The Rubicon Mineral Corporation Bungle: The Lessons We Can Draw

| About: Rubicon Financial (RBCF)

Summary

Beware Of Investing In Mining Projects That Are Based On Preliminary Economic Assessments.

Beware of Resource Estimations That Are Number Crunching Exercises Which Are Not Sufficiently Supported By Geological Features.

Be Skeptical During the Commissioning and Production Ramp Up Phase of Mining Projects; Management Tends Not To Be Prompt in Reporting Major Problems Or Set Backs.

When A Large, Knowledgeable Strategic Shareholder Sells Out, You Better Do To.

Executive Summary

On 11 January 2016 Rubicon Minerals Corporation ("Rubicon") (RBY), announced a revised resource statement for its Phoenix Gold Project ("Phoenix"), which indicated a drop of approximately 90% in its contained gold. The revision shocked the market and caused the share price to drop from C$0.14 to C$0.025. However, the price of C$0.14 already reflected a severe down grading of the share price, which was almost ten times higher at approx. C$1.4 in June 2015 when first gold had been poured.

This note reviews the history of the Phoenix Gold Project development, in particular the technical documentation related to the three latest resource estimates, to determine what lessons can be drawn from the history and whether an average investor could have detected earlier on that all was not well.

Rubicon's involvement in Phoenix started in 2002 and discovered in February 2008, after years of various exploration activities, the F2 Gold System, which became the focus of the mine development. Initially there was much excitement about the discovery of the high-grade deposits and the Association for Mineral Exploration, British Columbia (AME-BC), awarded the Phoenix Gold Project exploration team in 2010 the Colin Spence Award. The attractiveness of the project seemed to be endorsed by one of the major gold producers when at the end July 2011 Agnico-Eagle (NYSE:AEM) made a "strategic investment" of C$70 million, earning it a 9.2% shareholding in Rubicon.

However, the first sign less than one year later that all was not well was the announcement in April 2012 that Agnico-Eagle disposed a significant portion of its interest.

Because of the nature of the deposit, steeply dipping narrow high-grade zones following a favorable rock horizon, Rubicon was forced to spend considerable sums on getting underground access to reduce exploration cost. It refurbished existing underground infrastructure and further underground development for creation of exploration drilling platforms. Investment decisions were made on the basis of a number of resource estimations and preliminary economic assessments ("PEAs"). The decision to develop a mine was taken on the basis of a PEA instead of a feasibility study. Within a short time period and based on two different PEAs the approach to mining changes considerably from labor-intensive narrow vein mining to mechanized open stoping, which requires much greater continuity of the deposit(s). The capital spent by the time of first production (however not commercial production) in June 2015 had more than doubled compared to initial budget.

The last three resource estimates, carried out by three different parties, all took continuity of the very narrow, high-grade bodies for granted. Geologically mineralization had been demonstrated to be predominantly hosted by a particular rock type within which very high grade, but narrow drill intercepts were encountered. The resource estimation assumed that these intercepts were linked (first resource estimate) or were linking such intercepts based on geostatistical interpolation (the last two resource estimates).

The conclusions drawn from this review are:

· Beware when investing in mining projects that are based on PEA level studies.

· When an expert strategic shareholder such as Agnico-Eagle sells out, you better follow suit.

· Management is prone to gloss over problems and are not prompt in disclosing about developments that do not go as planned. A healthy dose of skepticism during the mine ramp-up phase is advisable.

· Beware about assuming the same mineralizing features and controls for a particular deposit as for other proven economical deposits in the same geological setting. Such characteristics need to be established from actually observed features. Don't assume.

· Resource estimation should primarily focus on whether the used wireframes are supported by geological information for the continuity and reflect the assumed control of mineralization.

· The lack of continuity between drill results presented in the cross section in the resource estimation reports should have been grounds for caution.

· Defining high-grade deposits simply on the basis of correlating high grades is not appropriate and should only be undertaken after geological observation has proven this to be applicable.

· An average investor would not have been able to pick up any potential problems with these estimation. Even when the reported amount of resources and average grade look high compared to what is presented in the illustrations, the bulk sample results would have provided much comfort that all was well.

