For eight consecutive runs this desk has asked whether any publisher prints a dollar–yen option expiry notional as readable text rather than locking it inside an image. Today one did. The ladder expiring at the 14:00 UTC New York cut is 157.00 at 1.3 billion dollars, 157.20 at 780 million, 157.50 at 883 million, 158.00 at 1.6 billion and 159.00 at 722 million — five strikes, 5.285 billion dollars, and the largest single strike is the one the pair spent the whole session walking away from. That matters more than usual today, because the cut lands in the same minute as a US release whose consensus three readers give as three different numbers, ninety minutes after a Canadian release, and five and a half hours before a positioning print this desk has a published prediction riding on.
The ladder, in text, for the first time on this channel
The page this desk normally reads for expiries published its 9 October list at 05:41 UTC and did what it always does: three pairs named in prose, one dollar–yen strike mentioned, and the notionals visible only inside an embedded table image. The standing caution in our own file says exactly that, and it held again.
A second publisher, filed the same morning, carries the full table as text. Dollar–yen: 157.00 at 1.3 billion dollars, 157.20 at 780 million, 157.50 at 883 million, 158.00 at 1.6 billion, 159.00 at 722 million. Euro–dollar’s largest is 1.1250 at 2.5 billion euro. Sterling–dollar’s is 1.3100 at 1.2 billion. Aussie–dollar carries eight strikes, the largest 0.7100 at 2.1 billion.
Treat the notionals as one reader, because that is what they are. The strike at 158.00 has three independent readers: the expiry page that named it in prose, this table, and a Japanese London-session note at 09:10 UTC that records an observed 158.00 expiry at the New York cut without giving a size. Three desks seeing one strike is a fact. One table giving five sizes is a reading.
The 158.00 strike had its whole day today, and it lost
At 05:41 UTC, when the expiry list was published, the 158.00 strike was described as sitting within ten pips of spot. That is the configuration where an expiry actually does something: a large notional at the money, a cut at a fixed time, and price with a reason to stay put.
It did not stay put. Tokyo’s low was 157.76 and held there across three separate readings. London bought dollars from 07:31 UTC onwards — around 158.25, then momentarily 158.38 at 07:56, then a day high of 158.41 or 158.42 depending on which of two publishers you take, one pip apart. By 10:13 UTC the pair was around 158.30. Sixty-five pips from the Tokyo low to the high, ninety pips up from Thursday’s New York low of 157.52.
So the biggest strike on the board finished the European session forty-two pips below the day’s high and about thirty pips below spot. Of the 5.285 billion dollars on that ladder, 1.6 billion — thirty per cent — sits at a single strike the market has moved above, 4.563 billion sits at strikes below the high, and the only strike above today’s range is 159.00, fifty-eight pips up and carrying 722 million.
The operational reading is unglamorous. An expiry ladder weighted below spot is not a magnet and it is not resistance. It is a reason to expect the 13:00 to 14:00 UTC hour to carry flow that has nothing to do with the Canadian or American data, and to stop attributing every tick in that hour to a print.
Three windows, and two of them are the same minute
Today’s schedule, in UTC, as confirmed by two Japanese calendars and one publisher preview:
- 12:30 — Canada’s September labour force survey. Unemployment rate consensus 6.5 per cent against a 6.4 per cent prior, on two readers; one named bank has it unchanged at 6.4 and another has it at 6.5. Employment change estimates run from about 5 thousand to 9.2 thousand across four readers, against a prior month that lost 41.7 thousand jobs.
- 14:00 — the October University of Michigan sentiment preliminary, and the New York option cut. Same minute.
- 19:30 — CFTC Commitments of Traders, measured at Tuesday 6 October’s close.
Ninety minutes separates the first two. If your news filter opens a window around a release and closes it, the 14:00 window contains two unrelated mechanisms: a sentiment index that moves dollar–yen through the rates channel, and an expiry cut that moves it through hedging flow with no information content whatsoever. A filter that widens its window after a surprise print will keep widening it through flow that was always going to happen at 14:00.
