The paper discusses the possibilities of predicting suicide and of evaluating suicide risk. Attempts at long-term prediction of suicide based on prospective follow-up studies have been unsuccessful in terms of too many false positive predictions. The author presents five areas of statistical risk factors for suicide: biological factors, psychosocial factors, personality traits, family history and psychiatric disorders. Different suicide rating scales are discussed as possible clinical tools which could help the clinician to evaluate risk of attempted suicide. The author presents several biological markers for suicide risk as discussed in the literature, but is sceptical to their clinical applicability. Finally, the clinical evaluation of suicide risk is presented in some detail. The author claims that intervention should still be based on a clinical evaluation.