Using initialization and validation conditions from hydrographic surveys of the Iceland-Faroe Front in August 1993, shipboard quasi-geostrophic model forecasts executed in real time are evaluated for quantitative skill in terms of anomaly correlation coefficient and rms error. The prototype dynamical forecasts are synoptically initialized from two observed initial states and validated against observations obtained 3--4 days later. The forecasts correlate with the validating observations, yielding anomaly correlation coefficients of 0.75--0.80, which beat persistence-of-day-zero forecasts by 0.07--0.30, depending on the region of interest and the initial state. Thus this quasi-geostrophic model is able to forecast the rapidly evolving currents of this front with quantitative skill. The forecast fields of stream function are also used to diagnose the physical processes of the frontal current variations. Energetic diagnostics clearly reveal that the frontal current evolution is controlled by baroclinic instability processes. Baroclinic wave disturbances at middepth serve to transfer available gravitational energy to kinetic energy, which then is transferred through the water column, resulting in the observed (and modeled) rapid changes in the frontal current. ¿ American Geophysical Union 1995