•Lightning Detection Network median error ellipses often misinterpreted and used in “binary” manner in forensic investigations.•A methodology to investigate the effectiveness of this practice is explained.•South African Lightning Detection Network data compared with photographed ground-truth lightning events.•The method demonstrates how median error ellipse do not always include the location of attachment. Lightning Detection Networks (LDNs) are sometimes used to investigate the possibility of lightning attachment to a specific geographic location for forensic investigations. When the time of the event to be investigated is unknown, the reported location of detected strokes and the measurement of their accuracy (median error ellipses) become the only LDN parameters available to determine whether lightning attached to a location or not. However, the information provided by the median error ellipses is often misinterpreted and the investigation only involves determining whether the median error ellipse includes the location of interest. In this paper, a reproducible methodology for investigating the effectiveness of using median error ellipses in this manner to determine if lightning attached to a specific geographic location is established. The methodology is applied to a case study – the Brixton tower, South Africa – and it is found that 40% of strokes detected as part of a flash that attached to the Brixton tower have median error ellipses including the location of the tower. Zero strokes that were detected as part of a flash that attached to another location had median error ellipses including the location of the tower. Concerns about using lightning events to tall towers for such an event are discussed.