Two publishers are carrying the same FX option expiry board for the New York cut, and they have put two different dates on it. One wrote it on 30 October and labelled it 2 November. The other published it this morning at 09:20 UTC and labelled it 3 November, in a headline that also calls it “today’s” cut. Five of the seven lines are identical to the notional. The two that are not grew, which is what a live board does. Somewhere in that disagreement sits $2.7bn of dollar–yen optionality struck either side of spot, and the question of whether it dies at 15:00 UTC today or at 15:00 UTC on election day is not settled by anything either page says.
The same board, published three days apart, with two different dates on it
A syndicated options overview dated 30 October 2026, 12:12, carries two boards: one for that day and one headed “Next Business Day, 2 November”. Its 2 November board reads EUR/USD 1.1150 at EUR 2.7bn, 1.1175 at EUR 660mln, 1.1200 at EUR 901mln and 1.1250 at EUR 654mln; USD/JPY 157.00 at $1.5bn and 158.00 at $1.2bn; USD/CHF 0.8300 at $952mln. Nothing in the other pairs.
A research desk’s expiry page published this morning, 2 November, at 09:20 UTC — and titled “Notable FX Option Expiries for todays NY cut: 3rd November 2026” — reads EUR/USD 1.1150 at EUR 3.2bn, 1.1175 at EUR 978mln, 1.1200 at EUR 901mln, 1.1250 at EUR 654mln; USD/JPY 157.00 at $1.5bn, 158.00 at $1.2bn; USD/CHF 0.8300 at $952mln; USD/CAD 1.3975 at $550mln.
Put them side by side. USD/JPY 157.00 at $1.5bn, USD/JPY 158.00 at $1.2bn, EUR/USD 1.1200 at EUR 901mln, EUR/USD 1.1250 at EUR 654mln and USD/CHF 0.8300 at $952mln are identical across both, to the last digit of the notional. EUR/USD 1.1150 went from EUR 2.7bn to EUR 3.2bn and 1.1175 from EUR 660mln to EUR 978mln, so EUR 818mln was added in the two strikes that were already the deepest. USD/CAD 1.3975 appears on the newer page and not the older one.
Five identical lines is not a coincidence between two independent boards for two different days. This is one board, observed twice, and the increments are new options written between Friday and Monday. We are not going to tell you what the people writing EUR 818mln of strikes below a rising market were doing, because we do not know, and the honest version of that sentence has no verb in it.
The date field is wrong in both directions, and we can show it
Before deciding which label is right, it is worth knowing how reliable the label is. We pulled the same publisher’s expiry page for 13th October 2026 and read its own byline. The page titled “Notable FX Option Expiries for today’s NY cut: 13th October 2026” was published on 14 October 2026 at 06:20 UTC. Its title date is one day earlier than its publication date.
The page titled 3rd November 2026 was published on 2 November at 09:20 UTC. Its title date is one day later than its publication date.
So the field is not a fixed offset from publication. It is one day early in one instance and one day late in the other, and both headlines contain the word “today’s”, which cannot be true of both. Neither page states a cut time at all. That is the whole finding about the date: it is not a date you can lean on, in either direction, and two observations are enough to establish that without being enough to establish a rule.
We have been here before with this class of problem, and the lesson has not changed. A vendor’s numbers and a vendor’s labels are separately reliable. The strikes on both of these pages cross-check perfectly. The date on top of them does not cross-check with the publisher’s own clock.
Which day it is, and why that matters less than you would expect
The weight of evidence is on today. A board written on 30 October and labelled 2 November carries these strikes; a page published this morning carries them again with two of them grown; and the one field that says 3 November is the field we have just shown to be wrong in both directions. We put it at today, we will not assert it, and if you are sizing off it you should know that the second-best reading has it twenty-four hours out.
Here is the part that makes the ambiguity survivable. Both candidate days have a United States data cluster in the same minute as the cut. Today at 15:00 UTC a calendar we read lists the October ISM manufacturing survey, three of its sub-indexes and September construction spending, all at once, with the 10am New York cut now at 15:00 as well. Tomorrow at 15:00 UTC the same calendar lists September JOLTS job openings, factory orders, factory orders excluding transportation, job quits and an economic optimism index — and tomorrow is the midterm elections.
