Collective sports comprise of cooperative and competitive organizational processes in which players have to be predictable for teammates and sufficiently diverse and unpredictable for opponents in order to satisfy the goal constraints. In highly stable, repetitive and thus predictable (cooperative or non-cooperative) environments the adaptive system may converge to a minimum unpredictability by minimizing the diversity of accessible patterns of functional behaviour. However, sport competition, akin to life itself, is not a highly stable and repetitive process based on negotiating stable environment. It creates conditions where highly demanding non-cooperative behaviour of the environment is rather rule than an exception. In such stochastically changing and non-cooperative environments, systems (players and teams) must develop high behavioural unpredictability potential to increase their fitness to the competitive environment. This means that the adaptation process rests on a tendency of permanent increase of the unpredictability potential which affords the ultimate goal, the survival (winning) of the system in sports environments. This state of affairs suggests that performance may be conceptualized in terms of unpredictability/diversity potential and that its development can be understood by the juxtaposition of three principles of adaptive systems: (a) relativity of organism-environment unpredictability; (b) satisficing diversity/unpredictability potential; and (c) tendency towards non-decreasing unpredictability/diversity. (a) Relativity of organism-environment unpredictability: Collective sports adaptation is a process of becoming fit to one's environment. Becoming fit means that the environment becomes more informative and less constraining for the player/ team and, on the other hand, the player/team becomes less informative, and thus more unpredictable for the environment. The relativity of organism-environment unpredictability is based on the conservation of bio-motor information principle (Hristovski, 1989). This principle says that performance variables, including motor abilities (effectivities), such as power, agility, speed, strength, endurance, etc., may be treated as entropy measures forming a state space in which training induces conversion of entropy into information. By increasing the effectivities,