The cognitive theories of depression emphasize the role of pessimism about the future in the etiology and maintenance of depression. The present research was designed for two reasons: to provide a clear demonstration that depressed individuals' predictions of the likelihood of future outcomes are more pessimistic than those of nondepressed individuals given identical information with which to make forecasts and identical conditions for forecasting, and to test two additional hypotheses regarding possible mechanisms underlying depressives' relative pessimism in forecasting: a social-comparison and a differential attributional-style hypothesis. We used a modification of the cue-use paradigm developed by Ajzen (1977, Experiment 1) and examined depressed and nondepressed people's predictions of the likelihood of future positive and negative outcomes for themselves and for others. The results provided strong support for pessimism on the part of depressed individuals relative to nondepressed individuals in forecasts for both self and others. In addition, whereas nondepressives exhibited a self-enhancing bias in which they overestimated their probability of success and underestimated their probability of failure relative to that of similar others, depressives did not succumb to either positive or negative social comparison biases in prediction. Finally, in line with the attributional-style hypothesis, depressed-nondepressed differences in subjects' cue-use patterns were obtained, especially in forecasts for self. The findings are discussed with respect to the mechanisms underlying predictive optimism and pessimism and the possible functions and implications of these predictive biases.