Corrected sampling variogram
for bait borehole

Bre-X's "quality control program" was based on cyanide leaching duplicate 750 g test portions taken from 10% of the core samples. Borehole BSEE198 provided 18 duplicate test results for gold. Selecting and cyanide leaching test portions of test samples is the analytical stage in the measurement chain. In ISO vernacular, the variance associated with this stage is the analytical variance. In geostat speak, however, the analytical variance metamorphosed into the nugget effect in gold ores and massive sulfides alike. The fact that the nugget effect and the sill value have the dimension of a variance may surprise those who are unaccustomed to the idiosyncrasies and intricacies of geostatistics.

This set of duplicate test results for gold gives an analytical variance of var(a)=50.92 (g/t)2 and is awarded 18 degrees of freedom. The first variance of the ordered set is var1(x)=49.29 (g/t)2. Since all values tabulated in F-distributions are greater than unity, the ratio between the highest variance and the lowest variance must be compared with tabulated F-values. In this case, the calculated value of F=var(a)/var1(x)=50.92/49.29=1.05 does not exceed the tabulated value of F0.05;18;400=1.63 at 5% probability. Hence, the analytical variance and the first variance term of the ordered set of test results are statistically identical. By implication, the difference between these variances is itself not a valid variance estimate. In other words, the intrinsic variance of gold in the largest phantom resource the world has never seen is statistically identical to zero.

Why did the ordered set of test results for gold in Bre-X's glory borehole display a highly significant degree of spatial dependence? The reason is that core samples from this borehole were sloppily salted with the same blend of pulverized ore and placer gold whereas those from others were salted with different blends of pulverized ore and placer gold. Some measure of order remained simply because 201 core samples, unlike a deck of 52 cards, cannot be shuffled!

Five boreholes with Bre-X's standard gold grades of 2-5 g/t and fifty duplicate test results would have been enough to prove that a salting scam was in progress at Busang. I submitted to the 2000 Mining Millennium Conference an abstract for a paper on this subject. I received no response.