There are two commonly accepted paradigms for organizing intelligence in robotic vehicles, namely, reactive and deliberative. Although these paradigms are well known to researchers, there are few published examples directly comparing their development and application on similar vehicles operating in similar environments. Virginia Tech's participation, with two nearly identical vehicles in the DARPA Grand Challenge, afforded a practical opportunity for such a case study. The two Virginia Tech vehicles, Cliff and Rocky, proved capable of off-road navigation, including road following and obstacle avoidance in complex desert terrain. Under the conditions of our testing, the reactive paradigm developed for Cliff produced smoother paths and proved to be more reliable than the deliberative paradigm developed for Rocky. The deliberative method shows great promise for planning feasible paths through complex environments, but it proved unnecessarily complex for the desert road navigation problem posed by the Grand Challenge. This case study, while limited to two specific software implementations, may help to shed additional light on the tradeoffs and performance of competing approaches to machine intelligence. © 2006 Wiley Periodicals, Inc.