This article investigates the effects of target characteristics on non-residential burglary sighting risks, and evaluates a sighting model framed by criminal event theories. It is based on a random sample of individual UK burglary incidents for which data were collected from officer questionnaires, victim interviews, site surveys and police records. The study measures guardianship ‘actively’ during burglary incidence and extends it beyond occupancy to include neighbours, passers-by and security patrols. The key influence on sighting is guardianship, conditioned by degrees of exposure, partly contingent on means of access to the site and premises. Neighbour guardianship is of over-riding importance in shaping sighting risks, particularly after dark, when guardianship from passers-by may also play a role. Guardianship from the occupancy of premises played no significant part in the sighting of burglars. At least half the explanation of sighting risks appears to lie in differences in target characteristics. Distinctive neighbour guardianship measures are identified.