LETTER TO THE EDITOR — Geostatistics blameless in mine failures

Your editorial and article implicating geostatistics in recent mine failures has it all wrong (“Challenging geostatistics” and “Recent History reveals trends in mine failures,” T.N.M., Sept. 28-Oct. 4, 1998).

The basis for over-estimated tonnages and grades has much more to do with unreliable assaying, inadequate sampling density and unconstrained geological interpretations than with the interpolation method. An experienced estimator will always have a backup resource estimate to compare with the computer-generated estimate. These should compare fairly closely.

A geostatistical or other computer-generated estimate will usually have slightly higher tonnage and slightly lower grade when compared with an estimate made by manual methods, such as the sectional method. A geostatistical estimate will compare even more closely with other computer-generated estimates, such as inverse-distance methods. If none of the methods compares in a global sense, something would be wrong with one or more of the methods.

You have really headed in the wrong direction by blaming geostatistics and alluding to “voodoo statistics.” Your efforts would be much better directed to estimator experience and the quality of the data and geological interpretation used.

In the case of Bre-X and other similar fiascos, it would not matter which estimation method was used. The assays used were fraudulent and the metal in the ground was, therefore, non-existent.

I hope I have clarified the issue. It is time to look at other reasons for mine failures and not reopen a debate that blames a popular interpolation method.

Ron Deptuck

Holland Landing, Ont.