Ofsted has frequently defended the judgements made during inspections by claiming that inspection ratings are reliable, as shown by the results from the collection of studies the inspectorate has conducted. I outline the inspectorate's view of reliability and problematise the studies that it has carried out, noting that these provide insufficient evidence to support Ofsted's claims. I describe a recent study, which utilised an alternative approach to studying the reliability of inspection outcomes, by way of introducing a short discussion of why the inspectorate needs to develop a more productive view of reliability that reflects better the inherent variability of inspectors' decision-making and judgements.