Lex is a machine learned referee system for 3v3 pickup soccer games. The Lex system includes smart goal posts, a soccer ball, unique machine recognized uniforms, an instruction manual, carrying bag and the machine trained referee.

Lex is designed to explore the question of decision automation in sports and whether a machine learned referee is actually more objective than a human referee. In the pickup domain, Lex explores how behaviors of players would change in the presence of a machine learned referee and what the limitations of this technology are. In exploring these questions Lex provides a critical view of the future of machine intervention in sports and insights into how we can better design this collaboration for the future.

Using color detection of the soccer ball, the goal posts record and display when a goal is scored. Each uniform has a QR identicon unique to the player printed on the socks and reversible jerseys. This unique identifier allows the referee to recognize when a particular player commits a fault, scores a goal or is given a penalty. These game statistics are tracked by the referee and can be later viewed on the LEX website alongside metrics of all the games taking place worldwide. The different calls made during the game are visualized by the LED patterns on the main frame of the referee. A legend of these calls are found in the instruction manual.