Bitcoin and the US Data Calendar

Research · Scheduled events · November 2026

We tested five scheduled events against how often Bitcoin made large price moves, under rules fixed before looking. One held up on both recent and older data: the hours after a US inflation release. The jobs report came close but did not clear the bar set in advance; producer prices and options expiry showed nothing reliable.

Two of these events, the inflation release and the Federal Reserve statement, were reported in When Bitcoin Moves Most. This note sets them beside three more scheduled events, tested the same way, and gives the significance figures behind all five.

The question

Scheduled US news is one candidate for when Bitcoin's larger price moves cluster. We asked a narrow question: in the hours around each scheduled event, were unusually large moves more frequent than Bitcoin's recent volatility would suggest?

An unusually large move means a price range bigger than four in five of the comparable ranges over the previous month, so it happens about one time in five. Each test compared the event hours with a model of recent volatility, time of day and day of week. It measured the size of moves, not their direction.

We tested five events: US consumer prices (CPI), Federal Reserve rate statements, the US jobs report, US producer prices (PPI) and Deribit's monthly Bitcoin options expiry. The data were 30-minute Bitcoin prices from January 2020 to mid-June 2026.

Each idea was tested once on January 2024 to mid-June 2026, and anything that passed was checked once, unchanged, on 2020 to 2023. The Federal Reserve statement is the exception: an earlier study had already looked at it on 2024 to 2026, so its fresh test was run on 2020 to 2023 instead.

What we found

One event held up on both periods, one on the older period only, and three did not.

A point is a percentage point of probability. For scale: if recent conditions implied a one-in-five chance of a large move, 25 to 30 points more is roughly one in two.

A year-by-year look at the CPI result, taken after the test, shows it near zero in 2020 and 2021 and present from 2022.

No event told us which way the price would move. Our earlier tests of direction, reported in When Bitcoin Moves Most, found no measure that predicted it.

Why the jobs report is not a finding yet

The jobs report produced the largest effect in recent data: large four-hour moves about 32 points more frequent, with 25 of 29 releases pointing the same way. It still does not count, for three reasons.

It missed the bar set in advance. On 2020 to 2023 the effect was about 10 points with p = 0.028, against a threshold of 0.025 fixed before the test. A near miss is a miss.

Part of it is Friday morning. 25 of the 29 recent jobs reports fell on a Friday; the other four were moved by a public holiday or the 2025 and 2026 government shutdowns. On ordinary Fridays with no release, the same 08:30 New York hours were already about 12 points above our model in 2024 to mid-2026. In the first half of 2026 the release hours were no higher than an ordinary Friday.

It looks like CPI, which is the problem. Like CPI, the effect was absent in 2020 and 2021 and appeared from 2022. Testing only the years since 2022 now would mean choosing the period after seeing the result.

The fair test is on releases that have not happened yet. We have logged it, to be fixed before any new data is seen, and will report the result either way once 36 new jobs reports have been released, counting from the first release after the test is locked. The lock date is not yet set; we will publish it when it is, and the report follows about three years later.

How we tested it

1. Dates from official calendars. CPI, jobs and PPI release dates and times from the US Bureau of Labor Statistics schedules, cross-checked against FRED. Federal Reserve statements from the Fed's meeting calendar. Options expiry from Deribit's own instrument records: every monthly expiry from December 2019 to May 2026, the last inside our price data, settled at 08:00 UTC on the last Friday.
2. Rules written down first. Each window, threshold and outcome was fixed and locked before any result was computed.
3. Compared with a shuffled market. Each result was set against thousands of copies of the data with the event windows slid to random dates, and against false calendars of the same weekday and clock.
4. Corrected for testing several ideas. Recent-data results had to clear a family-wide error rate of 10% across the four or five rules tested together in each study, using a step-down permutation correction. Older-period checks had to clear 5%, divided by the number of rules checked together: CPI was checked alone (bar 0.05); the Federal Reserve statement was tested with one other Federal Reserve rule, and the jobs report and PPI were checked as a pair (bar 0.025 each). With 2,000 shuffles the smallest p-value the test can return is about 0.0005, so the three results shown at 0.0005 mean no shuffle matched them, not an exact figure.
5. One look, then one check. Each idea was tested once on recent data (except the Federal Reserve statement, above); anything that passed was re-tested once, unchanged, on 2020 to 2023.
6. Reviewed for faults. A separate automated reviewer checked each design before it was locked and each result after it was run. The CPI result was also rebuilt by a separate automated process from a written description; all figures matched.

The View From The Edge

One date in the US calendar measurably stood out. Across five scheduled events, only the US inflation release showed a rise in Bitcoin's large moves on both recent and older data.

Why a holder would care. For a family, a company treasury or an executor moving a large amount of Bitcoin, the hours after 08:30 New York on CPI day are a known, scheduled period in which prices have moved more widely than usual, on our data mostly since 2022 (a split seen after the test). That is a fact about the calendar, not a view on price.

Four of the five did not hold up. The jobs report looked as strong as CPI on recent data and still failed the check we set in advance. Producer prices and options expiry showed nothing we could rely on. We report those results as fully as the one that held.

Size is not direction. Knowing when large moves cluster says nothing about which way they go, and none of these tests found anything that did.

Related reading: When Bitcoin Moves Most, The Alderlux Null Test and the Alderlux Research Library.

Basis and limits

This piece reports tests of whether scheduled US events and options expiry changed how often Bitcoin made large price moves. It is not a forecast, and it says nothing about which way prices move.

Range, not direction. Every result describes how often large moves happened, not which way the price went.

Small samples. Four of the five events have 29 occurrences in recent data (the Federal Reserve has 19), and 30 to 47 usable occurrences in the older period. The first CPI release of 2020 is lost to the model's warm-up, which is why 2020 to 2023 has 47 rather than 48. Only large effects, roughly 8 to 20 points, can be detected reliably.

Era dependence. The CPI effect is near zero in 2020 and 2021 and present from 2022, a split seen after the test. Why it switched on is not tested here, so it may also switch off; its weakest year since is 2026 to date, at about 16 points.

Clock effects. Our model knows the hour and the weekday separately, but not Friday mornings in New York, which were above it in recent data. That mainly affects the jobs report.

Expiry day, not settlement alone. The last Friday of the month also carries other futures settlements and month-end flows, so the expiry tests measure the whole day.

One market. Prices are Binance's BTCUSDT pair, quoted in a stablecoin, on one exchange. Other venues and the US dollar pair were not tested.

Automated checks. The reviews and the CPI rebuild were carried out by separate automated processes from written descriptions, on the same data. They confirm the calculation, not the choice of method.

Past periods only. Data from July 2026 onwards are held back for later checks.

Sources

1. US Bureau of Labor Statistics, release schedules 2020 to 2026: bls.gov/schedule
2. US Bureau of Labor Statistics, revised news release dates following the 2025 and 2026 lapses in appropriations: bls.gov
3. Federal Reserve, FOMC meeting calendars: federalreserve.gov
4. Federal Reserve Bank of St. Louis, FRED release and vintage dates: fred.stlouisfed.org
5. Deribit, public and historical instrument records: deribit.com and history.deribit.com
6. Binance Vision public market data, BTCUSDT 30-minute prices: data.binance.vision

Alderlux provides education and information only. We do not provide regulated financial advice, investment recommendations, legal or tax advice. Alderlux is not authorised by the Financial Conduct Authority. Past performance is not indicative of future results.