Seven official US releases, the print dates we hold since 1947 (months with no print dated in them: CPI 4, jobs 8, PPI 3, retail sales 2, PCE 80 — mostly a print that fell into the month next door, which then carries two; left unpaired: CPI 1, jobs 1, PPI 2, retail sales 2, PCE 21), and the dates already announced ahead. Each bar is the plain close-to-close move bitcoin made on a print day, beside what an ordinary day does.
+0.68% every print day on Oct-15 2026, the last day in this build, read from a public record, published by askthetape.com.
Seven numbers set the rhythm of the US calendar: consumer prices, producer prices, the jobs report, advance retail sales, GDP, personal income and outlays, and the statement the FOMC puts out after it meets. This page is their whole timetable — 5,125 prints on 4,676 calendar days, from 1947-07-20 to 2027-12-15, 72 of them already announced for the future — and beside it, what bitcoin's tape did on those days. We use the current agency time rules for older dates whose historical publication time has not been verified: the three statistical offices print at 08:30 in New York (4,844 of these prints), the FOMC at 14:00 (281). The most recent print was retail sales and PPI on 2026-10-15, and bitcoin closed +0.68% that day. The next one is 2026-10-28 at 18:00Z, FOMC. The bars are the plainest measurement available: the close-to-close move of the whole UTC day that contains the print, on 1,223 print days since 2010-07-30. Their median is +0.10%, the middle eighty per cent runs from -3.77% to +4.76%, and 52% of them closed higher. An ordinary day on the same tape — 4,711 of them — has a median of +0.13%, runs from -3.62% to +4.47%, and closes higher 53% of the time. So the direction is a coin toss on both. What does differ is SIZE: the typical print day finishes 1.64% away from the previous close against 1.52% on an ordinary day. That is the whole finding, it is small, and it is a description of the past, not a forecast. The last view leaves the tape alone and counts the calendar itself: how many official prints landed in each closed year, from 3 in 1947 to a median of 84 a year since 2015.
| leg | publication DATES from the FRED public release calendar, cross-checked against each agency's own published schedule (BLS, BEA, Census) and, for the FOMC, taken from the Federal Reserve's own calendar pages rather than from FRED. 12 of the 5,125 rows carry two independent sources in agreement, 4,784 come from FRED alone and 281 from the Federal Reserve. The tape beside them is our own Bitstamp daily-close leg. |
|---|---|
| cadence | the calendar is rebuilt weekly; a bar appears only once the print day's close has landed, and a print that happens today sits hollow on the zero line until it does |
| coverage, by release | CPI 965 dates 1949-03-24 → 2027-12-10 (next 2026-11-10) · PPI 618 dates 1978-02-10 → 2027-12-15 (next 2026-11-13) · jobs report 879 dates 1955-05-06 → 2027-12-03 (next 2026-11-06) · GDP 867 dates 1947-07-20 → 2026-12-23 (next 2026-10-29) · PCE 744 dates 1966-01-18 → 2026-12-23 (next 2026-10-29) · retail sales 771 dates 1966-01-10 → 2027-12-15 (next 2026-11-17) · FOMC 281 dates 1994-02-04 → 2027-12-08 (next 2026-10-28) |
| what prints next | 2026-10-28 18:00Z FOMC · 2026-10-29 12:30Z GDP · 2026-10-29 12:30Z PCE · 2026-11-06 13:30Z jobs report · 2026-11-10 13:30Z CPI · 2026-11-13 13:30Z PPI · 2026-11-17 13:30Z retail sales · 2026-11-25 13:30Z GDP · 2026-11-25 13:30Z PCE · 2026-12-04 13:30Z jobs report · 2026-12-09 19:00Z FOMC · 2026-12-10 13:30Z CPI · 2026-12-15 13:30Z PPI · 2026-12-16 13:30Z retail sales · 2026-12-23 13:30Z GDP · 2026-12-23 13:30Z PCE · 2027-01-08 13:30Z jobs report · 2027-01-13 13:30Z CPI · 2027-01-14 13:30Z PPI · 2027-01-15 13:30Z retail sales |
| 🩸 the bar is a whole UTC day, not a window around the print | our price history is daily closes, so the bar for a print day is the move from the previous close to that day's close — midnight to midnight UTC. The print itself lands at 12:30 or 13:30Z (08:30 in New York) or at 18:00/19:00Z (14:00 in New York), inside that day, and everything else that happened in the same 24 hours is inside the bar too. A tighter window would need an intraday leg across thirteen years, which we do not have; the page says so rather than passing the bar off as the reaction. |
