The NHL combines AWS cloud infrastructure with existing technology to offer innovative data analysis while increasing savings.
According to Forbes, the sports analytics market could reach nearly $4 billion by 2022. So, it shouldn’t be a surprise that Amazon Web Services (AWS), the cloud computing services company, is collaborating with sports leagues and teams to enhance team performances and viewing experiences. Two interesting examples include the National Hockey League (NHL), and the German soccer league, Bundesliga (Read more about the AWS Soccer story here).
In the United States alone, millions of people either watch or participate in ice hockey in one form or another. It is also the national winter sport of Canada and extremely popular in Europe and Russia. As for the NHL, it is considered the premier professional ice hockey league in the world.
As the official cloud, artificial intelligence (AI) and machine learning (ML) infrastructure provider of the NHL, AWS is looking to introduce its analytics and technology throughout the current season. “AWS’s state-of-the-art technology and services will provide us with capabilities to deliver analytics and insights that highlight the speed and skill of our game to drive deeper fan engagement,” said NHL Commissioner Gary Bettman.
Expanding the Data on the Ice With AWS
Initially, available data and graphics only included puck and player positioning, puck and player speed, shot and skating distance, time on ice, shift length and zone time. Thanks to AWS technologies, this season’s full rollout includes even more metrics: save percentage by distance, location and positioning, shot speeds, shot types and locations, team possession times, and zone entries and exits.
With the flood of new data, the objective is to enhance the fan experience for both those inside the arena and for those watching on television, via mobile devices or, in some cases, both, with a digital stream broadcast dedicated to showcasing puck and player information. The fans and broadcast partners will have access to information such as puck speed, player speed, player distance covered on the ice and precise information about each player’s ice time.
Another aspect has been the changing legal status of sports betting across different states. The NHL formed a partnership with MGM Resorts International, allowing the latter to offer its products to fans via the official NHL broadcast. MGM will have access to all the NHL’s enhanced proprietary data, which will allow it to offer customers specially tailored products.
“The new sports betting landscape presents a unique opportunity for fan engagement utilizing technology and data that are exclusive to our league,” announced Bettman. “As a leading global gaming operator and entertainment company, MGM Resorts is the perfect partner for us to begin our transformative entry into this space. Fan engagement, technological advancement and innovation are paramount to our progressive approach and will be at the forefront of everything we do.”
Data on Ice: AWS Provides Assists to the NHL’s Existing Technology
The NHL’s Puck and Player Tracking (PPT) technology being expanded this season was introduced at the 2019 All-Star game. PPT runs on the AWS cloud infrastructure, thus allowing better analysis of game play for fans, teams and media partners. The PPT system, developed by the league’s stats and broadcast partners SportsMEDIA Technology (SMT), tracks every movement of the puck during a game. “The Puck and Player tracking system can track pucks at a rate of 2,000 times per second in real-time with inch-level accuracy,” detailed Bettman at the 2019 NHL All-Star Innovation Spotlight. The PPT can also track each player 200 times per second.
A view of what the PPT system looked like in 2019.
HITS, also developed by SMT, has been the NHL’s official scoring system since the 2007-08 season. In 2019, SMT updated HITS to integrate with its OASIS (Organization of Asynchronous Sports Information Subsystems) platform. OASIS contains an AI engine called EIEIO—a tongue-in-cheek acronym for Eventing Intelligence Engine Inside Oasis. EIEIO takes full advantage of the live PPT data to enhance the NHL’s official stats and centrally manage data for the more than 1,300 games played every season.
By applying AWS’s machine learning services to game video and official NHL data through the Amazon Machine Learning Solutions Lab, the NHL will be able to develop and share advanced game analytics and metrics. This will include data from the PPT system as well as the Hockey Information & Tracking System (HITS) real-time stats.
