The author of this article is Dr. Sakthi Balan Muthiah, Professor, Department of Computer Science and Engineering, Shiv Nadar University, Chennai.
Sports analytics offers many interdisciplinary opportunities in education. It will become clear when we see the evolution of the topic. Let’s start by asking a few questions:
Which batter has the best strike rate in the IPL?
Where does the batter score most of the runs?
Does the batter’s strike rate decreases after playing three matches in a week?
Show me all the sixes hit by the batter over cover and extra cover.
Is the batter at a risk of getting injured in the next week?
Which fans would buy seasonal IPL tickets if they were offered some personalised discount?
The order of questions listed above signifies the evolution of sports analytics. These questions range from descriptive to diagnostic to predictive to prescriptive and then to commercial analytics.
Data-driven sports analytics has been evolving rapidly. The first instance of documented application of data analytics in sports was in the US Baseball League (MLB). During the late 1990s and early 2000s, US Baseball teams were competing to get the best players in their team. Teams with bigger budgets get the better chance to get star players but in contrast, the teams with less budget had to go with whoever was available. A team with limited budget called the Oakland Athletics wanted to get a team that can challenge and compete with the top teams. Billy Beane who ran the team used sabermetrics to get under-valued players so that they spend less but gain more – they were quite successful. Sabermetrics are statistical metrics defined by Bill James for the game of baseball that are more insightful than traditional metrics. Some of the metrics defined are on-base percentage, slugging percentage and so on. His work inspired the book, Moneyball which later got featured as a film too.
Descriptive analytics which were mainly based on historical data and statistical metrics, evolved to deeper analysis like tracking the player movements and positioning – like how much area Nadal covers on the tennis court. This diagnostic analysis in the mid of 2010s helped the analyst, coaches and players to dwell much deeper into the performance and make appropriate decisions. In 2018 and later, this transformed to a more integrated approach. Various modalities of the data were analysed to make decisions for fitness, optimal performance, medical, scouting and so on. With advancement in action recognition in videos in early 2020s, researchers used computer vision techniques to try automated video tagging in sports videos to search and retrieve specific segments of the videos for further analysis.
Automatic tagging and searching in videos, which are still evolving, can help answer queries like, “Show me all the backhand slices of Federer from the back of the court.” With advancements in sensor, IoT and AI, analytics has also evolved into prediction and real-time monitoring of players through wearables and biomechanics – for example, Str8bat is a smart device that makes a bat, smart cricket bat. It captures the bat dynamics during the play – it can capture finer aspects like the backlift angle, bat swing path, bat twist at the moment of impact and so on.
These insights are very crucial for the coaches to make finer adjustments when the batters play in different weather conditions and pitches. Analytics also extend further to the domain of attracting fans using various ticketing techniques and fan engagement analytics. Prescriptive analytics is getting more mature with the advent of new techniques and technologies. With the data collection of players becoming crucial, it also emphasizes the importance of safety of sports and medical data.
Sports analytics transforms data into actionable insights, and it enhances decision-making and certainty in preparation. But at the same time, sports remain rooted in the uncertainty of human performance and competition.