Actionable insights — Where I See Value in Gold, Uranium & Copper

34:03 1. Apply a hard size filter first — the $10 billion in-situ tier-one test

1. Before assessing management, jurisdiction or grade, size the deposit in dollars: multiply the in-situ recoverable reserves and resources by the current metal price.
2. Require a minimum of $10 billion, "but preferably $20 billion," for it to count as tier one — his stated bar for his own account.
3. For gold specifically, apply the parallel unit test: at least one million minable ounces. Below that, decline to rank at all.
4. Sanity-check the commodity market as well as the deposit: if annual global demand for the metal is tiny, no deposit in it can clear the bar, so the whole sub-sector is out by construction.
5. State the exception honestly: tier-two/three deposits can still perform as stocks, because "somebody can point to the big picture around a restricted material and appeal to investors' feelings as opposed to the way they think. They can sell a narrative as opposed to reality." Deciding not to play that game is a choice, not an analysis.

• Resource statements quoted in tonnes and grade but never in dollars — do the multiplication yourself.
• The gap between the resource statement and the reserve statement (the Mayfair problem): resources are "probably there," reserves are economically minable. Size on the second.

35:17 2. Make price part of the rating — downgrade on strength, upgrade on weakness

1. Keep one number that blends business quality and price, so a rating can change without any news.
2. When the share price rises against an unchanged business, cut the rating — the return available has shrunk.
3. When it falls back and the pipeline is intact, restore it. The work is already done; only the entry point changed.
4. Publish the reason explicitly ("based on price appreciation") so you can't later confuse a valuation downgrade with a deterioration in the business.

• Your own reluctance to downgrade a winner — the tell that the rating has become an identity rather than a measurement.
• Ratings that only ever improve after price rises: a momentum system wearing a value costume.

22:38 3. Close the sell side, then only act on down days

1. Decide, in advance and at the portfolio level, whether there is any price near current at which you would sell. If the answer is no, say so — that decision alone removes half of all possible actions.
2. Having done that, stop timing: "I'm not a timer… When I look at what the uranium business is going to do in the next 10 years, I'm really unconcerned with timing the market."
3. Set the only remaining trigger to be a bad tape, cause immaterial — "on days that are very bad days, when the NexGens or the Camecos sell off for whatever reason — the prime minister of Japan sneezed or something like that — I try to buy stock."
4. Deliberately ignore up days: "I have no interest in up days. It's just down days." Green tape carries no instruction for a holder who won't sell.

• A pre-set list of "buy on a bad day" names with the size you'd add, written before the bad day arrives.
• The distinction between a bad day (price) and bad news (thesis) — this rule only applies where the thesis is untouched.

16:51 4. Judge a miner on 20 years of reserve reconciliation, not on earnings

1. Remember that a mine is a depleting asset: every ounce sold is inventory permanently removed.
2. For each year, compare ounces produced against ounces converted from resource to reserve through the drill bit. A "positive reconciliation" means the company replaced more than it mined.
3. Run it over a full cycle — a decade or two — so that one good discovery or one high-price year can't flatter the record.
4. Then run the money version: how much reserve and resource did the margin from each ounce sold actually buy? If the answer is negative over 20 years, the company has been converting shareholder capital into production, not the reverse.
5. Use the result to decide whether management's capital allocation, not the metal price, is what you're underwriting.

• Reserve growth that comes only from a higher metal-price assumption in the reserve calculation rather than from drilling.
• Producers whose "growth" is acquisitions funded by issuing shares — check ounces per share, not total ounces.

11:58 5. Count the strategic shareholders — you can't run an auction with one bidder

1. On an exploration-stage company, read the register for industry holders (producers) as distinct from funds.
2. Score the benefits: adult supervision, and technical expertise the majors will often lend "at no charge."
3. Adjust the effective float — "major mining companies are seldom stock traders," so if two hold 10% each, the tradable float is much smaller than reported, which raises volatility in both directions. Size accordingly.
4. Then count the bidders. One strategic holder is a problem: rivals perceive it as having "a head start in any bidding process," which suppresses competition and the eventual premium. Two creates "dynamic tension."
5. Net it against the cost of doing business: in expensive terrain the continuous capital need can outweigh the sponsorship.

