Actionable insights — As Tech Trade Falters, Value Stocks Are Breaking Out

21:12 1. Trade the positioning extreme, not the narrative

1. Pull a speculative-positioning chart for the commodity (his is John Kemp's, "all kinds of variations on it" exist) and treat one line as the whole signal: "the red line is the key."
2. Read it inversely. High = "positioning in the futures market is very bullish, which means you want to go the other way… get out or go short." Low = "you want to be a buyer."
3. Buy the trough. On July 7 the line "hit the low point… and that's when oil was bouncing around 70 bucks a barrel."
4. Then use the same line to leave: it's climbing now, "and at some point it's going to get dangerously high — it would be time to exit."
5. Remember why it works mechanically: forced buying. A record short position "will eventually have to be repurchased," and because the trades "are mostly not human driven, they're computer driven" off headlines, the extremes overshoot in both directions.

• Spec-positioning/open-interest at multi-year lows in a commodity with a real supply deficit; record ETF short interest; the same chart turning "dangerously high" as your sell trigger.

4:40 2. Price the physical market, not the screen

1. Establish the size mismatch first: the derivative market "is something like 50 to 60 times the physical market," so the screen price is set by traders, not by barrels — "it really lends itself to the tail wagging the dog."
2. Go find an actual transacted physical price in the region of shortage. His: physical oil trading in Asia at $170/bbl in April–May, "not that far from 200" at one point, while the futures screen printed the 60s and 70s.
3. Treat the gap as the edge — and as the reason to disbelieve the consensus narrative built on the screen price.
4. Cross-check with operators rather than agencies: "it wasn't just Jeff [Currie], it was also Mike Worth, the CEO of Chevron and one of the senior Exxon executives" saying refilling is unavoidable (16:14). Discount the agency that has been "historically incredibly inept" — the IEA now forecasting "a super glut next year" into the worst shortage he's seen.
5. Separate transient flushes from structural tightness: the 130 million barrels released when the straits opened created "a near-term temporary glut" that fed the victory dances — and was "simply wrong" about the trend.

• Regional physical/spot premiums vs the futures strip; producer C-suite language on inventories; agency forecasts contradicting operator behaviour; one-off inventory releases mistaken for supply growth.

22:08 3. Screen for multi-year breakouts on undemanding valuations

1. Run the chart at a five-year scale, not a one-year one, and look for a price pressing against a ceiling that has rejected it repeatedly. "I love to look at multi-year breakouts."
2. Accept the trade before confirmation only when valuation gives you cover — then state it honestly: "admittedly this one hasn't happened yet, but I think it will. And when it does, it likely is going to run a bit more."
3. Check the valuation is "amazingly undemanding" on two measures, and ask which direction the denominator is moving: EOG at a PE of 10 with mid-range price-to-sales, and "you could argue that sales are going to be up substantially because of higher oil prices."
4. Treat optionality as free, never as thesis: "also a takeover candidate, though you never bet on that — that would be a kicker."
5. Generalise the pattern-recognition across asset classes: "it's not just with stocks and commodities — these breakouts work with the bond market too, and with currencies" (1:02:03).

• Five-year charts coiled under a repeatedly-tested ceiling; single-digit-to-low-teens P/E with a rising revenue driver; the same breakout logic applied to spreads and FX.

50:09 4. Buy breakouts on the pullback, not on the breakout

1. Identify the breakout — the ceiling cleared, "very bullish. It looks great."
2. Measure how stretched it is: "it is extended over the breakout points, extended over the 200 day moving average."
3. Then wait, deliberately: "I like to look at breakouts and then buy them when they pull back after they break out. The nirvana."
4. Use the pullback wait to verify the valuation is still there ("very reasonably priced 10 11 12 times earnings") rather than paying up for momentum.

• Distance above the breakout level and the 200-day; a first retest of the broken ceiling; valuation still intact after the pop.

28:01 5. Stack the discounts — value the fund and the market it references

1. Start with the fund's discount to what it holds: uranium spot is ~$85/lb while the Sprott trust is "effectively pricing at about 77" — a ~10% discount.
2. Then ask whether the reference price is itself the real price. In uranium it isn't: almost nothing trades spot — "almost all the volume occurs with utilities buying uranium to feed the reactors," on long-term contracts around $95.
3. Read the contract terms for the shape of the payoff: escalators, ceilings up near $150, floors "somewhere around current prices" — "you've got no downside and lots of upside over time."
4. Multiply the two layers: "you're getting a discount upon a discount… really a 20% discount from where most uranium is trading currently."
5. Require the sentiment confirmation — the discount only exists because "it's pretty out of favor"; when the fund trades at or above NAV, the edge is gone.

• Closed-end/physical trusts at unusually wide NAV discounts; a spot market that is thin relative to the contract market; contract terms with floors near spot.

