This study examines the role of non-severe depression as a psychological anchor against overoptimism. Using earnings forecasts from Estimize, we find that an increase in the proportion of the U.S. population with depression is associated with improved forecast accuracy among users. This effect is concentrated among forecasts that are optimistic and analysts who take longer time to issue forecasts, highlighting reduced optimism and slow information processing as economic mechanisms that explain our results. We also show that this effect is distinct from the influence of temporary seasonal depression or other sentiment measures on decision-making. Overall, our research establishes a link between depression and crowdsourced financial evaluations