A simulation algorithm is proposed that predicts the lighting energy performance of manually and automatically controlled electric lighting and blind systems in private and two-person offices. Algorithm inputs are annual profiles of user occupancy and work plane illuminances. These two inputs are combined with probabilistic switching patterns, which have been derived from field data, in order to predict the status of the electric lighting and blinds throughout the year. The model features four different user types to mimic variation in control behavior between different occupants. An example application in a private office with a southern facade yields that-depending on the user type-the electric lighting energy demand for a manually controlled electric lighting and blind system ranges from 10 to 39 kWh/ m(2) yr. The predicted mean energy savings of a switch-off occupancy sensor in the example office are 20%. Depending on how reliably occupants switch off a dimmed lighting system, mean electric lighting energy savings due to a daylight-linked photocell control range from 60% to zero. Crown Copyright (C) 2004 Published by Elsevier Ltd. All rights reserved.