Abstract: This paper presents several possible complexity-based measures of effectiveness (MOE) that address some of the limitations imposed by the use of attrition as a primary MOE in evaluating results of combat models. The considered measures are Shannon entropy (especially spatial entropy), fractal dimension, symmetropy, and the Hurst coefficient. Some of these measures have been used before (entropy, fractal dimension), others have not (symmetropy, Hurst coefficient). The suitability of the measures for the assessment of crowd confrontation operations (CCO) is discussed. CCOs are a prime example of a case in which attrition provides inadequate information. At the same time, crowds, due to their inherent complexity are an excellent test bed for the proposed measures. It is demonstrated that the complexity-based MOEs are effective in capturing the temporal dynamics of CCOs. Furthermore, the use of these MOEs allows for identification of emergent behaviours and patterns in the system dynamics. It is suggested that the changes in crowd dynamics due to the interaction with the law enforcement are consistent with phase transitions in critical systems, namely with a self-organized criticality.