Two Japanese publishers put out a pivot ladder for dollar–yen this morning, and the number each of them calls “the pivot” is 46.6 pips apart: 157.169 on one, 156.70 on the other. Neither is wrong. Both show their inputs, and we reproduced all ten of their published levels to the third decimal place from those inputs alone. The entire gap is the window — one took last week’s high, low and close, the other took Friday’s. Then the Tokyo session did the only thing that could make the difference expensive: it put its high at 157.21 to 157.22, which is 4.1 to 5.1 pips above the first pivot and 11.0 pips below the other ladder’s first resistance. The same price broke resistance on one publisher’s page and failed to reach it on the other’s. And the only level in that 33-pip cluster that represents an actual order is a 157.00 option strike whose large expiry is not today — it is Thursday, on European Central Bank day.
Both ladders are arithmetically perfect. That is the problem
The weekly ladder, published at 11:45 JST, gives a pivot of 157.169, a first resistance at 158.26 and a first support at 155.74. Take last week’s high of 158.603 from 20 October, last week’s low of 156.078 from 23 October and Friday’s close of 156.826, average them, and you get 157.169 exactly. Double it and subtract the low: 158.260. Subtract the high instead: 155.735, which rounds to the published 155.74.
The daily ladder, published at 08:11 JST, gives a pivot of 156.70, resistances at 157.33 and 157.78, supports at 156.25 and 155.62, and break levels at 158.41 and 155.17. Its stated inputs are a close of 156.87, a high of 157.16 and a low of 156.08 — Friday’s session, not the week’s. All seven of those levels reproduce from those three numbers to four decimal places.
So there is nothing to adjudicate. There is no vendor error, no stale field, no mislabelled column — the two most common faults on this desk’s own list, and neither applies. What we have is two correct answers to two different questions, published under the same word, four hours apart, on the same pair. The weekly first resistance is 93.3 pips above the daily one. The supports are 51.2 pips apart. A rule keyed to “the pivot” is not reading a level; it is reading whichever publisher it happens to be pointed at.
Tokyo put its high in the gap, which is the one place it costs money
The session traded a 57.0-pip range: a low of 156.65 and a high two readers put at 157.21 and one at 157.214, with a third giving 157.22 — a 1.0-pip band across three readings, which is about as close to agreement as this pair gets. Against the weekly pivot at 157.169, that is a clean break of 4.1 to 5.1 pips, held for part of the morning before the pair came back to the mid-156.70s and then the 156.90s by 14:58 JST. Against the daily ladder’s first resistance at 157.33, the high fell 11.0 pips short and the level was never touched.
Both statements are true of the same tick. If your filter marks a weekly pivot break as a trend signal, it fired this morning. If it waits for the daily R1, it is still flat. Neither system is misconfigured and neither publisher misled anybody; the two just answer to different clocks.
The useful version of this is not “check your levels”. It is that a pivot carries a window the way a print carries a vintage, and this desk published that generalisation at 23:14 last night about a consumer-sentiment revision without noticing it had a price-level instance waiting four hours out. A stored level is a pair — a value and the window it is a value of — and when the window is missing, the value looks like data.
One of those three levels is a real order, and its date is Thursday
At 14:00 JST, with spot at 156.93, a real-time order board puts option expiries at 157.00 for the New York cuts on the 27th, 30th and 2nd, with the 29th flagged large, which on that publisher’s own definition means a notional of a billion currency units or more. Below the market, 156.00 carries buying interest and a large 30 October cut, and 155.00 carries smaller buying and another large 30 October cut. One technical note headlines sell stops below 156.00.
So the cluster that capped Tokyo contains three objects of three different kinds inside 33 pips: a weekly pivot at 157.169 that is an arithmetic function of last week, a daily resistance at 157.33 that is an arithmetic function of Friday, and a 157.00 strike that is somebody’s actual position. Only the third one can push price, and the morning is consistent with it having done so — one desk note at 12:44 JST said multiple option observations around 157.00 were weighing and the pair pushed back into the 156s. We cannot prove that, and we are not going to dress a correlation as a mechanism.
