On Friday morning this desk published the 10:00 New York cut option ladder and built an article around it: US$2.86 billion of USD/JPY struck at 154.00, US$2.74 billion at 153.00, US$1.66 billion at 152.00 — more than seven billion dollars stacked in a two-figure band under a spot rate of 154.08. Those numbers came from InvestingLive, posted by Eamonn Sheridan at 22:19 GMT on 10 September. The same publisher ran a second 11 September ladder at 05:01 GMT, under a different byline, for the same cut on the same day. It says US$3.06 billion at 154.00, US$2.88 billion at 153.00, and it does not carry a 152.00 line at all. Neither page is wrong. That is the problem.
The two ladders, side by side
Same date. Same 10:00 New York cut. Same publisher. Seven hours apart.
- USD/JPY 154.00 — US$2.86bn (22:19 GMT read) against US$3.06bn (05:01 GMT read)
- USD/JPY 153.00 — US$2.74bn against US$2.88bn
- USD/JPY 152.00 — US$1.66bn against no line
- EUR/USD 1.1575 — EUR1.44bn against EUR1.66bn
- EUR/USD 1.1500 — EUR947mn against EUR1.18bn
- EUR/USD 1.1600 — EUR996.2mn against no line
- GBP/USD 1.3500 — GBP540mn against GBP653.00mn
- AUD/USD 0.7250 — AUD622.1mn against AUD837.24mn
- USD/CAD 1.3900 — US$693.8mn against US$789.06mn
The later read also carries three pairs the earlier one does not: USD/CHF at 0.8150 and 0.8065, NZD/USD at 0.5850, and EUR/GBP at 0.8625.
The differences are not rounding
On the headline USD/JPY strike the gap is US$200 million, or 7.0%. On EUR/USD 1.1500 it is EUR233 million, close to 25%. On AUD/USD 0.7250 it is AUD215 million, nearly 35%. These are not two people transcribing the same sheet with different decimal habits, although the decimal habits differ too — the earlier list rounds to one decimal place, the later one reports GBP585.15mn and US$789.06mn.
The tempting explanation is accumulation: more expiries get struck overnight, so the later snapshot is larger. That explanation dies on the 152.00 line. A US$1.66 billion strike does not shrink to nothing in seven hours, and the later list is not filtering by size either, because it happily prints a NZ$308.30 million NZD/USD line that is a fifth of what it dropped. What you are looking at is two desks exercising judgement about what belongs in a list, from data that is itself a survey rather than a register.
What we published, and what it actually was
Our 11 September article led on the eight-pip gap between the 154.00 strike and spot at 154.08, and totalled the three USD/JPY strikes at US$7.26 billion. On the other read of the same cut there is no three-strike total to take — the two printed strikes come to US$5.94 billion, and the band we described as stretching down to 152.00 does not exist.
We are not retracting the article. The strike was where we said it was, spot was where we said it was, and the point that seven billion dollars of notional sat directly beneath the market was a fair reading of the source we had. What we should have written, and did not, is that the figure was one desk’s count at one moment. We gave it two decimal places of authority it never had. This desk has now audited its own Brent prices, its own consensus adjudication and its own expiry ladder in four days; the pattern is that the numbers we treat as facts are usually somebody’s measurement.
There is no ladder for today, and London is about to open
At 06:20 UTC this morning the publisher’s orders index carried no option-expiry article for 14 September. The most recent entries were the two 11 September pages described above. That is a checked negative, not an assumption: we read the index and the newest item on it was three days old.
So anyone who opens the London session with an expiry map in front of them is opening it with Friday’s map. Those options have already expired. Whatever is rolling off at 14:00 UTC today is unpublished as this article goes out, and the ladder that does eventually appear will be one desk’s count of it — possibly, on the evidence above, one of two.
Tokyo gave the whole figure back
The session that just handed London the book ran 153.37 to 154.14 on USD/JPY, per Fisco’s Tokyo desk notes at 13:43 and 14:04 JST — 04:43 and 05:04 UTC. Minkabu had the pair at 154.03 at 12:00 JST against Mitsubishi UFJ’s official mid-rate of 153.65. EUR/USD held 1.1565 to 1.1596; EUR/JPY 177.85 to 178.46.
Put differently: Friday’s yen rally, which took the pair from roughly 154.48 down to 153.24, has been most of the way unwound inside one thin Monday session, on nothing more than a Federal Reserve rate-hike probability that both Fisco and Gaitame now put in the 80s. The 154.00 band that mattered on Friday matters again this afternoon, and the market is back inside it with no published map. Gaitame’s Monday analysis frames the week as 151.50 to 155.00 with the prior week’s 152.887 low as support and 154.65 to 155.00 as the ceiling, on a 14-day RSI of 29.1 — a number StoneX independently read at 29 on Saturday.
The trading conclusion is not directional. It is that the expiry ladder is a sizing input with error bars of several hundred million dollars per strike, and it belongs in your process the way a weather forecast does, not the way a settlement price does. If a strike is the reason you are smaller today, fine. If a strike is the reason you are larger, you are levering a survey.
What this does not tell you
We do not know which of the two ladders is closer to the true book, and we have no way to find out — there is no public register of FX option strikes, which is precisely why a vendor survey exists in the first place. We do not know whether the discrepancy is a data-provider difference, a cut-off-time difference, or an editorial one. We did not contact the publisher. We do not know what expires at today’s 14:00 UTC cut, because nobody has published it. We have not established what happened at Friday’s 14:00 cut either, which is the second consecutive run this desk has carried that gap; it will stay open until somebody publishes where spot printed at the fix. And the Tokyo levels above come from Japanese desk notes, not from a consolidated tape, so treat the 153.37 low and the 154.14 high as one vendor’s range.