· The acceptance of NI43-101 reports by authorities should put more emphasis in critically reviewing what the Competent Person publishes in support of their approach to resource estimation and operational aspects. For example, the plans and sections provided by SRK do not allow for comparison with actual drill results and do not illustrate continuity.

· Stock exchange authorities should not accept PEA reports when these are used to raise funding from public sources for the development of mining projects. It otherwise makes a mockery of the insistence on only using Measured and Indicated resources for inclusion in a mining schedule as companies can simply circumvent these through the use of PEAs.

Introduction

Rubicon announced on 11 January 2016 a revised mineral resource statement for its Phoenix Gold Project that shocked the industry. Compared to the 2013 resource statement it amounted to a decline of approximately 90% in contained ounces. Table 1 gives the comparison.

The resources statement was the basis for a Preliminary Economic Assessment, which included the Inferred Resources for the positive economic valuation on which basis the project was given the go-ahead. Had the Rubicon commissioned a pre-feasibility study, or a full feasibility study, it would have had to comply with the NI 43-101 obligation to only include Measured and Indicated resources in its evaluation. With the deposit at great depth it was probably deemed too expensive to upgrade the resources before being able to embark on such a study.

This report will review the history of the development of the Phoenix Gold Project to determine what lessons can be learned and whether shortcomings in the study could have been identified at the time which could have avoided this misadventure.

Looking back can of course be unfair in that it criticizes with perfect hindsight. This report will attempt to avoid this by concentrating on factual information available at the time when it was produced. This note has also been produced before the report on the revised resource estimate is available to the public, limiting the prospects for hindsight analysis.

Sources of Information

The main sources of information are the following:

· Geoex Limited, 11 April 2011, Mineral Resource and Geological Potential Estimates - F2 Gold System - Phoenix Gold Project NTS 52N/04, Red Lake, Ontario.

· SRK Consulting, 28 February 2014, Preliminary Economic Assessment for the F2 Gold System, Phoenix Gold Project, Red Lake, Ontario.

· Rubicon Minerals Corporation, Management Discussion and Analysis, Annual 2012-1014 and quarterly December 2014 - September 2015.

All illustrations in this note have been extracted from these reports, except for Figure 3.

Background to the Phoenix Project

Location

The Phoenix project (Figure 1) is located in within the Red Lake Mining Division of northwestern Ontario, Canada (Figure 1). The town of Red Lake is approximately 150 kilometers ("km") northwest of Dryden, Ontario and 265 km northeast of Winnipeg, Manitoba.

Figure 1

Location of the Phoenix Property in Ontario, Canada

Click to enlarge

The total area of mineral tenure is 510.4 hectares. The Mining Lease is for a standard fixed term. The current term has been extended to October 31, 2028. Prior to expiry of the extended term, an application must be made under the Ontario Mining Act for the Minister's consent to extend the leasehold for a further fixed term.

Rubicon's 100% interest in the property, comprising 25 Mining Licenses of Occupancy ("MLO") and the one Mining Lease, was acquired in two separate option agreements entered into with Dominion Goldfields Corporation ("DGC") in 2002. For these Rubicon paid C$0.8 million in cash, issued 260,000 shares to DGC and completed US$1.3 million of exploration work by 2004, meeting all the required payments and expenditures. The MLO's and the Mining Lease were subsequently transferred to Rubicon.

Rubicon' % interest in the Phoenix gold project is subject to a 2% net smelter return ("NSR") royalty on the majority of the water portions of the property to Franco-Nevada Corporation. Rubicon has the option to acquire a 0.5% interest in the NSR for US$675,000 at any time, in which case the NSR would be reduced to 1.5%.

History of Rubicon's Involvement

Rubicon has conducted geological mapping, approximately 22,000 m2 of trenching and stripping, 60,000 m of re-logging selected historic drill core, a high resolution airborne magnetic survey, a ground magnetic survey, a seismic lake bottom topographic survey, Titan 24 geophysical survey.

Drilling over a number of years at the Phoenix project area led to the discovery of the F2 Gold System in February 2008. The company cautions in its Management Discussion and Analysis for the year 2008 that "the results to date indicate that sub-parallel, high-grade gold lenses or shoots are developed within a robust gold-bearing structure that also hosts thick, lower grade intervals. These interpretations are preliminary in nature and relationships between the various styles of mineralization are complex."

