USD/JPY traded 155.00 overnight, the first time it has touched that figure since 7 September, and was back at 153.99 before Tokyo opened. A hundred and one pips out and a hundred and one pips back, inside one New York session, on the eve of the Federal Reserve decision and three days before the Bank of Japan. Not one link in the chain that did it was Japanese. It was a drone attack on a Saudi pipeline, a merchant ship struck in the Strait of Hormuz, the US ten-year yield crossing 5% for the first time since 2023, and then four social-media posts that unwound the whole thing.
The chain, in order, with nothing Japanese in it
Over the weekend a merchant vessel was struck in the Strait of Hormuz, with one death and three injuries reported by Iranian authorities. Saudi Arabia then shut its East-West pipeline — the seven-million-barrel-a-day line that carries crude across the kingdom to Yanbu on the Red Sea, and the one route that bypasses Hormuz entirely — after drone attacks. Losing the alternative at the same moment the strait itself became dangerous is the part that priced.
Brent crossed $108 a barrel for October and November delivery, up more than 3%. WTI for October rose more than 3% to around $103. The US ten-year yield went through 5%, printing 5.006% at 14:40 UTC, its first breach since 2023; the two-year reached 4.666%, the highest since July 2024. USD/JPY went with the yield to 155.00.
Then, between 15:05 and 16:31 UTC, the US President posted four times — on a Russia-Ukraine energy agreement, on an Iran deal, on crude flowing through Hormuz again, and on prices about to collapse. WTI unwound to close near $98.00. The ten-year came back to the 4.93% area. USD/JPY was at 153.99.
Tokyo takes the book at 154.36 on one vendor’s close and 154.39 on another’s. Japan contributed nothing to any of it.
Three publishers, three different highs for the same barrel
Here is the uncomfortable part of writing the paragraph above. This desk read three accounts of what WTI actually did, and they are four dollars apart on the session high.
One has WTI around $103 at the peak, up more than 3%. One has it in the mid-$104s before retreating. The third, writing after the close, gives a session high just above $100.50, a session low just beneath $97.00 and a close near $98.00 — which would mean the barrel never reached either of the other two figures at all. All three are describing 14 September. Two of them are sources this desk uses on most runs.
We are not going to pick one. The archive already carries an oil audit from 11 September where two vendors were three dollars apart on Brent inside a single morning, and this is the same failure a week later on a wider spread. The honest statement is the range: WTI’s high on 14 September is reported between roughly $100.50 and $104.50 depending on who you read, and it closed near $98.00 on the one reader that gave a close.
If you run anything that keys off an oil level — a correlation filter, a risk-off flag, a volatility regime switch — that spread is your error bar, and it is wider than most of the thresholds such a filter uses.
The Fed probability has the same problem, a day before the Fed
The Federal Open Market Committee convenes today and announces on Wednesday at 18:00 UTC. A 25 basis point hike is the consensus. How well priced it is depends entirely on which number you read.
One US market wrap puts it at roughly 85% after Friday’s hot CPI. One FX wire quotes 93%, attributed to a named terminal. One Japanese morning note says the probability rose above 90% during the session and then moderated, and a second note from the same publisher calls the hike nearly certain with a 3.75–4.00% target range. That last figure is single-sourced and we are labelling it as such.
Eight points of spread on a binary event a day and a half away is not a rounding difference. At 85% the surprise case is worth positioning against; at 93% it is close to noise. If your sizing rule reads a probability off a feed, note which feed, because this desk cannot get three of them to agree the day before the meeting.
The flat book got its first test and did nothing
On Sunday this desk published the Commitments of Traders report showing speculators had flipped 103,023 contracts in a single reporting week, from a net yen short of 92,227 to a net long of 10,796. We argued that the crowded side of the trade was gone, and that a fully priced Bank of Japan hike lands differently into a flat book than into a short one.
Last night was the first test of that and it is a clean one, because the impulse was external. A hundred-and-one-pip move up on a genuine supply shock and a 5% ten-year, and it round-tripped completely within hours. No follow-through, no acceleration through the figure, no stops cascading above 155.00. That is roughly what a flat book looks like: price goes where the macro pushes it and comes back when the push stops, because there is no position underneath it that has to be liquidated.
Remember what that report is, though. It is measured at the Tuesday close and published Friday, so the position we are describing is now six sessions stale and predates everything in this article. It is the last measurement available, not the current book.
What Tokyo is handing London
The forecast range from the morning note this desk reads first is 153.500 to 154.900, with 155.00 as resistance — the level that has now rejected twice in nine days — and support given as 153.00. A second Tokyo desk puts the same resistance at 155.00 but the support two figures lower, at 152.00, with 150.00 in view if it goes. Publish both; they are not the same map, and the gap between them is where a stop placed off one vendor’s chart sits inside the other vendor’s noise.
The week from here: the FOMC decides Wednesday at 18:00 UTC with the Summary of Economic Projections, Japan’s national CPI lands Thursday at 23:30 UTC in the same minute as a Reserve Bank of Australia speech, and the Bank of Japan decides Friday at about 03:00 UTC with a 25 basis point hike to 1.25% widely expected and Governor Ueda speaking about three and a half hours later. There is also a 20-year Treasury auction at 17:00 UTC today, which settles on Bank of Japan day.
The asymmetry worth naming: everything on that list is a Japanese or American policy event, and last night the yen moved a hundred pips on neither.
What this does not tell you
It does not tell you what oil did. Three publishers give three different highs and we have published the range rather than a number, which means any inference in this article that depends on the precise level of crude is weaker than it reads.
It does not tell you that the round trip proves the book is flat. A hundred-pip reversal is consistent with light positioning and also consistent with a headline being fully retracted by its own author within ninety minutes, which is what happened. We are offering it as a first observation, not a confirmation.
And it does not tell you where USD/JPY goes into the two decisions. The session ranges and vendor levels above are the map two Tokyo desks are trading from this morning. They are not a forecast, and on a week with the Fed and the Bank of Japan thirty-three hours apart, the honest position is that the range is the information and the direction is not available yet.
