Russia Has a New AI Soccer Analyst
Sberbank’s AI4SPORT team has developed PC-SSAS (Player-Centric State-Space Action Spotting), a model that can automatically identify passes, shots, tackles and the actions of individual players from match video. That opens up extensive possibilities for sports analytics and for how teams use those insights afterward.
The new technology turns match footage into a detailed digital timeline of the game, which can be used to build maps of passes, shots and tackles, evaluate individual players’ actions and analyze how attacks develop. According to the researchers, the system uses the Mamba architecture to process the temporal sequence of events, while a specialized training scheme improves its ability to recognize rare events.
What Makes the Model Different
PC-SSAS does more than find a relevant moment in match footage – it links that event to a specific player. The result is a detailed timeline built from the broadcast that can reveal each player’s contribution to the match. The key value is that the system generates this data automatically. That makes it useful for soccer analytics platforms, clubs, leagues and developers of sports technology.
AI Takes the Field
Sports analysts see considerable potential in AI and computer vision, while technology teams continue to build new tools around them. At the 2022 World Cup in Qatar, FIFA deployed a semi-automated system in which 12 cameras tracked players using 29 points on their bodies, while algorithms helped automatically determine offside positions. It became one of the first large-scale examples of computer vision being used directly in world soccer.
In 2023, Yandex and VSporte launched a Russian computer vision system for the Russian Premier League, or RPL. Using four 6K cameras and a neural network, the system began tracking players and automatically generating statistics on their movement, speed and other metrics. Starting with the 2023–2024 season, the platform became an analytics provider for the RPL.
The development of PC-SSAS continues Sber’s work on AI for sports. In July 2026, Sber AI and CSKA unveiled a technology for long-term identification of soccer players from video captured by just one camera, with the research earning Best Paper Award recognition.
The strength of Russia’s work in this field is also reflected in international competition: The AI4Football team behind the technology finished second among 19 teams in the international SoccerNet Player-Centric Ball Action Spotting Challenge 2026.
AI Helps Develop Human Talent
AI-powered sports analytics is advancing rapidly in Russia. Existing algorithms can already help teams structure their work and training more effectively. While early systems mainly answered the question of where a player was, newer models aim to determine who performed an action, what they did and when. Automated match analysis can build richer data profiles for individual players, helping teams make better use of each athlete’s potential.
Experts expect the technology to evolve further. Several AI4Football technologies could eventually be combined into a single pipeline: video, player identification and tracking, recognition of match events and automatic generation of statistics. Together, these capabilities could substantially simplify analytical work for coaching staffs. Experts also point to the technology’s international competitiveness: The team’s result at SoccerNet 2026 shows that Russian researchers are working at a high global level.
