Evaluation methodology that tests deployment rules by holding agents and objectives fixed while varying one rule to measure its causal effect on collective behavior.