposing guidelines and rules would allow for the NBA to have control over where the resources and focus would be. The league could put substantial effort towards health and safety and speed up development in this area. Altogether, this approach could help increase viewership and reduce injuries.

3. Effects on fans and the community:

AI in the NBA has an effect in the way fans interact with the game. With such high predictability capabilities and new insights on game play, there is a risk of the game becoming less exciting. Implementing rules and guidelines will aid in avoiding dilemmas like the 3-point shot as explored earlier and could help in preserving sports betting as we know it. However, free exploration of use does not mean we cannot preserve the game in other ways. For example, just as the NBA moved the 3-point line farther in the 1996–1997 season, the game can adapt to these new insights and trends. In addition, this new facet of competitiveness in the league might be exciting for fans and the community. Teams will also want to have the best technology and software capabilities, just like the War Room from IBM.

4. Summary of Analysis:

Regulation and AI is still a developing governance structure. There are no industries that have imposed rules and AI, until now, has taken a free exploration of use approach. In 2017, Canada launched the Pan-Canadian Artificial Intelligence Strategy providing CA$125 million to fund a national strategy. The strategy is intended to further build on Canada’s AI research-based ecosystem. One of its four goals is to develop global thought leadership on the economic, ethical, policy and legal implications of advances in artificial intelligence [13]. Before the NBA can impose rules and guidelines, there needs to be a clarification of the responsibilities of governments and sports leagues and until then, the NBA has no duty or legal responsibility to impose rules and guidelines [5].

Conclusion:

When considering fairness, health/safety of players and effects on society and the community, freedom of exploration of use of AI in the NBA is the only current viable solution. Other sports leagues are currently developing new AI capabilities and the NBA may fall behind if they don’t participate. The technology will be developed whether or not the league wants it. Advancement in technology is inevitable and humans cannot be stopped from gaining and wanting more knowledge. The NBA has already made huge advancements because of technology in areas of gear like basketball shoes and current sports analyzing techniques. AI will be no different and the NBA should embrace it sooner than later. A reactive approach would be smart for the NBA to pursue and to deal with problems as they emerge.

References:

[1] D. Cervone, A. D’Amour, L. Bornn, and K. Goldsberry, “POINTWISE: Predicting points and valuing decisions in real time with NBA optical tracking data,” in Proceedings of the 8th MIT Sloan Sports Analytics Conference, Boston, MA, USA, 2014, vol. 28, p. 3.

[2] P. Zuccolotto, M. Manisera, and M. Sandri, “Big data analytics for modeling scoring probability in basketball: The effect of shooting under high-pressure conditions,” International journal of sports science & coaching, vol. 13, no. 4, pp. 569–589, 2018.

[3] K. Goldsberry. “The NBA is obsessed with 3s, so let’s finally fix the thing.” https://www.espn.com/nba/story/_/id/26633540/the-nba-obsessed-3s-let-fix-thing (accessed.

[4] Y. Duan, J. S. Edwards, and Y. K. Dwivedi, “Artificial intelligence for decision making in the era of Big Data–evolution, challenges and research agenda,” International Journal of Information Management, vol. 48, pp. 63–71, 2019.

[5] P. Ding, “Analysis of Artificial Intelligence (AI) Application in Sports,” in Journal of Physics: Conference Series, 2019, vol. 1302, no. 3: IOP Publishing, p. 032044.

[6] B. G. “Toronto Raptors: Machine Learning as a Method for Improving the Roster Decision Process.” https://digital.hbs.edu/platform-rctom/submission/toronto-raptors-machine-learning-as-a-method-for-improving-the-roster-decision-process/ (accessed.

[7] P. D. Nibletto. “Toronto Raptors unveil a “digital war room”.” https://www.itworldcanada.com/article/toronto-raptors-unveil-a-digital-war-room/380686 (accessed.

[8] A. A. Tavor. “IBM Watson Tradeoff Analytics — General Availibility.” https://www.ibm.com/blogs/cloud-archive/2015/05/watson-tradeoff-analytics/ (accessed.

[9] E. Gent. “How AI is helping sports teams scout star players.” https://www.nbcnews.com/mach/science/how-ai-helping-sports-teams-scout-star-players-ncna882516 (accessed.

[10] S. Ogus. “The NFL And Amazon Want To Transform Player Health Through Machine Learning.” https://www.forbes.com/sites/simonogus/2019/12/06/the-nfl-and-amazon-want-to-transform-player-health-through-machine-learning/#27775c4a3c02 (accessed.

[11] F. Thabtah, L. Zhang, and N. Abdelhamid, “NBA game result prediction using feature analysis and machine learning,” Annals of Data Science, vol. 6, no. 1, pp. 103–116, 2019.

[12] M. Durborow, “How is Artificial Intelligence Improving the NFL?,” 2019. [Online]. Available: https://www.rebellionresearch.com/blog/how-is-artificial-intelligence-improving-the-nfl.

[13] T. A. Eduardo Soares, Ruth Levush, Gustavo Guerra,, “Regulation of Artificial Intelligence: The Americas and the Caribbean,” 2019. [Online]. Available: https://www.loc.gov/law/help/artificial-intelligence/americas.php.