Additionally, by leveraging Amazon Kinesis (AWS’s service for easily collecting, processing and analyzing video and data streams in real time) and machine learning services such as Amazon SageMaker (AWS’s service that enables data scientists and developers to build, train and deploy machine learning models quickly), the NHL will be able to audit its feeds to broadcast partners in real time. This capability will enable the NHL to create a smart monitoring system that will detect and automatically fix any potential feed issues.
The NHL is also building an enterprise video platform on AWS to aggregate video, data and related applications into one central repository. This will enable easier search and retrieval of archival video footage to facilitate newer and more detailed in-game analyses, predictions and video highlights. It will give broadcasters instant access to NHL content for syndication and licensing.
In addition, the NHL intends to use AWS Elemental Media Services to develop and manage a cloud-based HD and 4K video content delivery system. It will encode, process, store and transmit game footage from a series of new camera angles to provide continuous video feeds that capture plays and events outside the field of view of traditional broadcast cameras.
The Smart Puck Drops: Latest Tracking Technology Utilized By the NHL
The new hockey puck is fitted with a tiny embedded battery, a circuit board and 6 one-inch tubes that emit infrared light at a frequency of 60 Hz. The new pucks, though more expensive to manufacture at $40 each, are still made of vulcanized rubber and look the same as the old model —they weigh six ounces and are three inches wide and one inch thick. First, Soucy Baron, a rubber manufacturer outside of Montreal, produces the raw form puck. SMT then installs a battery, circuit board and light tubes within each puck.
After they are branded, the pucks are designated specific numbers for tracking purposes. The puck battery does not need to be rechargeable, as pucks are generally changed after every whistle to maintain their temperature. Hockey pucks are kept frozen before use to make them more solid, thus reducing bouncing and improving control. Hence, they will not be on the ice long enough to wear out their batteries.
Tracking the players posed a trickier problem. Due to the game’s speed, the close quarters of play, and the players’ bulky uniforms, it is more difficult to identify the players than in other sports like soccer or basketball. While there are plenty of distinguishing features of players visible in other sports, ranging from skin and hair color to differing footwear, hockey players only have their name and number displayed on the back of their jerseys. This could lead to incorrect identification and cause confusion in the data collection process.
The solution was simple yet elegant. Players were fitted with a rechargeable infrared tags placed in a pouch sewn into the back of their jerseys. Each player is assigned multiple tags per game to ensure that even if the players change jerseys, the replacement jersey will have a tag assigned to the player ID and the player can still be tracked.
Each of the 31 hockey arenas has 14 to 16 antennas installed in the rafters that receive the signals from all pucks and player tags, while four cameras support the tracking functionality.
The NHL has spent tens of millions of dollars over many years developing the system. Over the past couple of years, players and TV broadcasters experimented with the technology in the NHL All-Star Games to ensure that any kinks were ironed out before the major rollout this season.
According to Steve McArdle, NHL executive VP of digital media and strategic planning, the post-season experiments went smoothly, and clarified that it was more about familiarizing people with the system. “It’s about making sure the equipment manager is really appropriate when they put the tags in with the jerseys,” McArdle said. “This is about making sure the folks in the box know the right process to activate the puck before it’s given to the referee and put into play.”
Top Shelf Metrics: How the New System Benefits Stakeholders
Thanks to the AWS cloud infrastructure, the NHL is now able to maximize the potential of its existing PPT system. “The theory behind player and puck tracking was to give people insights into the game who maybe would learn how special the game is and would understand it a little bit better and so we started with the possibility of having broadcast enhancement,” said Commissioner Bettman. “But now in the era that we’re in, the opportunities were limitless. If you’re a millennial or a Gen Z in particular and you’re consuming sports differently than it’s ever been consumed before, we’re going to be right there for you giving you what you want.”
AWS is also working with other sports to fulfill their potential using its AI and ML cloud infrastructure. As a result, fans are becoming more knowledgeable and involved, teams are becoming more competitive, and leagues are becoming more efficient at marketing and cost savings. However, this is only the beginning. Sports fans around the world are looking forward to what new directions AWS will help sports take in the future. For a further look into that future, read this.
