The failure of tocolytics to improve neonatal outcomes in placebo-controlled trials has wrongly been interpreted as evidence that they do not work. While delivery is unequivocally prolonged by 24 hours, 48 hours and 7 days, the time gained was not exploited to optimise neonatal outcome. These trials typically studied women at relatively advanced gestational ages with predictably good outcomes, enrolled them in tertiary centres where they could not benefit from in-utero transfer, and had low levels of corticosteroid administration. No study has been powered to detect clinically meaningful differences that might be expected to accrue from 1-7 days prolongation of gestation. Despite this, Bayesian interpretation suggests that tocolytics do improve neonatal outcome. The largest placebo-controlled study showed clear trends towards better survival in fetuses <28 weeks, lower rates of cerebral palsy and higher Bayley mental scores. Meta-analysis of neonatal morbidity in the beta-agonist trials suggests a near-significant reduction in respiratory distress syndrome (RDS), together with trends towards reduced intraventricular haemorrhage, necrotising enterocolitis, and patent ductus arteriosus. Finally, there is the Orwellian analogy that tocolytics don't work, but some work better than others. Although calcium antagonists have not been evaluated against placebo, meta-analysis of comparative trials with beta-agonists demonstrate a significantly lower incidence of RDS and neonatal jaundice, presumably mediated through the reduced chance of delivery within 48 hours and 7 days. Development of tocolytics that are safe for mother and baby should facilitate adequately-powered placebo-controlled studies, which both focus on women most likely to benefit and capitalise on the 1-7 days gained.