We assessed the reliability of visual estimation of angles on computer images of radiographs, and compared their accuracy with the measurement of angles using computer software for ten distal radius fractures. We asked 73 clinicians to visually estimate the dorsal angulation on ten computerized radiographs of fractures of the distal radius. The reliability of these estimations was calculated. Their accuracy was compared to a 'gold standard' obtained by consensus agreement between three consultants measuring these angles using the software. Inter-observer reliability was calculated as ICC = 0.51 and intra-observer reliability as r = 0.76. The visual estimations were less accurate with a mean percentage error of 31% (range, 7-83%). As angulation increased the estimation accuracy improved. Although reliability and accuracy of such estimation was better for clinicians with greater experience, actual measurement was more reliable and accurate.