So the expiry lands on top of a release cluster whichever day it is. The sizing question — do not be carrying maximum exposure through a 15:00 UTC cut this week — has the same answer under both readings. The directional question does not, and that is exactly the question we are refusing.
The yen board is live and the euro board is spent
The two boards have opposite shapes and only one of them can do anything to you.
Dollar–yen has $1.5bn at 157.00 and $1.2bn at 158.00 — a hundred-pip corridor with spot inside it. One wire report has the pair steady at 157.64, which is 64 pips above the lower strike and 36 below the upper. A Tokyo report has it at 157.79–80 at 10 a.m. Tokyo time, which is 79 pips and 20 pips. Either way spot sits between the two largest strikes with hours to run, and $2.7bn is a real number on a day when the spot market has done very little.
Euro–dollar is the reverse. The board totals EUR 5.733bn and EUR 5.079bn of it — 88.6 per cent — is struck below a spot that has been above 1.1200 all session, with a session peak of 1.1237. The biggest strike on the board, EUR 3.2bn at 1.1150, is about 85 pips out of the money. Exactly EUR 654mln sits above spot, at 1.1250, some 16 pips up. A board that is seven-eighths spent is not a pinning force; it is an accounting entry. If you have a rule that widens stops or stands aside on “large expiries nearby”, it should be reading the distribution around spot and not the total.
The euro is where the day’s actual move is, for reasons that have nothing to do with the board: the dollar index eased about 0.5 per cent to 101.45 on one reading and sat at the low end of a 101.42 to 101.99 range on another, both describing a month low, with crude down about 4 per cent and election-sweep odds rising. Euro and Swiss franc led the G10. The yen underperformed by about 0.2 per cent.
Today’s high is eighty pips wide, and the pin question turns on it
We cannot tell you where the pair traded at its best today, and this is the number the whole expiry read depends on.
One publisher, at 10:29 UTC, describes dollar–yen as holding at the upper end of a 157.44 to 158.79 range. A Tokyo report, at 08:25 UTC, gives a morning low of 157.45 and a brief high of 157.99 just after noon Tokyo time. The lows agree to a pip. The highs are 80 pips apart.
The arithmetic goes against the higher figure, and it is the publisher’s own arithmetic. The same article says the yen is losing about 0.2 per cent on the day. Applied to the three Friday closes we published on Sunday — 157.487, 157.526 and 157.567 — that puts the pair between 157.80 and 157.88, which is 91 to 99 pips below the 158.79 its own range names. A third source has it steady at 157.64 the same day. On the evidence, 158.79 is not today’s high, and we are not publishing a high at all.
Why it matters in money: at a 157.99 high, the $1.2bn strike at 158.00 was missed by one pip and never tested, and the corridor is intact going into the cut. At a 158.79 high it was taken out by 79 pips hours ago, and anyone running a pin trade against it has been wrong since the European morning. Those are not two readings of one market; they are two different markets. Any inference you draw about pinning today is weaker than it reads, and the eighty pips is your error bar rather than ours.
One more field we will not reconstruct: the same Tokyo report has the pair down 24 sen on the week without stating the close it is measuring from, and the three Friday closes we published span eight pips. A change-on-the-week figure with an unnamed baseline is not a fact, it is a subtraction someone else did.
What this does not tell you
Every notional here is single-sourced in the sense that matters: two pages carry the same numbers, and we have given reasons to think they are the same board rather than two independent measurements of it, so agreement between them is not corroboration. No issuer publishes option expiry data, and the two sites that usually give us a second arrangement of it — an order-flow site and a retail research portal — both returned client errors and were never opened, for the fifth consecutive slot in one case. Neither page we did read states a time for the cut; 15:00 UTC is our own subtraction from 10:00 Eastern Standard Time, published here on 30 October and again this morning.
We have no option expiry board for tomorrow that is unambiguously tomorrow’s, which means we cannot tell you whether the strikes above would be replaced or repeated. We have read nothing about where resting spot orders sit today, so the corridor argument rests on strikes alone. The Finance Minister said this morning that the government remains ready to act against disorderly moves and that she could not say whether current levels are good or bad; the pair was largely unreactive, and we draw no intervention inference from it.
And the date question stays open. We have put the board at today on the balance of evidence from two secondary readers, one of which is demonstrably unreliable in exactly the field under dispute. That is an honest position and not a confident one.
