Objective: A variety of mathematical and trigonometric methods has been described for determining the position of the acetabular cups from conventional radiographs. However, these formulae are subject to unduly large inaccuracies. The aim of the study was to compare the reliability and the accuracy of radiological and CT-based determination of acetabular cup position. Materials and methods: The positions of acetabular cups of 31 patients were calculated in conventional plain radiographs of the hip using the method described by Widmer [Widmer K.H. A simplified method to determine acetabular cup anteversion from plain radiographs. J Arthroplasty 2004;19(3):387-90]. Further, in all patients computed tomograms of the pelvis were performed and the cup position was measured with the aid of a CT-based computer-assisted navigation software. As reference values inclination and anteversion of the cups were calculated in 3D reconstructions of the pelvis with the aid of an image processing software. Results: The radiological measurement as well as the CT-based method showed good intra- and inter-observer reliability and no significant difference in the calculation of the inclination (p = 0.409). However, CT-based determination of anteversion was significantly more exact than radiological measurement (p < 0.001). The calculation of the cup anteversion from the X-rays showed serious deviations from the reference method and a substantial error range (X-ray: mean deviation + 1.74 degrees, range - 16.6 degrees to +29.8 degrees, S.D. +/- 9.32 degrees; CT-based: mean deviation -0.74 degrees, range -6.6 degrees to +5.3 degrees, S.D. +/- 12.87). Conclusion: For any clinical problem or for clinical studies in which acetabular positions of acetabular cups have to be exactly determined, CT-based measuring methods are obviously the method of choice. Evaluations based only on conventional plain X-rays and calculation of the acetabular cup position using the formula described by Widmer must be regarded as unreliable, particularly, because of problems in measuring the anteversion. (C) 2005 Elsevier Ireland Ltd. All rights reserved.