How Artificial Intelligence and Cybernetics Are Changing International Sport

Children born at the beginning of the 2000s compete at the Winter Olympics in China. The youngest athlete is 15 years old, and she is already fighting for medals. As Sergey Kartashov (Sergejs Kartasovs), Senior Partner at IT company Roosh, recalls his years of youth, the Olympic Games were a high-profile event for almost everyone around back then. It took place at the turning point of the 90s. At that time, many Kartashov’s peers chose sports as a means of “escape” from the surrounding reality and its problems. Someone chose boxing and football, someone – chess and orienteering. Watching competitions on TV also brought a lot of pleasure.

Artificial Intelligence Certification provides an overview of AI ideas and workflows, machine learning, deep learning, and performance indicators in order to help learners unravel the mystique around AI and its business applications.

As generations change, so does the reality around us. But none of us could even imagine in our youth how much the impact of technology will change the sport soon. Today, we see how esports tournaments are gathering millions of people around the world, Artificial Intelligence is leading a football club to the championship, and computer programs have become indispensable assistants to sports analysts and coaches. Innovation uncovers new opportunities for athletes to become faster, taller, and stronger.

.

“Those currently competing for prizes in Beijing already have access to developments that their predecessors could only dream of. This is how utopia becomes a reality,” Sergey Kartashov notes.

VAEP, xG, fbref, and other abbreviations

Sergey Kartashov shared that he was lucky to watch a Champions League match with his friends who adore football (soccer) and are pretty savvy in it. He notes that not the game itself on the screen is as interesting as the discussions about it when watching the match with football-savvy people. The fan’s language is a different world, which can be not only interesting but also useful.

“During our viewing, now and then, I heard ‘now the judge will go to watch VAR,’ ‘it is interesting to see the heat map of the movement for the first half,’ ‘he has xG at a rock bottom.’ All these terms are directly related to the digitalization of elite sport,” Sergey Kartashov admits.

Digitalization makes it possible to enhance the training process, the physical and tactical training of football players. It also helps coaches, analysts, scouts, and physiotherapists with their work. Today, this area uses Artificial Intelligence, Data Services, all kinds of technological devices and cybernetics achievements.

One of the latest achievements of advanced technology is VAR, a video assistant referee that has been used in matches since 2018. Its operational principle is quite simple: each controversial episode of the game is displayed on the screen, where the chief referee can review the episode and make a decision. Technology like this has been used in hockey for a long time. In rugby, the analog of VAR is called TMO, a television match official.

Why does professional football require any data tools? They provide versatile game statistics for football analysts; help collect the necessary information both for the team at large and for each player in person.

Thirty years ago, a football analyst was a person with a pencil and a notebook, who was kept continuously making notes during the match. Now, Football Data Scientist, a programmer, is selecting and processing all the data. The basic skill for this specialization is the knowledge of the python programming language. It is used to make a football player’s radar, i.e., a visualized set of individual indicators: a distance covered per match, average speed, and a number of useful actions to the xG coefficient (expected goals). The radar comes to football from NBA, and xG has been actively used in recent years.

“In a nutshell, this model allows to predict the number of goals a team or a player could potentially score at a certain match’s timeframe. It considers dangerous provisions that threaten standards, and unrealized moments. That is, a player can score not much for objective reasons, but a high individual xG indicates that he is a useful and cool striker,” Sergey Kartashov (Sergejs Kartasovs) explains.

One of the services for obtaining analytical information is StatsBomb. It helps to get, for example, extensive information on Barcelona players and the clubs in the Spanish league in general. On Twitter, they often post the latest radars of individual players. However, a downside is that StatsBomb does not provide enough aggregated statistics for the top championships.

Fbref is an alternative free resource that offers many pre-calculated game metrics based on event data. Starting from the 2017-2018 season, it is possible to find aggregated metrics for all the biggest championships.

In 2019, at the international conference on data analysis KDD-2019, a group of scientists from the University of Leuven (Belgium), together with employees of Scisports, presented another metric—VAEP (Valuing Actions by Estimating Probabilities). Football (soccer) analysts consider it an effective way to evaluate the players’ performance on the field though it is not well-known to the general public yet. This metric is developed on a machine learning model that estimates each player’s on-the-ball action on the field and calculate how the game outcome would change as a result of a perfect action. In addition to mathematical algorithms, VAEP is based on the SPADL technology (Soccer Player Action Description Language), a language for describing the players’ actions, which might become the standard in the football analytics area in the future. Briefly, the player’s action score +0.05 indicates that as a result of the tactical and technical player’s action, the probability of scoring a goal increased by 0.05. Accordingly, a score of -0.05 means that the action taken increases the opponent’s chances of scoring a goal by the same coefficient. Based on such estimations, the rating of VAEP football players is made.

Athletes in bras, heat map of movement, computer detects offside

There is one more curious digital innovation in football—the heat map. Its task is to determine the players’ effectiveness individually and the team as a whole. It represents a schematic display in the form of multi-colored geometric shapes—fractals. For example, if a saturated “warm” color prevails on the right flank of the field, then it is through the right flank that the team lined up their attacks and combinations, and the right winger did the most work. The heat map provides a general picture of the game but still, it is compiled manually: an analyst with developed hand motility monitors the match broadcast and captures the activity of a particular team. Each successful action of a football player is “remembered” by pressing the keys opposite his number in a special program. Cameras at the same time monitor the moving speed of players. It is possible to get access to the necessary heat maps, for example, on the professional WyScout platform.

