Estimating predator abundance can be challenging. Many predators are inherently difficult to detect due to their low population densities, large home ranges and cryptic behaviour. Detection rates derived from camera traps are often used as indices of abundance, however many factors can influence a species’ detection rate and the extent to which it might reflect the species’ actual abundance. I investigated the relationships between detections, abundance and activity of two sympatric predators, the Tasmanian devil (Sarcophilus harrisii) and the feral cat (Felis catus). I used camera traps to detect devils and cats across the island state of Tasmania in southern Australia, where devil populations have progressively and variably declined since 1996 following the spread of the fatal devil facial tumour disease. Devil and cat detections on individual cameras were negatively correlated; however, this was unrelated to abundance. While cats and devils were detected at nearly all of the same sites, cats appeared to avoid devils over short distances, suggesting that negative relationships in detections at the camera scale may reflect fine-scale behavioural avoidance rather than suppression of abundance. These findings highlight the importance of understanding avoidance behaviour when designing camera surveys to detect predators and when using indices to infer interactions or numerical relationships among sympatric predators. These findings also provide a cautionary tale that highlights the need to consider alternative hypotheses to explain observed patterns, as the implications for species conservation and management outcomes could vary dramatically.