This paper supports the hypothesis that the spatial robustness of trip destination models depends on the extent to which activity sequences and the spatial configuration of activity locations are represented in the model. A destination-choice model is developed on the basis of the concept of detour time to take into account relative locations (i.e., distances between locations that may be combined in the same tour). The new model is incorporated in the activity-scheduling model of Albatross, an activity-based model of travel demand. The results indicate that using detour time instead of home-based travel distance substantially improves the prediction of spatial distributions of activities in a large sample of activity diaries.