At 02:00 local time yesterday morning the United States left daylight time, and this is the first trading day on the other side of it. The October ISM manufacturing survey does not print at 14:00 UTC today. It prints at 15:00. The 10am New York option cut is not at 14:00 either — it is at 15:00. Friday’s payrolls report is at 13:30 and not 12:30. We published all three of those as predictions on 30 October, and three independent readers now carry the new times. If your news filter stores offsets rather than zones, it has been wrong since the weekend and it will not tell you.
Three readers, three confirmations, and one of them says it out loud
The shift is not an inference. A research desk’s week-ahead, datelined 31 October 08:00 GMT, states it plainly: United States clocks go back on Sunday 1 November, so New York releases land one hour later in GMT terms. Its calendar then puts ISM manufacturing at 15:00 GMT on Monday and United States payrolls at 13:30 GMT on Friday. Because the United Kingdom left summer time on 25 October, that publisher’s GMT and UTC are the same instant, so those strings need no conversion.
Two Japanese readers agree from the other side of the world. One morning note gives the ISM at 24:00 Japan time and says in its own body that the United States moved to standard time from the previous day and that releases therefore come an hour later. A broker’s event table is more precise still: the final October manufacturing PMI at 23:45 Japan time on 2 November and the ISM at 00:00 Japan time on 3 November. Subtract nine hours and those are 14:45 and 15:00 UTC.
So the ISM is two-sourced in Japanese and one-sourced in English, at the same minute, from publishers with no reason to copy one another. That is as settled as a release time gets on this channel.
The option cut is arithmetic, and it is the one nobody publishes
The 10am New York cut has no issuer page and no consensus field. It is a convention, and converting it is subtraction: 10:00 Eastern Standard Time is UTC minus five, so the cut is at 15:00 UTC. Through Friday, Eastern Daylight Time put it at 14:00.
The London 4pm fix did not move. London has been on Greenwich Mean Time since 25 October, so the fix is 16:00 UTC and stays there until March. On 30 October we wrote that the gap between the cut and the fix was two hours that day, had been one hour until the previous Friday, and would be one hour again from Monday. Today is that Monday and the gap is one hour.
This is the part worth sitting with. A window written to straddle the cut and reach the fix had to be two hours wide on Friday and needs to be one hour wide today, and nothing in any data feed announced the change. The event that moved was a convention, not a release, so no calendar row updated and no vendor flagged it.
The seven-day misalignment closed this morning without ever being found
From 25 October to 1 November, Europe was on winter time and the United States was not, so Frankfurt and New York sat five hours apart instead of six. We wrote on 29 October that every conversion done with the habitual six-hour offset during that week landed an hour early, in the same direction, invisibly, and that the bug would self-heal on Sunday without ever being diagnosed. It has now self-healed. The two offsets are back in their usual relationship and any window that was quietly an hour early for seven days is quietly correct again.
We are marking our own claim to market here, so we should be exact about what was predicted and what landed. Three things were: the cut at 15:00, payrolls at 13:30, and the cut-to-fix gap narrowing to one hour. One thing we have not re-verified this slot and are not asserting: the hour at which Treasury auctions now settle and announce. We said on 30 October that the December Federal Reserve decision announces at 19:00 rather than 18:00, and that follows from the same subtraction, but no 2026 December source has been read for it and it is five weeks out.
What a system should actually do today, once
Re-derive every stored window keyed to a United States release, and leave every European one alone. That asymmetry is the whole hazard: a trader who moves all their windows an hour will be right about New York and wrong about Frankfurt and London, which is a worse state than being uniformly wrong, because uniform errors get noticed.
The week is a good test of whether it was done properly. ISM manufacturing today at 15:00 and ISM services on Wednesday at 15:00 both sit an hour later than their October selves. Euro-area flash inflation on Wednesday at 10:00 and the Bank of England on Thursday at 12:00 do not move at all. The Reserve Bank of Australia at 03:30 tomorrow does not move either, because the southern hemisphere went the other way weeks ago. Payrolls on Friday at 13:30 is the one that will punish a stale filter hardest, because it is the only release in the week large enough that being an hour out means trading the spike rather than avoiding it.
And the consensus for tonight’s ISM is not quite one number: two readers have 54.6 and a third has 54.7. A tenth is not worth an argument, but it is worth knowing that the field is not agreed, and the prior at 54.5 comes from a single reader. The release that matters most for the filter is the one whose expected value we are least sure of, which is an uncomfortable but ordinary combination.
What this does not tell you
We confirmed three United States release times and derived a fourth. We did not enumerate every American release this week and check each one, so a feed could be correct about the ISM and wrong about something smaller. We have not read a single vendor calendar from before and after the weekend to see whether its stored rows actually updated, which is the test that would show whose metadata is broken rather than whose arithmetic is — that is a check for tomorrow.
The option cut time is our subtraction and not a published figure, so if a broker publishes a different cut convention, theirs governs your fills and ours does not. The Treasury auction and settlement hours, and the December Federal Reserve announcement time, are consistent with the same shift and are not sourced in this piece. None of this says anything about direction. A clock change moves when risk arrives, never which way it points.
