Is there a best day of the week to buy Bitcoin?
We ran seven buying plans side by side: $100 every Monday, $100 every Tuesday, and so on through the week. That is dollar-cost averaging, or DCA, which just means buying the same amount on a fixed schedule whatever the price is doing. Each plan ran over the full Bitcoin price history, from Aug 18, 2010 to Aug 17, 2026. Same amount, same number of purchases, same fees. The gap between the best and worst day is 6.22%, and the day that “wins” changes every time you change the date range. That makes it noise. Below, we show you exactly where the noise comes from.
The short answer: no, and the data below is why
Over the full history the best weekday is Friday. Over last 2 years it is Thursday. Nothing about Bitcoin's market changed to cause that. A real day-of-week effect would show up in every window. A coin flip would not, and this behaves like a coin flip. The full-history figure is worse than merely noisy. About 64% of every schedule's Bitcoin came from its first 10 buys, all made in August 2010. And seven weekdays cannot share a start date, so the seven plans had to start on seven different days of that week. What the column ranks is those start dates, more than the weekdays. Publish the number, then ignore it.
Best day, full history
Friday
6.22% ahead of Monday. An artifact, see below
Best day, last 2 years
Thursday
0.55% ahead of Wednesday
Spread, last 5 years
0.23%
Best vs worst weekday
Purchases per schedule
834
$83,400 invested, identical for all seven
All seven weekdays, full price history
$100 per week, 834 purchases each ($83,400 invested), no fees. First buys land in the week of Aug 18, 2010; last buys land in the week of Aug 11, 2026.
| Buy day | BTC accumulated | Average cost / BTC | Value today | vs. average day |
|---|---|---|---|---|
| Monday | 21,043.70 | $3.96 | $1,337,595,433 | -3.01% |
| Tuesday | 21,160.76 | $3.94 | $1,345,035,867 | -2.47% |
| Wednesday | 22,291.26 | $3.74 | $1,416,893,803 | +2.75% |
| Thursday | 21,791.26 | $3.83 | $1,385,112,187 | +0.44% |
| Fridaymost BTC | 22,352.88 | $3.73 | $1,420,810,592 | +3.03% |
| Saturday | 21,699.03 | $3.84 | $1,379,249,595 | +0.02% |
| Sunday | 21,531.05 | $3.87 | $1,368,572,434 | -0.76% |
Read this table carefully. The ranking is an accident of one week in 2010. About 64% of the Bitcoin in every row came from that row's first 10 purchases. That is $1,000 of the $83,400 invested doing most of the work. Those buys all landed in August 2010, when the price was quoted in whole cents and one cent was a big slice of the whole price. Nudge that price by a single cent and the Bitcoin you walk away with changes by a double-digit percentage. That wobble is bigger than the 6.22% spread this entire table reports.
Then there is the second problem. Seven weekly plans cannot all begin on the same date, because only one weekday falls on any given day, so they start on seven days in a row. Over the full history those seven days sit inside that same August 2010 week, at those same penny prices. So the order of this table is largely a ranking of start dates, not of weekdays. The value column is the arithmetic of a fifteen-year schedule nobody actually ran. It is here for completeness, not as a target. The window table below is the honest test.
The winner changes with the window
Rank of each weekday by BTC accumulated (1 = most BTC) across four different lookback windows, each run the same way. If a weekday genuinely bought Bitcoin cheaper, its column of ranks would stay near the top. None of them do.
| Buy day | Full history | Last 10 years | Last 5 years | Last 2 years |
|---|---|---|---|---|
| Monday | 7 | 7 | 6 | 6 |
| Tuesday | 6 | 1 | 4 | 5 |
| Wednesday | 2 | 3 | 7 | 7 |
| Thursday | 3 | 2 | 3 | 1 |
| Friday | 1 | 4 | 2 | 4 |
| Saturday | 4 | 5 | 1 | 3 |
| Sunday | 5 | 6 | 5 | 2 |
| Best vs worst | 6.22% | 1.36% | 0.23% | 0.55% |
| Purchases each | 834 | 522 | 261 | 104 |
Each column is worked out from scratch: seven schedules, one per weekday, with the same number of purchases and the same total invested inside that column. The seven have to begin on seven days in a row, and so they also end on seven days in a row, since no two weekdays share a date. The windows overlap, so these columns are not independent samples in the statistical sense. That makes the reshuffling more damning, not less. Heavily overlapping data should produce similar rankings if the effect were real.
Why there is no best day, and why the question keeps getting asked
Bitcoin never closes. Stock markets have a real calendar: a weekend gap, an opening auction, index rebalancing on set dates. Bitcoin trades every hour of every day on hundreds of venues. Nothing about it makes one weekday reliably cheaper. If something did, exploiting it would cost almost nothing, so it would be gone within days.
