In Europe, the Light index is used for estimating the size of pneumothorax. It is based on the proportional relationship between the collapsed lung and the hemithorax. Other methods rely on absolute measurements on radiographs. The aim of this study was to compare the estimated size of pneumothoraces using Light index with that estimated using the CT volumetrics-derived Collins method. This was a secondary analysis of radiographs from a database of pneumothoraces treated conservatively. Radiographs were measured independently by two researchers according to the methods described by Light and Collins. Estimates of size derived by the two methods were compared using bias plot analysis techniques. The primary outcome of interest was agreement between pneumothorax size (as per cent) between the two methods. Sixty-one radiographs (27 patients) were analysed. The median age was 21 years and 76% were men. There was poor agreement between the methods. The average difference was 7.3% with the Light index underestimating the size compared with the Collins method. The 95% limits of agreement were wide (24% underestimation to 17% overestimation of size). The Light index as described does not accurately estimate the size of pneumothorax. New methods if accurate quantification of pneumothoraces is required clinically.