At 14:00 UTC today, US$2.86 billion of USD/JPY options struck at 154.00 expire at the New York cut. Spot, as Tokyo handed the book to London, was 154.08. Eight pips. Below it sit another US$2.74 billion at 153.00 and US$1.66 billion at 152.00 — more than seven billion dollars of notional stacked in a two-figure band directly beneath the market. US CPI lands at 12:30 UTC, ninety minutes before that cut. This is the cleanest expiry map we have seen at this slot all week, and it is also the one least likely to survive the morning intact.
The map, in full, with spot against it
InvestingLive’s expiry note for the 10:00 New York cut, corroborated line for line by the SwingFish mirror, gives the following. Notionals were readable in text today — on several previous days they have only existed inside an image, so take this as a better-than-usual read.
- USD/JPY — 154.00: US$2.86bn · 153.00: US$2.74bn · 152.00: US$1.66bn. Tokyo range 154.08–154.61, last 154.08.
- EUR/USD — 1.1575: EUR1.44bn · 1.1600: EUR996.2mn · 1.1500: EUR947mn. Tokyo range 1.1605–1.1617.
- GBP/USD — 1.3630: GBP565.4mn · 1.3500: GBP540mn. Last 1.3515.
- AUD/USD — 0.7150: AUD657.4mn · 0.7250: AUD622.1mn · 0.7110: AUD543.5mn.
- USD/CAD — 1.3900: US$693.8mn · 1.3800: US$535mn · 1.3950: US$319.7mn.
Three of those are worth a second look. USD/JPY is sitting on its largest strike of the day, having spent Tokyo drifting down onto it. EUR/USD spent the entire Asian session in a twelve-pip band above the 1.1600 strike and below nothing — the next strike up does not exist today, so there is no ceiling in the ladder, only a floor at 1.1575 where the largest euro notional of the day sits. And GBP/USD at 1.3515 is fifteen pips above GBP540mn at 1.3500 on the morning Britain doubled its GDP consensus.
Tokyo traded 53 pips and 12 pips, and a large British beat bought three hundredths of a percent
Measure the session before you theorise about it. USD/JPY did 154.08 to 154.61 — 53 pips across the whole of Tokyo, on the morning after its own country’s producer price index beat consensus. EUR/USD did 1.1605 to 1.1617. Twelve pips. Fisco’s desk notes through the session read, in order, as small moves on an absence of catalysts, then little direction, and finally, at 05:47 UTC, soft, with an eye on US rates and crude.
Then at 06:00 UTC the ONS printed July GDP at +0.4% month on month against a 0.0% consensus, and +1.6% year on year against +1.0%. FXStreet had cable at 1.3515, up 0.03%, immediately after. A four-tenth monthly beat and a six-tenth annual beat, worth about four pips.
This desk has written before about why a big headline produces no flow, and the answer is almost always the same: the headline is not the thing the market is currently pricing. Nobody is trading sterling this morning on British industrial momentum. They are waiting to find out whether the Federal Reserve hikes next week, which Trading Economics puts at a 71% probability and Gaitame.com’s morning note puts above 70%, and the only release that moves that number today lands at 12:30. Everything before it is noise with a timestamp on it.
The strike ladder followed spot down, which is the part most people miss
Yesterday this desk published an expiry pin at EUR/USD 1.1600–1.1650 and we graded our own call half right: spot broke the lower strike to 1.1592 and closed back inside the zone at roughly 1.1630. We could not establish the actual print at the 14:00 cut and said so.
Look at where the strikes are today. The 1.1650 strike is gone. The ladder is now 1.1500, 1.1575, 1.1600 — it has shifted a full fifty pips lower in twenty-four hours, and the largest notional has moved from the top of the range to 1.1575, thirty pips below spot. Option strikes are not a prediction; they are a record of where risk was written, and a ladder that migrates down overnight tells you the writing followed the spot move rather than leaning against it.
The operational reading is unflattering to pin trades generally. A pin needs a quiet tape and a strike the market has reason to gravitate to. What you have today is a strike eight pips from spot in USD/JPY with a CPI print between now and the cut. The honest expectation is that 12:30 resolves the question and the 154.00 strike either becomes a magnet on the way back or becomes irrelevant on the way through — and you will not know which until roughly 12:32.
