Actionable insights — Live with the Oak Bloke

1. The screen: a blown-out sector or country, plus a catalyst, plus a four-to-five-year clock

1. Start with what has been thrown away wholesale rather than what is individually cheap — a sector nobody will finance, or a country nobody will underwrite. Wholesale abandonment is what creates prices that ignore the assets underneath.
2. Require a catalyst you can name. His list is short and concrete: a new CEO, a policy change in the country, a divestment that refocuses the business, the start of a spending cycle. "Cheap" without a catalyst is a value trap; the catalyst is what converts a discount into a re-rating.
3. Check the asset can survive the wait — cash generation now, an oligopoly or licence position, no financing cliff. The catalyst decides the upside; solvency decides whether you are still there for it.
4. Underwrite four to five years, not four to five quarters. Both of his named multibaggers took five. If your timeframe cannot absorb that, this method does not work for you.
5. Expect the path to be jagged and plan the re-entries in advance (insight 2).
6. Prefer situations small or awkward enough that professionals cannot participate — "this is stuff that institutions can't mess with."

• The catalyst arriving on schedule versus sliding indefinitely — a spin-off that keeps being "under review" is the tell. Whether the discount is closing through NAV growth, buybacks or a re-rating (all three is the Georgia Capital case). And the falsifier: a blown-out sector where capacity is being added rather than starved, which means the abandonment was correct.

2. Plan for "several bites at the apple" — treat the drawdown as scheduled, not as news

1. Enter with a partial position, having written down in advance that a multi-year thesis will go through at least one violent, unrelated-looking sell-off. "Nothing is linear."
2. When the sell-off comes, ask one question only: does the news touch the thesis, or the geography of the thesis? Proximity is not exposure.
3. If it does not, add. Say so publicly if you publish, because the discipline of restating the thesis at the low is what keeps you honest.
4. Look for a mechanism by which the scary event makes the asset better, not merely unharmed — that converts a hold into a buy.
5. Keep dry powder proportional to the number of bites you expect, not to your conviction on day one.

• Whether the drawdown is accompanied by an operating deterioration (guidance, contracts, financing) or only by narrative and proximity — only the first invalidates. Also watch your own behaviour: if you cannot add at the low, the original size was wrong (insight 5).

3. Price your structural edge: time arbitrage, because nobody is asking you why it is down 30%

1. Write down explicitly what you have that a professional does not. His answer is not information — it is the absence of an investment committee, a quarterly report card, and redemption risk.
2. Deliberately fish where that edge pays: positions too small, too illiquid, too foreign or too slow for a fund to hold. "This is stuff that institutions can't mess with… they're not going to mess with this."
3. Set the holding period to match the edge. If your advantage is that you cannot be fired for a bad year, then a five-year thesis is the product you should be buying.
4. Refuse trades that require you to be right quickly — that is competing where the professionals are strongest.
5. Corollary for anyone with a shorter clock: this whole framework degrades, and the honest response is to size smaller, not to shorten the thesis.

• Self-deception: "time arbitrage" is only an edge if the thesis is actually intact. Set a review trigger (a named operating metric or catalyst deadline) so patience cannot silently become denial. And watch for the edge disappearing — when an ETF or a large fund starts buying your illiquid corner, the discount you were being paid for is gone.

4. Benchmark against a full-cycle record, not a run — the 20%-a-year test

1. To judge anyone (including yourself), get the cumulative record across at least one complete cycle, not the current run. "Everybody's a genius when the market's going up."
2. Ask specifically about the down leg: "what kind of a drawdown do they take? What's their ability to navigate a bear market or crisis?"
3. Anchor on the number the greats actually compound at: go to "the first page of any Berkshire Hathaway annual report or Fairfax Financial… What's the average? 20% a year."
4. Use it as a filter on your own expectations: if a strategy requires 100%+ annually to work, it is a lottery ticket wearing a spreadsheet.
5. Distinguish process from outcome before crediting anyone (including yourself) with skill — the $10,000 bill you found on the street was a windfall, not an edge.
6. Watch for the shooting-star pattern in newsletter and social-media track records: rapid rise on one theme, then burnout when the wave lands. "It's like surfing… now you got to paddle back out."

