RI 6406 Using Unequal Sample Interval Lengths and Weilrhted Averages in Estimating Grade of Ore for Bedded Deposits

This report presents the results of an initial investigation into some problems resulting from the use of assays representing unequal sample interval lengths and the use of the weighted average grade of ore in evaluating bedded mineral deposits . A mathematical proof and examples are given to show that the unweighted and weighted assay averages are equivalent when there is no correlation between assay value and sample interval length but that these averages are different when there is correlation between assay value and sample interval length . The amount of this difference between estimates is a function of the degree of correlation . One set of sample and assay data , based on 11 stratigraphic sections of the Phosphoria formation in Idaho , indicates that the relationship between the amount of difference between estimates and the correlation factor is approximately linear . If this correlation relationship can be established , within acceptable limits of precision , as a consistent attribute of deposits , the estimate of weighted average grade of ore could be predicted directly from the unweighted estimate of grade of ore by using the relationship as a prediction factor . Randomness tests on the sample and assay data from the Phosphoria formation , using the mean - square - successive - difference method , indicate that correlation between assay value and sample interval length ( each sample interval length represents an individual bed or stratum of the formation ) , apparently has little effect on the independence between successive assays , between successive sample interval lengths , and between the products of assays and sample interval lengths . Therefore , these data can be analyzed statistically as long as the original sample sections have been selected at random for sampling .