Measurements from the Geoscience Laser Altimeter System (GLAS) aboard NASA's ICESat satellite (2001 launch) will be used to estimate the secular change in Antarctic ice mass. We have simulated 5 years of GLAS data to infer the likely accuracy of these GLAS mass balance estimates. We conclude that ICESat will be able to determine the linear rate of change in Antarctic ice mass occurring during those 5 years to an accuracy of ~7 mm/yr equivalent water thickness when averaged over the entire ice sheet. By further including the difference between the typical 5-year trend and the long-term (i.e., century-scale) trend, we estimate that GLAS should be able to provide the long-term trend in mass to an accuracy of about ¿9 mm/yr of equivalent water thickness, corresponding to an accuracy for the Antarctic contribution to the century-scale global sea level rise of about ¿0.3 mm/yr. For both cases the principal error sources are inadequate knowledge of postglacial rebound and of complications caused by interannual and decadal variations in the accumulation rate. We also simulate 5 years of gravity measurements from the NASA and Deutsches Zentrum f¿r Luft-und Raumfahrt (DLR) satellite mission Gravity Recovery and Climate Experiment (GRACE) (2001 launch). We find that by combining GLAS and GRACE measurements, it should be possible to slightly reduce the postglacial rebound error in the GLAS mass balance estimates. The improvement obtained by adding the gravity data would be substantially greater for multiple, successive altimeter and gravity missions.