Amazon Web Services, Inc. (AWS), an Amazon.com, Inc. company (NASDAQ: AMZN), and the National Hockey League (NHL) confirmed the launch of Face-off Probability. Face-off Probability is alive, in-game NHL stat that will be highlighted as a graphic that immediately displays the chances of a player winning a face-off and ownership of the puck and presents them on screen for fans watching the broadcast of the game.
According to the NHL, Face-off Probability is being highlighted as part of NHL Edge IQ. By using AWS’ machine learning services on game footage and official NHL data, it can create and distribute analytics and insights like Face-off Probability, which provide fresh in-game analysis and forecasts to improve live broadcasts and immerse spectators in the game.
Furthermore, the NHL’s new ML-driven Face-off Probability model, which is driven by AWS technology and developed in collaboration with the AWS ML Solutions Lab, utilizes cloud capabilities like analytics, storage, serverless computing, and media services to decide a player’s likelihood of success and then displays that probability as a graphic on the live game broadcast for viewers.
Face-off Probability Model
In addition, when the game is over, the Face-off Probability model starts determining victory probabilities for the players it thinks will take the next face-off. The probability is determined by factors such as the number of players on the ice, the location of the face-off, and the current game condition. Face-off Probability is one of the first machine learning (ML)-driven analytics generated by NHL Edge IQ, which is supported by Amazon Web Services.
The Face-off Probability model is also adaptable, changing forecasts as the game situation changes. If a player is kicked out of a face-off due to a rule violation, the model adjusts the forecast to the new matchup using a real-time puck and player monitoring streaming sensor data with sub-second latency.