• Records that begin at a market low; strategies whose worst historical drawdown has not yet been tested in live money; and single-theme concentration masquerading as skill. The cleanest tell is whether the manager's process would have produced the same trades in a different regime.

5. Size a binary by the sleep test, then scale in as it de-risks

1. Classify the position honestly. Reactivation, permitting, financing and political-reopening trades have binary outcomes — they either happen or they do not.
2. Open small, explicitly small enough to be wrong. The first purpose of the position is not profit; it is that "you take a small position, then you're forced to watch it."
3. Add only against observable improvement — the situation "heals or gets better" — and preferably on a pullback, never on the good headline itself.
4. Apply the physical check on total size: "if you're laying in bed looking at the ceiling fan and your stomach is hurting and you're worrying, you can't sleep, then your position's too big. Sell it down to where that's not happening."
5. Pattern-match against your own completed trades of the same shape before sizing; prior reps in the same structure are the only real edge on a binary.

• The de-risking events you named in advance actually printing (a licence, a first cargo, a financing closed) rather than the story merely getting louder; and your own sleep. Losing sleep is the signal to cut size before the thesis is tested, not after.

6. Jockey investing: buy the allocator plus the tax asset, not the current portfolio

1. Find a listed vehicle whose stated purpose is to allocate capital rather than to operate one business — a post-restructuring holdco, a cash shell with a mandate.
2. Judge the jockey: has this person done this deal type before, in this geography, and did they arrive by cleaning up someone else's mess? "He came in, sold a bunch of stuff" is a stronger signal than a plan.
3. Count the structural assets the vehicle carries that a normal acquirer does not — net cash, and especially accumulated tax losses, which make future profits worth more inside this wrapper than outside it.
4. Check the alignment and the backer: who owns the control block and what is their time horizon.
5. Ask what the vehicle lets you own indirectly that you would refuse to own directly — that is often the whole point.
6. Accept the concentration risk explicitly: this is a bet on one person's judgement, so the exit trigger is the jockey leaving, not a bad quarter.

• Deals actually closing and at what multiple; the tax losses being used rather than merely disclosed; cash burn on overhead while the allocator waits; and any sign the jockey is being pushed by the control shareholder. Key-man departure is the thesis-ending event.

7. The country screen: growth you do not have to be clever about, and 54 different Africas

1. Look for a regime change in incentives, not a regime change on a map: a successor who "realizes that I can steal more if the pie grows" is a more reliable reformer than an idealist.
2. Demand a growth rate high enough that security selection stops mattering: "I don't have to be a genius if Uzbekistan's going to grow their economy at 7 to 8% ad infinitum… you're casting a net into a larger school of fish."
3. Prefer good demographics and low debt — the two things a Western allocation cannot buy at home.
4. Disaggregate the region. "In Africa there's 54 countries. There's a difference between Rwanda and Equatorial Guinea. One's investable, one's not… francophone Africa with all the coups is not the same as Kenya or Tanzania." The same for Central Asia: Uzbekistan and Kazakhstan are open; Turkmenistan will not even issue a visa.
5. Separate cheap from investable. Cheap-and-uninvestable goes on a watch list with a named condition for promotion, exactly where Venezuela sat before it became a position.
6. If the direct route is closed, take the listed one — a London or New York-listed fund or holdco is the keyboard version of getting on a plane.

• The reform programme surviving a leadership change (this is a one-life-expectancy risk in most frontier markets); GDP growth actually reaching listed companies rather than state champions; currency convertibility and repatriation rules; and the promotion condition on any cheap-but-uninvestable name — for the UK, a policy shift, not a lower price.

