The eye movement system reacts very systematically to visual transients that are presented during the planning phase of a saccade. About 50 to 70 ms after the onset of a transient, the number of saccades that are started decreases, a phenomenon that has been termed saccadic inhibition. Saccades started just before this time window are hypometric compared to regular saccades, presumably because the presentation of the transient stops them in mid-flight. Recent research investigating the properties of repeated saccades to fixed locations found that these early saccades were additionally faster than expected from the main sequence relation, suggesting that a saccadic dead time during which saccades can no longer be modified does not exist. The present study investigated the properties of saccades to random locations in a guided saccade task. As expected, early saccades starting just before the saccadic inhibition dip in frequency were hypometric. Their velocity profiles implied that these saccades were actively stopped after reaching peak velocity. However, the peak velocities of these saccades did not generally deviate from the main sequence relation. The question whether an active stop of early saccades is incompatible with the idea of a saccadic dead time is open to debate.