In the confusion that has arisen from the demonstrable inability of a certain type of mathematical model to predict the time course of the SARS-CoV-2 pandemic, many have taken to reciting – with a variable mixture of glee and sympathy – that ‘all models are wrong but some are useful’.