Minsoft is currently doing research into estimation of the the mean grade of an orezone from samples
using a log-probability plot.
The main issues are that the mean of the samples has to be corrected for a number of variances
which exist to give the mean or head grade of the ore zone.
The variances to be corrected for are;
 - distribution variance e.g. if the distribution is lognormal, it must be corrected for by the variance of the
 - volume variance, the samples have a different variance to the blocks which each sample represents.
It is noted that the slope of the log probability plot is related to the variance of the data.
The process of correcting for the variances has been carried out mathematically
which gave a similar result to the graphical method devised.
The mathematical methods used, used a number of assumptions such as the use of formuleas
for lognormal distributions, however it is hoped at a latter date to use a numerical processing
technique which will only use actual data without the assumptions of any distribution types, and
then relate this technique to the graphical technique devised to verify its mathematical basis, and
to develop a technique to estimate means or head grades of ore zones, which is a major problem
faced by the mining industry today, and relates to techniques such as grade cutting which is still
a primitive science. Previous work by Sichel was examined, but is shown to still be inaccurate,
in cases, as a technique for estimating means from sample data.
Sichel's t-estimator can be a usefull technique for estimating the means of data from sample
populations, and hence also for estimating topcut's, however it is hoped to improve on this

This document maintained by Material Copyright © 2013 Minsoft Computing ABN 72 863 916 427