As every coach and player knows, it takes a lot of knowledge to master the play of many sports. Often the difference between outstanding play and average play in a sport is the result of keen knowledge of the sport—its rules, tactics, strategies and competitive advantages—more so than mere athletic ability. Playing smarter can be more important than simply playing harder.
And as every fan knows, it takes knowledge of the sport to appreciate the play of the sport and the skill of the teams and players themselves. Without that knowledge, a spectator can find watching a fast-moving and complex competitive sport confusing. To be a fan, a spectator must acquire knowledge about the sport in order to enjoy watching the play. In fact, a measure of the devotion of a fan is the depth of knowledge about the sport, it’s teams, players and coaches that the fan has learned.
But important knowledge content within a sport can be hard to acquire, despite a library of books and fan magazines. Innovations and changes in equipment, coaching and the rules of play of the sport make knowledge of the sport a moving target. Coaches, players and fans alike need to continually update their understandings of the sport. In today’s fast moving digital world, new ways to manage the knowledge content of sport are needed.
Adding to the challenge of sport for the coach, player and fan is the rapid change in sports data collection and distribution. While scoring and player performance statistics have always been important in sports, the growing ability to collect data from the players and the field of play have added new dimensions to sports data collection and analysis. Radio frequency digital data links can instrument almost anything from racing cars to players’ footwear. Video surveillance of the area of play can capture small changes in player movements and team positioning in near real-time. Wireless networks can analyze and distribute the data collected during play to hand held devices for the coaches, players and fans.
This astonishing ability to collect data, however, creates a new challenge for sport—what does the data mean? It is more than just player and team statistics. Buried in the data are the subtle differences between the outstanding player and team behaviors and those that are more ordinary, even for the same player or team. But, to find this information lying hidden in the vast amount of data that is or can be collected will require a better tool than player or team statistics.
Fortunately, the field of knowledge engineering may hold the key to enhancing sport for the players, coaches and fans. The term knowledge engineering can be defined as the structured effort to organize knowledge so that it is accessible and usable by other entities. The result of knowledge engineering is a knowledge base of structured representations of knowledge. To some degree, familiar activities such as writing a book or preparing a course of instruction involve informal knowledge engineering. Knowledge engineering came of age as a formal discipline with the development of machine intelligence. In order to structure knowledge so that a computer could use it, our understanding of knowledge and its formal representations took a big leap forward. In fact, it is just this ability to provide sports knowledge in a form usable by a computer that enables the exploitation of sports data which today is possible to collect but hard to analyze in depth.
The first impression of knowledge engineering is that it somehow involves rules of logic. Can the nuances of a sport be reduced to rules in this manner? It seems unlikely that rules could describe a good first serve into the wind in tennis or a good pass on a soggy rugby field. Modern knowledge engineering has grown beyond the notion of rules, however. An example of this is the way in which the situation of play and the strategy, tactics and actions of the players are often represented separately within a knowledge base. Situations are used to represent the state of play: where the players are and the status of their resources. The strategy and tactics knowledge describes what can be done within the rules of play to achieve specific goals and sub-goals that form a network of connected tactics. A third set of knowledge describes how players can perform a sequence of movements that carries out a tactical step, such as a centering pass. When these three types of knowledge are combined, a dynamic, contextual understanding of the play of a sport can be achieved, moment by moment, from the data that is collected.
This computer-based understanding of play, however, is much easier to maintain than a book or video library. Using knowledge engineering, new situations can be described, new strategies and tactics can be defined and new techniques for player actions can be created, allowing the computer to immediately understand the changes. When rules or equipment for a sport change, outdated knowledge can be easily corrected or made obsolete within the knowledge base, along with any other knowledge that is indirectly impacted by the changes. Unlike a book, a new knowledge base can be transmitted over the internet and immediately put into use inside a computer.
Of course, the goal of this knowledge is to enhance the sport for the benefit of the players, coaches and fans. Once the knowledge is inside a computer, there are many roles it can play in a sport.
During the play of the sport, future knowledge-based systems may be able to analyze the play of a team and its competitors and provide suggestions for changes in strategy or tactics based on the actual unfolding situation. An example of this role of knowledge is the Racing Associate concept from Sportronix (www.sportronix.com), which is intended to provide advice to a racing team during the conduct of a race. The knowledge in the system concept uses telemetry data from the race cars of a team and scoring data from the track systems to understand the changing race situation, including the health of the car and the driver’s behaviors. It then generates strategy and tactics suggestions in near real time to the pit crew chief to enhance the race outcomes for the team.
For coaches, computer-based knowledge of play could allow a better analysis of player and team behaviors, including the refinement of strategies, tactics and techniques that maximize the employment of player individual capabilities. This would allow a coach to separate out issues that are the result of specific performance of techniques from those issues that are related to the player or team’s larger understanding of strategy and tactics in play. This information has the potential to increase the effectiveness of coaching and player development.
In the future, combining a knowledge-based tool with an existing video analysis tool, such as the ProSuite software offered by Dartfish (www.dartfish.com), may allow the analysis of player behaviors to move from the practice field into the actual play of the game. The knowledge of the actual situation and tactical play could enable a refined analysis of the player game video, leading to more effective coaching. For example, the analysis of a golf swing could take into account specific information about the lie and terrain on the course, location of hazards, wind conditions and the location of other competing players’ current scores.
One of the most important uses of sports knowledge engineering is support for the fan. The economics of sport depend on the interest and excitement of the fan, and only through knowledge about the sport does the sspectator become the fan. By making knowledge about a sport more accessible to the fans, the loyal audience of a sport is greatly increased.
One way that sports knowledge can be used to drive fan understanding and loyalty is through the personalization of sports content delivered to the individual fan. Major steps in this direction are already underway by companies like LiveClips (www.liveclips.com) that offer personalized video clips of sport events, delivered to personal video devices either during or after the play of the sport.
Not all fans are equal in their understanding of the play or of the teams and players in a sport. For novice fans, it may be challenging to describe what kinds of video clips are interesting and how the clips should be delivered. In many cases, additional text or voice commentary may be needed to explain the clip content to the novice fan. Experienced fans however may have less need for commentary or annotations within the clip.
By combining the video clip capability with a future sports knowledge system that includes knowledge about types of spectators and fans, a personalized commentary at the depth needed by the fan could be created. This could increase fan connection with the sport, and may add significantly to the experience of new fans that are still learning about the sport.
A second important use of sports knowledge that supports the fan is the development and refinement of sports video games. Players of sports video games often become strong fans of the sport later in life. But in order to create a realistic setting for a sport as a video game, all of the player and coach models must show solid sports strategy, tactics and performance. Underlying the success of these sports video games is a strong knowledge engineering effort that must capture and represent the nuances of a sport in order to implement an interesting and involving video game.
There are many reasons for embracing the discipline of knowledge engineering within the world of sport. Yet this topic is new to sport, and will require thought leadership to adopt the practices of knowledge engineering. But for those that do, there are significant benefits. The ability to quickly and correctly understand and exploit the situation in sport, and to allow the fan to share in the nuances of the play could elevate a sport, a team or a player to new levels of fan interest, excitement and devotion. This enhanced understanding of a sport, even by a novice, is the goal of sports knowledge engineering.