A combined discrete element method (DEM)-Monte Carlo simulation algorithm is used to investigate intra-tablet coating thickness variability for six pharmaceutical tablet shapes. The DEM simulations are used to collect tablet orientations when the tablets enter a specified spray zone. The Monte Carlo simulations are used to "coat" the tablets assuming a uniform spray and orientations chosen randomly from the distribution generated by the DEM simulations. The simulations demonstrate that for the non-spherical tablet shapes investigated here, the coating thickness variability, defined using a coefficient of variation, decreases with the square root of the number of coating trials, then approaches an asymptotic value. The greater the degree of a tablet's preferred orientation, the larger the asymptotic coefficient of variation. In addition, the number of trials required to reach this asymptote decreases for larger asymptotic values. These findings are consistent with theoretical analyses. For the range of parameters investigated, increasing the pan speed decreases the asymptotic coefficient of variation, but increasing the fill level had no consistent effect. The asymptotic coefficients of variation did not correlate with tablet sphericity, aspect ratio, or band-to-tablet area ratio, but the results suggest that a measure of a tablet's symmetry may provide better results. Modifying the angle of the coating spray so that a larger portion of the band is coated during each coating trial results in smaller asymptotic coefficients of variation than when the spray is oriented normal to the tablet bed surface. In addition, the relative order of the tablets with the smallest asymptotic coefficients of variation changes when the spray orientation is changed. (C) 2011 Elsevier Ltd. All rights reserved.