Share of days wider than the same week's typical non-event day
- CPI release days
- Non-event days of the same weeks (base rate)
Hover or tap a column.
October 27, 2026 · 3 min read
CPI Mornings in NQ: Is the First Hour Really Wider?
On 59 of 114 CPI release days since 2017, NQ's first regular-session hour was wider than the typical non-event day of the same week: 51.8%, against 50.0% for a non-event day. The regular session opens an hour after the release, and its first hour looks much like any other.
CPI mornings have a reputation: the inflation number lands, and the Nasdaq opens into a wild first hour. The release is at 8:30 am ET, though, an hour before the regular session opens. The question is whether the first hour of the session, 9:30 to 10:30 am, is still wider than usual once the overnight market has had an hour to take in the number.
The base rate
We took every CPI release day from 2017 on and compared each one with the non-event days of the same week: days with no CPI release, Fed statement or payrolls report. Comparing within the week matters, because a volatile month makes every day look eventful. A day counts as wider when its first-hour range beats the median of the other non-event days in its week. By construction, a non-event day does that about half the time. That half is the base rate.
| Measure | Day type | Days | Wider than the week's median | 95% interval | Average range |
|---|---|---|---|---|---|
| First hour | CPI release day | 114 | 59 (51.8%) | 43–61% | 0.91% of the open |
| First hour | Non-event day (base rate) | 452 | 226 (50.0%) | 45–55% | 0.82% of the open |
| 10:30 am to close | CPI release day | 114 | 65 (57.0%) | 48–66% | 1.28% of the open |
| 10:30 am to close | Non-event day (base rate) | 452 | 225 (49.8%) | 45–54% | 1.21% of the open |
The first hour of a CPI day beat its week's typical quiet day 51.8% of the time, against 50.0% for the quiet days themselves. The interval runs from 43% to 61%, comfortably across the base rate. The median CPI first hour was 5% wider than its week's median quiet day. The average is higher than the median, at 0.91% of the open against 0.82%, because a few CPI mornings were very large. A typical CPI morning, in the first regular-session hour, looks much like any other morning that week.
Is it chance?
We slid the CPI labels along the calendar by every offset of at least 26 sessions and recomputed the gap each time. For the first hour, 74% of the 514 shifted versions produced a gap at least as large as the real one: the result is what chance alone produces. The rest of the session does a little better, 57.0% against 49.8%, but 20% of shifted versions match that too, and it was one of three measures tested.
The same comparison by era, with 95% intervals (top: first hour, bottom: 10:30 am to close)
- CPI release days
- Non-event days
- 50%
Hover or tap a row.
The two halves of the sample disagree. In 2017–2021, CPI first hours were wider than their week's median quiet day only 25 of 57 times (43.9%), and the median CPI first hour was 7% narrower. In 2022–2026, when inflation was the market's main story, it was 34 of 57 (59.6%), and the median was 16% wider. Even in that era the interval, 47–71%, reaches below the base rate. The rest-of-day window shows the same split: 50.9% in the first era, 63.2% in the second.
What it does and does not say
The data does not say CPI days are quiet. It says the part most people picture, a burst at the cash open, does not show up reliably: the overnight session trades the number from 8:30 am, and the regular session opens an hour later. This post does not measure the 8:30 am move itself.
Direction is a coin flip as well. The first hour closed higher on 56 of 114 CPI days, 49.1% (interval 40–58%), against 51.5% on non-event days (47–56%). The data says little about which way a CPI morning goes, and little about how far.
- Clock: check which window the stat measures. A release at 8:30 am and a session that opens at 9:30 am are different events.
- Base rate: compare the event day with ordinary days from the same week, not with all days.
- Era: the sign here changed between 2017–2021 and 2022–2026. Check whether the next version of the stat holds in both halves of its own sample.
Data and method. E-mini Nasdaq-100 (NQ) futures, regular-session 1-minute bars (9:30 am to 4:00 pm ET), 3 January 2017 to 19 October 2026; sessions with a contract roll inside them, and sessions without a full first hour, are left out. CPI release days are the Bureau of Labor Statistics release dates; the number comes out at 8:30 am ET, before the regular session opens. A non-event day has no FOMC statement, CPI release or payrolls report. The first-hour range is the high minus the low from 9:30 to 10:30 am, as a share of the open; the rest-of-day range is the same from 10:30 am to the close. Each day is compared with the median of the other non-event days of its own Monday-to-Friday week, and a day needs at least 2 such days to count. Intervals are 95% Wilson intervals. The shift placebo slides the CPI marks along the series of CPI and non-event sessions of CPI weeks by every offset of at least 26 sessions; p is the share of offsets producing a gap at least as large. Three measures were tested: the first-hour range, the rest-of-day range and the direction of the first hour.