Fifteen years after its resurrection, pallidotomy for Parkinson's disease (PD) and dystonia has once again been supplanted, this time by deep brain stimulation (DBS). Did this occur because pallidotomy was not effective or safe, or because DBS was found to be more effective and safer? This review focuses on the evidence-and its quality-supporting the effectiveness and safety of pallidotomy for PD and dystonia, and the comparative effectiveness and safety of DBS of the subthalamic nucleus (STN) and globus pallidus pars interna (GPi). Discussed first are the determinants of "level 1" recommendations, including the confounding effects on interpretation of randomized clinical trials (RCTs) that fail to control for patient bias (i.e., placebo effects). Although several RCTs have been performed comparing unilateral pallidotomy to medical therapy, GPi DBS, or STN DBS for PD, none controlled for patient bias. Comparison of these trials to estimate the placebo effect, and examination of retrospective case series, suggests that the true effectiveness of unilateral pallidotomy is 20% to 30% reduction of 'off' total motor UPDRS scores, which is similar to the effects of unilateral GPi DBS or STN DBS, but less than bilateral STN DBS. At experienced centers, safety of unilateral pallidotomy appears equivalent to unilateral DBS, but bilateral DBS is likely safer than bilateral pallidotomy. Whereas there have been no RCTs of pallidotomy for dystonia, two double-blind, sham-controlled RCTs of bilateral GPi DBS were performed. Nevertheless, limited uncontrolled series suggest that bilateral pallidotomy is similar to GPi DBS in effectiveness and safety for dystonia. Thus, pallidotomy was not rejected because of lack of effectiveness or safety, and it remains a viable alternative in situations where DBS is not available or not feasible.