Reliable discrimination between sinus tachycardia (ST) and pathologic tachycardia has been a major problem for automatic implantable antitachycardia devices. In patients whose sinus response to activity is as rapid or faster than their pathologic tachycardia (rate crossover), these unsophisticated devices deliver the programmed tachycardia response to either the pathologic or sinus tachycardia. Over a one-year period, 50 Intermedics Intertach Model 262-12 antitachycardia pulse generators were implanted to evaluate the specificity of a new group of tachycardia recognition algorithms. Patients were subjected to exercise testing and noninvasive programmed stimulation to demonstrate the efficacy of this new approach. The five recognition algorithms tested were various combinations of the following criteria: high rate (HR), sudden onset (SO), rate stability (RS), and sustained high rate (SHR). False positive rates (tachycardia response inappropriately triggered by ST) were as follows: HR (93%); HR + SO (3%); HR + RS (63%); HR + (RS or SHR) (87%); HR + HS + SO (8%). Pair-wise significance testing between HR only and HR + SO (p less than 0.001), HR + RS (p = 0.01) and HR + SO + RS (p less than 0.001), demonstrated a significant reduction in the rate of false positives through the use of the sudden onset and rate stability criteria in concert with the standard high rate criterion.