Seven buckets, one dataset. Split any jumpy price history into seven arbitrary groups and one of them will come out on top. That is arithmetic, not a discovery. The real test is whether the same group keeps winning when you change the date range. Here it does not.
The gap is smaller than your fees. Over last 5 years the best and worst weekdays differ by 0.23%. Now compare that to what an exchange charges. Commission plus the spread, meaning the gap between the price you buy at and the price you could sell at, commonly runs 0.5% to 1.5% per purchase. Moving from a 1.49% fee tier to a 0.4% one changes your outcome by more than every weekday decision you will ever make. Unlike the weekday, that saving is real and it repeats.
Consistency beats timing. What really decides a DCA outcome is how many purchases you actually make. A missed month costs you a whole contribution. Waiting for the “right” day is how missed months happen. Pick the day after payday, automate it, and stop thinking about it.
How these numbers were computed
For each window we find the first UTC date on or after the window start whose weekday matches. A weekly schedule then runs from that date: $100 per purchase, priced at that day's closing price, zero fees, using the same engine as the main calculator. We cap the number of purchases at the lowest of the seven, so all seven schedules invest exactly the same total. All dates are bucketed in UTC. Exchange candles are UTC-aligned, and doing this in local time would shift the whole schedule by a day for anyone west of Greenwich. Prices are Coinbase daily candles from July 2015 onward, and real daily market prices from blockchain.info for August 2010 to mid-2015. Value today uses the latest spot price, so every row is valued at the same price and the ranking by value is identical to the ranking by BTC. Figures refresh once every 24 hours; the data covers Aug 18, 2010 to Aug 17, 2026. Full detail on the methodology page.
Common questions
Is there a best day of the week to buy Bitcoin?
No, not one you can rely on. We ran seven identical $100/week schedules from Aug 18, 2010 to today. The gap between the best and worst weekday is 6.22% of the Bitcoin accumulated. More importantly, the winner is not stable: Friday leads over the full history, but Thursday leads over last 2 years. A real edge would show up in every window. This one does not.
Which weekday came out ahead over the full history?
Friday, by 6.22% over the worst weekday (Monday). That number does not mean what it looks like it means. About 64% of every schedule's final Bitcoin came from its first 10 purchases. Those buys were made in August 2010, when Bitcoin was quoted in whole cents and a single cent was a big slice of the entire price. So one cent of rounding moves the Bitcoin you end up with by a double-digit percentage, which is more than the 6.22% spread being reported. The seven schedules also begin on 7 different days, because no two weekdays share a date. What the full-history table really ranks is which of those start dates caught the cheapest penny prints. It describes one week in 2010. It does not predict anything about weekdays.
Why do people say weekends are cheaper for Bitcoin?
The usual story is that big trading desks and bank transfers shut down at the weekend, so fewer buyers and sellers are around and prices drift. But Bitcoin trades every hour of every day. Any weekend discount that stuck around would be free money, so traders would close it almost immediately. In the numbers on this page the weekend days do not consistently beat weekdays across time windows.
How much money is a weekday difference actually worth?
Over last 5 years, $26,100 invested produced a best-to-worst gap of $89.55 across the seven weekdays. A single 1% exchange fee, or one missed contribution, dwarfs that.
What should I optimize instead of the weekday?
Four things you control actually move the outcome. How much you put in. How long you keep going. The fees you pay. And whether you stop buying during a drawdown, meaning a long stretch when the price sits well below its peak. Pick whichever weekday you are least likely to forget, usually the day after you get paid, and automate it.
How are these numbers calculated?
Each schedule buys $100 at the daily close on its anchor weekday, in UTC, with no fees. Every schedule makes exactly the same number of purchases (834 over the full window), so total invested is identical. The weekday is not the only thing that differs, though. Seven weekdays cannot share a start date, so the seven schedules begin on 7 consecutive days. In a recent window that does not matter. Over the full history those days fall in August 2010, and about 64% of every schedule's final Bitcoin came from its first 10 purchases. So the full-history ranking mostly reflects which start date caught the cheapest days, not which weekday is better. The figures are recomputed once a day against fresh prices.
Run your own numbers
Pick any weekday you like. It barely matters. What does matter is the amount, the fee, and how long you keep going. Change all three in the calculator and see what the same history does to your actual plan.
Open the calculatorKeep reading
Not financial advice. This page is a historical simulation published for educational purposes. Past performance does not guarantee future results, and a pattern found in historical data is not a prediction. That is the entire point of this page. Bitcoin is volatile and you can lose money.