Autotomy, the self-amputation of limbs or appendages, is a dramatic anti-predator tactic that has repeatedly evolved in a range of invertebrate and vertebrate groups. In lizards, caudal autotomy enables the individual to break away from the predator's grasp, with the post-autotomy thrashing of the tail distracting the attacker while the lizard makes its escape. This drastic defensive strategy should be selectively advantageous when the benefit (i.e. survival) exceeds the subsequent costs associated with tail loss. Here, we highlight how the position of autotomy along the length of the tail may influence the costs and benefits of the tactic, and thus the adaptive advantage of the strategy. We argue that most studies of caudal autotomy in lizards have focused on complete tail loss and failed to consider variation in the amount of tail shed, and, therefore, our understanding of this anti-predator behaviour is more limited than previously thought. We suggest that future research should investigate how partial tail loss influences the likelihood of surviving encounters with a predator, and both the severity and duration of costs associated with caudal autotomy. Investigation of partial autotomy may also enhance our understanding of this defensive strategy in other vertebrate and invertebrate groups.