This article draws upon recent youth research in some of Britain's poorest neighbourhoods (in Teesside, north-east England). It stresses the importance of a qualitative, biographical and long-term perspective in attempting to understand drug-using and criminal careers (and wider youth transitions) and points to some difficulties in applying — straightforwardly — influential models of risk assessment and prediction to individual biographies. In a context of deep, collective disadvantage, most research participants shared many of the risk factors associated with social exclusion in early adulthood. Yet the majority did not pursue full-blown criminal or drug-using careers and the research struggled to identify background factors that seemed to play a causal role in separating out more 'delinquent' transitions from more 'conventional' ones. Youth biographies were marked by flux; they did not roll on deterministically to foregone conclusions. Unpredictable 'critical moments' turned transitions in unpredictable directions; sometimes towards crime, sometimes away. The article concludes that there is danger in criminal career research — as in studies of youth transition — in prioritising individual level explanations at the expense of an assessment of the 'risks' presented by sociospatial and historical context.