Tuesday, January 30, 2018

Pretium - an admission of failure?

As Otto mentioned (link) the grades 'that should not be mentioned' mined during Q4 at the Valley of the Kings Wangs* was a lot lower than planned - 8.4 g/t vs 15.4 g/t, a 45% difference.

*I'm pandering to my Chinese overlords

The market didn't go yippee!

down 35%

Fortunately, the Brucejack web-page (link) contains all the answers to those irritating questions about grade, production and so on.....

I was interested that they included a set of guidance figures for H1, 2018.

Steady state - like Brownian motion and the infinite probability drive. The last sentence is very intriguing...

A 50K or 33% (or 25%  if you are using the upper end of the prediction) swing is massive, that equates to a variance of up to 8,333 ounces per month.

However, this isn't too surprising if you are mining a highly variable deposit with many erratic, inconsistent ultra high-grade veins in a low grade host rock. So, why don't we have some fun, and interrogate the information that we've been given.

I want to calculate the planned head-grades for H1, 2018 to see how they compare to the 2016 reserves (link).

  • Recovery =  95% - slightly below average recovery for 2017
  • Mill Throughput = 2850 tonnes per day - slightly less than average throughput in 2017
  • Mill utilization = 100% or 181 days
  • Production - Pretium PR
    • Top end = 200,000 ounces
    • Bottom end = 150,000 ounces
Step 1 - Recovery
I'm going to apply the recovery, 95%, to the production. This will tell us how much gold they'll mine before the 5% is lost through the milling process.
  • Bottom end = 150,000 / 0.95 =  157,894.74 ounces mined
  • Top end = 200,000 / 0.95 = 210,526.32 ounces mined

Step 2 - calculating the tonnage
  • 2850 x 181 =  515,850 tonnes milled
Step 3 - Back calculating the head grade
  • Bottom end
    • 157,874.74 / 515,850 = 0.306 oz/t
    • 0.306 oz/t = 9.52 g/t Au
  • Top End
    • 210,526.32 / 515,850 =  0.408 oz t
    • 0.408 oz/t = 12.69 g/t Au

Even if Pretium meet the top end of their guidance, the planned head-grade at Brucejack will be >20% lower than expected, which poses some interesting questions:
  • Are Pretium deliberately aiming low so that so that they don't over promise and under deliver?
  • Are we seeing and indirect acknowledging that there is a problem with the deposit?
  • Whoever designed the mill did a great job - it is over-performing, both with throughput and recovery!
However, there is no cause for alarm, Pretium are telling us that everything is going so well (ignoring teething problems like mining lower grade rock and producing less gold - nothing important), they've decided to make the mill bigger (link). We all know that bigger is better, but is this quick 'paper over the cracks' fix to bring short-term production up to levels stated in the feasibility study?

They admit to albeit indirectly, in the PR with the new grade control drilling and poor reconciliation,   is it wise to increase the mill capacity if they still don't fully understand the gold distribution?

They may find themselves in a position where they have a hungry hungry mill that they'll either start to shove any ol' crap through it (i.e. low grade development rock) or use bulk mining techniques (excessive dilution) to keep it fed.

Often this starts the death spiral for an operation, with decreasing head grades leading to decreasing production, lower revenues and so on. 

2018 is going to be an interesting year...


  1. Last at $ 8.68 and this is getting serious. Is the problem chronic or acute? If it's the later where's the buy line? Just about any other mine on earth would be overjoyed with 8.5 gms/ton.

    As an aside AG take a freakin bow. You called this bang on many, many moons ago. We are genius through it rubbing off.


  2. The only Gold mine I know of thats Cash-flow positive after just 2 quarters.

    1. But, if they had reached their targets outlined in the FS, how much more gold and cash would they have? They do have a massive debt to pay.

  3. AG,

    Thanks for reviewing Pretium.
    Nice sleuthing to uncover the grade implications.

    However, given that Pretium have hit bonanza type grade in exploratory drilling outside the current resource areas - is it not also possible the mill enhancements will be very much needed ?

    Robert Quartermain sits amongst the 20% of people that produces 80% of the results - I'm currently seeing this as a buying opportunity.