During 2009 the company dewatered and rehabilitated an existing shaft and underground workings to enable drilling and access to the deposit from underground. The historic shaft was extended to a depth of approximately 335 m with a drift established on the 305m level directed towards the F2 Core Zone.

In November 2010 Rubicon published an initial NI43-101 compliant inferred mineral resource by Geoex Limited ("Geoex"), of which an amended version had to be resubmitted in March 2011 with changes required by the British Colombian Securities Commission. The 2010 resource estimation was carried out on the basis of 166,886 m of drilling completed between February 2008 and 32 July 2010. The declared resources amounted to 3.5 million ounces.

The resource estimation was followed by the publication of a preliminary economic assessment ("PEA") by AMC Mining Consultants (Canada) Ltd ("AMC"), based on a revised resource estimation of almost 2.8 million ounces of which almost 0.5 million in the Indicated category. The estimation was based on results from 259,000 m drill hole core completed up to 28 February 2011.

It seems that on the basis of this PEA the company committed itself to mine development announcing on 19 October 2011 that it was going to spend C$55 million to define sufficient additional indicated mineral resources to support a minimum of five years of potential production at rates envisaged in the PEA of AMC. In order to achieve this, the current shaft was being extended from 338 m to 610 m below surface to allow better access for development. In addition, approx. 600 m of underground development on three levels (244 m, 305 m and 585 m) was planned to support further development and potential future mining and also to provide new drill access for infill and delineation drilling. A target date for "potential production" was set for the fourth quarter of 2013.

Early in 2012 the company raised C$200 million to finance project development and extended the headgear and embarked on mill construction. However, the first sign of trouble was the announcement in April 2012 that Agnico-Eagle has disposed of a significant portion of its shareholding in Rubicon, less than one year after taking a "strategic investment" in the company.

In January 2013 Rubicon announced that it has appointed SRK to complete an update of the mineral resources and to undertake optimization studies. The level of confidence was again set at only PEA level, probably in order to be able to incorporate Inferred resources in its economic evaluation.

The results of the 2013 SRK study were announced in June 2013 and gave favorable economics based on production of 2.2 million ounces. The resource estimation for this report included 3.4 million ounces, of which 1.1 million ounces in the Indicated category, had not expanded much from the AMC estimate, despite much more drill information: 499 surface diamond drill holes for a total of 229,164 m and 573 underground diamond drill holes (428,710 m). Of these 355,611 m were on the F2 system on which the development was focused.

By early 2014 it had become evident that the cost of development would far exceed original estimates. In February a gold streaming agreement of US$75 million was entered into with Royal Gold, and a C$115 million placement in March. On 19 June 2014 Rubicon announced that the project was "on schedule and on budget" showing a table that indicated C$205.2 million had been spent and C$167.5 million, including contingency, remained to completion.

Fortunately, the press release on 23 June 2014 disclosed that results from infill drilling "demonstrate continuity of mineralization in areas of the deposit where we had anticipated continuity."

By 14 April 2015 the project remained on schedule for initial production by mid 2015, but its estimated cost to completion was raised by C$44 million to C$417 million, with as reasons given delays in underground development.

In June the mine poured its first gold and prepared to start stoping operations, but a fall of ground in a cross cut and complications with the haulage system were given as reasons in August that only a trial stope was producing that would yield 20,000 tonnes by September.

More problems were encountered when on 30 September the company was ordered by the Ministry of Environment and Climate Change ("MOECC") to temporarily suspend mill operations at the mine to treat elevated ammonia levels, discharge sufficient water from the tailings management facility, and to upgrade this facility, under specific timelines. Why it took the company until 5 October to release this news is unclear.

Finally, on 3 November 2015 Rubicon announces that it suspended underground activities, because it had discovered during trial stoping that the F2 Gold Deposit was much more geologically complex compared to its understanding from historical drilling. The decisions is somewhat strange in the light of the reconciliation provided in the September 2015 quarter MDA for trial mining results, which are reproduced in Table 2.

The table shows that there were differences compared to model, but none so dramatic to warrant suspension of mining activities.

