Discussing Exploration Risk Video

Francis MacDonald of Kenorland recently did a great talk about exploration generally and project generation in particular.  

I think this video is important enough that I'm going to do a timestamped discussion of it.  These are basically just me sharing my personal notes.  Before I jump in it's worth saying I think the talk is great.  Francis knows way way more about this stuff than I do.  I also like Kenorland enough that I bought shares.  But I'm not going to agree 100% with everything in the talk, even though I think there's a lot to love here.

1:14 Project Risk

I think it's a pretty good model of the different kinds of risk on a project.  What ties all of these together are things that prevent the creation of a profitable mine.  Because ultimately the value is the present value of the future cashflows from a profitable mine.  In early stage expoloration by far the biggest percentage risk is that you don't find anything that could become a resource that could become amine.

5:29 Success Curves

I really like how he maps this to similar industrial success curves.  I've been muddling over how eerily similar mineral exploration is to venture capital.

VCs pick winners only 2.5 percent of the time

out of more than 4,000 VC investment rounds annually, the top 100 generate between 70 and 100 percent of industry profits

among the very top performing VCs, 4.5 percent of invested capital generates 60 percent of their funds’ returns


6:45 Projects are measured in Area

It's not number of projects, but there is a spatial component.  How much land area.  What he talks about more later on is that land area is also a liability.  Besides staking/acquisition costs and holding fees to actually cover area with geophysics or soil sampling is quite expensive.  There is also a 3d depth component where a company like Kodiak copper can come in to a relatively well drilled area and drill deeper and make a new discovery.  

6:55 Project Tiers

It used to be Tier 1 was 500,000 oz/year for 10 years, 5Moz total.  It also assumed lowest quartile costs.  Increased difficulty and time to permit have increased how big a project needs to be to justify going through that process.  Meanwhile it's getting harder to make new discoveries.  In Canada, Australia, Nevada, and similar it's very rare to be able to walk up to an outcrop, hit it with a hammer, assay the chunk you broke off, and drill right underneath.  

7:33 1 Million Hectares Per Tier 1 Discovery

This is money.  Screenshot it, print it, stick it on your wall.  63 >10Moz AuEq deposits in Canada.  61,000,000 hectares of mineral claims in Canada.  968,000 heacares of mineral claims per Tier 1 deposit.  

8:27 0.5 Million Hectares Per Tier 2 Discovery

Also money.  127 >5Moz AuEq deposits in Canada.  480,000 hecares of mineral claims per Tier 2 deposit.  I also notice that as the size of the deposit doubles the number of deposits of that size halve.  

8:42 264k Hectares Per Tier 3 Discovery

This is the third slide in a row where I'm just like why isn't anyone else doing this level of analysis?  This is great.  231 >2Moz AuEq deposits in Canada.  264,000 hectares per.  Slight typo here as he meant Tier 3 deposits.  Also for symmetry it would have been great if the cutoff here had been 2.5Moz so each tier was double the previous tier.  Still, the relationship of doubling the deposit size and halving the number is roughly true.

10:11 Prospectivity Mapping - Not All Mineral Claims Are Equal

Prospectivity mapping 14% of the land area captures 82% of the deposits.  Yet 87.8% of the staked ground is outside of prospectivity corridors.  So 585% more likely to have a deposit in the prospective ground compared to outside the prospective ground.  Now, this is a little circular because the ground with deposits in it is how we decided the ground might have characteristics to have deposits in it, so of course that ground will score well.  Ideally you'd train your prospectivity data twice on each half of the deposits and see how much the data both overlaps in prospectivity scoring and how well it predicts the other half of the deposits.

13:20 Prospective Ground Still Low Odds

The percentage of ground hosting deposits even in highly prospective area is still really low.  Though of course there are still more unfound deposits.  It's also convenient that the solution here is Kenorland's approach.  It's a valid approach, one I endorse.  But there are other approaches.  One such approach is to go to parts of the world where you can walk up to outcropping rock and hit it with a hammer.  You can do that in Columbia, Ecuador, Mongolia, and a few other places.  Of course you'd rather build a mine in Canada.

14:59 Exploration Maturity Via Grid Mapping

I like this model a lot.  500m grid, 500m of drilling in that grid, pretty low chances you'll find a >2Moz deposit drilling more in that grid.  

16:45 Canadian Malartic

It's worth noting Canadian Malartic was staked by a project generator and optioned to Osisko.  Golden Valley spun out Abitibi Royalties, both were later bought by Gold Royalty Inc ($GROY) for a handsome profit for shareholders.

