In the stop-signal task, subjects should withhold their response in a choice reaction time task when a stop-signal, usually a tone, is presented. Successful stops have been associated with event-related potentials (ERPs) featuring a larger frontocentral positivity relative to failed stops. The functional interpretation of this stop-P3 has been disputed, because stop-ERPs are distorted by overlap from ERPs elicited by preceding go-stimuli. We effectively removed confounding potentials with the ‘adjacent response filter method (ADJAR)’. Confirming an interpretation in terms of response inhibition, the stop-P3 remained and overlap removal resulted in a more anterior distribution. As a new finding, the N1 was larger on trials with successful stops, which suggests that inhibitory performance at least partly depended on the ability to switch attention to the stop-signal. Finally, the parietal P3 tended to peak earlier for successful than for failed stops. This is in line with the Horse Race Model, which states that faster stop-processes have a higher chance of winning the race against the go-process.