It is clear that management was not totally open about the problems it encountered.

The lessons that can be learned for investors from this history:

· Beware when investing in mining projects that are based on PEA level studies

· When an expert strategic shareholder such as Agnico-Eagle sells out, you better follow suit.

· Management is prone to gloss over problems and not prompt in disclosing about developments that do not go as planned. A healthy dose of skepticism during the mine ramp-up phase is advisable.

Geology of the F2 deposit

The F2 gold system lies within the Phoenix gold project boundaries and comprises a northeast-trending, west-dipping sequence of ultramafic to mafic volcanics ± intrusives, felsic intrusives and minor sedimentary rock types. According to SRK, based on extensive mapping, trenching, diamond drilling, and geophysical surveys a very consistent geological sequence could be defined that can be correlated along the length of the property for over 4 km. Figure 2 shows the geological map for the project area.

Figure 2

Geological Map of the Phoenix Property

Click to enlarge

The geological map does confirm the consistency of most geological units, but the main host to the gold mineralization, High Titanium ("HiTi") basalt (in dark green color on the map), is only present in narrow bands east and southeast of the Phoenix Shaft and the individual bodies do not extend over more than a few hundred meters.

The property is interpreted to largely represent limb domains parallel to structures of the oldest folding phase (called F1) that trends north-northeast. This strong structural fabric through the area is considered part of the East Bay Deformation Zone ("EBDZ"), which represents a very large structural zone or "break" separating two major geological domains. The Phoenix Gold Project area straddles the EBDZ, which at the project area is to the southeast in sharp structural contact with a later folding phase (called F2) domain, where northwest trending (F2) fold axes are perpendicular to the EBDZ. On the map in Figure 2 the F2 domain area is to the southeast of the Phoenix Shaft.

In the area of the existing mine shaft, the F1 foliation and the geological sequence dip approximately 50 degrees to the northwest whereas towards the southeast, in the area of the F2 gold system which occupies the core of the EBDZ, the foliation dips are sub-vertical to steep northwest. Within the F2 gold system, the F1 stratigraphy is folded by later F2 folding. This F2 deformation is observed as broad, open folding or warping of the F1 stratigraphy (Figure 3).

Figure 3

Isometric View of the F2 Structures Affecting the F1 Structures

Click to enlarge

Gold mineralization in the F2 gold system is characterized by vein and sulphide replacement styles, which are preferentially hosted along the boundaries of two main rock types - the HiTi basalts and felsic intrusive rocks - with additional mineralization associated with crosscutting structures. Gold, however, is distributed through all of the adjacent rock types, with the majority contained within the HiTi basalt.

Two main areas of mineralization were recognized: the Main Domain with gold in the HiTi Basalts and felsic intrusives and the Hanging Wall Domain with gold along the footwall contact of the hanging wall basalt (light green color in Figure 2 close to the Phoenix Shaft), where it is in (structural) contact with variably talc altered ultramafic rock.

The discussion on geology and mineralization correspond closely for the 2011 and 2013 technical reports and both reports point to a strong relationship with other deposits within the Red Lake camp "including host rock types, preferential basaltic host to gold mineralization, and similar structural history."

Resource Estimation by Geoex in 2010-2011

Continuity of Mineralization

The start to the discussion of the methodology used to estimate the resources points out the pitfalls to assessing the gold content of gold bearing structures using diamond drilling. The standard operating procedure for decades by experienced gold developers was "drill for structure" and "drift for grade". Most of these type of deposits in Canada were discovered on the basis of significant surface showings and were initially explored by shallow shafts and drifting on the vein. However, currently exploring for new Archean gold mines is usually in overburden or water-covered areas and must rely completely on drilling to define new zones of sub-cropping gold mineralization. Similarly, in-mine exploration at depth and along strike commonly must rely on drilling of wide-spaced holes to provided indications of lateral or vertical extensions of known ore bodies.

Given this insight one would expect that resource geologists would be aware that the ability to correlate between gold bearing structures is of prime importance before embarking on any number crunching. The 2011 report provides tables including all gold intersections and tables with composites that meet the postulated minimum economic criteria of being in excess of 5 g/t Au and 10 meter x g/t (i.e. a minimum of 5 g/t over at least two meters). An additional criterion was used of a minimum grade per tonne of gold of 5 g/t over an interpreted true width of 1.2m.