4.2.2 Malartic CHL Property
The property known as the Malartic CHL Prospect originally comprised 10 contiguous mining titles. The total surface area of these claims (numbers 72283 through 72292) amount to 388.64 ha. The claims were originally registered in the name of Golden Valley Mines Ltd. (“Golden Valley”) but were transferred to Abitibi Royalties upon its creation and listing in March 2011.
In 2006, Golden Valley entered into a mining option agreement with Osisko regarding the 10 claims. In return for cash payments totalling $150,000 and a work commitment of $2,000,000, Osisko would be vested with a 70% interest in the property. Golden Valley would then be carried to production with a 30% interest in the property, then transferred to Abitibi Royalties. Osisko became the operator of the joint venture and the option was exercised by Osisko in 2011. A joint venture between Abitibi Royalties (30%) and Osisko (70%) was deemed to be formed upon Osisko’s exercise of the option.
In March 2015, after the Osisko Transaction, Abitibi Royalties sold its 30% interest in the Malartic CHL Prospect to CanadianMalartic GP in exchange for 459,197 common shares of Agnico Eagle and 3,549,685 common shares of Yamana, with a value of approximately C$35 million (based on the respective closing prices of such shares on the TSX on NI 43-101 Technical Report – CanadianMalartic Mine – March 2021 47 February 20, 2015, the date immediately prior to the public announcement by Abitibi Royalties of entering into the letter of intent), and 3% NSR royalties to Abitibi Royalties and Osisko Gold Royalties Ltd. (“OGR”) on the Malartic CHL Property. Both royalties are guaranteed by hypothecs. Otherwise, the mining titles are free and clear of all hypothecs, or any other encumbrance, including royalties, whether for minerals, metals, extracts or by-products.

17:35 Geophysics Fail

OMG, this is amazing.  VMS are one of the deposit types considered to be a good target for geophysics.  Xstrata flew MEGATEM from 2001-2006.  2,179,403 hectares flow, 180,000 line kilometers, 40,000 EM anomalies, 2,500 selected anomalies, 268 diamond drill tests, no econmic disoveries.  Exploration is hard.

18:40 Geochemistry Fail

This is pretty honest self-reflection.  I'll not pile on.  Still could find something here, but tough drill campaign.  Suffice it to say expoloration is hard.

20:00 Ateration Fail

Alteration and breccia with no significant discovery.  Expoloration is hard.

20:45 - Structure Fail

Exploration is hard.

23:23 Estimating Exploration Costs Method 1

This is really great.  Drill costs will vary widely.  I tend to use $300/m as my default if you can avoid a helicopter.  His $250/m is ballpark the same.  The discovery costs are super interesting. $725m for >10Moz AuEq, $383m >5Moz, $192m >2Moz.  He calls out the issues with this data set, but it's still super helpful.  Wonderful dataset.

25:51 Estimating Exploration Costs Method 2

Wow, great find.  $320m (2022 USD) per Tier 1 Discovery, $111m if you net out cash generated by divesting smaller deposits.

26:58 Estimating Exploration Costs Method 3

These three slides in a row have really shown three different ways to look at expoloration costs that are each phenominal, and together really have left me with nothing to add.  $1.55b per >10Moz AuEq discovery.

28:46 - He's covered this slide in a previous expoloration talk (linked in the Required Reading section of this website).  

30:13 - Great list of companies.  This makes me want to put together a project generator index both to keep track of interesting companies for myself and to see how project generators have performed compared to other things like the $GDXJ.  There are of course differences in this list.  EMX is turning into a royalty company, buying existing royalties, though their project generation is still active.  Altius is also more of a royalty company now.  Globex considers themselves a land bank, which they define as a project generator that isn't also an operator.  Some companies are missing like Val-d'Or.  Some have exited like Golden Valley.  But it's an impressive list, at 100 it's way more than I've found.

31:00 - That's a lot of mining companies.  Anytime you wonder why a company you are invested in doesn't move on some obvious news, realize that people probably just missed the news.  1 >10Moz, discovery for every 350 companies.  Using ratios from earlier slides that means 1 >5Moz for every 175 companies, 1 >2Moz for every 95 companies.

32:50 - This is one of the more important points, which is that project generators are much more likely to be involved in a major discovery.  But I think the data set is too small to be very accurate.  As an example of the data set being too small Mark Creasy is an individual and not a pubicly traded corporation, so removing him to make an apples to apples comparison reduces the number 25%.  It would be good to see if this holds for >5Moz and >2Moz sizes as well.  However, the general point seems valid if it is 10x or 20x

34:06 - This is a great example.  I'd like to see a few more examples of how the project generator stock compared to the exploration company that optioned a property with a major discovery.  It's worth noting that prior the 1st drill hole in Porvenir, which Cornerstone has no interest in, Cornerstone stock was often outperforming Solgold.  So why does Cornerstone's much smaller stake do as well as Solgold's much larger stake?  One reason I've talked about before is benefit weighted spend.  