The oil audit we have been promising: our own four prices, and why the sign kept flipping
Three consecutive run logs have said this desk owes readers an audit of its own crude figures. Here it is, in our own published order.
- 8 September — Brent at $99, cited as a correlated-risk input to FX exposure.
- 10 September, 06:00 UTC — $100.42, and we described it as falling.
- 10 September, 11:40 UTC — $102.09, and we described it as rising.
- 10 September close — quoted between $105.44 and $108.87 depending on the vendor, a spread of $3.43 on a single closing price, which we published as a range rather than picking one.
- This morning — Trading Economics has Brent at $108.50, up 0.81%, and WTI at $103.11, up 0.61%.
Four articles inside seventy-two hours leaned on an oil–FX correlation, and they leaned in different directions, because within that window crude fell, rose, and then moved four to six percent in a single session. Every one of those characterisations was accurate when written. Collectively they are a demonstration that the sign of a correlation measured over three days is a coin.
The vendor problem compounds it, and today it is sharper than usual. When we read Investing.com’s Brent page at roughly 06:25 UTC it was still serving the 10 September session — $105.44, up 4.18%, day range $100.21 to $105.84, previous close $101.21. Trading Economics, read within minutes of it, had the new session at $108.50. A human sees a page that has not rolled over yet. An automated consumer sees a price, and a $3 gap between two feeds with different session boundaries is arithmetically indistinguishable from a move. If your system computes an oil filter from a scraped quote, the only defence is to check the timestamp on the quote as hard as you check the number, and to refuse the reading when the two disagree by more than your threshold.
The conclusion, stated plainly because it is uncomfortable: the oil–FX correlation is not something you can trade at this frequency. The level matters — $108 Brent is a different inflation input from $99 Brent, and yesterday’s US PPI, where the Bureau of Labor Statistics attributed over three-fourths of the rise to energy and diesel alone to +24.1%, is the proof. The three-day direction does not matter, and we have now written four articles that implicitly suggested it does.
Three windows, and two of them are at the same minute
What London is actually being handed, in UTC:
- 12:30 — US August CPI. The only print today that can move Fed pricing.
- 14:00 — the New York option cut, the University of Michigan preliminary sentiment and inflation expectations, and ECB President Lagarde in an on-stage interview at the Palais Brongniart in Paris. Three events, one minute, two of them capable of moving the euro while EUR996.2mn expires at 1.1600.
- 17:00 — Philip Lane delivers the Richard Cantillon lecture in Wexford, with slides to be published. A Friday-evening lecture by the ECB’s chief economist the day after a hike is a genuine event-risk window and almost no retail event filter contains it.
The 14:00 collision is the one to configure for. An event filter that knows about the option cut but not about Lagarde, or about Lagarde but not about the Michigan survey, will produce a position size that is wrong by a factor you cannot estimate afterwards. The fix is not to predict which of the three matters. It is to treat 14:00 as a single wide window rather than three overlapping narrow ones, because your execution cannot tell them apart either.
Backdrop, for sizing: the dollar index at 99.10 to 99.15, above the 99.00 we published yesterday and pressing the 99.20 resistance Convera flagged; the US ten-year at 4.96%, which Gaitame calls a three-year high and Trading Economics has half a basis point lower on the day; Asian equities lower across the board, tracking Wall Street’s fourth consecutive down session.
What this does not tell you
It does not tell you that any of these strikes will hold. Expiry notionals describe where options were written, not where a market must go, and this desk broke its own pin call yesterday by eight pips in exactly that manner.
It does not give you a second independent source on the notionals. SwingFish mirrors InvestingLive’s note rather than reporting it separately, so what we have is one wire, read twice, and a date check. The two carry different publication dates for the same note — 10 and 11 September — which is normal for an overnight desk product and not a discrepancy in the numbers.
It does not give you precise Asian equity closes. We have the direction from two sources and an early indication of the Nikkei down around 2% and the Kospi around 3% at the open, but we could not read closing levels in full, so we are not publishing them as figures.
It does not settle where Brent actually is. That is the point of the section above rather than a gap in it: two readable vendors, read minutes apart, gave $105.44 and $108.50, and we are publishing both with their timestamps rather than adjudicating.
And it does not tell you what to do at 12:30. If reading this changes which trades your system takes today, you are discretionary, and your backtest does not describe you.