37:10 6. Walk the substitution chain, then check the exporter

1. Name the shocked commodity and then its substitute. Hormuz threatens LNG as well as crude, and LNG is what Asian power stations burn instead of coal — so "Asia is going to have no choice."
2. Check the substitute's own supply side for a second, independent squeeze: Indonesia, "the largest exporter of thermal coal… in the world," is "drastically reducing their coal exports" to keep the fuel domestic — the same resource-nationalism move he flags in China's oil behaviour.
3. Buy the leg that has been "pounded" hardest rather than the one already working — "look at that pounding that it's had this year, like so many commodities… the whole complex got sold off very very hard and I think created just a tremendous buying opportunity."
4. Take the idea from a specialist and say so: credit to Trader Ferg, who is "really bullish on these coal companies and he's made a great case for it."
5. Act before the write-up if the catalyst is live: he issued "just a blurb of recommendation… on Monday because I was concerned they would start moving given what was going on in the Middle East," with the detailed note following.

• Export restrictions by a dominant supplier; substitute-fuel price ratios at extremes (thermal coal vs LNG); the most-pummelled member of a complex that has sold off as a group.

10:05 7. Harvest the volatility instead of being expelled by it

1. Accept the asset's nature before you size it: "oil markets [are] extremely volatile. So you get these big rallies, you get these big sell-offs."
2. Pre-commit to buying declines, because the alternative is capitulation: "if you don't take advantage of the volatility, when the declines come you go, 'Oh gosh, what a fool I was, and I'll never touch energy again.'"
3. State it as a rule: "if you're going to be in energy, just like if you're going to be in precious metals, you've got to embrace volatility. Most people just have the hardest time with it… and it drives them away at the wrong times."
4. Run the sell side symmetrically. After a monster move, scale out rather than call a top — "I've been gradually liquidating some of my futures contracts just because it's had such a big move" — while keeping a residual for the tail case ("I still [have] a lot of contracts left because I do believe we could get a blow-off top").
5. Do the mirror image into manias: on silver's run toward $120 it "was a great time to do gradual dollar cost averaging into the blowoff" (1:16:56) — the same technique, reversed.

• Your own urge to abandon a thesis after a drawdown; parabolic price action as a scale-out trigger; position sizes small enough that a normal 30–40% swing doesn't force a decision.

51:32 8. Rotate out of the bubble instead of out of the market

1. Diagnose concentration, not "the market": ~45% of US market cap in AI and AI-adjacent stocks, ~70% of the NASDAQ. That is a sector problem masquerading as an index problem.
2. Check the historical base rate before assuming everything falls together. From 2000–02 the S&P fell ~50% and the NASDAQ ~80%, but "for most of that bear market you still would have been up if you'd been in value stocks" — only July 2002 took everything down at once.
3. Note that the laggards' bear market usually precedes the top: most stocks, "especially smaller companies," had been falling for two years from spring 1998 to spring 2000 (49:39).
4. Fund the new positions from the crowded winners: harvest AI gains "and then roll it into some of these other areas of the market that have been unloved but are now experiencing breakouts." "Absolutely… it's more likely to be the great rotation than the great meltdown."
5. Budget for the behavioural cost: "it's against human nature… we tend to want to keep our winners and buy more of them," and by the time a broken trend is obvious "you've thrown in a lot of good money after bad."
6. Then screen the destination on valuation: per GMO, value is very cheap and deep value "unbelievably cheap"; "it's amazing how many companies out there that are pretty good companies are trading at 10 times earnings or less — and some of it's because they're viewed to be AI victims" (53:47).

• One theme's share of index market cap; the breadth of the laggard bear market already in place; sectors that rose through the last comparable unwind; quality companies at ≤10× earnings tarred as structural losers.

38:59 9. Rebuild the earnings number before you believe the multiple

1. Find where the profit actually came from. Of Alphabet's very good quarter, "almost two-thirds… have come from the gains on their venture capital investments in these entities which are big customers" — booked in other income.
2. Name the loop: "it is a circular financing type of situation" — the vendor funds the customer, the customer buys the vendor's capacity, and the vendor books both revenue and a mark-up gain.
3. Strip the non-recurring line even though the accounting is proper: "they're required to book these gains, but you need to really make an adjustment and not just look at these headline earnings numbers because they are very unsustainable."
4. Add back the cost that hasn't arrived: the depreciation on hundreds of billions of data-centre capex "has really yet to hit the P&L." Do both adjustments and (per Fred Hickey) a teens-to-twenties P/E becomes "like 65."
5. Apply the same skepticism to a low P/E: when price-to-sales is extreme while P/E is single-digit, the margin — not the multiple — is the anomaly. Micron's "margins are just off the charts… rising based on scarcity" not innovation, so the earnings are the peak, not the run-rate (41:28).
6. Then check what the crowd is paying for a cyclical on peak earnings: CAT at ~32× earnings and 6× sales, "way above anything it's ever seen before" — "take some partial profits at least" (57:16).

• Share of earnings in "other income"/investment gains; vendor financing of major customers; capex-to-depreciation lag; single-digit P/E alongside record price-to-sales; cyclicals at growth multiples.

