This is a systematic study of the social background of Danish males convicted for the first time of lethal violence, either actual or potential (e.g. unlawful killers, attempted homicides, negligent homicide, grievous bodily harm, n = 125). Using registers, the paper addresses the following question: Do young men, convicted of a lethal violent crime (either actual or potential), have the same kind of risk factors related to social disadvantage as other first‐time convicted violent offenders (n = 1,849) and first‐time attempted suicides or completed suicides (n = 476)? The paper describes three separate analyses of the total 1966 birth cohort followed through a 13‐year period from age 15 to 27 (n = 43,403). In each case the discrete‐time Cox model is used to analyse associations between the relatively rare response events and the relatively rare stress factors. Results suggest that all three groups of subjects have a similar exposure to risk conditions, but also that there are important differences in the predictors for the three groups when the risk factors are analysed one by one. So, for example, the experience of domestic violence during adolescence is a strong predictor of males' later violent behaviour but a less strong predictor of suicidal behaviour. In contrast, being battered and being neglected during childhood more strongly predict later suicidal behaviour than violent behaviour. The implications for prevention are considered.