To study whether absolute (m/s) or relative (body lengths/s) speed should be used to compare the vulnerability of differently sized animals, we developed a simple computer simulation. Human 'predators' were asked to 'catch' (mouse-click) prey of different sizes, moving at different speeds across a computer screen. Using the simulation, a prey's chances of escaping predation depended on its speed (faster prey were more difficult to catch than slower prey of the same body size), but also on its size (larger prey were easier to catch than smaller prey at the same speed). Catching time, the time needed to catch a prey, also depended on both prey speed and prey size. Relative prey speed (body lengths/s or body surface/s) was a better predictor of catching time than was absolute prey speed (m/s). Our experiment demonstrates that, in contrast to earlier assertions, per unit body length speed of prey may be more 'ecologically relevant' than absolute speed. Copyright 1998 The Association for the Study of Animal Behaviour.