• A single major creeping toward a blocking stake — the point at which sponsorship becomes a ceiling on the takeout price.
• Whether the strategic holders' own project pipelines are full or empty; an empty pipeline is what makes a bidder motivated.

43:14 6. Never judge a commodity price in aggregate — split it by producer type

1. When a headline price looks distressed, ask whose cost curve you're comparing it to. The same price can be below cost for one producer and free money for another.
2. Sort producers into primary (the commodity is the product) and associated / by-product (it comes up alongside something more valuable).
3. For the by-product group the relevant cost is close to zero: with $80 oil in the Permian, "your gas is free… any price north of a nickel makes money."
4. For the primary group, compare against the specific field's cost of production — "$2.65 if you're a primary gas producer in a gas field like the Hugoton is below your cost of production."
5. Then pick your exposure by basin oversupply rather than by the national price.
6. Separately, ask whether persistent underinvestment converts a cyclical low into a future shortage — and be willing to upgrade the verb when the arithmetic supports it.

• Basin-level differentials rather than the benchmark price; takeaway capacity announcements that erase a local discount.
• The producers still funding sustaining capital at the bottom — they're the ones who own the shortage when it arrives.

41:44 7. Separate "great asset" from "good investment" — and let price be the deal-breaker

1. Assess the asset and the sponsor on their merits and say so out loud, so that admiration doesn't leak into the valuation.
2. Then compare the market capitalisation to proven and probable reserves and resources — what has actually been demonstrated, not what the geology suggests.
3. If the gap is unjustifiable, decline regardless of the quality of the people involved: "I'm just not willing to pay up like that without a lot more by way of reserve and resource."
4. Keep the name on a watch list with an explicit re-entry condition (drilling that converts resource to reserve, or a much lower price).

• A promoter's track record being offered as the valuation case; newly listed vehicles priced on the sponsor's name.
• Resource conversion rates in subsequent drill programmes — the only thing that closes the gap honestly.

24:08 8. Handicap the geology you can't model — data density does not fix discontinuity

1. For an underground narrow-vein deposit, read the reconciliation: how actual infill grades compare to the block model. Large local variability is the warning.
2. Give explicit credit for data density and for a deliberately over-financed build — cushions that let a company survive a model miss.
3. But separate two questions: do we know the deposit? and can it be mined to plan? A structure that "shrinks and swells a lot" fails the second even when the first is well answered.
4. Check management ownership as a secondary factor — ~1% means the downside is asymmetric between them and you.
5. Decide by style-fit, not by disdain: pass without claiming the company will fail, and record who was right last time you disagreed with that team.

• Quarterly reconciliation disclosure once production starts — mined grade versus modelled grade is the moment of truth.
• Dilution and mining-width assumptions in the study; narrow veins fail on the metres, not the grams.

25:48 9. Identify the dependency that can never be disclosed — then refuse to forecast the timeline

1. Ask what must happen before the project restarts, and whether that event is one the company is legally or practically forbidden from describing.
2. If it is, accept that no amount of monitoring will help — "you aren't going to be able to monitor that. Nor am I" — and stop assigning dates.
3. Learn the euphemism that will stand in for the announcement, so you can recognise it when it arrives: "we are satisfied with the current security arrangements… and as a consequence of that we're resuming major construction."
4. Look instead for revealed-preference signals from people with better information than you: who is willing to take the job? Senior local hires after a security catastrophe "would suggest that people who know believe that it is either safe or becoming safe."
5. Size the position for an unknowable schedule rather than pretending to know one.

• Senior local hires, insider buying, and contractors mobilising — actions rather than statements.
• Company language shifting from "evaluating" to "satisfied with current arrangements."

