Actionable insights — The Most Convex Trade of My Career

1. Triangulate forward inventories to front-run the draw

1. Don't wait for the weekly DOE print — build a forward picture of inventories from independent data: producer output data, announced refinery runs, ship/tanker signaling (boats heading to the Gulf of America), and import/export flows.
2. Net the supply against the demand side to project where stockpiles land weeks ahead — the draw is visible before it shows up in the headline number.
3. Lean on the rare desks that actually do this work (energy talent is thin after 15 years as a 3-4% index backwater; he names HFI Research, Energy Aspects, Sparta, Kayrros as the sources he cross-checks).

• Refinery runs rising, imports falling, exports rising simultaneously (the three accelerants); tanker counts to the US Gulf; the weekly DOE confirming the forward model.

2. Separate inventory destocking from real demand destruction

1. When you see declining imports, don't assume demand is falling — falling imports can simply mean a country is drawing down stock (destocking) rather than consuming less.
2. Cross-check consumption with hard activity data (connected-vehicle / transaction counts), not just headline volumes; global import declines are backward-looking and partial.
3. Deflate any spending-based demand read by price. A "-3% y/y" gasoline figure built from credit-card spend overstates the drop: when prices spike, consumers buy at cheaper stations and lower grades, and more pay cash inside to dodge the >3% card fee — all of which depresses card spend without depressing real demand.

• Import troughs that turn up; transaction counts diverging from dollar spend; analysts citing card-spend without price-deflating it.

3. Use operational-minimum tank bottoms as the shortage trigger

1. Know the floor for each inventory: physical minimums you can't draw below — linefill, minimum refinery stocks, tank deadstock. (US commercial crude ~370mmbbls; Cushing ~17-20mmbbls; gasoline ~200mmbbls; distillates ~100-110mmbbls.)
2. Project the draw path toward those floors; the moment stocks approach the operational minimum without demand destruction is when outright shortages / fueling lines become possible — a step-change, not a gradual tightening.
3. Respect the sequence: the inventory nearest its floor breaks first. Gasoline (esp. PADD1 / East Coast) leads crude by weeks.

• Cushing nearing 17-20mmbbls; PADD1 gasoline nearing operational minimum; SPR discharge rates degrading as the caverns deplete (the DOE's own T4 RFP warning).

4. Spot the futures→options→dealer-short-gamma squeeze setup

1. After a VaR shock, watch whether participants degross and shift from futures into options (to cap stop-out risk) — that thins prompt-futures liquidity and pushes the risk onto dealers.
2. If that flow leaves market makers short calls, and volatility is low, you have squeeze fuel: a rally forces dealers to buy futures into an illiquid market to hedge rising delta and vanna, accelerating the move (the same mechanic as the semiconductor delta/vanna squeeze).
3. Confirm the powder is dry by checking that outright speculative net-long is low — if the elevated net-long is mostly options exposure, a pop on dealer hedging is more violent, not less.

• Thinning prompt-futures liquidity; rising option open interest vs futures; low realized vol; spec net-long near multi-year lows while options length stays elevated.

5. Fade crowded retail positioning in a broken inverse ETF

1. Find the leveraged/inverse ETF the retail crowd is using to express the opposite of your view, and track its asset/flow trajectory, not just its price.
2. Read persistent inflows into a structurally decaying product (a -2× daily inverse fund) despite heavy losses as a one-sided, over-positioned sentiment extreme — contrarian fuel for the move they're fighting.
3. Cross-check against the "smart" side: who's on the other side of the trade in the CoT data?

• Inflows into an inverse/levered ETF that keeps losing; CoT producer/merchant net-long rising while "other reportable" cuts length; commercials (Exxon/Chevron commentary) buying what synthetics sell.

6. Separate "inevitable" from "imminent" — then size for the convergence

1. Distinguish a thesis that is baked-in / inevitable (the math can't avoid it) from one that is imminent (about to be marked). They're often not the same — the gap is the "frustration window" where you can be right and still marked against you (the Big Short analogy).
2. Express an inevitable-but-not-yet-imminent view with convex, time-tolerant structures — long-dated calls with positive carry (roll yield in backwardation) — so the wait costs little and the payoff is outsized.
3. Watch for the catalyst that collapses the gap (a non-linear "oh sh*t" awareness moment), and re-size as "weeks/months" compresses to "days/weeks."

• A single visible catalyst ("Tom Hanks has Covid" moment); the draw appearing in the weekly print; backwardation deepening so the call structure earns carry while you wait.

7. Discount the headline driver everyone is trading

1. When a geopolitical headline (here, Iran / the Strait of Hormuz) dominates the tape, test whether your thesis actually needs it — run the numbers with the headline risk fully removed.
2. If the conclusion holds either way, treat the headline as noise and ignore the crowd waiting for a "deal" / dip — "when everyone is waiting for something, the market won't let them in."

• Consensus positioned for a headline-driven dip that never comes; the math being unchanged in your "what if the risk resolves benignly?" scenario.