The repeatable method: find the accessible-inventory floor, then read the paper market's positioning distortions that keep price from reflecting it — until it snaps.
1. Find the "empty is not zero" floor before you trust an inventory number
The repeatable method
- Take the headline commercial inventory (e.g. ~460mmbbls US crude ex-SPR) and subtract the portion that is not drawable: linefill to keep pipelines moving (~150mm), tank bottoms, and minimum refiner working stocks.
- Set the true operational floor (~350-370mm), not zero, as the constraint. The distance from current level to that floor — not to zero — is the real runway.
- Overlay the draw rate (tanker counts signalling the Gulf, refinery run rates into driving season) to date when you hit the floor. "The math is the math."
Here: ~460mm now, floor ~350-370mm, refiners ramping → sub-400mm "by sometime in July very easily… the margin is days/weeks, not months."
Watch for
- Draws accelerating toward the operational floor; tankers "ballasting" (arriving empty = taking, not delivering); SPR flow slowing so commercial stocks do the heavy lifting.
2. Refuse to extrapolate a buffer-drawdown linearly
The repeatable method
- Recognize the human bias: we price a slow, steady draw as if it continues linearly, when a depleting buffer produces a non-linear/exponential endgame (the stadium at 3% full is minutes from disaster).
- Separate inevitable from imminent — the outcome can be baked in by the physics while the timing still surprises; don't let a quiet price talk you out of the setup.
- Size for a discontinuous repricing (outright + convex calls) rather than a smooth grind, because the reaction comes "gradually, then suddenly."
Here: "price expresses itself non-linearly, the market becomes discontinuous, and shortages emerge seemingly out of nowhere" — awaiting the "Tom Hanks Has Covid" common-knowledge flip.
Watch for
- A market "staring at the sky" while the buffer nears its floor; the recency/anchoring bias of "where's the crisis?"; the crowd mistaking inevitable for not-imminent.
3. Map weak hands via broken levered/inverse ETF flows AND their open-interest footprint
The repeatable method
- Track assets/flows in the -2×/2× daily products; a structurally-doomed vehicle attracting a wave of assets is capitulative crowding on the wrong side (vol-drag + backwardation roll guarantee decay).
- Convert the fund's holdings into contracts and compare to open interest in each expiry — especially the illiquid back of the strip that commercial hedgers have abandoned.
- When a single retail product owns a material share of a thin contract's OI, treat it as a real market-moving force and a coiled contrarian spring.
Here: SCO $100mn→$1.1bn despite a >50% price drop, short ~10k in each of Aug26/Dec26/Jun27 = ~8% of Jun27 WTI open interest; UCO roundtripped to $400m; USO below $1.8bn.
Watch for
- Ballooning assets in the structurally-losing vehicle; its contract count as a % of a thin back-month OI; commercial hedgers absent from the far curve making it fragile.
4. Read a futures→options migration as a degrossing/illiquidity tell
The repeatable method
- Decompose aggregate non-commercial length into its futures vs options components over time.
- When total length is near a prior high but the futures share has shrunk (replaced by options), read it as VAR-constrained players unable to carry large futures notionals — a "stock-replacement via calls" analog and a sign of forced degrossing after a vol shock.
- Expect thinner, more gap-prone price action (and negative-gamma dealer hedging) as a consequence.
Here: length near last July's high but "the proportion of futures is far less, having been replaced with options" as implied vol relaxed and VAR limits bit.
Watch for
- Rising options share of positioning; desk firings/restructurings after a vol shock; illiquidity + tweet-driven whipsaw feeding on itself.
5. Trace a policy-driven hedging flow to explain "why didn't it move?"
The repeatable method
- When price action defies the fundamentals, look for a mechanical hedging flow from a recent policy action (here an SPR release structured as a loan — return 24% more barrels in 2Q27).
- Trace the required hedge: awardees are long prompt / short a 1-yr-forward commitment, so they must sell front months and buy the out-year — pressuring the prompt and spreads.
- Factor in the timing frictions (a multi-day allocation lag leaving traders "flying blind") that let headlines/interventions land while the hedge can't be executed — then fade the distortion once it washes through.
Here: the 92.5mmbbl SPR loan-swap, a rumored ~3-day allocation lag, and overnight tweets/BoJ-MoF rumors "vaporized the profitability" and capped crude late-week.
Watch for
- A structured release with a forward return obligation; a lag before awards; spread/prompt collapse into an illiquid overnight session; the flow "washing through" as the fade window.
6. Use the Brent-vs-WTI managed-money split to price jurisdiction risk
The repeatable method
- Compare net managed-money length in Brent vs WTI. A persistent skew to Brent says specs are pricing a US-specific policy risk (an export ban) rather than a pure supply/demand view.
- Express the bullish thesis through the instrument that dodges the jurisdiction risk (Brent/BNO) rather than the one exposed to a domestic ban (WTI/USO).
Here: "the net long among managed money in Brent dwarfs their length in WTI… participants know the ultimate risk is jurisdiction, and they are voting with their feet" — his own book is "mostly reflected through Brent (BNO)."
Watch for
- A widening Brent-over-WTI positioning skew; export-ban chatter (and official denials — "nothing is official until it's officially denied"); producers/CoT confirming the tilt.