The first thing that stands out when reviewing these tables is the generally very narrow width of the interpreted true width of the intersections. Figure 4 shows the frequency distribution generated by me for the 161 intersections that meet the criteria for inclusion in the resource estimation.

The arbitrarily chosen intervals were 0-0.5 m, 0.5 - 1.0 m, 1.0-2.0 m, 2.0 - 5.0 m, 5.0 - 15 m and larger than 15.0 m. For graphing purposes the arithmetic midpoint of the ranges were taken and for the three samples with widths above 15 m (16.1 m, 25.3 m and 44.2 m), the midpoint was arbitrarily taken as twice the bottom value.

Figure 4

Frequency Distribution of Interpreted True Widths of Gold Intersections

Click to enlarge

That only 8% of the intersections had true widths in excess of 5m should have raised the question what type of mineralization could be responsible for such abrupt bulging of width along strike. It would have been reasonable to suspect that the wide intersections were caused by intercepts from drilling along crosscutting structures.

Yet, on the basis of average drill spacing of 55 m down to 500 m below surface, spacing between 60 m and 70 m between 500 m to 1000 m below surface and 100 m spacing below 1200 m below surface, the conclusion was that:

The Author has reviewed all of the technical data relative to the resource estimate and concludes there is sufficient data to reasonably interpret the geology of the vein systems in plan and sectional views and to prepare resource calculations based up industry standard polygonal long section analysis."

The confidence attached to the continuity was based on the close association with lithologies - 90% of the gold is within HiTi basalts and felsic intrusive, which could be traced over hundreds of meters. There is a bit of a logical jump here as continuity of lithology does not immediately imply continuity of individual mineralization. Geoex refers to level plans and cross sections in Figure 20 to 23 in its report to illustrate interpreted continuity. In Figures 5 are partial reproductions these.

Figure 5 shows the southern portion of two levels plans in the 2011 resource estimation report, which are vertically 400 m apart, showing the felsic intrusive in red, the HiTi basalts in green and gold lodes in bright red. The location of the two cross sections in the Figure 6 and 7 of this report is also indicated.

Figure 5

Level Plan at 5075m

Level Plan at 4675m

Click to enlarge

The details are difficult to discern, but upon enlarging the illustrations one can see that the intersections and lodes do not well overlap and that in the northern portion (not shows here) one lode is interpreted to trend very oblique to the normal trend.

Figure 6 shows the interpretation for cross section 50010N.

Figure 6

Gold Intersections in Cross Section 50010N

Click to enlarge

What is immediately apparent is the somewhat contrived correlation of intersections. For example, the deepest two intersections are connected without good geological reason and the second deepest intersection has a hole in close proximity for which no results are presented. Why the second deepest intersection at the footwall contact should correlate with the intersection higher in the middle of the HiTi Basalt (and for which no result is given) is also unclear.

Figure 7 has a cross section 180 m further south from the section in Figure 6.

Figure 7

Gold Intersections Section 49830N

Click to enlarge

On first impression the continuity and grades look very attractive, but upon review of the individual intersections the following is evident:

· Hole 306-03, the deepest hole on the cross section - intersection has a true width of 0.6 m.

· Hole F2-84, second deepest hole - intersection has sub-economic grade and intersected the deposit very obliquely.

· No intersection reported for hole above F2-84.

· F122-10 is the holes with the widest reported intersection in the database. Reported true width is 25.5 m.

· F2-80-W2 is under a very obliquely drilled and has a true width of only 0.8 m (!) within 50 m above F122-10.

· Hole 122-21 has a sub-economic intersection.

On the basis of the presented level plans and the two cross sections it is hard to endorse the conclusion in the 2011 resource estimation report that:

"These level plans and cross sections illustrate that high grade gold intercepts can be correlated over vertical distances of greater than 300 meters and horizontal distances of greater than 100 meters which, in the opinion of the Author, is very positive for lode gold systems at this stage of exploration."

