The ability to not execute (i.e. to inhibit) actions is important for behavioural flexibility and frees us from being slaves to our immediate sensory environment. The antisaccade task is one of several used to investigate behavioural inhibitory control. However, antisaccades involve a number of important processes besides inhibition such as attention and working memory. In the minimally delayed oculomotor response (MDOR) task, participants are presented with a simple target step, but instructed to saccade not to the target when it appears (a prosaccade response), but when it disappears (i.e. on target offset). Varying the target display duration prevents offset timing being predictable from the time of target onset, and saccades prior to the offset are counted as errors. Antisaccade error rate and latency are modified by alterations in fixation conditions produced by inserting a gap between fixation target offset and stimulus onset (the gap paradigm; error rate increases, latency decreases) or by leaving the fixation target on when the target appears (overlap paradigm; error rate decreases, latency increases). We investigated the effect of gaps and overlaps on performance in the MDOR task. In Experiment 1 we confirmed that, compared to a control condition in which participants responded to target onsets, in the MDOR task saccade latency was considerably increased (increases of 122–272 ms depending on target display duration and experimental condition). However, there was no difference in error rate or saccade latency between gap and synchronous (fixation target offset followed immediately by saccade target onset) conditions. In Experiment 2, in a different group of participants, we compared overlap and synchronous conditions and again found no statistically significant differences in error rate and saccade latency. The timing distribution of errors suggested that most were responses to target onsets, which we take to be evidence of inhibition failure. We conclude that the MDOR task evokes behaviour that is consistent across different groups of participants. Because it is free of the non-inhibitory processes operative in the antisaccade task, it provides a useful means of investigating behavioural inhibition.