AI x Sports: How Machine Learning is Transforming Sport Analytics
Tuesday, April 16th. Golden State Warriors versus Sacramento Kings. Klay Thompson chucks up 10 three-pointers and misses all of them.
It’s the most misses consecutively from the Warriors in a postseason game since at least the 1967–68 season.
Let’s side-track a bit. X formerly known as Twitter, has a new AI integration called “Grok” which makes trending banners on X based on the current news.
However, Grok is still in its early stages and it’s very prone to mistakes…
This was the trending headline in X. Simply saying Klay was accused of vandalizing houses in Sacramento with bricks. This left fans cracking up.
If you aren’t an avid basketball fan, bricks mean misses in basketball slang. So when everyone on X was saying stuff like “10 bricks in a row is crazy” or “No way he shot 10 bricks in Sacramento”, Grok was interpreting it as Klay throwing bricks at houses
Sports and AI!?
Even though you’ve probably seen Tesla driving themselves or used ChatGPT for an essay, there are many more applications in AI, especially in the sports world where immense manpower is needed.
Currently, sports stats are determined by referees and officials on the field. Constantly observing, looking for foul calls, yellow cards, and travels.
To go into more detail, it’s legit people with usually a paper with boxes that they check when a certain event happens. So in basketball, a box would be checked for steals, points, and assists but also turnovers, travels, and fouls.
This not only takes up a lot of resources but also a lot of time.
Even though some bits of technology are used, for example, in baseball, pitch tracking systems can automatically record each pitch’s speed, trajectory, and location.
If you rewind to the 2022 World Cup, they kicked off the tournament using a new high-tech soccer ball with a sensor that collects spatial positioning to make better and more accurate calls.
What is the Tech and how does it even Work?
To make the process of gathering the data, certain sporting departments use a subsection in machine learning called Computer Vision.
Computer Vision is exactly what you think it sounds like. Allowing computers to see and interpret what they see.
For instance, training the computer vision model to record every time a touchdown is scored would save a lot of resources.
Here’s how it works:
For the sake of this process, let’s imagine we want to record how many times Austin Matthews scores a goal.
If you want to remember the following steps, use the acronym Instant Pizza From the Delivery Robot Crew.
- Image Acquisition (Instant): The simplest part, obtaining an image of what we want to find. So in this case, a picture with the view of the goal and the players.
- Preprocessing (Pizza): Once the images are captured, they often undergo preprocessing to improve their quality and make them easier to analyze. This might involve tasks like resizing and cropping.
- Feature Extraction (From): Next, the computer extracts important features from the images. These features could include edges, corners, and textures. These features are what the model will learn from.
- Recognition of Patterns (Robot): With the extracted features, the computer compares them to patterns it has learned from training data. This is where machine learning comes into play. Machine learning algorithms, such as neural networks, are trained on large datasets containing labeled images.
- Classification (Crew): Based on the learned patterns, the computer classifies objects in the images or detects specific features. Classification involves assigning a label or category to the entire image, so identifying whether Austin Matthews scored a goal or not.
Top Companies Working on Bringing AI to Mainstream Sports
You may think this is a very new thing but there are already companies and people working in this space, thriving to bring it out the general public.
International Olympic Company: 2 days before this article was published (on April 19th, 2024), the IOC unveiled its strategy for leveraging AI in preparation for the upcoming Paris Olympics, slated to commence in nearly 100 days. Their plan encompasses harnessing AI to discover talented athletes, personalize training routines, and enhance fairness in competitions by refining judging processes.
NFL x AWS: This season, the NFL has been working with Amazon Web Services to push out a new portal for their players, Digital Athlete.
Digital Athlete utilizes AI and machine learning (ML) to forecast potential injuries by analyzing plays and body positions of players. The platform gathers data from players’ RFID tags and 38 5K optical tracking cameras positioned around the field, capturing 60 frames per second.
Where else can Machine Learning be used in the Sports Industry
Sports analytics isn’t the only area in the sporting industry where machine learning can be used. Here are two prime examples:
Predictive Stats: Rather than recording the current stats, predictive AI models are capable of learning from the databases on previous games and predicting the next possible outcome. A lot of people used this for the March Madness bracket too.
Injury Prevention and Rehabilitation: Machine learning models can analyze specific biomedical features such as blood information, performance data, and other relevant information to identify injury risk factors, monitor player health, and develop personalized rehabilitation programs. This proactive approach can help prevent injuries, accelerate recovery, and prolong athletes’ careers.
