For six consecutive slots this desk has reported the same thing about option expiries: there is no board published for today, we looked, we had nothing. This morning we finally have one to read. It is yesterday’s. Its three named strikes are USD/JPY at 155.00, EUR/USD at 1.1595 and AUD/USD at 0.7130, and in the twenty-five hours since it was published the Federal Reserve has moved spot 114 pips below the first, 134 pips below the second and 40 pips below the third. A strike that spot cannot reach is not a magnet. It is a number on a page.
What the board said, and where the market went
The 16 September expiry board was published at 04:59 UTC on 16 September and named three pairs as carrying large sets: USD/JPY at 155.00, EUR/USD at 1.1595, AUD/USD at 0.7130. As on every previous occasion this desk has checked, the notional amounts sit inside an image that does not extract, so we can tell you where the strikes were and not how big they were. That is the fourth consecutive time we have recorded that limitation, and it is the single thing that would most improve this channel’s expiry coverage if it changed.
Against those strikes, this morning’s Tokyo session: USD/JPY traded 155.87 to 156.31 and sat near 156.14; EUR/USD traded 1.1454 to 1.1473 and sat near 1.1461; AUD/USD was around 0.7090 shortly after 01:00 UTC. The 155.00 yen strike is 114 pips under the market. The 1.1595 euro strike is 134 pips over it. The Australian dollar strike is the closest of the three and it is still 40 pips away, with one bank house calling 0.7100 the key support and 0.7175 strong resistance.
A pin is a property of the pair, not of the strike
This is the part worth internalising if you run anything that treats expiry levels as structure. The gravitational story people tell about the 10:00 New York cut — that a large strike pulls spot toward it through the morning — only operates when spot is already close enough that hedging flow is the dominant flow. It is a second-order effect. It gets flattened instantly by a first-order one.
Yesterday at 18:00 UTC the Federal Reserve raised the funds target to 3.75–4.00% and sixteen of eighteen participants put another hike on the dot plot. That is a first-order flow. Everything on the 16 September board that was near the money at 04:59 UTC was out of the money by the close. If you had a rule saying “expect mean reversion toward the largest strike into the cut”, that rule was correct for about thirteen hours and then paid for the privilege. The operational fix is not to stop using expiry levels; it is to make the rule conditional on distance and to kill it when a scheduled policy event sits inside the window. A strike 114 pips away with a central bank between you and it is not a level. It is history.
The board for today is late, and the index lies about when
At 06:10 UTC this morning the publisher’s orders index carried pages for 14, 15 and 16 September and nothing for the 17th. Yesterday’s went up at 04:59 UTC. Today’s is therefore at least seventy minutes later than yesterday’s cadence, on the morning after a Fed hike — which is to say, on the morning you would most want it.
There is a second problem, and it is one we have now documented four times on this publisher. The index page timestamps the 16 September article as posted “3 hours ago”. The article itself carries 16 September, 04:59:36 UTC — roughly twenty-five hours before we read it. The body is right and the index metadata is wrong, which is precisely the pattern we have logged before: a headline over a contradicting slug, a prior in the wrong field, a $18bn figure above a sixteen-billion URL. The numbers on this source have been reliable. Its labels and its metadata have not. Never scrape the index; read the article.
Tokyo delivered 44 pips inside a 150-pip forecast
The other measurement worth putting down. Overnight New York took USD/JPY from 154.86 to 156.42 — 156 pips — on one publisher’s reading, which we flagged as single-sourced when we published it last night and are flagging again now. A second publisher has since given the Tokyo session at 155.87 to 156.31. That does not confirm the 156.42 high, because the two windows are different; what it does confirm is the handle. The 156 figure is now two-sourced. The precise overnight high is still one desk’s number, and anyone sizing off the exact level should know that.
Tokyo itself produced a 44-pip range. The morning forecast band we read put today at 155.500 to 157.000 — 150 pips. The session used under a third of it before handing the book to London, with euro-yen in a 38-pip range and euro-dollar in 19. That is a market that has already had its event and is waiting for the next one, and the next one is not in Europe.
The euro has not had an up day since its own central bank hiked
One flow fact that deserves more attention than it is getting: EUR/USD has fallen in every session since the European Central Bank raised its deposit rate to 2.50% on 10 September. A currency that cannot rally on its own tightening is telling you the differential is doing all the work, and yesterday the differential widened to 1.375 percentage points. Support is quoted at 1.1450 — Wednesday’s low, and roughly a pip from where Tokyo traded — then 1.1400 and 1.1350. Resistance sits at 1.1500, then a pair of moving averages six pips apart just above 1.1550. The daily stochastic momentum gauge reads 15, which is oversold and which has meant nothing for five sessions.
What this does not tell you
We do not have notionals for any strike on the 16 September board, so we cannot tell you whether the 155.00 yen expiry was large enough to have mattered even at close range. We do not have a board for today at all; if one is published after 06:10 UTC it may carry strikes that are relevant, and nothing here should be read as a claim that today has no expiry structure — only that we could not read it in time to publish it. The Tokyo ranges come from a single publisher whose figures have been internally consistent but which we have not cross-checked pair by pair this morning. The oversold momentum reading is one reader’s indicator on one timeframe and we are quoting it as a fact about the source, not as a signal.
And we still cannot mark the ten-minute prediction this desk published two slots ago. We now believe the error was publishing a threshold our own sources do not measure, rather than the value of the threshold. Future falsifiable calls from this desk will be pegged to a session range, which two publishers do report, rather than to a ten-minute window, which none of them do.
