It is often desirable to predict with confidence the performance of a full-scale mill Using good quality laboratory data. In the present paper, the sensitivity of Austin's scale-up procedure to batch grinding data was investigated. To this end, the effects of breakage function and selection function parameters on the product of a full-scale mill in open circuit were simulated. An array of numbers was randomly generated around the average value of the batch milling parameter considered and within a predefined range. Then, the mill products corresponding to the array of values were calculated. After analysis, Austin's scale-up model was found to be more sensitive to the selection function than the breakage function. Selection function parameters should therefore be measured with great care when modelling open mill circuits. (C) 2017 Elsevier B.V. All rights reserved.