Validation is critical for models that are intended to be used in practice, and is required by some organizations that wish to use models of adversary behavior. Various models have been proposed in the literature to better understand intelligent-adversary behavior, and the strategic interactions between defenders and intelligent adversaries (such as terrorists or other attackers). However, few if any such models have been validated to date because of the sparsity of empirical data, which limits the applicability of the models in practice.
Zhuang, Jun, Vicki Bier and Seth Guikema. 2016. "Introductions to Adversary Behavior: Validating the Models." Risk Analysis 36 (April): 650-652. http://onlinelibrary.wiley.com/doi/10.1111/risa.12621/full