Bitcoin DCA start date heatmap
Every cell below is the same plan: $100 on the first of every month, no fees, starting in January of the row's year and running for the column's number of years. The number is the annualized return over that window, meaning the yearly rate of growth that gets you from what went in to what came out. It is measured at the Bitcoin price on the day the window ended, not today's price. The exact method is XIRR, which accounts for the money going in a bit at a time instead of all on day one.
Read down a column and you see how much the start year mattered. Completed 1-year windows landed anywhere from -73.2% to +4,278% a year. Completed 10-year windows sat between +62.2% and +176%. The 6 completed 10-year windows start a year apart and overlap: only 1 of that length fits in the data without sharing a year.
Annualized return by start year and holding period
| Started | 1 yr | 2 yrs | 3 yrs | 5 yrs | 7 yrs | 10 yrs |
|---|---|---|---|---|---|---|
| 2011 | +690% | +307% | +1,023% | +240% | +298% | +176% |
| 2012 | +223% | +1,656% | +319% | +179% | +148% | +143% |
| 2013 | +4,278% | +165% | +100% | +233% | +101% | +73% |
| 2014 | -58.4% | +24.9% | +71.7% | +81.5% | +111% | +70.1% |
| 2015 | +143% | +142% | +405% | +101% | +119% | +80.7% |
| 2016 | +197% | +696% | +94.7% | +133% | +55.4% | +62.2% |
| 2017 | +1,857% | +24.5% | +50% | +99.4% | +53.6% | In progress |
| 2018 | -73.2% | +9% | +114% | +17.7% | +56% | In progress |
| 2019 | +40% | +208% | +131% | +49.3% | +45.8% | In progress |
| 2020 | +493% | +150% | -6.7% | +59.6% | In progress | In progress |
| 2021 | +18.2% | -52.3% | +25.5% | +32.4% | In progress | In progress |
| 2022 | -60% | +57.5% | +82.3% | In progress | In progress | In progress |
| 2023 | +137% | +121% | +45.8% | In progress | In progress | In progress |
| 2024 | +116% | +16.4% | In progress | In progress | In progress | In progress |
| 2025 | -19.7% | In progress | In progress | In progress | In progress | In progress |
Legend: annualized return
- 200%+ a year
- 100% to 200%
- 50% to 100%
- 20% to 50%
- 0% to 20%
- Negative
- Window not finished yet
Color is only a shortcut. Every completed cell prints its own number, so the grid reads the same with the colors ignored.
Every completed window in dollars: invested, end value and total return
| Window | Invested | Buys | End value | Total return | Annualized |
|---|---|---|---|---|---|
| 2011–2012 · 1 yr | $1,200 | 12 | $4,374 | +264% | +690% |
| 2011–2013 · 2 yrs | $2,400 | 24 | $14,188 | +491% | +307% |
| 2011–2014 · 3 yrs | $3,600 | 36 | $780,675 | +21,585% | +1,023% |
| 2011–2016 · 5 yrs | $6,000 | 60 | $467,681 | +7,695% | +240% |
| 2011–2018 · 7 yrs | $8,400 | 84 | $14,510,203 | +172,641% | +298% |
| 2011–2021 · 10 yrs | $12,000 | 120 | $31,675,049 | +263,859% | +176% |
| 2012–2013 · 1 yr | $1,200 | 12 | $2,404 | +100% | +223% |
| 2012–2014 · 2 yrs | $2,400 | 24 | $145,462 | +5,961% | +1,656% |