| 🩸 print days against ordinary days, measured on this build | print days: n = 1,223, median +0.10%, p10 -3.77%, p90 +4.76%, typical distance from the previous close 1.64%, 52% closed up. Ordinary days: n = 4,711, median +0.13%, p10 -3.62%, p90 +4.47%, typical distance 1.52%, 53% up. The difference in direction is nothing; the difference in size is about a tenth of a percentage point. These are overlapping, non-independent draws from one market in one era — a description of what those days looked like, never a rule. |
| by release, measured on this build | CPI n=206 median +0.12% (p10 -4.64, p90 +4.96, 52% up) · PPI n=204 median +0.01% (p10 -3.31, p90 +4.13, 51% up) · jobs report n=199 median +0.15% (p10 -3.99, p90 +4.00, 51% up) · GDP n=219 median +0.00% (p10 -3.35, p90 +4.20, 51% up) · PCE n=196 median +0.39% (p10 -3.19, p90 +5.01, 55% up) · retail sales n=205 median +0.16% (p10 -3.74, p90 +6.26, 52% up) · FOMC n=137 median +0.30% (p10 -3.79, p90 +5.06, 58% up) |
| 🩸 a day with two prints is one bar | 137 of the scored days carry more than one release — GDP and PCE share a morning, and some months put CPI and the FOMC in the same week. The «every print day» view holds one bar per DAY and names what printed on it in the table; each per-release view holds its own days, so the same bar can appear in two of them. Adding the per-release counts does not give the total, and is not meant to. |
| 🩸 the holes we know about | the longest silence between two prints since our tape begins is 28 days, 2025-09-26 → 2025-10-24. Before 1994 the FOMC did not announce its decision on the day it met, so that row starts there and no earlier, and decisions taken between meetings by conference call from 1994 to 2012 are missing from it. Announced future dates reach the end of 2026 for the monthly series — the 2027 agency schedules are not published yet — and 2027-12-15 for the FOMC. One release people watch, the ISM survey, is not here at all: its publisher answers with a captcha instead of a page, and a time we cannot quote is a time we do not print. |
| 🩸 months with no print dated in them | between each monthly release's first and last print, the months that carry no print date: CPI 4, jobs 8, PPI 3, retail sales 2, PCE 80. Most are a print that fell into the month next door, which then carries two dates. Left unpaired once each month with two dates is matched to one empty neighbour: CPI 1 (2025-11); jobs 1 (2025-10); PPI 2 (2004-02, 2025-10); retail sales 2 (2019-01, 2025-10); PCE 21 (1994-02, 1995-02, 1995-11, 1995-12, 1996-02, 1997-01, 1998-01, 1999-01, 1999-09, 2001-01, 2001-09, 2004-10, 2005-07, 2007-01, 2009-01, 2010-01, 2010-09, 2013-10, 2015-01, 2019-01, 2025-10). GDP is quarterly and the FOMC meets eight times a year, so neither is counted this way. |
| 🩸 the clock, for the old dates | the DATE of every print is the publisher's own. The TIME is today's rule applied to the whole history — 08:30 New York for the three statistical offices, 14:00 for the FOMC statement — converted to UTC on each date, which is why the same release reads 12:30Z in summer and 13:30Z in winter. For dates from the 1940s to the 1990s that hour is the rule, not a quoted fact; the date is. |
| what differs | a calendar you can read against the tape. We confirmed 12 dates against two sources and identify the single source used for the remaining dates. The publication minute is carried in UTC and in New York time, the dates the agencies have already announced are here, and every number beside them is measured on our own daily leg with its sample size printed next to it. |
| thinner here | no forecast values, no consensus, no surprise — this is WHEN things print, not what they will say. The reaction is measured over a whole UTC day, because that is the resolution our price history actually has. And nothing here is scored per release beyond the plain distribution: 137 to 219 days is a small pile for any one of them. |
| policy | no export · dates only, never a forecast of the number itself · read it here · read on the site · not exported: this page is drawn from someone else's archive |
| this build | data to Oct-15 2026 |
| policy | no backfill · no export · read it here · read on the site · not exported: this page is drawn from someone else's archive |
The chart itself is drawn in the browser: turn JavaScript on to hover a day, pick a window and draw your line. The numbers above come from the last build, Oct-23 2026 21:24Z.
What the tape says today · what's new · all 111 metrics