    PS The Heart of Gold was powered by an Infinite IMprobability drive... just in case your reference was to the HHGTTG :)

    1. Hello,

      It was a hitchhikers reference, and I could do with a nice cup of tea.

      The issues with bonanza type mineralization is that they are hard to predict, which was the base of the comments made by Strathcona (http://www.pretivm.com/news/news-details/2013/Pretium-Resources-Inc-First-Bulk-Sample-Cross-Cut-Processing-Results/default.aspx) leading to their withdrawal from the project.

      What the concern is, that if Pretium have misinterpreted (or been overly optimistic) the gold distribution and 'created' a large, bulk mine-able deposit, where in actuality it is a series of irregular ultra-high grade veins that can only support a smaller operation is that by increasing the mill capacity and the mining rate you compound the problems as there may not be sufficient veins that can be mined to produce enough material to support the mill.

      the next 5 months are going to be key. The good news is that the Q4 financial results will be great as they had almost 30K oz gold they that they hadn't sold from Q3.

  4. While your geology is bang on, I question everyone's sanity here - at the low end of all this - say 9g/ton - and pulling back spending a bit once they get through this 'episode' - the mine, at $1,275 gold will free cash flow (FCF) somewhere around $260mm/year. HELLO? The EV of the company ($350mm LOC, streams and the current market cap = $1.9bn usd = 7.3x FCF. I've never seen such a world of haters...death spiral??

    1. You make some huge assumptions including the production rate can be maintained at 9gpt and also "pulling back spending." It's likely costs are going to be higher than projected and they may find it very difficult to delivery 9gpt to the mill at 3800tpd (there could be bottlenecks upstream of the mill). An argument could be made for FCF more like $100-$150 million per year. At that rate debt repayment starts to become iffy.

    2. according to the FS by Tetratech, they should be able to maintain almost twice the grade, but reduce the mine life to 13 years (excluding discovery or new resources).

      I'm using the logical argument - bigger mill = more production as all variables stay the same (Armchair expert theory)

  5. Thank you for the article TAG,
    Could please explain what grade is the breakeven grade for this company, considering their sustaining costs and debt payments?

    thank you,

    1. How is it possible to know this at the current point? We don't know their development and production cost at 3800tpd but guaranteed it will be higher vs FS.

  6. I think we are all second guessing PVG - what if this? what if that? and mental masturbation is a poor substitute for facts. The only people who truly know the facts, down deep in the minutia of the detail's details, are management and insiders and about all they can be criticized for is not getting ahead of the story in their press release and interviews. Where, pray tell us, is Quartermain in all this, as this truly is HIS show piece swan song of a project? I'm not satisfied by the answers provided so far by the company, but I'm still sitting on my hands and holding my shares. PVG has a lot to prove in the next 3 or 4 quarters and I sure hope they do!

    1. Hello Yodie,

      I hope that this is just an early ramp-up hiccup, maybe development was slower than expected and that high-grade areas will come on stream later than expected. For companies this is a double-edged sword - do you tell the market, and get crucified, or hope you can catch-up and hope that no-one notices.

      The fact in the Q4 production figures they 'forgot' to tell us the grade (just tonnes mined and gold produced) showed that they knew there was a problem, but didn't want anyone to know. If the problems continue, I wonder if they'll just tell us how much gold they producted.

  7. I realize AG is doing some quick, back of the envelope calculations, however, I think a better attempt at a correct calculation would have been helpful. After "Step 3, back calculating the head grade", the concept of there being dilution is completely missed. The feasibility study does assume and base it's costs on there having been 15% dilution, that is the ore is mixed with 15% waste rock at a grade of 0 g/t before going to the mill. Therefore, the head grade hoped for is 13.6 g/t as was clearly stated in the technical reports. The higher end of guidance 12.69 g/t is actually reasonably close to this, it's only 6.7% away, not a huge gap of 20% that was incorrectly stated. This is the only part of the web page that I actually read. If the conclusion is not correct, then no need to read any further.

    Regards, Paths ( my id from the stockhouse pvg board)

    1. Paths,

      Read the FS more closely, look carefully at table 16.8. the 13.6 g/t Au headgrade relates to year -1 where they mine 80,000 tonnes at 13.6 g/t.

