Editor's Notes: This is the first article in a three-part series examining the evolution of AI in sports.
Previously a futuristic fantasy that could capture communications of unattainable data and turn it into actionable insights, AI is no longer the next frontier in sports analysis.
Now it's standard.
According to Peter Ziemes, a senior lecturer at the University of New Hampshire and founder of the University of Sports Analytics Lab, there are hierarchies within different sports, with teams within each sport using AI in their own way, but deployments to analyze data in ways previously impossible are now spread across all major sports.
“Everyone is using AI right now,” he said. “There are still different degrees of access to data, and while the things that teams choose to access are still different, it is broad.
Currently, AI, which represents the widespread evolution of analytics in sports, is still in a rather early stage.
As Openai announced the launch of CHATGPT in November 2022, a sudden and substantial improvement in Generated AI (Genai) technology has similarly enabled the widespread use of AI in sports.
Before Genai, and more recently, the emergence of agent AI has been limited by the human capital needed to manage, model and analyze data, as well as the constraints of the systems used to intake, integrate and prepare data. Most of the same data currently used to notify AI models and applications was there to be analysed in reports and dashboards, but it was impossible to utilize it all.
Genai reduced, eliminated, reduced and eliminated some of the previous barriers to determine players, scout enemies and strategies using AI. And we opened up a future filled with new ways to make decisions that will determine the outcome of a particular game, including personnel, training, strategic planning, and in-game choices.
[AI in sports] It was inevitable. Especially in the sports industry where only a certain number of teams are trying to win a championship, every team is about to take a step forward. Addison HunsickerSenior Manager of Football Analysis at Philadelphia Union
“[AI in sports] Addison Hunsicker, senior manager of soccer analysis for the Philadelphia Union of Major League Soccer, said, “We hope that every team will stay one step ahead, especially in a competitive industry, in a sport where there are only a certain number of teams looking to win a championship.”
The beginning
Sports was driven by statistical analysis long before someone called it analysis.
The power hitter was placed in the middle of the baseball lineup for his home run prowess. The running backs handed out football frequently for the yards they accumulated, and frequently handed out plays designed for the highest shooting rate, where the scoring skaters were placed in the center for the impact of scoring.
The analysis has become more refined over the decades. For example, as baseball evolves, lineup decisions and pitching moves are statistically demonstrated that batters are better off against left-handed pitchers, and conversely, pitchers play batters with the same dominant hand.
However, until the turn of the 21st century, true analysis, which is sophisticated, came to sports.
Auckland A is a pioneer and found that certain statistics, such as on-base percentages and walks, are underestimated and use their realizations to their advantage. Over the next 20 years, the analysis expanded and spread across all major sports.
The analysis spread quickly to baseball as sports are the copycat industry and as the team follows the lead of other successful people. Basketball and soccer followed suits, with the Houston Rockets among the analytics pioneer in the late 2000s and the international football club AC Milan invested in the same decade in the same time.
Soccer was perhaps the slowest professional sport to widely adopt analysis.
Kelvin Beachum, the Arizona Cardinals offensive lineman who began playing professionally for the Pittsburgh Steelers in 2012, was not exposed to analysis until several years of his career.
“I think 2013-14 was when analysis started to get things going,” he said. He said that most of the data used by NFL teams is for player rep rather than improving player performance, scouting enemies and making strategic decisions.
Finally, in the late 2010s, players began to get information about their opponents' tendencies and their performance, gaining benefits and improving their play, Beachum continued.
“The past five or seven years have become something [provide] So I can see the defensive purpose I'm going to face over the next few weeks, see where they line up, what defensive tendencies they have,” he said. But when [analytics in the NFL] It started, it was simply to evaluate individuals. ”
Meanwhile, AI had emerged as a way to advance sports analysis by the early 2020s, expanding teams and leagues to machine learning and predictive modeling. Baseball and soccer have a lack of action compared to others, but sports that feature a burst of intense activity have been loaned out to AI.
“Not all teams were doing that,” said Zaimes of UNH. “Not every team has made progress, but technology existed.”
For example, Kamara is strategically located throughout the baseball stadium, and images of each pitch capture hundreds of images, which allow you to know the spin rate of the ball.
In 2020, the Minnesota Twins adopted such metrics compiled over time to build simulation models that help teams quantify how well pitches occur in a particular location, using data management and model development tools from data ablics from data management and analytics vendors.
Similarly, the Bundesliga, the top German soccer league, partners with AWS in 2020 to capture millions of data points per match and combine it with historical data from machine learning models to calculate the probability of shots taken in a match, such as the likelihood of shots taken in a match.
However, this use of AI to acquire strategic benefits has been limited until the past few years.
After teams like Auckland A in the early 2000s pioneered the use of analytics in sports, AI is now taking statistical analytics to a new level.
Barriers to AI
Lack or technical expertise has been a hindrance for many teams.
All of these, especially those below the top level in sports, can't invest in the entire sector dedicated to data science. Even major league soccer, according to Hunsicker, who joined the union in 2020, said the lack of personnel in the top US soccer leagues, not to mention AI, was a major barrier to analysis beyond basic statistics.
“In 2020 there was really nothing, no data infrastructure, no architecture that puts everything in one place,” he said. “I had lots of bosses and sent files back and forth. [to analyze data]. It was all stored manually on spreadsheets, presentations, and OneDrive on our own laptop. ”
The technology available itself is another obstacle, and AI and machine learning development platforms may not be able to handle the data volume needed to properly teach the model to provide accurate results.
“A lot of that was data processing speed, the number of metrics that we could actually process,” Zaimes said. “There were all the advanced analytical metrics, but no one had a processing mechanism to address large datasets to make predictions.”
Another obstacle was isolated data, according to Beachum.
“There are so many different silos right now,” he said. In particular, a company called Catapult collects running and Forceplate tests to determine the amount of force players to use during movements such as sprints, jumps, body monitor data, and neurological data. “There are so many different platforms, so many different use cases, but nothing brings it all together.”
Recent advances in technology have reduced the need to have a team of experts to build and analyze AI models, helping to improve some data silos and processing speeds. And now, AI is becoming more widespread.
“There's a race to understand how to incorporate genai and apply it to use cases,” Hunsicker said.
Eric Avidon is a senior news writer at Informa TechTarget and a journalist with over 25 years of experience. He covers analytics and data management.
