A study comparing the accuracy of the SRK II, SRK/T, Hoffer Q, Holladay 1, T2 and Super formula in predictive refractive outcomes in children undergoing primary IOL insertion into the capsular bag following cataract surgery. A total of 377 eyes in 377 patients (<13 years of age) were included. Predictive outcomes for the formulae were subcategorised based on age and axial length. Mean predictive error was hyperopic with all formulae. Hoffer Q and Holladay 1 generated the lowest absolute errors followed by the Super formula. SRK II showed the greatest error (P<0.001). Results of particular note are that in the 0-24 month age group all formulae were deemed unsatisfactory. Age at time of surgery was an independent factor significantly contributing to the refractive surprise. Their discussion compares and contrasts numerous other papers in this area. There is no overall consensus; there are non-controllable variables including effective lens position, A-scan induced AL errors, corneal K measurement inaccuracies and capsule changes. The main strengths of this study are the large cohort and selective criteria (only one eye included if bilateral surgery, trauma, corneal scarring, no PFV).