From the archive · Monday, August 17, 2026
Peace made a new high and passage a new low over a weekend with no cash session — and the storage band reversed
The strait stopped being a news story and became a storage story
Chokepoint priced as a condition, not an event
There was no cash session to express a view over the weekend, so the weekend expressed it somewhere else. That turns out to be the most useful thing about it: with the exchanges shut, what moved was our probability layer, and what it did was reorganise the entire Hormuz question.
It moved in two directions at once, which is the part worth sitting with. The ceasefire market went to the strongest reading it has ever carried. Simultaneously, two dated event markets — separate contracts, on named days — settled at certainty for successful strikes on shipping on 13 and 14 August. A brief that treated peace and passage as one trade would have to call one of those marks wrong. This brief has argued since 9 August that they are two trades, and over the weekend the Iranian foreign minister said the same thing in public: the Oman negotiations and the reopening of the strait are separate issues, with reopening tied to American compliance with a June agreement rather than to a truce. Both 15 August deadlines expired without a signature and settled at effectively zero.
The genuinely new development is not that separation, which is by now well established. It is that the strait moved from the diplomatic column into the physical one. The market on American crude reserves falling to 285 million barrels by month-end had ticked DOWN on Friday, and this brief published that tick on Saturday as evidence against its own reading. Across two weekend marks it reversed to the highest level it has carried, the largest single move anywhere in the complex, alongside weekend reporting that American reserve levels are at a forty-year low and a serious question in the financial press about how many months of conflict global stocks can absorb. The numbers, and the five corroborating access markets, are in WHAT WE SEE.
The counter-arguments get the same treatment. If this were a demand story, copper would have led it; copper fell across the six sessions in which the barrel rose, and the American consumer contracted. If it were a war-premium re-rating, the diplomatic markets would have fallen with the physical ones; they went the other way. And the strongest case against us is a fact, not an argument: the last reported weekly American crude stock change was a large build. We have named Wednesday's release as the test, because a view that cannot be checked on a date is not research.
What follows from all this is less about the barrel than about which instruments can carry it. An event has an expiry and an option can hold it. A partially reopened waterway under a toll is a condition, and conditions live in freight, insurance, refining margins and term premium — which is where the risk has been migrating while equity volatility went on getting cheaper. The same reasoning applies to the AI build, now financed rather than funded: when the marginal dollar is borrowed, the binding constraint is a credit spread, and the index inherits it through its largest names.
Methodology footnote. Probability bands are read from the dated day-by-day series only; the stored previous-value column is a longer lookback and would manufacture single-session collapses that did not happen. Bands whose deadline has passed are labelled settled and are never quoted as current. Positioning figures are the maximum-exposure contract per commodity in the weekly report and cover one reported category of trader rather than all buyers. The cash tape throughout is Friday's settled close, the last session to have settled; the weekend marks are from markets that trade continuously.