Cutting or Capping of High Assay
Values
William E. Roscoe, Ph.D., P.Eng.
Roscoe Postle Associates Inc.
Presentation to TGDG Seminar
February 7, 2012
Toronto Denver London
Vancouver Quebec City
experienced | collaborative | knowledgeable | responsive www.rpacan.com
Overview
Why cut or cap high assay values?
Emphasis on gold
Cutting curve for producing mines
Methods for non-producing projects
What not to do
Conclusions
experienced | collaborative
passionate | ambitious | knowledgeable
| collaborative | responsive
| knowledgeable www.rpacan.com
www.scottwilson.com
Grade Cutting
Why cut or cap high assay values?
Assay grade distributions tend to be strongly skewed
Dominated by low values with lesser high values
Range can be several orders of magnitude
For gold, distribution can have an erratic tail
For coarse gold, erratic high values can be caused by
presence of a gold particle
High assay values can have a disproportionately large
influence on the average grade
High values must be treated to reduce their influence
Most common treatment is cutting or capping
experienced | collaborative
passionate | ambitious | knowledgeable
| collaborative | responsive
| knowledgeable www.rpacan.com
www.scottwilson.com
Grade Cutting
Producing Mines
Cutting curve a graphical approach
Plot average cut grade vs cutting level
Use samples that represent material mined and
processed from a specific volume or volumes
Use calculated head grade for the same volume or
volumes of material mined and processed
Determine cutting level from head grade plotted on
cutting curve
Use a minimum of 500 or 1,000 samples
experienced | collaborative
passionate | ambitious | knowledgeable
| collaborative | responsive
| knowledgeable www.rpacan.com
www.scottwilson.com
Grade Cutting Cutting Curve
12
Uncut
10
Average Cut Grade g/t Au
8
Head Grade
0
0 10 20 30 40 50 60 70
Cutting Level g/t Au
experienced | collaborative
passionate | ambitious | knowledgeable
| collaborative | responsive
| knowledgeable www.rpacan.com
www.scottwilson.com
Grade Cutting
Non-Producing Projects
Zoom in to high end of simple histogram to identify erratic
tail
Cumulative probability plot to identify erratic tail
Decile plot to assess effect of capping on highest
percentile of data set
Cutting curve to assess effect of capping on average
grade
experienced | collaborative
passionate | ambitious | knowledgeable
| collaborative | responsive
| knowledgeable www.rpacan.com
www.scottwilson.com
Grade Cutting
Univariate Statistics of Gold Assay Data Set
All sample lengths 1 m
N = 404
Mean 1.57 g/t
Median 0.30 g/t
Maximum 99 g/t
Variance 44.74
Standard Deviation 6.69
Coefficient of Variation 4.26
experienced | collaborative
passionate | ambitious | knowledgeable
| collaborative | responsive
| knowledgeable www.rpacan.com
www.scottwilson.com
Grade Cutting
Gold Assay Data Set Highest 15 Values in g/t
99.00
65.29
40.73
26.92
23.10
19.34
15.50
12.90
12.00
10.30
9.79
9.79
9.13
8.13
7.80
experienced | collaborative
passionate | ambitious | knowledgeable
| collaborative | responsive
| knowledgeable www.rpacan.com
www.scottwilson.com
Grade Cutting
Au Histogram
450
400
350
300
Frequency
250
200
150
100
50
0
5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100
Upper Limit of Bin in g/t Au
experienced | collaborative
passionate | ambitious | knowledgeable
| collaborative | responsive
| knowledgeable www.rpacan.com
www.scottwilson.com
Grade Cutting
Au Log Histogram
140
120
100
Frequency
80
60
40
20
0
0.01 0.03 0.1 0.3 1 3 10 30 100
Upper Limit of Bin in g/t Au
experienced | collaborative
passionate | ambitious | knowledgeable
| collaborative | responsive
| knowledgeable www.rpacan.com
www.scottwilson.com
Grade Cutting Zoom on Histogram Tail
Au Histogram
30
25
20
Frequency
15
10
Cut?
0
5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100
Upper Limit of Bin in g/t Au
experienced | collaborative
passionate | ambitious | knowledgeable
| collaborative | responsive
| knowledgeable www.rpacan.com
www.scottwilson.com
Grade Cutting Log Probability Plot
experienced | collaborative
passionate | ambitious | knowledgeable
| collaborative | responsive
| knowledgeable www.rpacan.com
www.scottwilson.com
Grade Cutting Decile Analysis
Au Percentile Analysis
45%
Cut to 10
Cut to 20
40%
Cut to 30
Cut to 40
35% Cut to 50
Uncut
30%
Percent of Metal Contained
25%
20%
15%
10%
5%
0%
2.42 2.82 3.22 3.79 4.48 5.29 6.68 9.04 11.90 22.99
Top 10 Percentile Au (g/t))
experienced | collaborative
passionate | ambitious | knowledgeable
| collaborative | responsive
| knowledgeable www.rpacan.com
www.scottwilson.com
Grade Cutting
Total Contained Metal Removed at Various Cutting Levels
36%
34%
32%
30%
28%
26%
24%
22%
20%
Gold Metal Removed
18%
16%
14%
12%
10%
8%
6%
4%
2%
0%
Cut to 10 Cut to 20 Cut to 30 Cut to 40 Cut to 50 Uncut
Cutting Level
experienced | collaborative
passionate | ambitious | knowledgeable
| collaborative | responsive
| knowledgeable www.rpacan.com
www.scottwilson.com
Grade Cutting Cutting Curve
1.80
Au Mean Cut Grade at Various Cutting Levels
1.60
1.40
1.20 Mean Cut
Grade (g/t)
Mean Cut Grade (g/t)
1.00
0.80
0.60
0.40
0.20
0.00
Cut to 10 Cut to 20 Cut to 30 Cut to 40 Cut to 50 Uncut
Cutting Level
experienced | collaborative
passionate | ambitious | knowledgeable
| collaborative | responsive
| knowledgeable www.rpacan.com
www.scottwilson.com
Grade Cutting
Histogram suggests 25 g/t cutting level
Log probability plot suggests 10 g/t Au cutting level
Decile analysis suggests 15 g/t cutting level
Recommend 20 g/t cutting level
experienced | collaborative
passionate | ambitious | knowledgeable
| collaborative | responsive
| knowledgeable www.rpacan.com
www.scottwilson.com
Grade Cutting
DO NOT:
Use a certain percentile of the data set such as 95%
Use the log histogram no detail on erratic tail
Use too many bins in the histogram too much detail
Cut or cap composites instead of assays
High assay values can get smeared out in the composites
experienced | collaborative
passionate | ambitious | knowledgeable
| collaborative | responsive
| knowledgeable www.rpacan.com
www.scottwilson.com
Grade Cutting
Conclusions
Choose a cutting or capping level with care
Use more than one method if possible
For producing mines, make sure the sample data
represent the mined volume
Review the effect of cutting on the average grade and
contained metal
Consider alternate methods of reducing the influence of
erratic high assays on the mean resource grade
Be aware that these may not have the same effect as capping
experienced | collaborative
passionate | ambitious | knowledgeable
| collaborative | responsive
| knowledgeable www.rpacan.com
www.scottwilson.com
Roscoe Postle Associates Inc.
William E. Roscoe, Ph.D., P.Eng.
Tel: (416) 947-0907
Email:
[email protected] Toronto Denver London
Vancouver Quebec City
experienced | collaborative
passionate | ambitious | knowledgeable
| collaborative | responsive
| knowledgeable www.rpacan.com
www.scottwilson.com
