This study provides a review of the literature on sporting production functions with an emphasis on different input-output specifications and alternative estimation procedures. Empirical evidence is reported for English association football on the robustness of estimates of coaching efficiency to changes in estimation methods and the definition of team performance and playing talent inputs. A measure of player quality based on predicted start-of-season transfer values is developed. It is found that the estimation of coaching efficiency is sensitive to the choice of time-invariant efficiency models versus time-varying and inefficiency effects models. It is also found that the results are little affected by different measures of team performance but are highly sensitive to the use of an ex post financial expenditure input measure. Ex ante input measures based on start-of-season player characteristics or predicted transfer values are recommended as more appropriate on both theoretical and empirical grounds.