DCA vs. lump sum — few beginner questions get asked more often and answered with numbers less often. "Spread your capital, it lowers risk," say some. "Time in the market beats timing the market, buy it all now," say others. Both are opinions. Here is the distribution instead: across 346 historical entries into Bitcoin, the lump sum won in 60 % of cases and delivered a higher median return — but cost noticeably more nerves. This isn't a matter of belief; it's a measurement.
How we measured it
The setup is deliberately plain, so the result doesn't hang on hidden assumptions:
- Fixed capital = 1. Both strategies deploy the same amount.
- Lump sum: everything on the entry day.
- DCA: equal amounts in weekly tranches across a 90-day deploy window.
- Both then hold to a 365-day horizon from entry.
- We slide that entry in weekly steps across the entire daily BTC history (Dec 2018 to Jul 2026) → 346 resolvable windows.
Look-ahead free: every purchase uses only the price on its own day, and the outcome is always measured forward. No retro-picked "ideal" entries.
The result
| Metric | Lump sum | DCA |
|---|---|---|
| Median return (365 d) | +52.6 % | +45.2 % |
| Share of positive windows | 68 % | 68 % |
| 25th percentile (bad cases) | −16.8 % | −8.8 % |
Two numbers carry the whole story:
- The lump sum has the higher median return (+52.6 % vs. +45.2 %) and led in 60 % of windows. On average, DCA does not beat the lump sum.
- But DCA halves the downside: in the bad cases (25th percentile) the lump sum lost −16.8 %, DCA only −8.8 %.
DCA doesn't buy a higher return. It buys a calmer loss side.
Why it works this way
The reason is unremarkable: markets rise on average. Deploy everything at once and you're fully invested on day one, capturing the early rise in full. Spread it over 90 days and you buy part of the position at higher prices while some capital sits on the sidelines — both drag the mean return down. That same spreading, which costs return in a rising market, cushions an entry made right before a drop: not all your capital hit the top. That is the trade-off, nothing more — less return for less timing risk.
What about the regime at entry?
The obvious follow-up: does the picture change if you condition on the cycle state at entry — say, only deep fear? The honest answer: the data can't carry a verdict here. Across the whole history there are only two independent "deep fear" episodes:
| Episode | Lump sum (365 d) | DCA (365 d) |
|---|---|---|
| 2018-12-01 | +76.4 % | +101.4 % |
| 2025-02-27 | −22.2 % | −26.9 % |
Two observations are an anecdote, not a base rate. Cycle bottoms and tops occur only a handful of times per decade — the tool's regime overlay shows these episodes but explicitly warns about the tiny sample. The directional pull is there (in one of the two bottom-ish cases DCA led clearly), but building a rule from it would be exactly the error we call out everywhere else.
Limits of the test
- ~1.7 cycles. The daily score history starts in late 2018 → just under two halving cycles. Enough for a robust unconditional base rate, not enough for reliable regime claims.
- Overlapping windows. The 346 entries sit a week apart; adjacent 365-day windows share history and are not independent. This is a descriptive base rate, not a confidence interval.
- BTC only in this run. For ETH and SOL the history is thinner still.
- The result depends on the deploy window. Over 30 days instead of 90 the return gap shrinks — and so does the downside protection. There is no universal "DCA window."
A sober conclusion
If you enter purely for return and can stomach the swings, the lump sum served you better historically. If you'd rather fear the bad entry timing less, you pay a few percentage points of mean return for a halved downside. Both are defensible — they simply answer different questions. "DCA is safer" is true; "DCA is better" is not.
What Backtesting Arena adds here
Instead of an opinion, the free decision tool DCA vs. Lump Sum gives you the distribution itself: pick the asset (BTC/ETH/SOL), the deploy window and the hold horizon, and see the median, hit rate and downside of both strategies side by side — plus the anecdotal regime overlay with a visible small-sample warning. If you'd rather work through the related "buy now or wait for the dip?" question, that lives in the Dip Decision Tool. Everything is computed look-ahead free from frozen cycle scores. It's a base rate, not a recommendation — the decision stays yours.
FAQ
Is DCA better than lump sum for Bitcoin? On expectation, no: across 346 historical entries the lump sum had the higher median return and won 60 % of windows. DCA was better only on the downside (−8.8 % vs. −16.8 % at the 25th percentile).
Then why is DCA recommended so often? Because it lowers timing risk and is psychologically easier to stick with. Both are real — it's just a risk advantage, not a return advantage.
Does this hold outside Bitcoin? The mechanism (rising market → lump sum ahead on average, DCA with a shallower downside) holds generally for long-term rising assets. The specific numbers here come from BTC history since late 2018.
How long should the deploy window be? There is no universal window. The longer you spread, the greater the downside protection and the greater the return you give up in a rising market. In the tool you can compare 30/90/180 days.
Does it help to DCA only during fear phases? The data isn't sufficient for a verdict — there are only two historical "deep fear" episodes. That's an anecdote, not a rule.
Is this a buy recommendation? No. It's a historical base rate over two entry strategies — not a price forecast and not a recommendation.