38:26 10. Assume the good trade reverses — size it so a two-sigma move can't end you

1. Write down why the trade works structurally (for the yen carry: borrow at ~1% in a currency structurally depreciating, invest at ~5% in one appreciating against it — "an extremely reliable game for 20 years").
2. Then assume it stops working temporarily, because "markets do" reverse "for brief periods of time."
3. Size so that a second- or third-standard-deviation move is survivable. His rule on that trade: "I ran that trade on sort of 50% equity."
4. Recognise the failure mode: the better and more reliable a trade has been, the more leverage gets applied to it — "if you had the ability and the stupidity to leverage that trade 30 or 40 or 50 to one… you got stopped out."
5. Keep the historical template in view: Long-Term Capital Management was right about convergence and still died, because "there was a hiatus and they were so leveraged that they couldn't take a second or third standard deviation event."

• Crowding metrics on the funding trade; the point at which a carry trade's justification is its own track record.
• Your own leverage creeping up as a strategy proves itself — the mechanism, not the market, is what kills.

40:48 11. Attribute the share-price move to the right asset before you extrapolate it

1. When a diversified miner jumps, don't credit the headline commodity by default. Read the quarter and find which asset actually delivered.
2. Check whether the news in the "obvious" commodity was in fact bad — if so, the market is telling you something different from what the narrative says.
3. Weight operational results against the frictions they had to overcome (transport costs, political constraints) — beating those makes the result more impressive, not less.
4. Only then decide whether the move changes your view of the company or merely re-prices one division.

• Segment-level results and by-product credits in the quarterly, not the headline commodity's spot price.
• Logistics constraints being lifted (rail access, port capacity) — often a bigger swing factor than grade.

32:59 12. Refuse to publish a number where you have no edge — but still take the exposure if the thesis is structural

1. Audit your own historical accuracy by sector, not just your returns. Where the record is poor, say so.
2. Withhold a rating in that sector — an opinion you don't trust is worse than none, because it invites others to size on it.
3. If the exposure is still worth having, justify it at the industry level rather than the company level: deferred sustaining capital must eventually be spent, whoever spends it.
4. Then own the largest, most established names in that sector, where company-specific judgement matters least.
5. Where the industry thesis has a specific geography, prefer the operator with that specialisation.

• Industry-wide capex guidance and rig/frac-fleet utilisation — the confirmation that the catch-up has begun.
• Your own opinions in sectors where you can't name your last three calls.

46:31 13. Read the financing tells — an extended warrant is a funding warning

1. When a junior extends expiring warrants "for no reason," supply the reason: the warrants are out of the money and the company needs the exercise cash it was counting on.
2. Read it as a statement about the treasury, not about optimism: "yeah, they need the money."
3. Follow the money's destination before assuming it funds work in the ground — he ends the chain bluntly: "the warrant into stock and the stock into cash and the cash into salary for them."
4. Cross-check general-and-administrative expense against exploration spend before participating in any subsequent raise.

• Warrant-extension press releases with no accompanying exploration news; repeat extensions on the same tranche.
• G&A as a share of total spend across two or three years.

3:07 14. Expect the paperwork — plan bullion purchases around the $10,000 reporting rule

1. Know the rule before you transact: any cash transaction over $10,000 is treated as suspicious "in and of itself," and the institution must file a Suspicious Activity Report.
2. Expect it to apply whether you are withdrawing your own money or using cash to buy gold — it's the cash, not the purpose, that triggers it.
3. Treat the filing as routine rather than as an accusation — "I have done it hundreds of times."
4. If asked the purpose, answer accurately and plainly. His example: "to purchase physical gold to hold outside the United States," written on the form because holding gold offshore is not illegal. He never heard back.
5. Understand the institution's incentive: under-filing is itself a red flag for the bank, which is why the alerts are automatic and impersonal.

• Structuring — deliberately breaking a purchase into sub-$10,000 pieces is itself an offence; the point is to expect the report, not to avoid it.