      If you look at the following row, year 1's production you see that they were planning to mine 839,000 tonnes at 15.4 g/t Au

      If you look at the Reserves for Brucejack (http://www.pretivm.com/projects/brucejack-overview/reserves-and-resources/default.aspx)
      they are 15.6Mt @ 16.1 g/t Au and LOM production headgrade is 14.1 g/t Au, so I've assumed that they have factored in the dilution into the FS calculations.

    2. Ok but that is somebody making an attempt to imagine the scenarios and decisions being made four years in advance. As a basis for comparison, I would prefer to use the diluted values as calculated in the Dec 2013 Snowden report, page 133, Table 15.2 where the diluted grade is listed as 15.1mt at 13.6 g/tonne. I would also be almost tempted to view 1200 to 1230 as development ore, since they are extending into a lower grade for the purpose of establishing a sill pillar location.


    3. I think you have this backwards. The sill pillar is at certain intervals (every 120 meters) and so they happen to split the deposit into more-or-less equal thirds. What you presume to be lower grades at the 1200 level is lower drill density because they are not going to mine the sill for a while so they didn't feel the need to drill that level out fully. If anything, the fact they are not able to reconcile grades while following the zones down from the more densely drilled level (1230) to 1200 level is an indication of a problem with the model (i.e. poor vertical continuity of high grade domains).

    4. The sill pillars at 120 intervals are temporary, and will be completely removed, when it fits the sequence. There are two main sill pillars that act as boundaries of the "core of the VOK", shown (sort of) on page 15-9.
      I think it was lower drill density because, the original plan was to locate the sill pillar at 1230(??) Which may have been shifted down to 1200. This all makes sense to me, I don't see why people online are not viewing things with calmly.

      Later, Paths

    5. Yeah I already said the sill pillars are temporary. I don't know what you mean by "original plan" but the FS has the sill pillars at 1170-1200. Not all areas have stopes continuing vertically from level to level and the plan could be adjusted to access stopes based on the target grade, which is why some stopes would have started at the 1230 level. The reason there is not a lot of drill density is the nature of the fan pattern. It spreads out from the drill station. And since they aren't going to mine the 1170 to 1200 level for a long time it made no sense to have a drill station at the 1200 or even 1230 level. So the 1200 level stopes are informed (not very well) by infill drilling from the 1260/1270 level drill stations.

    6. The bottom sill pillar and maybe the one at the top of the middle VOK, in reality will most probably be permanent. The excavation will be very large, and the rock strength is moderate or less, Q' 4-10. Not practical to try to return there later.

    7. Correction Q'= 10-40 into the category of good rock strength, but not in the very good category, may well by that 1200 down to 1170 levels may not be mineable. Paths,

    8. It's certain mining costs will be higher due to smaller stopes, ore dilution and incremental ground control. May only be worthwhile to extract the sills at higher gold prices or if a selective mining method is developed.

    9. Yes, but not just higher cost, the sill pillars below and above the entire middle vok, have a good chance of being stressed past yield strength and becoming physically impossible to mine later, some inferences to that in the report. Paths,

  8. This comment has been removed by a blog administrator.

  9. Dilution appears to be the problem. In the PR, they emphasize drill density and grade control.
    These are narrow highly variable veins. To mine selectively you need tight infill drilling and good grade control to get the grade otherwise dilution and ore loss will kill it.
    Looks like they are considering less selective mining. You hope to recover all the metal in the erratic veins by taking more rock. This is at the expense of grade through the mill.
    I would argue that the resource classification might need reviewing, 20-25% variability on a panel of 6 months production is huge for measured, it's barely indicated.

    1. It's not veins, it's a corridor of low grade stockworks dotted with high grade stringers. In a few places the stockworks themselves have a grade above cutoff but in the vast majority of mining blocks it is the high grade stringers that make a block presumably economic. The problem is the stringers may be subject to controls that limit their spatial distribution in ways that the model has not incorporated. As a result, stringers may not actually exist where the model has assumed them to be and conversely there may be economic concentrations of stringers in areas where drilling has missed them. The solution is to mine on a selective scale with a focus on the highest concentration of stringers (based on infill drilling) and along the way try to understand the local controls which will presumably allow you to chase the stringers into those areas where they continue to be concentrated and avoid those areas where they peter out.