Moreover, it should be noted that, with much drilling from surface, many intersections were under a very oblique angle. Almost half of all intersections were interpreted at an angle of 30 degrees or less.

Table 3 gives average statistics for the intersections from holes form surface drilling and underground drilling.

Upon closer inspection, all six holes that had interpreted true widths of more than 10 m are from underground drilling. When these are excluded the average width drops to 1.74 m, close to that of surface drilling and the weighted average grade increases to 16.43 g/t, also very close to surface drilling.

Choice of Estimation Methodology

Geoex choice of polygonal long section analysis as the most appropriate resource estimation method was based on the fact that, whereas sufficient drilling was carried out to warrant an inferred resource estimate, statistically insufficient data was available for each of the 22 Zones and their subzones) to apply rigorous statistical analysis. In the Geoex's opinion polygonal long section analysis is also most appropriate for "structurally complex gold systems such as the F2 Gold System". Separately an initial block model analysis was completed by Rubicon and was reviewed and accepted by Geoex as a crosscheck of the polygonal model results and assumptions.

Not surprisingly, quantitatively the polygonal and block model methodologies produce similar total grade-tonnage results since they are based on the same spatial database and any variance between the two methods would be related to differences in interpolation and geometric assumptions. With Geoex relying on the outlines provided by Rubicon, the volume within which it estimated the resources are similar for both methods.

Grade Capping

The assumed continuity also influenced Geoex decision to discount grade capping.

Geoex motivated this as follows:

strong continuity of high grade mineralization is demonstrated in areas where sufficient drilling has taken place. Accordingly, treatment of the data as one coherent, log normally distributed data set is not valid and top cutting is thus not warranted at this time.

This is followed by a statement that insufficient data does not allow for statistically motivated grade caps:

there are currently well over 30 sub-zones present each of which requires adequate sample density to develop an appropriate approach to top cutting. Top cutting should be reconsidered once more data, including bulk samples, are collected allowing for enhanced interpretation and proper zone allocations."

In other words, grade capping is advisable, but can only be carried out once sufficient data is available to set a cap. One would think that ANY cap would be then preferable than no cap at all.

When it came to reporting the results Geoex did verify what effect the 10-5-2 capping strategy, which was used for many years by Goldcorp (NYSE:GG) and Placer Dome at the Red Lake Mine, would have on grade. It is an empirical capping strategy that caps gold values greater than 10 oz/tonne to 10 oz/tonne, those between 5 and 10 oz/tonne to 5 oz/tonne, those between 2 and 5 oz/tonne to 2 oz/tonne. Values less than 2 oz/tonne remain uncapped.

The estimation results of Geoex based on longitudinal polygonal estimation and on block modeling are shown in Table 4, and which assigns only an Inferred category to the numbers.

It is surprising that all the average grades in the table, both uncapped and capped, exceed the weighted average grade of 15.2 g/t in Table 2 using uncapped grades for a true width of at least 1.2 m, which has supposedly also been applied by Geoex.

The conclusion is that both estimates give similar results, particularly when using capped values. It demonstrates that the various estimation techniques are different methods of averaging grades and that the amount of resource is determined by the volume within which the estimating is carried out.

The lesson that can be learned from this report is:

· Beware about assuming the same mineralizing features and controls for a particular deposit as for other proven economical deposits in the same geological setting. Such characteristics need to be established from actually observed features. Don't assume.

· Resource estimation should primarily focus on whether the used wireframes are supported by geological information for the continuity and reflect the assumed control of mineralization.

Resource Estimation by AMC in June 2011

A total of 511 boreholes, totaling 239,340 m, were used in the modeling. Again the independent competent person, AMC, accepted Rubicon's interpretations of lithologies, mineralization controls and geology domains. The description in the AMC report of the mineralization closely follows previous descriptions, including statements on the continuity of the mineralization.

The report includes the following cross-section, which is "through the well drilled section of what has been termed Domain 1, showing the geological interpretation by Rubicon of the F2 Core Zone mineralization."

Figure 8

Cross Section 49991N Through the F2 Core

Click to enlarge

The cross section shows a definite concentration of gold intersections in the HiTi basalts, but how the individual high gold intercepts could have been interpreted as delineating discrete high-grade bodies that can be traced over tens of meters, let alone hundreds of meters, is unclear.

