Cut-off grades: statistical estimation and reality

"This paper illustrates the different ways grade-tonnage curves can be generated f or a mineral deposit. It emphasizes systematic discrepancies between reality, estimates and actual production according to different estimation methods and grade control programs. It does not offer a universal solution to the prediction of grade-tonnage curves. Hopefully, it will inspire caution to mineral economists who use estimated grade and tonnage data in the establishment of an optimum cut-off grade for a mining operation.IntroductionThe value of a mineral development venture often depends on a grade-tonnage curve. Statistics can help. The idea is to divide mentally the orebody into small blocks called selective mining units or SMUs. A SMU is an idealized version of the quantity of rock the grade of which will be tested against the cut-off at the time of mining. It could be as small as a truck load or a shovel load in very selective operations like some uranium mines where radiometric probing of each truck or car is performed. More typically, in an open pit operation, it will be the smallest zone flagged as ore from blast hole data, e.g. a 5 m by 5 m by 10m block, if blast holes are on a 5 m grid and bench height is 10m. Having defined the SMUs, the grade-tonnage curve determination amounts to evaluating the grade distribution of all the SMUs that can be mined in the depo sit. If we know this distribution we simply apply the cut-off to it and we get the number of SMUs above the cut-off (i.e. ore tonnage) and their average grade. .Assessment of the grade distribution of SMUs from exploration drill hole data is not an easy exercise and much has been published on the subject (David, 1972; Marechal , 1972; Huijbregts, 1975; Marechal, 1975; Guibal and Touffait, 1982). In this paper, we use a gold mine example to illustrate the influence of SMU size on the results. With the same example we also show that ""common-sense"" methods of grade-tonnage curve determination can sometimes be misleading."