To determine the optimal ventricular stimulation sequence, an 11-step programmed electrical stimulation (PES) protocol was completed, even if a ventricular arrhythmia (VA) was induced with earlier steps. The protocol consisted of one, two, and three extrastimuli during sinus rhythm (SR), and at two drive pacing rates (VP1 and VP2) plus rapid burst and ramp pacing. By analyzing the 79 completed protocols that induced the clinical arrhythmia, the following were determined: (1) the frequency of induced clinical and nonclinical VA with each stimulation step; (2) the yield ratio (YR) of each step, defined as the probability of inducing clinical versus nonclinical arrhythmia; (3) the cumulative yield of induced clinical and nonclinical arrhythmia with two widely used stimulation sequences, i.e., triple extrastimuli delivered early in the stimulation protocol (MMC sequence) and triple extrastimuli delayed until after double extrastimuli failed to induce the clinical arrhythmia (B sequence); (4) the relative efficiency of these sequences were determined. The percentage of induced clinical and nonclinical arrhythmia with SR + 3 extrastimuli, VP1 + 2 extrastimuli, and VP2 + 2 extrastimuli were (53%, 5%), (36%, 5%), and (41%, 9%), respectively. The cumulative yield of induced clinical VA with the MMC-type sequence reached 55% by the third step of the protocol, whereas 50% was attained only at the eighth step of the B-type sequence. The cumulative percentage of induced nonclinical VA with either sequence was similar during the early steps of the protocol. The MMC sequence was more efficient, requiring overall 36% of potential steps for clinical arrhythmia induction, compared with 48% for the B sequence (P less than 0.001). For questionable arrhythmia states, e.g., syncope of unknown origin and nonsustained VT, a modified sequence is proposed that may further reduce the induction of uninterpretable arrhythmias.