Using reflective learning to master opponent strategy in a competitive environment

Cognitive models of people interacting in competitive environments can be useful, especially in games and simulations. To be successful in such environments, it is necessary to quickly learn the strategy used by the opponent. In addition, as the opponent adjusts its tactics, it is equally important to quickly unlearn opponent strategies that are no longer used. In this paper, we present human performance data from a competitive environment. In addition, a cognitive model that uses reflective learning is introduced and compared to the empirical findings. The model demonstrates that it is possible to simulate learning in an adversarial environment using reflection and provides insight into how such a model can be expanded.