The posterior condylar offset (PCO) has been proposed as a determinant of a postoperative range of motion after total knee arthroplasty, although there is no consensus. This study aimed to demonstrate the error introduced by forcing the femoral rotation to overlap both condyles for the “true” lateral X-ray projection for the PCO measurement. We hypothesize that the angular discrepancy between the posterior femoral cortical reference plane and the posterior condylar axis plane due to rotation invalidates the acquisition of reliable measurements on X-rays. We have measured the PCO in 50 “true” lateral X-rays and compared it with the medial and lateral condyles PCO's assessed on a computed tomography-scan-based three-dimensional (3D) model of each knee. PCO based on the 3D imaging differed significantly between the medial (25.8 ± 3.67 mm) and lateral (16.59 ± 2.92 mm) condyle. Three-dimensional PCO values differ significantly from those determined in the radiographic studies. Also, the mean values of the medial and lateral condyle PCO measurements differed significantly (p < 0.001) with all PCO measurements on radiographs. We have identified a difference between the posterior cortical plane and the posterior condylar axis projections, both on the axial plane with a mean value of 11.23° ± 3.64°. Our data show an interplane discrepancy angle between the posterior femoral diaphyseal cortical and the posterior condylar axis plane (due to the femur's necessary rotation to overlap both condyles) may invalidate the 2D X-ray PCO assessment as a reliable measurement.