The Federal Reserve’s October meeting opens today, Tuesday 27 October, and the decision comes on Wednesday 28 October at 2:00 p.m. ET, with Chair Kevin Warsh’s press conference at 2:30 p.m. ET. Traders put the odds of a hold at around nine in ten, but a hike in December is priced as likely: CCN, citing CME FedWatch via MarketWatch on 26 October, shows about 9% for a hike this week and 73% for December. Bitcoin enters the meeting at about $79,470 (Coinbase, 07:04 UTC), roughly 0.7% below Monday’s close, which raises a question traders ask every cycle: does bitcoin do anything on the day before the decision?
Informational only, not investment advice. See the full disclaimer at the end.
The answer in one paragraph
No — not on average, and not in a way that survives a statistical test. Across the 38 FOMC meetings from January 2022 to September 2026, bitcoin’s return on the first day of the meeting (the day before the decision) averaged −0.18%, with a standard error of 0.44 percentage points and a median of −0.03%. It rose on 18 of the 38 days. The average absolute move was 1.85%, against 1.80% on all days. If you drew 38 random days from the same period, you would see an average at least as low as −0.18% about 27% of the time. In plain language, the day before the Fed looks like any other day.
Video: a 16 October 2026 preview of this meeting from an independent channel. Opinion, not our analysis.
Method
We used the decision date of each scheduled FOMC meeting from 26 January 2022 to 16 September 2026, 38 meetings in all. The 22 August 2025 notation vote is excluded, and 37 of the 38 decision days are Wednesdays (the exception is Thursday 7 November 2024). “Day 1” is the calendar day before the decision, which is the first day of the two-day meeting. Returns are daily close-to-close changes in Yahoo Finance’s BTC-USD series (UTC days), from 2 January 2022 to 25 October 2026, 1,758 daily returns in all. As a check, the same data reproduce the decision-day figures from our earlier study, How Does Bitcoin Trade on Fed Decision Days?: +0.49% with a standard error of 0.51 points. One caution about timing: a UTC day closes at 00:00 UTC, which is 7:00 or 8:00 p.m. ET depending on daylight saving, so the decision-day return includes the afternoon reaction and the day-1 return does not.
Results: nothing stands out on day 1
| Window | Meetings | Mean return | Std. error (pts) | Median | Up days | Mean absolute move |
|---|---|---|---|---|---|---|
| All days (benchmark) | 1,758 | +0.06% | 0.06 | −0.03% | 867 | 1.80% |
| Tuesdays (benchmark) | 251 | 0.00% | 0.16 | −0.05% | 119 | 1.92% |
| Day 1: the day before the decision | 38 | −0.18% | 0.44 | −0.03% | 18 | 1.85% |
| Decision day | 38 | +0.49% | 0.51 | +0.14% | 21 | 2.23% |
| Day after the decision | 38 | −0.46% | 0.50 | +0.03% | 19 | 2.20% |
| Two days to decision close (day 0 close → decision-day close) | 38 | +0.29% | 0.60 | +0.54% | 22 | 2.94% |
Source: Bitcoin Mastery calculation on Yahoo Finance BTC-USD daily closes, 2 Jan 2022–25 Oct 2026. Fed meeting dates from the Federal Reserve calendar. “Up days” counts positive returns.
Three things stand out. First, the day-1 average is smaller than its own uncertainty: a mean of −0.18% with a standard error of 0.44 means the data cannot distinguish it from zero. Second, volatility is unremarkable: the standard deviation of day-1 returns is 2.74%, against 2.64% on all days, and the share of days with a move of 3% or more in either direction is 8 of 38 (21%), against about 18% in the full sample (roughly 6.8 of 38). Third, the decision day is where the larger moves live: its mean absolute move is 2.23%, and the day after is 2.20%, both above the 1.80% benchmark, although with 38 observations each that gap is also within the noise.
The range is where the risk sits
An average of zero does not mean nothing happens. The eight day-1 moves of 3% or more were: −8.34% (19 March 2024), −5.02% (30 April 2024), −3.34% (20 September 2022), −3.26% (15 September 2026), −3.14% (11 June 2024), +3.34% (13 December 2022), +3.64% (17 September 2024) and +9.05% (6 November 2024). The last is the day after the US presidential election, and it drives the sample. Removing it (a post-hoc choice, which we flag) turns the other 37 meetings into an average of −0.43% with a standard error of 0.38 points and 17 up days: slightly more negative, and still not statistically significant. Whichever way you cut it, there is no reliable edge, and one extreme day can change the sign of the average.
Has the pattern changed? Two eras
| Period | Meetings | Day-1 mean | Std. error (pts) | Up days | Mean absolute move |
|---|---|---|---|---|---|
| January 2022 – December 2023 | 16 | +0.24% | 0.41 | 10 | 1.25% |
| January 2024 – September 2026 | 22 | −0.49% | 0.71 | 8 | 2.28% |
| 2026 only (post hoc, tiny sample) | 6 | −0.95% | n/a | 2 | n/a |
Source: Bitcoin Mastery calculation as above. The split at 2024 is our choice, made after seeing the data; treat it as description rather than a test.
The more recent period is both lower on average and noisier, with a mean absolute move of 2.28% versus 1.25% in 2022–23. Six of the 22 meetings since 2024 had day-1 moves of 3% or more. But the standard error of 0.71 points means the −0.49% is no more distinguishable from zero than the full-sample number. In 2026, bitcoin fell on the day before four of the six meetings, including −3.26% on 15 September, the day before the hike; six observations is a story, not a statistic.
Chart: bitcoin over one month, TradingView widget (live data).
What this does and does not tell you about tomorrow
- It does not say bitcoin will fall or rise today. A −0.18% mean with a ±0.44 standard error covers a range from roughly −1.1% to +0.7% at two standard errors, and individual days have moved ten times that.
- It says day-before-Fed risk is ordinary. The average size of move is within 0.05 points of any other day, so the case for treating Tuesday as special is weak.
- The surprise happens on the decision day. Decision days and the day after are the windows with above-benchmark absolute moves, consistent with our earlier finding that bitcoin’s reaction depends on what is surprising, not on the date.
- This meeting is a bigger test than average. A hold is the base case, but the December-hike pricing, Microsoft and Alphabet reporting that evening, and Thursday’s PCE put three catalysts in 28 hours. We covered that in our Fed-week preview, and our range-compression study showed bitcoin’s 30-day range had narrowed to about 10% before the week began.
Limits
- Only 38 meetings: the standard errors are wide, and a five-point move on one day shifts the mean by 0.13 points.
- Daily closes at 00:00 UTC. The decision (18:00 UTC) falls inside the decision-day candle, so the day-1 window does not contain it.
- Yahoo Finance and Coinbase closes differ by small amounts, so the figures here will not match Coinbase candles exactly.
- Several day-1 observations coincide with other events (the 6 November 2024 observation follows a US election), and we have not tried to separate them.
- Splits by era and the election exclusion were chosen after looking at the data, which makes them descriptive only.
We will add the 28 October meeting to this table once the decision-day candle closes, and report whether Tuesday’s move, whatever it turns out to be, changes any of these averages.
Disclaimer: This article is for informational purposes only and does not constitute investment advice. Bitcoin and other cryptocurrencies are volatile and you can lose some or all of the money you put in. Nothing here is a recommendation to buy, sell or hold bitcoin, any exchange-traded fund, any listed security or any other asset, and the technical levels, probabilities and scenarios discussed are descriptions of published data, not forecasts. Do your own research and consult a licensed financial advisor before making investment decisions.