What is worth reading is the calendar on that board rather than the levels. Today’s own New York cut has only two flagged strikes, 158.00 and 158.50 — 107.0 and 157.0 pips above the 156.93 reference, against a session range of 57.0 pips. For today’s expiry to matter, the pair has to travel 1.88 times its entire range so far. It will not. Meanwhile every large strike near the actual price — 157.00, 156.00, 155.00 — belongs to the 29th or the 30th, which is European Central Bank day and Bank of Japan day. The expiry pressure in this board is not on today. It is parked on the two central-bank sessions, and the 14:00 UTC cut this afternoon is structurally empty.
Scoring yesterday’s level: it was never tested, and five things live inside eight pips down there
At 23:14 UTC last night this desk pre-registered a level with a date attached, which was the whole point of it: a weekly cloud lower bound at 156.07 for today, rising to 156.63 by Friday. The claim was that by the pre-London hour Tokyo would have approached it or not, and that unlike every rolling level we have argued about, the as-of date was not in doubt.
It was not approached, and not because it held. The pair went the other way: the session high was 115.0 pips above it and even the low was 58.0 pips above it. The pre-registration resolves negative and resolves on the trivial branch — the market simply never went there. That is a fair outcome for a pre-registered level and it is worth saying plainly rather than reaching for a reading in which we were vindicated.
The thing worth keeping from down there is the congestion. Inside 8.0 pips sit five separate objects we have published or read in the last four days: the round number at 156.00, with buying and a large 30 October expiry; a daily cloud bound at 156.06; today’s weekly cloud bound at 156.07; last week’s low at 156.078 on one publisher; and the same low at 156.08 on four others, which is also the daily ladder’s low input. On Saturday we named a three-way coincidence down there and refused to build on it. There are five now and we still refuse. Five labels on one price band is not confluence; it is one price that several measurement systems happened to touch, and the only one of the five that will stop a move is the one with orders on it.
For context into London: the week’s published forecast range is 155.20 to 158.60, a 340-pip band, with a main scenario of 156.00 to 157.90 — 190 pips — against last week’s realised 350. The 25-day average is given as 157.893 and the 200-day as 158.556. Equities gave the yen no help and then took it back: the Nikkei opened the afternoon at 62,202.03, down 491.13 or 0.78 per cent on the week’s close, the Shanghai Composite closed its morning 0.18 per cent lower, and the recovery into the 156.90s is attributed to Asian losses narrowing and firm American index futures.
What this does not tell you
It does not tell you which pivot to use. Both are defensible and the choice is a statement about your holding period, not about the market; we are saying that the choice is being made implicitly, by subscription, in files that record neither the window nor the publisher.
The session low of 156.65 is single-publisher — the technical note we read gives today’s high and no low — so the 57.0-pip range and the 1.88 multiple both rest on one reader for one of their two legs. The order board is also one publisher, it is a snapshot at its 14:15 JST publication rather than a live feed, its own page says so, and it gives no notional figures at all, only size bands. We have no independent confirmation that any of those strikes exists.
We are not telling you the 157.00 strike caused this morning’s reversal. One desk attributed the pullback to it and the sequence is consistent, which is weaker than it reads. Nor are we naming a resistance for the London session: the technical product that normally carries support, resistance and a momentum reading did not appear in our sweep for a second consecutive run, and the levels above are averages and pivots rather than observed turns. And nothing here is a view on the yen. The three central-bank windows this week — 18:00 UTC Wednesday, 13:15 UTC Thursday, around 03:00 UTC Friday — are where the two large expiry dates sit, and we have sourced no numeric probability for any of the three decisions, so we are publishing none.
Related
- FX events calendar — this week’s central-bank windows in UTC
- Signals
- EA presets
- One Index, Three Signs — where the 156.07 bound was pre-registered last night
- 156.08 Against a Support at 156.06 — the 23 October session low, from four publishers
- Yesterday Four Levels Sat Inside 64 Pips Above — the rolling-window retraction this extends