In published photos from athletes’ training, it is possible to notice strange black tops worn over uniforms or on a naked body. The people jokingly called them “bras.” This device is called the Apex Athlete Series, a GPS tracker manufactured by STATSports. It is most widely used in professional football. The black top fixes a tracker on the player’s back and the software transmits and processes data about the player’s actions in real time: how many kilometers he runs, what his pulse and heart rhythms are, what the body’s reaction to the power struggle and fatigue is.

“Besides, the tracker allows collecting data for the heat map as it shows which part of the field the player was most active during the match. Today, such devices are applied not only by football players but also by rugby players, sprinters, and marathon runners,” Sergey Kartashov (Sergejs Kartasovs) says.

In addition, at the World Championships in Qatar this year, computers will record offside positions. Twelve cameras will be installed under the roof of the stadium, and they will record the players’ actions at a speed of 50 frames per second and, in case of offside, transmit a signal to a referee. The unique technology was already tested at the Arab Nations Cup in December.

Artificial intelligence helps to become a champion

Digital technologies in team sports are not only associated with collecting and processing information. Now, AI assembles a lineup of an entire football club. In 2018, British billionaire Tony Bloom bought Union, a moderate Belgian FC from the first division. He developed a program that made it possible to muster up a renewed Union lineup from “no-names”: low-cost football players and free agents. That is, the new owners spent almost no money on transfers, and artificial intelligence reassembled the team from those worth a penny or completely cost-free players. For example, the program found their current top goalscorer somewhere in the recesses of the third German league. The saved money was invested in the stadium reconstruction and the club infrastructure development.

“This reminds me of the plot of the Hollywood movie Moneyball with Brad Pitt: there, too, with limited financial capacity, they used mathematical calculations to assemble a baseball team with little-known players. The story with Union also looks like a movie because the owners’ crazy idea worked, and now the club sensationally ranks first in the top division of the Belgian championship. A little bit more and the championship is in the bag. AI is proving its effectiveness,” Sergey Kartashov says.

Tennis players with iPhones, NBA digital arena, and ice hockey blockchain

Digitalization has impacted not only football but other sports as well. Last spring, Tennis Australia, the governing body of Australian tennis, signed a two-year agreement with SwingVision, an AI-powered video analysis app that uses a smartphone camera to track the movement of a ball and athletes. Over 2,000 Australian coaches got a complimentary subscription. The app works offline; one should install an iPhone or iPad behind the court to record live video while SwingVison provides feedback. Thus, the players get feedback on the shot direction, body position, and footwork. Speed, a shot type, and spinning are also tracked. According to the developers, this year they are going to launch an application for Android.

Professional sports keep on suffering from the impact of the pandemic. Before the new year, the World Junior Ice Hockey Championship in Canada was canceled after two rounds played due to the coronavirus outbreak. Many competitions are still held with no audience. The NBA club Toronto Raptors solved the problem of empty stands with the help of new technologies. They developed a “digital arena” in the club’s mobile application to connect the team and its fans. Using the “digital arena,” it will be possible to communicate in a general chat room, monitor statistics, take quizzes, make predictions on the match, and buy tickets with donations to a charity fund. Toronto management states this is necessary so that fans around the world could feel united despite quarantine restrictions. Now, other NBA representatives, Portland and San Antonio, are going to acquire similar “digital arenas.”

And the New Jersey Devils NHL club works with fans using blockchain. Devils were the first in the league to entering into a partnership agreement with Socios, a creator of digital tokens for fans from the Maltese-based blockchain company Chilliz. Such unique tokens make it possible to choose the best players in the match, vote in polls and take quizzes, get exclusive club media content, prizes and gifts. This is done, according to the authors of the idea, to enhance communication between the club and the fan community. Although, of course, international sport is always big business. Therefore, such innovations as “digital arenas” and fan tokens are primarily aimed at marketing purposes.

Hot winter in Beijing

In early February, the Winter Olympic Games started in Beijing.

“When it comes to the Olympics, you know all the big news even if you are not interested in it. And now, the media landscape is filled with news from the capital of China. And you can see right now how innovative technologies are changing winter sports,” the Senior Partner of Roosh says.

As an example, Sergey Kartashov tells about Jim Richards, an American biomechanics professor, who developed software for figure skaters based on 3D rendering. 40 motion sensors are attached to a skater’s body, and 10 infrared cameras allow building 3D models of athletes on the computer screen. Thus, the training process is improved because the coaches see all the smallest details of the skaters’ movement around the rink. In addition, the riskiest elements to perform can be simulated without putting the Olympians at risk.

A large number of winter sports initially imply a large amount of computer technology that records the speed and superiority at the finish line. For example, bobsleigh. Specifically for this discipline, Omega has developed a technology for recording the sled speed. Each bob is equipped with a speed sensor, an acceleration sensor, and a 3D sensor gyroscope that determines the cornering speed. All data is collected in real-time and displayed on the screen.

As for Canadian skiers, tech company Treadsport Training Systems developed back in 2014 a VR simulator imitating mountain slopes. The program records all the movements of the athletes and their physical condition. Later, coaches will use this data to make adjustments to the training and recovery methods for the sportsmen.

“All this demonstrates us not only the development of sports and sports technologies but also how technologies permeate all aspects of our life and become an integral part of it, allowing us to be Citius, Altius, Fortius!” Sergey Kartashov comments.