1:00:12 10. Use the credit market as the early-warning system — and know which gauge is broken

1. Start from the premise: "so often what gives you advanced warning when a bubble is going to pop is the credit market."
2. Check the headline gauge and be willing to report it clean: investment-grade spreads to Treasuries "are very tight… no warning sign at all coming from investment grade bonds, to be clear."
3. Move to the gauges that aren't index-managed. Private credit yielding ~9% vs junk ~7.3% is an unusually wide gap, and "default rates in private credit are soaring." A single marked-down portfolio (100 → 81 overnight, via Gundlach) contradicts the "our book is rock solid" defence — "something is rotten in Denmark."
4. Refuse the blanket verdict but keep the signal: "there's obviously really good credits… I don't want to panic" — yet "it is one of those warning signals going off right now." Verdict: "at least some fire."
5. Watch a within-junk quality spread rather than junk-vs-Treasuries: BB ("high-grade junk… less than 1% defaults per year") against CCC ("right on the door of default"). That gap has "widened out dramatically… about a 300 basis point" move and looks "on the verge of a multi-year breakout. I think they will break out" (1:02:03).
6. Read single-issuer bonds as a solvency verdict on a story stock: "Oracle bonds are trading almost like junk" despite a near-trillion-dollar market cap — "they really leveraged up… trashed balance sheet" (47:00).
7. Track the long end separately: the 30-year Treasury at 5.18% "is a breakout", rising despite soft growth and tame inflation because "foreign central banks have been backing away," shrinking the buyer pool. A run to 6% "would be really bad news for the stock market" (1:03:07).

• IG spreads (as the gauge that lags), private-credit vs high-yield yield gaps and default rates, BB-vs-CCC breaking multi-year resistance, single-name bond spreads on leveraged AI borrowers, the 30-year through 5.5% toward 6%.

1:20:58 11. Read fund flows inversely — apathy in a profitable sector is the setup

1. Pull annual fund flows for the sector ETF, not price. For GDX/GDXJ the largest outflow year on record was last year.
2. Compare the flows to the fundamentals over the same window. The miners had "tremendous returns… reporting really good profits" — so the money leaving carried no information about the business.
3. Invert the usual reading. Big inflows after big gains are "a big warning sign like it was back in 2016"; big outflows after big gains are "a lot of investor apathy" — the condition that precedes a re-rating.
4. Have a mechanism for the anomaly rather than shrugging at it: early in the year the metals worked while the miners didn't, so "a lot of precious metal oriented investors bailed on the miners."
5. Then require an improving fundamental, not just cheapness. Mining "has been very poorly run for a long time" but is "starting to emulate the energy industry" on capital discipline — the same shift that transformed energy shareholder returns (1:22:57).
6. Distinguish structural demand from speculative demand on the other side of the trade: measure central-bank gold buying in tonnage, not dollars, to test whether "they aren't really increasing purchases, it's just the price."
7. Check the historical analogue actually applies before assuming a decade of dead money: 1980 was "an artificial peak… the Hunt brothers," and its crash coincided with Volcker driving real rates to extremes — "that's not happening this time. And we're in a period of monetary debasement, which we weren't at that point" (1:19:00).

• Sector ETF flows diverging from sector earnings; inflow spikes as tops and outflow records as bottoms; capital-discipline language in mining/energy management; central-bank buying measured in tonnes; whether the bearish historical analogue shares this cycle's monetary regime.

1:06:34 12. Forecast the intervention, then own what it debases

1. Assume officials act, and rank their tools by political cost. First the disguised ones — sell short-dated Treasuries into strong demand and buy the long end ("the operation twist"), or "get the banks to step in and buy with zero reserve requirements."
2. Then name the step past that: "if things really get bad… I believe that the Treasury is going to buy stocks" — index funds, not individual securities.
3. Ground the forecast in precedent rather than assertion. Hong Kong's Monetary Authority bought equities in the late 1990s "and it worked beautifully"; the logic is Bagehot's dictum — in a crisis the state lends at a high rate against good credit — extended from bonds to shares: with the market down 30–40% the earnings yield is very high, so "they can make a killing" (1:07:43).
4. Check your own track record on the method: he called the Fed buying corporate bonds "seven or eight years ago" for "the next crisis, which turned out to be COVID."
5. Then invest the consequence, not the event. Endless games to escape the debt hole (Grant Williams' "project Zimbabwe") mean "there's going to be this persistent bid under precious metals."
6. Watch the offsetting flow that could break the passive bid: Japan repatriating capital (Korea's playbook, "a trillion dollars easily"), with the yen ~50% undervalued on purchasing-power parity, hedge funds heavily short, and a BOJ rate hike as the plausible trigger — a yen rally being "a real nasty hit to the AI trade" per Felder's mirror-image chart (1:09:22).

• Yield-curve twists and reserve-requirement tweaks as the first tells; official equity-purchase precedents; deficit-to-GDP trajectories into a recession; BOJ short-rate moves and yen positioning; PPP gaps (the Big Mac index) at record extremes.