| 2012–2015 · 3 yrs | $3,600 | 36 | $64,552 | +1,693% | +319% |
| 2012–2017 · 5 yrs | $6,000 | 60 | $206,550 | +3,342% | +179% |
| 2012–2019 · 7 yrs | $8,400 | 84 | $798,932 | +9,411% | +148% |
| 2012–2022 · 10 yrs | $12,000 | 120 | $9,984,046 | +83,100% | +143% |
| 2013–2014 · 1 yr | $1,200 | 12 | $15,900 | +1,225% | +4,278% |
| 2013–2015 · 2 yrs | $2,400 | 24 | $7,741 | +223% | +165% |
| 2013–2016 · 3 yrs | $3,600 | 36 | $12,515 | +248% | +100% |
| 2013–2018 · 5 yrs | $6,000 | 60 | $426,647 | +7,011% | +233% |
| 2013–2020 · 7 yrs | $8,400 | 84 | $229,699 | +2,635% | +101% |
| 2013–2023 · 10 yrs | $12,000 | 120 | $535,090 | +4,359% | +73% |
| 2014–2015 · 1 yr | $1,200 | 12 | $769.30 | -35.9% | -58.4% |
| 2014–2016 · 2 yrs | $2,400 | 24 | $3,052 | +27.1% | +24.9% |
| 2014–2017 · 3 yrs | $3,600 | 36 | $9,233 | +156% | +71.7% |
| 2014–2019 · 5 yrs | $6,000 | 60 | $38,617 | +544% | +81.5% |
| 2014–2021 · 7 yrs | $8,400 | 84 | $306,393 | +3,548% | +111% |
| 2014–2024 · 10 yrs | $12,000 | 120 | $465,855 | +3,782% | +70.1% |
| 2015–2016 · 1 yr | $1,200 | 12 | $2,007 | +67.3% | +143% |
| 2015–2017 · 2 yrs | $2,400 | 24 | $6,853 | +186% | +142% |
| 2015–2018 · 3 yrs | $3,600 | 36 | $101,527 | +2,720% | +405% |
| 2015–2020 · 5 yrs | $6,000 | 60 | $56,663 | +844% | +101% |
| 2015–2022 · 7 yrs | $8,400 | 84 | $384,147 | +4,473% | +119% |
| 2015–2025 · 10 yrs | $12,000 | 120 | $769,957 | +6,316% | +80.7% |
| 2016–2017 · 1 yr | $1,200 | 12 | $2,278 | +89.8% | +197% |
| 2016–2018 · 2 yrs | $2,400 | 24 | $39,417 | +1,542% | +696% |
| 2016–2019 · 3 yrs | $3,600 | 36 | $11,818 | +228% | +94.7% |
| 2016–2021 · 5 yrs | $6,000 | 60 | $100,375 | +1,573% | +133% |
| 2016–2023 · 7 yrs | $8,400 | 84 | $57,902 | +589% | +55.4% |
| 2016–2026 · 10 yrs | $12,000 | 120 | $316,104 | +2,534% | +62.2% |
| 2017–2018 · 1 yr | $1,200 | 12 | $8,493 | +608% | +1,857% |
| 2017–2019 · 2 yrs | $2,400 | 24 | $3,040 | +26.7% | +24.5% |
| 2017–2020 · 3 yrs | $3,600 | 36 | $7,149 | +98.6% | +50% |
| 2017–2022 · 5 yrs | $6,000 | 60 | $54,708 | +812% | +99.4% |
| 2017–2024 · 7 yrs | $8,400 | 84 | $54,671 | +551% | +53.6% |
| 2018–2019 · 1 yr | $1,200 | 12 | $629.84 | -47.5% | -73.2% |
| 2018–2020 · 2 yrs | $2,400 | 24 | $2,629 | +9.5% | +9% |
| 2018–2021 · 3 yrs | $3,600 | 36 | $14,369 | +299% | +114% |
| 2018–2023 · 5 yrs | $6,000 | 60 | $9,328 | +55.5% | +17.7% |
| 2018–2025 · 7 yrs | $8,400 | 84 | $59,093 | +603% | +56% |
| 2019–2020 · 1 yr | $1,200 | 12 | $1,448 | +20.6% | +40% |
| 2019–2021 · 2 yrs | $2,400 | 24 | $9,527 | +297% | +208% |
| 2019–2022 · 3 yrs | $3,600 | 36 | $16,777 | +366% | +131% |
| 2019–2024 · 5 yrs | $6,000 | 60 | $19,532 | +226% | +49.3% |
| 2019–2026 · 7 yrs | $8,400 | 84 | $42,018 | +400% | +45.8% |
| 2020–2021 · 1 yr | $1,200 | 12 | $3,592 | +199% | +493% |
| 2020–2022 · 2 yrs | $2,400 | 24 | $7,145 | +198% | +150% |
| 2020–2023 · 3 yrs | $3,600 | 36 | $3,242 | -10% | -6.7% |
| 2020–2025 · 5 yrs | $6,000 | 60 | $24,511 | +309% | +59.6% |
