Low completion rates for apprentices and trainees have received considerable attention recently and it has been argued that NCVER seriously understates completion rates. In this paper Tom Karmel uses NCVER data on recommencements to estimate individual-based completion rates. It is estimated that around one-quarter of trade apprentices swap employers during their apprenticeship. Taking this into account, Karmel estimates completion rates for individuals, ranging from 39.2% for the food trades to 64.2% for electrotechnology and telecommunications trades workers. He notes that employer churn is an issue, with the worst occupations being hairdressing and the food trades. (Contains 2 tables and 1 figure.)