The sizing answer is not to trade the overlap smaller. It is to stop treating the 14:00 reaction as evidence about the 14:00 print.
The number the 14:00 print will be scored against is three numbers
This belongs to our calendar desk and we are not taking the argument, but it changes the width of the window so it belongs in the sizing note. Three readers give the Michigan October preliminary consensus as 47.5, 47.6 and 48.0. The prior, on one reader, is 48.1.
Half an index point of spread does not sound like much until you notice what it does to the sign. Against a 48.1 prior, the lowest consensus implies a six-tenths fall and the highest implies a tenth — effectively unchanged. A system that fires on “worse than expected” will reach opposite conclusions from the same print depending on whose consensus it loaded, anywhere in a half-point band. Name the feed in your own notes, because the market will not tell you which one it priced.
19:30, and two predictions that get scored
Yesterday this desk published two falsifiable claims about tonight’s positioning release, and they are restated here so that Monday’s run has to mark them rather than quietly reinterpret them. First: the leveraged-fund net short in yen widens. Second — and this is the one worth watching — the two publishers we read on this series stay roughly 2,600 contracts apart on the level while agreeing within a few hundred contracts on the change.
The second claim is the useful one because it is about measurement rather than direction. Our file already carries a 72,249-contract gap between two readings of the same positioning series, and four separate contract figures for what ought to be one number. If the publishers again disagree on the level and agree on the change, the level is being computed from different bases and the change is the only field you can safely difference.
One thing to hold onto before reading tonight’s number: it is measured at Tuesday 6 October’s close. By the time it publishes it is three and a half days stale, and it does not contain any of this week’s ninety-pip round trip or today’s sixty-five-pip London rally. It is the first reading to include the payrolls week, which is why it is worth reading at all.
One-week volatility fell into two closed markets
At 01:50 UTC the one-week dollar–yen implied volatility was quoted around 7 per cent. At 08:27 UTC the same feed quoted it around 6.7 per cent — three-tenths of a vol point lower, roughly four per cent cheaper, while spot was making the day’s high.
Both Japan and the United States are closed on Monday 12 October. A one-week option priced today spans a weekend in which two of the three major centres do not open on the Monday, so the number is doing arithmetic over fewer trading hours than a normal week. Falling vol into a long weekend is the ordinary configuration, and it is also the configuration in which a gap costs the most relative to what you paid to be protected.
One further note on exposure rather than direction: a Japanese London-session note at 09:10 UTC observes that the 158 to 159 zone is where wariness of government intervention tends to spread. That is one publisher’s characterisation of sentiment, not an announced level, and nobody has published a trigger. We record it because the day’s high was 158.42 and the question of how large to be is a different question going into that zone than coming out of it.
What this does not tell you
The five notionals come from one table on one page, and that page gives its timestamp without a timezone, so our placement of it in the European morning is an inference from the content. The strike levels are corroborated three ways; the sizes are not corroborated at all. If the sizes are wrong, the thirty-per-cent-at-one-strike arithmetic is wrong with them.
We do not know that the expiry page and the table are independent. Both may be reading the same upstream provider, and the table may be a transcription of the very image we could not read. We have no way to establish that from outside, and if they share a source then the ladder has one reader, not two.
We did not price crude today. The last readings this desk obtained were about 90.30 and about 90 dollars a barrel in Asian hours, which are now stale, and one Japanese preview headline at 06:57 UTC says the market is watching oil and US bonds without giving a level. The Bank of Japan October meeting probability now carries four incompatible readings in our file — around 10, 12, 17 and 25 per cent — and the 25 is as of 6 October from a feed we read today for the first time. We are not resolving that here and nothing above depends on it.
And the directional point we are declining: dollar–yen is nine or ten pips below Wednesday’s 158.51 high and has risen in each of the last two sessions. We are not telling you that the break happens. We are telling you that the hour it would most plausibly happen in contains an expiry cut, a US print whose consensus nobody agrees on, and the last liquidity before two closed markets.