The modeling used wireframes for the HiTi basalts and Felsic Intrusive provided by Rubicon, whereas the definition of the mineralized domains was undertaken interactively between AMC and Rubicon. This generated twelve mineralized domains based on a low gold threshold of 0.1 g/t gold. The isometric view shown in the AMC report of these domains indicate that these were wide and probably reflect the favorable lithologies. It is the next step, the outlining of the high grade bodies within these favorable lithologies, that is not clearly discussed and motivated. The discussion simply proceeds to discussing the number crunching, which appears to include all data within these low grade domains. In other words, the statistics used for variance, grade capping, search parameters were all based on data in the low grade wireframe and the determined trends assumed to define the high grade structures. Figure 9 shows the block model projected on plan and Figure 10 gives the vertical projection.

Figure 9

Block Model Projected on Plan

Click to enlarge

Figure 10

Block Model Projected on Vertical Projection

Click to enlarge

Immediately apparent from the illustrations is that the individual high grade bodies in orange, red and purple appear wide and with dimensions of +100 m both along strike and vertically. This could the effect of projection, but this nowhere pointed out.

The limited extent of the high-grade bodies is only clear from illustrations comparing the drillhole grade intercepts with block model grades. Figure 11 shows a plan view with the grid coordinates defining 10sm intervals.

Figure 11

Plan View Comparing Drill Data and Model

Click to enlarge

Also apparent from the illustration above is that the drill data do not very evidently support the dimensions of the bodies and their block model grades (e.g. refer to the high grade blocks in the body on the left and colors of the borehole intercepts). A similar picture can be shown for a vertical comparison.

It is therefore curious that a comparison between two bulk samples collected by Rubicon on the 305 m level and the weighted average from drilling for broad zones from within which the bulk samples were extracted gave favorable confirmation of the estimated grade (refer to Table 5).

There is no detail on how the bulk sample was exactly collected.

The resource statement of AMC estimated as a cut-off grade of 5 g/t Au, dated June 2011, is reproduced in Table 6.

The lesson that can be learned from reviewing this report is that:

· The lack of continuity between drill results presented in the cross section of Figure should have been grounds for caution.

· Defining high-grade deposits simply on the basis of correlating high grades is not appropriate and should only be undertaken after geological observation has proven this to be applicable.

· An average investor would not have been able to pick up any potential problems with this estimation. Even when the reported amount of resources and average grade look high compared to what is presented in the illustrations, the bulk sample results would have provided much comfort that all is well.

Resource Estimation by SRK in 2013

Introduction

A total of eight geologists and resource geologists were involved in the 2013 resource estimation: four for three dimensional ("3D") modeling, one for domain model review, two for the geostatistical analysis and to carry out the resource estimation, one to review the geostatistical analysis and one for overall review of the estimation.

With hindsight, this shows that there is no comfort in numbers. Apparently nobody has raised concern about the validity of continuity of mineralization and the approach to estimation.

Resource Estimation Database

SRK had the benefit of a substantially expanded database, now comprising 820 holes, of which 618 holes from underground. Underground fans had been drilled approximately 25-35 m apart horizontally, with boreholes collared in 10o vertical increments on section and crossing the F2 core mineralization. In addition underground workings could be visited to inspect exposures of the mineralization.

Continuity of Mineralization

As for AMC, SRK prepared a three-dimensional lithological model based on sectional interpretations of geology provided by Rubicon and used wireframes provided by Rubicon, but this time representing a higher (0.5 g/t) gold threshold to define domains hosting gold mineralization.

Within the gold mineralization domains, narrower, high-grade domains characterized by continuous grades above 3.0 g/t gold were also created by SRK. According to SRK, the high-grade domains essentially represent high-TiO2 basalt units, whereas the lower grade domains generally occur in more felsic rock, surrounding the high-TiO2 basalt units.

SRK defined 56 gold mineralization domains (31 high-grade and 25 lower grade domains) that were used to constrain mineral resource modeling. According to SRK, these domains represent a refinement of the gold mineralized geometries and an overall improvement of continuity of the modeled gold mineralization, which constrains the gold interpolation in the mineral resource model. Also according to SRK, these refined wireframe geometries closely reflected the interpretation at the time of the underlying geology framework.