What has Changed to Make this Happen?
Something you might be wondering is that why is this happening now?
If you’ve been keeping up with the tech news for the past couple of years, artificial intelligence is making headlines day in and day out. From the ChatGPT boom in November of 2022 to OpenAI CEO, Sam Altman getting fired in November 2023 to even Nvidia’s stock going down a few days.
This recent spike in AI development propelled a lot of startups and companies forward by providing them with more venture capital and more resources to grow their platform.
Eventually, people started wondering, where else can they apply AI?
Next thing you know, social media platforms like Instagram and X have their own AI search options, insane AI robots, and it even found its way into the healthcare industry with applications such as more accurate diagnostics.
What are the Major Setbacks to using AI in Sports?
Even though these AI integrations into sports will be very revolutionary, there are still many challenges that hold AI back from reaching its full potential.
The main one is the ethical considerations.
Just the thought of AI replacing the jobs of tens of referees and officials who make the calls alone. Ethical concerns also come with data security issues. What if the AI algorithms get messed with and become biased towards one team?
This would not only ruin the experience for both fans and players but also make us regret giving that much lenience to the AI models.
This is one of the biggest debates going around: Should artificial intelligence be left alone to manage things on its own or not?
If you do, there could be severe consequences but if you don’t, you will be behind and using a lot of resources while everyone else will probably be using the easier methods with AI.
Key Takeaways:
- ML Transforms Sports Analytics: Machine learning (ML), especially through computer vision, revolutionizes sports analytics by automating data collection.
- Manual Methods are Inefficient: Traditional data collection methods in sports are resource-intensive and slow, highlighting the need for automation.
- Tech Examples: Technologies like pitch tracking in baseball and sensor-equipped soccer balls enhance accuracy in sports decision-making.
- Computer Vision’s Role: Computer vision streamlines data collection by acquiring, preprocessing, and analyzing sports images, aiding in performance tracking.
- Broad Applications: ML extends beyond analytics to predict outcomes and prevent injuries, leveraging player data and health information.
- Ethical Challenges: Ethical concerns like job displacement and algorithm bias pose hurdles to AI integration in sports, necessitating a balance between innovation and responsibility.
Works Cited
“Auston Matthews Stats, News, Bio.” ESPN, www.espn.com/nhl/player/_/id/4024123/auston-matthews.
avcontentteam. “All You Need to Know about Sport Analytics in 2023.” Analytics Vidhya, 3 Aug. 2023, www.analyticsvidhya.com/blog/2023/08/sport-analytics/#:~:text=Integration%20of%20Artificial%20Intelligence%20and. Accessed 19 Apr. 2024.
Bailey, Andy. “Klay Thompson Landing Spots If Golden State Warriors Reset.” Bleacher Report, www.bleacherreport.com/articles/10117291-klay-thompson-landing-spots-if-golden-state-warriors-reset. Accessed 19 Apr. 2024.
Dowsett, Ben. “The World Cup’s New High-Tech Ball Will Change Soccer Forever.” FiveThirtyEight, 22 Nov. 2022, www.fivethirtyeight.com/features/the-world-cups-new-high-tech-ball-will-change-soccer-forever/.
“How AI Is Helping the NFL Improve Player Safety.” CIO, www.cio.com/article/1306736/how-ai-is-helping-the-nfl-improve-player-safety.html#:~:text=This%20season%2C%20the%20NFL%20has. Accessed 21 Apr. 2024.
Khanna, Ayesha. “Council Post: Can AI Score Big in the Future of Sports? Five Key Trends Shaping the Industry.” Forbes, www.forbes.com/sites/forbestechcouncil/2023/09/27/can-ai-score-big-in-the-future-of-sports-five-key-trends-shaping-the-industry/.
“Olympic Organizers Unveil Strategy for Using Artificial Intelligence in Sports.” AP News, 19 Apr. 2024, www.apnews.com/article/olympics-ai-artificial-intelligence-sports-30d3a33d4be893e1495e63fc1116aecc.
Hello! I’m Yakshith and I’m a high school student and a brain-computer interface (BCI) enthusiast. I hope you learned something new from this and if you have any sort of feedback or suggestions, don’t hesitate to reach out. Thank you for reading and don’t forget to check out my X, Substack, LinkedIn, and YouTube!
