Wednesday, December 5, 2012

Asking McGill's Brass

Here's an emessage that was transmitted to McGill University on Thursday, August 12th, 2005. Why is it that not a single response has been received? Geostatistical software converted Bre-X's bogus grades and Busang's barren rock into a massive phantom gold resource. McGill Office of Academic Integrity ought to assess whether or not the variance can be stripped off the distance-weighted average AKA kriged estimate. Not a single response has ever been received!


When Sir Ronald A Fisher was knighted in 1953 each weighted average had its own variance. When geostatistics was hailed as a new science in the 1960s the distance-weighted average was reborn as an honorific kriged estimate but lost its variance. Statistically dysfunctional geologists and mining engineers were as ecstatic as nascent geostatisticians simply because any set of measured values, determined in samples selected at positions with different coordinates, defines an infinite set of distance-weighted averages-cum-kriged estimates. It is beyond comprehension that the infinite set of variances was replaced with the kriging variance of a subset of the infinite set of kriged estimates.

Professor Dr Michel David, who taught at McGill University in the 1970s, wrote the first textbook on geostatistics, in which he cautioned, "This is not a book for professional statisticians" and predicted, "... statisticians will find many unqualified statements..." Dr David was a dedicated defender of geostatistical doctrine until he passed away.

Google "bre-x bogus barren" to find out how the junk science of interpolation without justification converted $30,000 worth of placer gold into the world's largest phantom gold resource. Look under "Documentation/Book reviews" and read what I wrote about David's textbook and a few others. Geostatisticians assume spatial dependence, interpolate by kriging, smooth to perfection and rig the rules of mathematical statistics. In contrast, mathematical statisticians at McGill University accept the following inviolable rules:

1 One-to-one correspondence between weighted averages and variances is sine qua non,
2 Each and every distance-weighted average-cum-kriged estimate has its own variance,
3 Kriging variances of subsets of infinite sets of kriged estimates are invalid.
   
Surely, scientific integrity and fair play demand that McGill's students not be taught stochastic modeling with pseudo variances.

Yours truly,
J W Merks

This emessage was sent to the following Officers, Faculty and Staff:

    Mr Richard W Pound, Chancellor
    Dr Heather Monroe-Blum, Principal & Vice-Chancellor
    Professor Dr David Wolfson, Chair, Mathematics and Statistics
    Department Members, Mathematics and Statistics
    Dr Ferri Hassani, Webster Chair Professor and Director of the Mining Program
    Academic Staff, Mining Engineering
    McGill Office of Academic Integrity
    McGill Reporter