Predictions based on the Gypsy Moth Event Monitor were compared to remeasurement plot data from stands receiving gypsy moth defoliation. These stands were part of a silvicultural treatment study located in northern West Virginia that included a sanitation thinning, a presalvage thinning and paired no-treatment controls. In all cases the event monitor under predicted the initial mortality for all stands in terms of trees per acre. Prediction errors, with regards to trees per acre, were largest in stands receiving the heaviest defoliation. In terms of basal area prediction, those stands receiving the heaviest defoliation had predictions which were too high in the 4- to 5-year period following initial defoliation, after which prediction error diminished. For stands receiving light defoliation, predicted basal area was lower than in observed stands and this error increased with the length of the projection period.