The Best Days to Buy Bitcoin

Buying Bitcoin on Holidays Delivers 4x Higher Next-day Returns

In this study, we explored the forward return of buying Bitcoin on different calendar days over the past 13 years (May 1, 2013 to May 8, 2026), focusing on single day returns. Without accounting for holidays, Mondays and Wednesdays lead the week with an identical average next-day return of +0.38%, while Thursday is the only day of the week to average a negative return at -0.09%.

The most surprising finding is how dramatically the US holiday effect stands out against the rest of the data. US holidays register a +0.77% average next-day return compared to the non-holiday average of +0.19% — holidays outperformed non-holidays in 11 out of 14 calendar years in the dataset.

The Best Holidays to Buy Bitcoin 

New Year's Day is the strongest single holiday in the dataset, averaging a +2.01% next-day return across all 13 January 1st observations, with an 84.6% win rate — meaning Bitcoin was up the following day in 11 of 13 instances. Columbus Day matches that 84.6% win rate with a +1.70% average return, while Christmas delivers +1.46% on a 53.8% win rate.

The New Year's Day effect likely captures a broader January momentum effect well-documented in traditional markets, where fresh capital allocations at the start of the year lift risk assets. For Bitcoin specifically, the transition from December tax-loss selling to January positioning may amplify this dynamic. The price on January 1st has ranged from $313 in 2015 to $93,507 in 2025 — and across that entire range, the pattern of next-day strength has held with notable consistency.

Not All Holidays Are Equal

Two US holidays average negative next-day returns: MLK Day at -0.84% and Independence Day at -0.26%, both with win rates below 50%. MLK Day's average is dragged significantly by a single outlier — January 15, 2018, when Bitcoin fell -18.65% the following day during the onset of the crypto bear market. Even adjusted for this, MLK Day remains the weakest holiday in the dataset.

Veterans Day warrants a separate caveat. Its average of +1.75% appears strong, but its win rate is only 46.2%. Three large positive outliers (November 2014 at +14.3%, November 2015 at +10.4%, and November 2024 at +10.2%) are responsible for the elevated average, masking a result that is otherwise unreliable. Median return and win rate are better guides for Veterans Day than the mean.

Weekday vs Weekend: The Difference Is Negligible

Weekdays average a +0.21% next-day return compared to +0.22% for weekends — a difference of just 0.01 percentage points, which carries no practical significance. Unlike equity markets, Bitcoin never closes, which means the structural forces that typically drive weekend effects in traditional finance do not apply in the same way. Saturday and Sunday behave similarly to Tuesday and Friday — neither meaningfully better nor worse than the weekly average.

Over a Full Year, When You Buy Barely Matters

The short-term day-of-week effect dissolves almost entirely at a 365-day holding period. All seven days of the week cluster within a narrow band of 142.15% to 144.56% average annual return — a spread of just 2.4 percentage points across the full range, negligible relative to Bitcoin's annual volatility. For long-term holders: the day of purchase has no meaningful impact on a one-year outcome.

The holiday effect persists slightly longer. US Holidays average a 157.12% return when held for one year versus 142.96% for non-holidays — a 14 percentage point gap. However, this should be interpreted with caution: at a one-year horizon, the result largely reflects whether a given holiday fell in a bull or bear market year, rather than any genuine holiday-specific effect. The genuine holiday signal lives in the short term.

A Theoretical Note on Holiday DCA

If an investor had chosen to concentrate their Bitcoin purchases exclusively on US federal holidays since 2013 — buying once on each of the approximately 11 annual holidays — they would have captured the +0.77% average next-day return on each purchase versus the +0.19% baseline. Over 149 such purchases, that represents a meaningfully higher short-term entry quality. In practice, most investors buy more frequently than 11 times per year, making this a framing exercise rather than a replicable strategy. The more actionable takeaway is simply that if a planned purchase coincides with a US holiday, history suggests it is not a bad day to execute.

Methodology

This study analyzed daily Bitcoin price data sourced from CoinGecko, covering May 1, 2013 to May 8, 2026, comprising 4,753 observations. Each data point represents the Bitcoin price in USD at the 00:00 UTC daily snapshot. All day-of-week and date references reflect UTC time, which means a "Monday" purchase in this study corresponds to UTC midnight — approximately Sunday evening for US Eastern Time users. This is noted for transparency but does not materially affect the findings.

Forward returns were calculated using exact calendar date lookup: the N-day return for a given date is computed as (price on date + N calendar days) / (price on date) − 1. Observations at the tail end of the dataset where no future price existed were excluded from return calculations.

US federal holidays were identified using the official US federal holiday calendar. Only canonical holiday dates were included; "observed" variants (e.g. a holiday shifted to Monday when the canonical date falls on a weekend) were excluded to avoid double-counting and to keep the analysis anchored to the actual calendar date. All 10 federal holidays each appear exactly 13 times within the study window

Win rate is defined as the percentage of observations where the forward return was positive. All averages are arithmetic means unless stated otherwise.

Summary: The Best and Worst Days to Buy Bitcoin

Raw Data

Table 1: Day-of-Week Returns

Table 2: Weekdays vs Weekends

Table 3: US Holidays vs Non-Holidays