Application of a supramaximal electrical twitch to the voluntarily contracted muscle is used to assess the level of muscle activation. Large variability in the interpolated twitch torque (ITT) has been observed when repeated stimulations are performed. It is hypothesized that this variability in ITT is caused by the stochastic nature of the timing of twitch application relative to pulses of voluntary excitation trains. Two experiments were performed on 12 subjects each to test this hypothesis. For the first experiment, a single twitch was superimposed on a train stimulation at different time intervals relative to the train pulses. For the second experiment, single, double, triple, or quadruple twitches were applied on a voluntarily contracted muscle. The ITT critically depended on the time point of twitch application: a single pulse applied 5 ms before a train pulse consistently evoked higher ITTs than all other stimulation conditions. Furthermore, variability of the ITT decreased as the number of applied twitches increased. The results support the hypothesis that a large part of the variability in the ITT may be caused by the timing of the superimposed twitch relative to the motor unit trains. The variability may be reduced by increasing the number of superimposed twitches.