This paper is concerned with the modeling of the behavior of players operating in a competitive environment that is characterized by interactions amongst players or groups of players. Thus the paper describes a new, online, hierarchical, probabilistic modeling architecture that is based on hidden Markov models (HMMs). For the purpose of online behavior recognition, a probabilistic decision tree is implemented that accepts HMM behavior probabilities of player and effectively segments their behavior-with-time trajectories. This allows the location of important points in time where behavior changes occur. Furthermore, the hierarchical nature of the system allows individual player classification results to be used towards the modeling and classification of higher-level tactical behaviors of groups of players, as defined within an application envelope. The system is applied in a relatively simple "air-patrol" scenario and system simulation performance results are provided in terms of certain useful metrics.