The National Hockey League’s Edmonton Oilers are favourites to return to the Stanley Cup Final for a second consecutive season, despite starting the season by getting outscored 15-3 in their first three games. It wasn’t until an overtime victory against the Utah Hockey Club on November 29 that they raised their record to 12-9-2, securing more wins than losses for the first time and maintaining that positive record from there on.
The numbers don’t lie.
When you review the Oilers' box score after a game to check on Leon Draisaitl, the NHL’s leading goal scorer and front-runner for the 2024-2025 Hart Memorial Trophy, you’ll see the usual statistics: goals, points, shots on goal, minutes played, and plus/minus, among others. What you might not notice, however, is the intricate web of advanced analytics being tracked and studied far beyond those raw figures – patterns, probabilities, and predictions that are helping teams like the Oilers improve their performance, identify weaknesses, and gain a competitive edge on the ice. Beneath the surface, cutting-edge data and metrics are quietly reshaping how NHL games are played, won, and watched.
Meghan Chayka is a renowned Canadian data scientist and co-founder of Stathletes, the NHL’s go-to data performance and analytics supplier. Stathletes is a privately owned firm with over 200 employees, projecting over $10 million in revenue in the next fiscal year.
Stathletes, operating within a rapidly expanding sports analytics market, offers industry-leading hockey insights across 31 hockey leagues in 34 countries and has successfully scaled to the highest professional levels in North America and Europe.
The company was founded in 2010 following Chayka’s part-time venture into sports analytics. She and Neil Lane started filming her younger brother, John Chayka, during his Ontario Junior Hockey League games. Using a video analysis method, they collected data to help him improve his game, producing various statistical data points in real time. Today, that same video analysis approach has merged with new software tools and scalable data products and is utilized by every NHL team on a nightly basis.
Stathletes' analytics software, powered by machine learning, captures and analyzes game footage to provide detailed data on movements, actions, and performance metrics. Their proprietary tracking software gathers event data at a resolution 100 times greater than conventional methods and tracks millions more data points per game, helping teams minimize unpredictability and make informed decisions to improve player development and in-game strategies. Consider it to be examining the statistics that support the statistics.
Through their StathletesTV product that integrates data and video, each data point connects directly to a video timestamp. This allows coaches and analysts to efficiently review footage of specific plays or player actions.
“There are times when a player’s struggles become clear in a data set, such as changes in zone entry patterns. If they previously attacked through the middle with speed but shifted their style, data and video can help highlight a successful past play, showing how attacking with pace and creating off the rush made them effective,” said Chayka.
Predictive modelling is a key application in Stathletes' approach to understanding the game, tracking many statistics that are not visible to the human eye all at once.
“Expected Goals is one of the [most useful] metrics. It quantifies scoring chances by considering variables like pass location, goalie position, shot speed, type of shot, and how good a scoring opportunity is.”
“Player context also matters; a shot by Auston Matthews differs from one by a replacement-level player. For goalies, metrics such as Goals Saved Above Expected are valuable. Traditionally, goalie performance was hard to assess, but now we can track aspects like their stance, movement, and how they make saves, whether with one pad down or using other techniques.
“Passing models help us understand how players move the puck. In the past, it was hard to quantify playmaking in the box score, even though players like Mitch Marner and Matthew Tkachuk are known for creating a lot and being smart about how they pass and make plays. Now, with metrics like Shot Assists [passes that lead to shot attempts], we can give value to those players and better quantify how much their playmaking helps in a game.”
Although it’s easy to trust the science, concerns have been raised about excessive reliance on analytics, potentially diminishing the human element of performance and decision-making, along with the failure to account for intangible factors like leadership, motivation, work ethic, team chemistry, or adaptability – aspects of an athlete’s repertoire that, perhaps, can’t be quantified.
“A lot of times, the work ethic of professional players sets the framework for their continued progression and focus. So I think it’s a big part of the game, and I definitely don’t underestimate that,” Chayka said. “Are there ways to quantify [intangibles]? I mean, yes. You can quantify whatever you want, more anecdotally, by working with coaches, scouting and player development staff, or sports science departments, and even going through written reports while having some way to aggregate them.
“Just letting a computer play the game isn’t fun, but I think there’s a downside risk of not using analytics to your advantage.”
The global sports technology market size is projected to reach USD $61.4 billion by 2030, according to Yahoo Finance – a reflection of the increasing demand for enhanced fan experiences and the adoption of data analytics. Chayka views sports analytics as a key component of the broader “sports tech” industry, increasingly intersecting with tech giants like Meta, Amazon, and others. She believes the future of sports analytics will focus on broader tech applications rather than just data analysis.
“Automation, including computer vision, AI, and machine learning, makes it easier to collect and share data, allowing people to access it on their own. This leads to more consumer-led applications where fans can track their favourite teams or players, or find new ways to consume broadcasts. A specific example could be creating 3D forms with computer vision, like watching Carlton the Bear play, which might appeal more to kids than real-life broadcasts.”
The numbers may not lie, but they are only part of the story. Sports analytics, once viewed as a limited toolset used occasionally to understand performance, are now redefining the sports landscape. With AI advancements still on the rise, players, coaches, and fans alike can continue to expect significant changes in how they consume and engage with their favourite sports.
Follow: @aleksa_cosovic
