Dollar–yen has spent the Tokyo session doing the one thing it has not done for a fortnight: trading through the floor. One publisher’s morning note, out at 22:39 UTC yesterday, put the forecast range at 157.500 to 158.600 and named support at about 157.55, last week’s low. The session low printed 157.53 — two pips through that support and three pips above the bottom of the publisher’s own forecast band. Along the way the pair took out 157.80 at 03:15 UTC, which was already below Tuesday’s 157.83 and Wednesday’s 157.84, so the two previous days’ lows went in the same move. And yet the range discipline held: 157.53 to 158.25 is 72.0 pips, comfortably inside a yen, which makes this the fourteenth consecutive business day of sub-one-yen trading on the same publisher’s count. The floor broke and the compression did not. That combination is what this note is about, because it is the configuration that punishes a breakout rule and a mean-reversion rule in the same session.
Two pips through the support, three above the forecast floor
The precision here is uncomfortable rather than impressive. A publisher gave a floor at 157.55 and a range bottom at 157.500, and the market stopped between them, 2.0 pips below one and 3.0 pips above the other. That is not a forecast being vindicated; it is a 5.0-pip window being landed in, which is inside the noise of any of these numbers.
What it does establish is direction. Monday was the last time this pair showed dollar selling of any size, and the 157.80 print at 03:15 UTC was flagged by one reader as the lowest since then. Thirty pips below Tuesday’s low and thirty-one below Wednesday’s is not a large move by any normal standard — it is roughly half a day’s range in a market that has had no days — but it is the first time in three sessions that the bottom of the structure moved at all. If you have been selling rallies into 158.20 and buying 157.85 for a fortnight, the second half of that trade stopped working this morning and the first half still does.
One source problem, stated plainly: the publisher giving us the session range at 05:51 UTC prints 157.53 to 158.25 in its range field and says in its text that the pair fell to around 157.30. Those are 23.0 pips apart and we cannot reconcile them. We have used the range field, which is the more structured of the two, and every distance in this note carries that gap as an error bar. If 157.30 is right, the break of support is 25 pips rather than two and the session range is 95 pips rather than 72 — still under a yen, but only just.
Day fourteen, and why the count is more believable than its source
The fourteen-day figure — thirteen completed days plus today in progress — comes from one publisher and we have not found a second reader for it. That would normally mean labelling it and moving on. In this case we can do slightly better, because this desk has independently measured the last three of those days from other sources: 78.3 pips on 20 October, 55.0 pips on 21 October, and 72.0 so far today. All three are under a yen, they average 68.4 pips, and none of the three readings came from the publisher making the claim.
That does not confirm the length of the streak. It confirms the mechanism the streak describes, over the only window we can check. There is a real distinction: we can say with confidence that this pair has been trading in a two-thirds-of-a-yen box for at least three sessions on independent data, and we are relaying the thirteen-day count as that publisher’s claim. Anyone sizing a breakout on the strength of a long compression should be sizing on the three days they can verify, not the fourteen they have been told.
The same publisher’s forecast range is worth one more line, because it quietly concedes the point. A band of 157.500 to 158.600 is 110.0 pips wide. If the day realises its own forecast at both extremes, the streak ends — the publisher’s central case is a range 10 pips too wide to continue the sequence it is reporting. As of 05:51 UTC the session has used 72.0 pips, or 65.5 per cent of that band.
Four resistances inside 64 pips, and a cap nobody can state to better than four pips
This is the structural reason the upside has been going nowhere. With spot at 157.92 at 03:20 UTC, the levels stacked above it were the 21-day average at 158.02, the 10-day at 158.23, the Ichimoku conversion line at 158.25, and the 200-day at 158.56. That is 10.0, 31.0, 33.0 and 64.0 pips away respectively — four distinct technical levels inside a 64-pip ceiling, in a market averaging 68 pips a day. There is no room between them to get a trend started before something else is in the way, which is a tidier explanation of a fortnight of chop than any narrative about positioning.
The 200-day is the one everyone is watching and it is the one nobody can pin down. Three readings are now in our file: 158.56 from the indicator page this morning, 158.57 from another publisher yesterday, and “about 158.60” from that same publisher today. A 4.0-pip spread, from two sources, on one of them disagreeing with itself a day apart. For most purposes four pips is nothing. For a cap that the week’s high has been grinding against, four pips is the difference between a clean rejection and a false break, and anyone with a rule that fires on a 200-day touch should know that the level it is firing on is a band rather than a number.
For completeness, the rest of the envelope: the 21-day Bollinger bands at two standard deviations sit at 158.84 and 157.20, a width of 164.0 pips. The session low came within 33.0 pips of the lower band. The 100-day average is 147.0 pips above spot and the Ichimoku cloud base is at 155.65 — both far enough away to be irrelevant to anything happening this week.
The driver is Tesla and the Gulf, not Japan
Nothing Japanese did this. The Tokyo-hours tape reads as a straight risk-off cascade with an equity trigger: Nasdaq futures down about one per cent by 02:37 UTC and 1.2 per cent by 04:06 after Tesla cut its 2026 capital expenditure target to 20 billion dollars from more than 25 billion; the Nikkei down more than 700 yen at the morning close, about 1,000 in the afternoon, roughly 1,300 by 04:19 and about 1,500 by 05:33; Korean equities down three per cent. US yields fell after hours on safe-haven buying. Yen strength is the output of that, not the cause.
The geopolitical leg arrived on top of it. Reports at 23:00 and 23:56 UTC yesterday had the United States preparing options against Iran, and at 04:50 UTC a report that Iran’s Revolutionary Guard had attacked four oil tankers in the Gulf. Crude pushed back above 99 dollars in after-hours trading by 05:25. We are not publishing a crude level: the two readings we have of yesterday’s close are 2.25 dollars apart and this desk’s rule on a geopolitical session is three readers and a published band, which we do not have.
The practical consequence is about which correlation you are actually exposed to. A yen position taken this morning is an equity-volatility position and a Gulf-headline position wearing a currency ticker, and both of those legs can reverse inside an hour on a wire correction — which is exactly what happened to the Gulf embassy story yesterday evening, when the State Department narrowed the order and most of the move faded. If your yen exposure and your index exposure are both long risk-off this morning, you have one position, not two.
What this does not tell you
It does not tell you the range is over. A 72-pip day that broke the previous floor by two pips is not a breakout; it is the box moving down slightly. The level that would settle it is 157.20, the lower Bollinger band, 33 pips below today’s low, and nothing we have seen suggests that is in play before tonight.
It does not give you a close. This note is written during the Tokyo afternoon and the session low is a few hours old; the London session has not started and the 12:30 UTC claims print and the 11:00 UTC Turkish decision both sit ahead of it. The 157.53 low may not survive as the day’s low, and the 72.0-pip range is a running figure, not a result.
And the two weakest claims in it are the ones to discount first. The fourteen-day count is one publisher’s, corroborated in mechanism over three days and not in length. The session range is from a publisher whose own text contradicts its own range field by 23 pips. We have not reached a single primary exchange or settlement source in this piece; every level is a secondary reading. Nor do we have today’s option expiry board: the 22 October ladder was not published anywhere we could find it this morning, and the two pages search returned under that title were both from a previous year, caught on their URL slugs and not opened.
Related
- FX events calendar — the 11:00 and 12:30 UTC windows ahead of the London session
- Signals — range-versus-trend classification when daily ranges compress
- EA presets — correlated exposure across yen crosses and index risk
- 1KPIPS News — this morning’s note on the 200-day average and the week’s high