Valuing Prospect Generators Part 3 - Benefit Weighted Spending
A lot of prospect generators will talk about partner funded exploration. For example here’s Orogen Royalties (a prospect generator) talking about partner funded exploration of $3.2 million in 2021. Orogen Project Generation Business Unit Update - News | OrogenOrogenSo, what do we do with that infor…

Solgold had to cover 100% of the costs through a bankable feasability study, but Cornerstone benefitted from their undiluted 15% for every dollar spent adding value to the project.  A little more since Cascabel also got some Solgold shares as part of the option, an amount diluted down to 6.9% now.  Solgold meanwhile benefitted only 85% for every dollar spent, and had to raise every dollar they spent by issuing more shares diluting existing shareholders.  

I'll do a dedicated post on converting carried equity stakes to equialent equity in the company doing the carrying, but the short version is that you have to account for the expected dilution in the other company during the carried period.  So if Solgold diluted their existing shareholders 50% to pay for exploration/development (I have no idea the real number) to bankable feasability study then Cornerstone's 15% carried stake would have been worth 30% of Solgold's equity at option time.  

The choice of which is a better metric, benefit weighted spending or equivalent equity conversion that accounts for dilution, is really project/company specific.  Neither is perfect.  But in all cases a 15% carried interest is worth more than 15%, there is value in the free carry, it's just trying to figure out how much the carry is worth.

36:56 - 1.4 million hectares in Alaska and Canada, 2.2% of all mineral claims in Canada.  That's kind of a jaw dropping number.  In order to pre-drill screen all their land at $35/hectare they will need to spend more than their current market cap.   They've managed to get partners to pay for some of the pre-drill screening, but they have also paid for a limited amount of actual drilling themsevles.  If their claims are average claims they'd expect 1.5 >10Moz deposit, 3 >5Moz deposits, and 5.5 >2Moz deposits.  They'd expect more or less than that depending on the quality of their pre-staking prospectivity mapping and the effectiveness of their screening process.  

On Frotet/Regnault Kenorland is now a 20% contributing partner and as the operator gets a fee for 10% of the exploration costs, no that nets out to 10% net contribution for 20% of the benefit.  They also get 20% of whatever tax benefits there are for exploration.  That's pretty decent.  Using their 20% stake, a resource size of 2Moz-5Moz, and a per ounce value on the resource of $25-$100 per oz you end up with a range of value of $10m - $100m USD.  That's their most advanced project, which is a reminder that explorers and project generators are impossible to accurately value.  At the time of this writing Kenorland's total market cap is just over $28m USD.  So if you hear Kenorland say they are trading for less than the value of their stake in Frotet/Regnault it is an uncertain but reasonable claim.

39:00 Summary

I totally agree with points 1-3.

I really want to address point #4, because I think it's wrong and I don't think Francis actually believes it.  If there was no difference in risk/reward between explorecos and project generation and it was just about shots on net as an investor I could just invest a smaller amount in a larger number of explorers to get the same number of shots on net as a project generator.

Both explorecos and project generators value comes from the present value of the future cashflow of a profitble producing mine.  With explorecos there are a lot of costs to get to a producing mine.  Most of that cost in the early stages is drilling (mostly diamond, but also rc in some disseminated deposits).  Later it is studies and construction costs.  After you account for the time value of money, risk premium, cost of capital the value of an exploreco is the present value of all those future cashflows from exploration, development, construction, & mining.  And since most exploration goes nowhere near an actual producing mine you have to add all the failed exploration into those costs too.

The simplified version for explorecos is that the value of what you found is worth more than it cost you to find it. I call this wealth creation.  

Wealth generation is the most important part.  It is what separates good explorecos from bad explorecos and is one of the main factors seprating good project generators from bad project generators.  If you spend a lot of money and don't find anything the outcomes is bad regardless of how you structured things.  If you spend little money and find a huge deposit that goes on to become a profitable mine the outcome will be good.

Project generators are different from explorecos in one important way, they can benefit disproportionately from the wealth creation of others.  You can think of this as creating a call option on others success, avoiding a bigger percentage of costs than the percentage of value given up to avoid those costs, and structurally getting a bigger share of the profits relative to their share of the costs.