8. Outside your circle of competence, rent judgement — the ETF-plus-13F-correlation screen

1. Admit the bandwidth constraint before the analysis, not after: "how many girls can you kiss?… Is it in your circle of competence?" A sector that would take a lifetime to learn goes in the "too hard pile" — unless you change the instrument.
2. If you want the cycle rather than the company, buy the cycle: take a foothold via an ETF while you work, on the view that the sector runs in eight-to-ten-year waves and you are near a trough.
3. Identify the genuine specialists in that sector — funds whose whole existence is that one field.
4. Pull their 13F filings (the quarterly US disclosure of institutional holdings) and look for overlap: "what are the correlations across all these portfolios that everybody likes?" Consensus among specialists is a shortlist, not a recommendation.
5. Use an AI to do the aggregation — he names it explicitly: "you could use Claude for this… upload all these 13Fs or have it do it."
6. Then do the part that cannot be outsourced: understand why the five overlapping names are held. "Then you can maybe have enough bandwidth to kind of understand why that is."

• 13Fs are a quarter stale, show US long equity only, and omit shorts and private positions — so treat overlap as a research queue, never a portfolio. Watch for crowding (the same five names across every specialist is also a crowded exit) and for the sector's cycle actually turning rather than merely being cheap.

9. When a trend gets politically expensive, buy its second and third derivative

1. Take a consensus trend everybody is already buying at the first derivative (the chips, the model companies, the hyperscalers).
2. Ask what physical inputs it consumes and where those are actually constrained — for AI: power, water, and land you are permitted to build on.
3. Look for the point where the constraint turns political, because that is where the bottleneck becomes durable rather than merely tight. "It becomes a populist issue with the voters" once household electricity bills rise.
4. Then ask who owns the scarce input and collects regardless of which operator wins — landlords, royalty holders, water-rights holders, toll collectors.
5. Prefer owners with no operating risk and no need to pick the winner. "Nobody knows" which model company survives; everyone needs the acreage.
6. Sanity-check that the counterparties are already transacting: signed or advanced siting deals separate the thesis from the story.

• Announced siting deals converting into signed contracts with cash terms; water-rights litigation and aquifer depletion data; whether behind-the-meter rules loosen (which would restore the grid option and remove the scarcity); and valuation — a landlord thesis this well known can already price in the deals.

10. Express a commodity view where the leverage actually is — and know when the equities stop being free options

1. Establish which phase of the cycle you are in. Early phase: nothing is being produced, so every developer is a story with nothing to disprove — "all visions of sugar plums… a come bet," and a basket works.
2. Once companies must actually produce, switch instruments: "now these companies actually have to perform and the disappointment has set in." Operational reality separates the survivors, and you are unlikely to pick them.
3. Before buying any producer, find its realised price in the accounts and compare it to spot and to the term price. Legacy contracts can hand you all the operating risk and a fraction of the commodity upside.
4. If the producers fail that test, own the commodity itself through a physical trust — and buy it mechanically off its discount to net asset value, not off the news.
5. Alternatively own the bottleneck nobody can replicate (conversion, processing, storage) rather than the ore body everybody can option.
6. Write down the condition that would end the trade — for him, a major finally committing billions to build new supply — and until it appears, treat sell-offs as adds. "Use the volatility, as Rick Rule says… as your friend. When they sell off, you have to buy more."

• The trust's discount/premium to NAV (a sustained premium means it is issuing units and buying material, which is the bullish regime, not the buy point); producers' realised prices converging toward term as legacy contracts roll off — that is when the equities become interesting again; a major committing capital to greenfield supply; and, for the theme's end, Germany reversing its nuclear position.

11. Force objectivity with a structural checklist — Porter's five forces (host)

1. Before valuing a company, score the industry on five axes: power of buyers, power of suppliers, threat of new entrants, threat of substitutes, and rivalry among incumbents.
2. Treat a favourable score on one axis as necessary but not sufficient — "nobody will finance a new mine" is a real moat and tells you nothing about demand.
3. Count every party with a claim on the cash, not just the suppliers: governments, regulators, and monopoly infrastructure between you and the customer.
4. Accept a mixed scorecard as the useful output. The purpose is to state the weaknesses out loud so a low earnings multiple cannot silently do the arguing.
5. Re-run it when a leg changes — a substitute becoming cheaper is the fastest way a structurally protected industry stops being protected.

• Rail and port throughput (for a bulk exporter the logistics monopoly is often the binding constraint, not the ore); the cost curve of the nearest substitute; and any change in the financing embargo — if capital returns to the sector, the one favourable leg disappears.