Risk assessment can be a valuable aid to decision making but, having become fashionable, is in danger of being considerably oversold. It is applied to very different circumstances that bear only the broadest resemblance to each other. Each is likely to require a different methodology. Thus, to a lawyer, risk management relates largely to the avoidance of legal liability, the quantity surveyor will see it as an aid to controlling a budget, the HSE will relate it to accident avoidance and the structural engineer to maintaining the stability of structures. Such themes are old; only their collective title is new. Nevertheless, codification of some hazards and risks is developing new methodologies which have spawned extensive guides and manuals. All human endeavour includes some risk, the analysis of which may in one case be scientifically and/or statistically based and, in another, calls on judgment, instinct or prejudice. Even assessments scientifically arrived at are rarely exact. Inevitably, therefore, most risk assessments are subjective and need to be recognised as such. Awareness of the benefits and limitations of such assessments has become important to engineers. This paper takes an overview of these: it does not aim to be another manual; its conclusions are strategic and not specific to structural engineering.