The outcome of treatment for amblyopia and the factors that affect this are not well understood. A major reason for this has been the exclusion from previous large studies of a sometimes unknown number of patients because of failure to comply with treatment. This paper analyses the outcome of amblyopia treatment in a retrospective review of the orthoptic records of a cohort of 961 children treated for amblyopia at seven centres who first attended in 1983. The final visual acuity was recorded by Snellen or matching methods in 894 children (93%). Of these, 48% achieved 6/9 or better, 35% less than 6/9 but better than or equal to 6/18, and 17% achieved less than 6/18. The outcome was best for pure anisometropic amblyopia, intermediate for pure strabismic amblyopia and least good for mixed strabismic and anisometropic amblyopia with a final visual acuity of 6/10.2, 6/12.8 and 6/14.8 respectively. While the age at start of treatment did not correlate with final visual acuity both poor initial visual acuity and poor compliance were associated with poor outcome. The main factor affecting the outcome of amblyopia treatment is the initial visual acuity. Comparison with the literature suggests that the results of treatment in this country may be falling far short of what would be possible in ideal circumstances with unlimited resources.