Actionable insights — Watch China and the Crude-to-Diesel Margin

0:00 1. Back out the implied crude price from the product price — the gap tells you which leg is dislocated

1. Take the refined product that is making the headline (diesel, gasoline, jet) at its actual price.
2. Apply a normal refining margin to work backwards to the crude price that product level would imply.
3. Compare that implied crude price to spot. If they agree, the shock is in the barrel — a supply/demand story on crude. If the implied price sits far above spot, the barrel is fine and the entire dislocation lives in the conversion margin (the crack).
4. Trade the leg the diagnostic points at: the spread, or the barrel that a mean-reverting spread will bid — not the level that is already making the headline.

• Product prices at records while the crude benchmark sits mid-range — that divergence, not either price alone, is the entry signal. Re-run it any time a refined product, a metal cathode or a processed intermediate makes a record without its feedstock following.

0:56 2. Watch the marginal processor, not the marginal consumer — the same actor moves both legs

1. When a spread blows out, ask who owns the conversion capacity between the two legs — the refiner, the smelter, the processor — not who buys the finished product.
2. Check whether the dominant processor has stepped out. If it has, its absence pushes the input price down (it stopped buying) and the output price up (it stopped producing) simultaneously — one behaviour, two opposite price moves, and the spread is the residue.
3. Expect the same actor to close the gap: whoever owns the capacity is the one who can capture the margin, and capturing it is what destroys it.
4. Therefore position for the input to be bid as capacity restarts, rather than for the output price to keep running.

• Chinese refinery run rates and teapot utilisation; and the general form — in any processed commodity, identify who holds the processing share (China in copper, aluminium, steel, critical minerals) before attributing a spread to end demand.

1:14 3. Use the processor's local futures price as the early tell that it is coming back

1. Find the exchange the dominant processor actually buys on (Shanghai for China, not just Brent/WTI).
2. Track that local contract against the international benchmark. A local premium is the processor bidding for feedstock before the restart shows in run-rate or trade data.
3. Read the size of the margin it is chasing as the urgency measure — the bigger the crack, the faster and harder the restart.
4. Front-run the restart in the input leg and expect the spread to normalise over weeks, not quarters.

• Shanghai crude futures vs Brent/WTI; the diesel crack narrowing off its record; Chinese product export quotas and refinery restarts confirming the tell.

2:19 4. A cartel without spare capacity is not a cartel — price the group by its physical optionality, not its communiqués

1. Stop reading producer-group meetings as the source of market power. Ask one question instead: does the group hold spare capacity it can actually deliver?
2. If not, treat its quota decisions as non-events — by definition it cannot punish a defector or discipline a price, so headlines about output being "kept unchanged" carry no information.
3. Trace what would have to be true to restore the power: for spare capacity to matter it must be deliverable, so audit the export routes — the physical choke points between the barrels and the buyer.
4. Score whether those constraints are improving or worsening versus a few months ago, and put a calendar on the fix. A constraint that is growing means the group's influence is still declining, whatever it announces.

• OPEC spare-capacity estimates (not quota headlines); Hormuz, Bab el Mandeb and Suez transit volumes and insurance rates; the East-West pipeline's throughput as the bypass of record.

3:42 5. Separate the two clocks — the fix that takes months from the fix that takes years

1. For each proposed remedy to a supply constraint, date it honestly: restarting idle capacity is a months problem; building new physical infrastructure is a multi-year problem.
2. Start the months clock only from the day the underlying conflict is resolved, not from today — "bare minimum if we solved it tomorrow" is the correct framing, and nothing has been solved.
3. Do not let a credible long-dated fix compress a near-dated position. A route that arrives in 2027 does nothing for the next two quarters of pricing.
4. Expect the eventual release to be fast, not gradual, and size for it: once capacity is available the producers "want to take this up as fast as they possibly can."

• FIDs and construction starts on Gulf bypass pipelines (the 2027 clock); ceasefire/de-escalation headlines as the day the 6-month clock actually starts; and the near-term weeks clock on Chinese refinery restarts.