Intchains Group Ltd V.ICG

From Integra tech report, High Grade Capping !

[Marine2]

14.2.4. High Grade Capping

Many low grade mineralizations, typically precious metals like gold exhibit sample grade distributions with outliers on the high side. It is common in the industry to remove some of the highest (aberrant) values from the assay distribution prior to compositing samples. The main objective of this process is to make sure that the erroneous value could not affect the local grade estimation.

However, in most cases, those values are real and correspond to a small quantity of very rich material likely to be found in some mining blocks. Hence the mean grade of the entire deposit should take those values into account since, with hundreds or thousands of samples, it is most likely that the proportion of this material is not over-represented in the distribution of samples. The real problem occurs with local estimates of mining

2015 NI 43-101 Technical Report and Mineral Fall 2015 Resource Estimate Update on the Lamaque project December 15, 2015 Integra Gold Corp. 199

www.innovexplo.com

blocks around the outliers: now one outlier with a very high grade in a group of 20 samples certainly do not indicate a small proportion of very rich material in the block. On the other hand, it is not because the author has not found any rich sample around a block that they not exist. Hence, whenever there is some evidence of the presence in very small quantities of some real very rich zone, block estimates close to outliers should be reduced and block estimates with no outliers around might be increased a little. With the development of statistical methods for estimating grade, this process became less important. In fact Indicator Kriging, and simulation techniques involving normal transformation of score are very efficient to deal with distribution containing outliers. On the other hand, the presence of some very high values may make the variogram very difficult to establish.

If there is no sampling problem (i.e. very high grade duplicate well), there are no theoretical or scientific ground to cap high sample data. As explained above, outliers are a problem in local (block) interpolation with traditional interpolation method (including kriging) in the sense that blocks close to outliers are likely over-estimated but, if there are thousands of even hundreds of samples, in most cases, they do not bias the global estimates (i.e. the average grade of the deposit).

More than often, the author consider that high grade capping is more psychological than anything else: if you haven’t cap your sample data in a gold deposit, people will not trust you.