Hong Kong has made considerable progress in reducing landslide hazards from man-made slopes. As a consequence, Hong Kong has recently commenced the systematic evaluation of landslides from natural slopes. This paper discusses the assessment approached adopted, limitations with the approach and the problems with landslide hazard assessment in general by means of a case study. The paper concludes that the current approach, whilst suitable for its original purpose, i.e. the rapid evaluation of the likely magnitude of landslide hazards at the review stage of a development, has limitations when used for systematic landslide assessments at existing developments. It is suggested that landslide hazard assessments to quantify “top events”, e.g. loss of life, would be an improved approach which would allow simplified QRA to be undertaken. However, such an approach requires a significant level of engineering geological and engineering geomorphological input. In particular, the careful derivation of appropriate values of landslide magnitude and frequency, based on expert judgment and making use of available data, knowledge and experience.