To the Editor:We appreciate the letter from Dr Philipp 1 in response to our article "Placement of Stereotactic Electroencephalography Depth Electrodes Using the Stealth Autoguide Robotic System: Technical Methods and Initial Results." 2 The authors are critical of our results, indicating that they do not support our conclusions. Rather, they suggest that the Autoguide system is both unsafe and inaccurate. We respectfully disagree.Our article was written as a technical report, describing a new method, and the early modifications that we adopted. It represents our initial learning curve. This is by no means an authoritative clinical series. Despite this, Philipp et al 3 compared our findings to several large series. These are simply not valid comparisons. They compare our results with those of Cardinale which included 500 patients. They also reference a report by Miller et al 4 of 152 patients using a standardized method and a meta-analysis with 2624 patients. 5 Comparison of these large series with a technical report of 9 patients during an initial period of learning is simply not justifiable.Furthermore, their entire safety concern is based on a single hemorrhage in the first patient implanted, a case in which we acknowledge the drilling method recommended by the manufacturer was problematic. Had we chosen to exclude the first patient and only report on the subsequent 8, and we would have missed the opportunity to point out this important technical issue to other users. Rather, we would have reported a 0% complication rate, seemingly indicating that the system is the safest one every reported in the literature! This would be equally ludicrous. This is the danger of statistical inference in low sample sizes. Based on their previous publications, 6 Philipp et al should be well aware of this danger. The goal of this article was to provide technical guidance for other users to increase their accuracy and reduce rates of complications. Any other reading of this article is incorrect.Furthermore, it is likely that true stereotactic electroencephalograhy hemorrhage rates are grossly underreported. By example, the Cleveland Clinic group (included as part of the analysis reported here as proof of our low safety 5 ) has reported a per the electrode complication rate of 0.18%, and only 3 patients having any sort of hemorrhage at 2.4%. 7 However, a blinded review of their own hemorrhagic complications in 549 consecutive patients noted a true hemorrhage rate of 19.1%, of which 2 (2%) were symptomatic and a death rate of 0.2%. 8 In other words, the exact same group, looking at their own data in 2 different ways, reported a hemorrhage rate that changed from 2.4% to 19%. This does not imply that the authors are dishonest.