| 2021–2022 · 1 yr | $1,200 | 12 | $1,315 | +9.6% | +18.2% |
| 2021–2023 · 2 yrs | $2,400 | 24 | $1,213 | -49.5% | -52.3% |
| 2021–2024 · 3 yrs | $3,600 | 36 | $5,208 | +44.7% | +25.5% |
| 2021–2026 · 5 yrs | $6,000 | 60 | $13,274 | +121% | +32.4% |
| 2022–2023 · 1 yr | $1,200 | 12 | $755.02 | -37.1% | -60% |
| 2022–2024 · 2 yrs | $2,400 | 24 | $3,989 | +66.2% | +57.5% |
| 2022–2025 · 3 yrs | $3,600 | 36 | $10,382 | +188% | +82.3% |
| 2023–2024 · 1 yr | $1,200 | 12 | $1,979 | +64.9% | +137% |
| 2023–2025 · 2 yrs | $2,400 | 24 | $6,092 | +154% | +121% |
| 2023–2026 · 3 yrs | $3,600 | 36 | $6,795 | +88.8% | +45.8% |
| 2024–2025 · 1 yr | $1,200 | 12 | $1,868 | +55.7% | +116% |
| 2024–2026 · 2 yrs | $2,400 | 24 | $2,824 | +17.7% | +16.4% |
| 2025–2026 · 1 yr | $1,200 | 12 | $1,067 | -11.1% | -19.7% |
How much did the start date matter?
This is the same data read column by column. "Spread" is the gap between the best and worst completed window of that length, in percentage points of annualized return.
| Hold length | Worst | Best | Spread | Completed windows | Ended below zero |
|---|---|---|---|---|---|
| 1 year | -73.2%2018 | +4,278%2013 | 4,351 points | 1515 share no years | 4 of 15 |
| 2 years | -52.3%2021 | +1,656%2012 | 1,709 points | 147 share no years | 1 of 14 |
| 3 years | -6.7%2020 | +1,023%2011 | 1,030 points | 135 share no years | 1 of 13 |
| 5 years | +17.7%2018 | +240%2011 | 222 points | 113 share no years | 0 of 11 |
| 7 years | +45.8%2019 | +298%2011 | 252 points | 92 share no years | 0 of 9 |
| 10 years | +62.2%2016 | +176%2011 | 114 points | 61 shares no years | 0 of 6 |
Windows in a column start one year apart, so the longer ones overlap, and the small figure under each count is how many windows of that length fit end to end in the data. The 6 completed 10-year windows start a year apart and overlap: only 1 of that length fits in the data without sharing a year. The year under each worst and best figure is the start year it belongs to. Read the bottom rows as a description of one stretch of history, not as a sample.
What the grid actually shows
Below zero, column by column. Windows that ended worth less than was paid in: 4 of 15 at 1 year, 1 of 14 at 2 years, 1 of 13 at 3 years, and none of the completed 5-, 7- and 10-year windows. Nothing about the plan changed between the best and worst row of a column. Only the start year did.
The spread, shortest hold to longest. The gap between the best and worst start year was 4,351 points for 1-year windows and 114 points for 10-year windows. It was not narrower at every step: the 7-year column is spread wider than the one before. Each extra year adds purchases at prices the earlier ones never saw, so one entry price counts for less of the average cost. The long columns also hold fewer windows, those windows overlap, and each had to start long enough ago to have finished, so a narrow spread there is not a like-for-like comparison with the shortest column.