SRK remarked in its 2013 report that

the geological information is sufficiently dense to demonstrate the continuity of the stockwork gold mineralization associated with a northeast-trending sequence of interbedded ultramafic, high-TiO2 basaltic rocks and felsic sills and dikes. The confidence in the geological model is good."

From the discussion on geology, mineralization and resource estimation the reader does not really get a good understanding of the dimensions and geometry of the bodies that will have to be mined. However, under the Planned Mining Methods section of the SRK report is mentioned that, since AMC completed its report in which it envisaged labor intensive narrow vein mining,

subsequent structural geology work has led to a revised interpretation of the structural controls on mineralization that has resulted in a different interpretation of the mining widths that are expected. The current mining plan envisions a complex deposit that is relatively discontinuous, somewhat disseminated in nature, has weak visual indicators, and a strong nugget factor."

Under the characteristics that needed to be considered for the mining method are listed among others:

· The mineralized zone is 150 to 200 m wide in section, up to 1,000 m along strike and discontinuous to the extent that much of this area is below the design cut-off grade.

· mineralized zones above the conceptual mine design cut-off grade vary in true width from less than 1 m to greater than 30 m.

· mineralized zones above the conceptual mine design cut-off grade can pinch and swell rapidly along strike and along dip.

· Grade continuity in the mineralized zones above design cut-off grade is generally variable, which is indicative of a strong nugget effect.

Obviously, the SRK mining engineer did see problems with continuity.

Resource Estimation

These 56 domains were subsequently combined into three groups based on their spatial orientation: Main, Main 45, and Hanging Wall ("HW"). The Main 45 domain is a domain proximal to the Main domain and is characterized by a unique structural orientation.

In the end the HW domain added virtually nothing to the resources and the Main 45 domain very little.

This time the variogram analysis was based on modeling the grades within the high-grade domains, with the results applied also to the low-grade domains. The block modeling approach was similar to that used by AMC and the resource statement in Table 7 gives total resources which are only 20% larger than of AMC, but with a much higher proportion (almost 140% higher) in the Indicated category.

It should be noted that the plans and sections SRK provided in its Appendix D for comparison of original borehole data with mineral resource block data on plans, sections, and down the plunge of the mineralization, do not provide borehole intercepts and are basically meaningless to outside parties.

SRK refers to the results of the same two bulk samples as AMC when comparing their grade model with actual material mined. Curiously SRK chose to quote the assay results of 10 tonne subsamples rather than the grades of the full 1,000 tonnes (net of the 10 tonne sub-sample) processed in a pilot plant test, which was used by AMC.

Table 8 gives the comparison. The SRK report do not use the same references as AMC of WLB2 and F2-Core for the bulk sample sources, but use "A" and "B" with A predominantly drawn from the Main 45 domain and B from the Main domain. However, given the same quoted grades provided by Rubicon, the AMC and SRK tables refer to the same material.

SRK drew comfort from the results that its mineral resource model grade compared low to the bulk sample grades from various sources. The discrepancy for the B-sample warranted further investigation and verification on whether the material came from the blocks modeled by SRK.

The description of the sampling protocol focuses on handling after collection and the only reference to the sample collection itself was: "extraction from underground of 1,000-tonne minus 230, material from a specific zone". No detail on the exact zone, the width of the zone samples, whether continuous, or from several places, etc.

The lesson that can be learned from reviewing this report is that:

· The absence of plans and sections showing how interpretations by the Competent Person compares to drill results makes it impossible for the reader to make up its own mind.

· Despite certain specialists within SRK recognizing the highly variable nature of the high-grade deposits, the resource geologists still took continuity for granted in their resource estimation.

· Again, an average investor would not have been able to pick up any potential problems with this resource estimate, relying on the expressed confidence of the Competent Person about continuity and attaching an Indicated category to a large portion of the resources, in addition to drawing comfort from the reported bulk sample results.

Disclosure: I/we have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours.

I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.

Editor's Note: This article covers one or more stocks trading at less than $1 per share and/or with less than a $100 million market cap. Please be aware of the risks associated with these stocks.