We read with interest the study on intraoperative aberrometry (ORA System, Alcon Laboratories, Inc.) by Cionni et al.1 They calculated the prediction error with the ORA as the difference between the manifest refraction spherical equivalent [MRSE] predicted intraoperatively for the intraocular lens (IOL) implanted and the MRSE postoperatively achieved with the IOL implanted, which is basically the standard definition, differing only in the order.2 But the authors calculated the preoperative formula–estimated prediction error as the difference between the MRSE predicted by the IOL formula used for the preoperatively planned IOL and the MRSE that would have been achieved if the preoperatively planned IOL had been implanted.1 This is not the conventional way to calculate a prediction error. This is simply because there is no direct method to establish with certainty what the refraction of a pseudophakic eye could have been if another IOL had been implanted. Any calculation that is made will assume that a change of 1.0 diopter (D) in the IOL plane will cause around 0.7 D of change in the spectacle plane, but it depends, of course, on the actual pseudophakic anterior chamber depth, which is variable. The differences will not be very significant in most eyes, but it is important that the concept be clear.2 The appropriate way to calculate the prediction error of any biometric method always uses as a reference the actual postoperative refraction of the pseudophakic eye with the lens that was implanted, and compares it with the prediction that the biometric method would do for the lens that is inside the eye.2 That way, no assumptions are required. In addition, when calculating the prediction error in the proper way, it is not really important if the IOL of the power indicated by the ORA System was or was not implanted, or if the one that had been calculated with a biometric formula in the preoperative examination was used, and this much facilitates the understanding of the results by the reader, because the performance of the biometric method is more clearly indicated. With this prediction error, all statistical measures of central tendency, position, and dispersion (including the mean, the median, the percentage of eyes within ± 0.50 D, the maximum error) can be determined using either arithmetic or the absolute value, depending on the particular purpose of a given analysis, without the need to divide the data into subgroups according to whether the IOL agreed or not with what was planned preoperatively. Another great advantage of this standardized approach, based on real measurements, is that any biometric formula can be evaluated, whether or not the surgeon actually used it, because knowing the preoperative biometric data of a given eye (axial length, keratometry, phakic anterior chamber depth, corneal diameter, and lens thickness), knowing the type and power of the implanted IOL, and having the data of the postoperative spherical equivalent, the prediction error can be calculated for any biometric formula without difficulty.