Where the largest figures sit. In every column the highest figure belongs to a start year of 2013 or earlier. On Jan 1, 2011 the price in our series was $0.30. The top rows describe a starting point at that price. The later rows are nearer to recent prices, and they are history too, not a forecast.
Every window here ended in the past. A cell fills in only once its full holding period has run out, and it is valued at the price on that end date. That matters more than it sounds. Valuing a completed 2013–2016 window at today's price would credit a three-year hold with everything that happened in the decade afterwards. Cells whose window is still running say "in progress" instead.
One thing this grid cannot tell you is what any future window will do. It covers a single asset over a single 15-year stretch, in which the price in our series went from $0.30 on Jan 1, 2011 to $88,738 on Jan 1, 2026. The long windows overlap each other heavily. It describes what happened. It does not predict.
How each cell is calculated
- •$100 buys on the 1st of each month, on UTC calendar days, for the full window: 12 purchases per year of hold.
- •The stack is valued at the Bitcoin price on the window's end date, never at today's price.
- •The percentage is XIRR: the yearly rate at which every dated purchase plus the closing value nets to zero. It is not the total return divided by the number of years, because the money went in gradually.
- •No fees, no taxes, and no spread (the gap between what an exchange sells at and what it buys at) are modeled. Real returns would be lower.
- •One price per UTC day. From Jul 20, 2015 that is Coinbase's daily candle close. Before it, it is blockchain.info's market price, an average across exchanges stamped at 00:00 UTC, which tracks the previous day's close, so purchases and window-end values in those years are priced about a day early. If Coinbase's closes cannot be loaded the page keeps its last computed grid; it does not switch to another series. The methodology page covers the price sources and their limits.
- •The grid recomputes daily. A window fills in once its end date has a daily closing price, which is the day after its anniversary. The data currently runs to Oct 11, 2026.
Run your own numbers
This grid fixes the amount at $100 a month and the cadence at monthly. Change the amount, the cadence, the fee and both dates in the live calculator and it backtests against the same historical price data.
Open the calculatorCommon questions
Does the start date matter for Bitcoin DCA?
Across the completed 1-year windows in this grid, annualized returns ranged from -73.2% to +4,278% per year. Across the completed 10-year windows they ranged from +62.2% to +176%. The gap between the best and worst start year was 4,351 points for 1-year windows and 114 points for 10-year windows. It was not narrower at every step: the 7-year column is spread wider than the one before. The 6 completed 10-year windows start a year apart and overlap: only 1 of that length fits in the data without sharing a year. This is a description of what happened, not a forecast.
How is the annualized return in each cell calculated?
Each cell simulates $100 bought on the first of every month for the whole window. The resulting Bitcoin is then valued at the market price on the window's end date. The annualized figure is XIRR: the yearly rate at which every dated purchase, plus the closing value, nets out to zero. It exists because the money went in gradually rather than all on day one, which a plain "total return divided by years" would get wrong. Historical windows are never valued at today's price.
Why do some cells say "in progress"?
A cell is only filled in once its full holding period has elapsed and its end date has a daily closing price. A 10-year window that started in 2017 has not finished, so showing a partial figure next to completed 10-year windows would be misleading. Until then a cell is marked in progress.
Do these historical returns predict future Bitcoin returns?
No. This grid records what one mechanical schedule produced over one stretch of the past, during which the price in our series went from $0.30 on Jan 1, 2011 to $88,738 on Jan 1, 2026. Nothing here should be projected forward. Past performance does not guarantee future results, and you can lose money.
Keep reading
- All DCA scenariosFixed amounts and cadences, worked out year by year.
- Lump sum vs DCABoth strategies, every start month since 2010.
- Best day to buy BitcoinSeven weekday schedules and a week-by-week test.
- Methodology & data sourcesWhere the prices come from and where the model falls short.
- DCA and tax lotsWhy buying regularly creates hundreds of separate cost bases.
- Why BitcoinThe case, the counter-case, and the risks.
Not financial advice. This is a historical simulation for educational purposes; past performance does not guarantee future results. Bitcoin is volatile and you can lose money.