Suicide rates in most Western countries have not decreased in the last decade, a finding that compares unfavorably with the progress made in other areas, such as breast and skin cancers, human immunodeficiency virus, and automobile accidents,1 for which the rates have decreased by 40% to 80%. Preventing suicide is not easy. The base rate of suicide is low, making it hard to determine which individuals are at risk. Our current approach to the epidemiologic risk factors has failed because prediction studies have no clinical utility—even the highest odds ratio is not informative at the individual level. Decades of research on predicting suicides failed to identify any new predictors, despite the large numbers of studies.2 A previous suicide attempt is our best marker of a future attempt, but 60% of suicides are by persons who had made no previous attempts. Although recent studies in cognitive neuroscience have shed light on the cognitive “lesions” that underlie suicide risk, especially deficits in executive functioning, we have no biological markers of suicide risk, or indeed of any mental illness.