At 06:16 UTC this desk published a session low of 157.53, called it two pips through a named support, and said the fortnight-long compression had survived anyway at 72.0 pips — day fourteen under a yen. It named 157.20 as the level that would settle whether the range was over. By the London afternoon the low was 156.75. That is 45.0 pips below the falsifier we named, 78.0 pips below the low we published, and it makes the day 150.0 pips from high to low. Day fourteen is the last day of the sequence, and the Tokyo session is where the information was least useful.
The handoff, in figures: 72 pips became 150 in four and a half hours
Take the day as two sessions. Tokyo ran 158.25 to 157.53 — 72.0 pips, comfortably inside a yen, and a 2.0-pip break of the 157.55 support one publisher named overnight. Europe then took 78.0 more pips out of it: 156.83 by 08:15 UTC on one publisher’s intraday note, 156.75 by 10:30 on another’s range field. The second session produced more range than the first, and it produced it below everything the first session was being measured against.
The arithmetic that matters to a position is not the low. It is the ratio. This desk independently measured the three preceding days at 78.3, 55.0 and 72.0 pips, a mean of 68.4. Today is 150.0 — 2.19 times the recent average day, in a market whose own morning forecast band was 157.500 to 158.600, or 110.0 pips wide. The realised day is 136.4 per cent of that band, and the low sits 75.0 pips beneath its floor.
One publisher’s intraday note puts the fall at 1.42 yen from 158.25, which reconciles exactly to 156.83 and is the one piece of internal corroboration available on the magnitude. A Reuters-sourced wire at 10:27 UTC called it the dollar’s largest single-day fall against the yen since late September, with the pair at 157.27 at the time of writing — which is 52 pips above the low already set, so treat that level as a snapshot, not a close.
The falsifier was named and it broke — which is the only reason this is scoreable
This morning’s piece did one thing right and it is worth isolating, because it is cheap to repeat. It did not merely say the range might end. It named the number: 157.20, the lower band of a 21-day two-standard-deviation envelope, 33.0 pips below the low at the time. Below that, the range question is settled; above it, the compression stands.
156.75 is 45.0 pips below 157.20. There is no interpretation to perform. Compare that to the alternative version of the same claim — “the floor is under pressure” — which would have been equally true at 157.53 and at 156.75 and would have told you nothing about which. A pre-registration you can lose is worth several you cannot.
What it also means is that the structural stack we published at 06:16 is now the wrong way round. We listed four resistances inside 64 pips above spot and argued there was no room to start a trend before something was in the way. True, upward. Downward there was nothing named until the Ichimoku base at 157.19 — taken out, 44.0 pips above the low — and then a 10-day envelope floor at 156.65, which the low is still 10.0 pips above. The next named level below is the cloud bottom at 155.65, 110.0 pips away. The congestion was all on one side and we only counted that side.
Every yen cross went, which is what distinguishes this from a dollar move
Euro-yen is the cleaner read. The Tokyo session traded 177.09 to 177.75; the European session traded 175.87 to 176.66, and one publisher records the pair below 176.00 for the first time since 5 November of last year. That is 122.0 pips below the Tokyo low, on a 79.0-pip European range. Sterling-yen broke 210.40, down more than a yen and a half from the Tokyo open on the same publisher’s wire. Meanwhile euro-dollar traded 1.1216 to 1.1243, a 27.0-pip range, and a Reuters-sourced wire had it steady at 1.1234.
So the dollar leg is the small part. A pair-by-pair breakdown with a 150-pip dollar-yen day, a 122-pip euro-yen extension and a 27-pip euro-dollar range is a yen story with a dollar component, not the reverse. If your exposure is spread across yen crosses on the assumption that they are partly independent bets, today is the configuration in which that assumption costs the most: one factor moved and every leg paid.
The causes on offer are the same list the Tokyo session was already carrying, plus one new one. Falling US yields narrowing the rate differential, and a Bank of Japan meeting on 29 and 30 October with a December increase broadly expected, are the yen-specific legs. The risk-off legs are the Tesla capital expenditure cut, Tokyo equities closing down 1,713 yen and Hong Kong down 3.03 per cent, crude near 99 dollars after a reported attack on four tankers in the Gulf. And the new one is France: a Reuters-sourced wire puts the French ten-year around its highest since 2002 on debt and weekend unrest, and one Tokyo publisher names French fiscal and political instability outright as the session’s driver. We have been carrying the French leg unread for nine runs. It arrived as a yen bid.
What a range classification is worth at the moment you make it
Here is the uncomfortable part, and it is aimed at this desk’s own output. At 06:16 we published a correct measurement, correctly sourced, with the right error bars and a named falsifier — and the classification it supported was obsolete four and a half hours later. The measurement has not become wrong. The 72.0 pips were real at 05:51 UTC. What expired is the inference.
A mean-reversion rule sized on a 68-pip average day met a 150-pip day. A breakout rule that required the 157.20 band got its signal roughly eight hours after the floor it was watching first cracked, by which point 45 pips of the move had already happened. Both rules behaved as designed. The problem is that a range classification is a statement about a window, and the window a Tokyo close gives you closes at the Tokyo close. We wrote this morning that the floor broke and the compression did not; the honest update is that the compression was still breaking while we were writing, and we published its survival in the hour before it ended.
The operational version: if a compression count is load-bearing in your sizing, re-derive it at each session handoff, not once a day. Fourteen days is a fact about history. It was never a forecast, and it stopped being a description at about 07:00 UTC.
What this does not tell you
It does not tell you where the day closes. Nothing here is a close: the 156.75 low is one publisher’s range field at 10:30 UTC, and the New York session, the 12:30 initial claims release and the 14:00 option cut are all still ahead. The low may extend or the day may revert; both leave the 150.0-pip range intact and neither is predicted here.
The fourteen-day count itself was always one publisher’s, and we said so this morning. We have independently measured only the last three terms. Ending a sequence we could not fully verify is a weaker result than it sounds: what we can defend is that today is 2.19 times the three-day average we measured ourselves, and that today exceeds one yen on two publishers’ figures. The length of what ended remains that publisher’s claim.
The day’s high of 158.25 comes from one publisher’s structured range field and is corroborated only by a second publisher’s arithmetic reconciling to the same number. The low rests on two publishers at two different times, 156.83 and 156.75, which is consistent with a low that extended rather than a disagreement — but neither figure has a second reader at its own timestamp.
And the causal list is not a weighting. Five candidate drivers were in circulation during this session and nothing we read apportions the move between them. The French leg in particular is named by one Tokyo publisher and implied by one wire; Moody’s reviews France tomorrow, which will be the first thing resembling a test.
Related
- FX event calendar — the 12:30 claims print and the 14:00 cut still ahead of this move
- Signals — range-versus-trend classification and when it expires
- EA presets — sizing on an average daily range that just doubled
