Ship trackers counted five or six vessels a day through the Strait of Hormuz at the start of this month. The US Treasury Secretary says at least ten million barrels a day are moving, and fifteen to seventeen million on a good day. The President says thirty ships a night, forty on Tuesday. These are not rounding errors between two estimates of the same thing — they are different quantities, published by different institutions, and the gap between the smallest and the largest is a factor of six. Every dollar-inflation premium currently sitting in the long end of the US curve is priced off one of those numbers. The 30-year bond reopening at 17:00 UTC today will clear against whichever one the buyer believes.
Six ships, or forty. Both figures are on the record.
Al Jazeera assembled the competing counts on 3 September, and the spread is the story. Kpler put transits at six vessels on the Wednesday, eleven on the Tuesday, five on the Monday. Lloyd’s List Intelligence averaged twelve transits a day over 26 August to 1 September, and about fourteen non-Iranian-linked ships a day over 17–23 August. PortWatch has averaged seven vessels a day across the whole six-month conflict. Against that, Treasury Secretary Bessent’s “at least ten million barrels” and the White House’s thirty-to-forty ships a night. The pre-war baseline everyone agrees on is roughly one hundred ships and twenty million barrels a day.
We are not in a position to tell you who is right, and we are not going to pretend otherwise. What we can tell you is that the counting methodologies differ — whether Iranian-linked tonnage is included, whether partially laden vessels count, whether a transit is a crossing or a port call — and that a market cannot arbitrage a definition. If you are running an oil-sensitive currency book, the input variable is not measured. It is estimated, by five bodies, with a five-fold dispersion, and the estimate you happen to read is a function of which outlet you opened.
The barrel counts disagree with each other inside the same outlet, from the same tracker
This is the part that should bother a systematic trader more than the political dispute. On 20 August, Al Jazeera published Kpler’s series: 2.3 million barrels a day from April to 17 June; 6.1 million barrels a day through the US–Iran memorandum period from 17 June to 18 August, which is 374 million barrels over sixty days; and a 2025 pre-war baseline of about fifteen million barrels a day. On 27 August, the same outlet published a pre-war baseline of about seventeen million barrels a day and a current figure of roughly nine million.
Fifteen against seventeen for the same historical baseline, one week apart. Six-point-one against nine for roughly the same present. Neither piece is wrong; they are almost certainly measuring crude only versus crude plus condensate, or the strait itself versus the wider Gulf, over slightly different windows. That is exactly the point. A twenty percent disagreement about the denominator is invisible in prose and fatal in a ratio. If your model normalises current flow against a pre-war level, the choice between fifteen and seventeen moves your disruption estimate by four percentage points before you have observed anything at all.
Transits fell ninety-five percent and barrels fell about half — that is not a contradiction, it is the answer
Two separate windows produce the same headline collapse in ship counts. Kpler’s ninety-five percent figure, reported via Gulf News, rests on 116 crossings between 1 and 19 March against a normal-times baseline of about 120 transits a day. Al Jazeera’s 27 August piece covers 15 July to 23 August and gets there differently: more than a hundred vessels a day before the war, just five after, an almost ninety-five percent decrease. Different windows, different arithmetic, same answer on transits.
And yet the volume series in the same reporting only falls by roughly half. Both cannot describe the same physical event unless the barrels per remaining transit went up sharply and some of the oil stopped using the water. The port-call data says exactly that: Kuwait down eighty-six percent, the UAE down sixty-nine, Qatar, Iraq and Bahrain down sixty-six to sixty-eight — and Saudi Arabia down fifteen. Saudi Arabia has pipeline capacity to the Red Sea. The others do not.
So “Hormuz traffic” and “Gulf crude supply” are two different trades wearing one headline. A vessel-count feed will tell you the chokepoint is closed. A barrel feed will tell you supply is impaired but functioning. If you have a news filter keyed to the word Hormuz, it is firing on both and treating them as confirmation of each other. They are not confirming each other. They are measuring different things, and one of them — the barrel series — is the one oil futures actually trade.
Which brings you to 16:00 and 17:00 UTC today
Brent is $102.09, up 0.87% on the day; WTI is $97.51, up 1.52%, per Trading Economics this morning. InvestingLive published a piece at 11:01 UTC arguing that oil is the thing driving the rise in long-term yields and that everything else is noise, with the 10-year above 4.8% and dated Brent quoted at $114.26 on 9 September. Note that last number against the futures price: a twelve-dollar spread between the physical assessment and the screen is itself a measurement gap, and it is the physical one that a refinery pays.
Today you get exactly one audited oil number, and it is not the one that matters. The EIA Weekly Petroleum Status Report lands at 16:00 UTC — noon Eastern, delayed from its usual Wednesday slot by Labor Day, confirmed on the EIA’s own holiday schedule page and by the trade press. That number counts barrels sitting in tanks in Cushing, Oklahoma. It does not count barrels leaving the Gulf. Then at 17:00 UTC the Treasury reopens the 30-year against an August stop of 5.216%, into a long end that InvestingLive says is being driven by oil.
The honest description of today is that the market will price a thirty-year US inflation premium at 17:00 using an oil-supply input that five institutions measure five different ways, having received at 16:00 a precise weekly figure about a different quantity entirely. That is not a reason to have a view. It is a reason to size as though your input has a wide error bar, because it does, and the error bar is published.
What this does not tell you
We do not know which transit count is correct, and nothing here should be read as saying the official figures are wrong and the trackers right, or the reverse. We have not seen Kpler’s, Lloyd’s List’s or PortWatch’s underlying data; we have read reporting that cites them. Bloomberg has carried Hormuz flow figures we could not read behind its paywall and have therefore not used or cited. A CNN piece on the same measurement problem is disallowed to us by robots and was not read, so it is not in the sources below.
We also cannot tell you the size of today’s 30-year reopening. Treasury’s own tentative auction schedule carries dates but no offering amounts, and the one figure we found yesterday was single-sourced; we flagged it as indicative then and we are not repeating it as fact now. And nothing above forecasts the oil price. The argument is about the dispersion of the input, not the direction of the output.
Related
- FX Events — the dated calendar, in UTC
- Signals
- EA Presets
