DariX F0 Square

13 years of daily BTC data, every rolling-window scenario computed.

3 answers: when to buy, when to sell, and why buying BTC right now is a worse idea than it feels.
Bitcoin has completed 4 cycles. Every one ended in a 77-93% drawdown, followed by a new ATH within three years. Given that, how to deploy capital into this asset is not a marketing question. It's a math problem.
https://x.com/nobrainflip/status/1976341269921296653
The consensus advice splits into two camps.

- HODL — buy whenever, never sell.
- DCA — never lump-sum, spread over months.

Both are simplifications that don't survive the data.
I ran every rolling-window combination of lump-sum (LS) vs DCA on 13 years of daily BTC prices — five DCA lengths, three holding horizons, 5% cash yield. ~400,000 scenarios.
Three answers are in this article:
When to lump-sum — and when not to.When to sell — with specific triggers that have worked 3 cycles in a row.Why BTC at −41% today is the worst entry zone in its entire history — not the best.
The conclusions are not what CT is saying. Here's the evidence.
1. The Vanguard question, applied to Bitcoin
In 2012, Vanguard published the definitive paper on this problem: Dollar-Cost Averaging Just Means Taking Risk Later. They tested rolling 10-year windows across US, UK, and Australian equities and found LS beat DCA ~67% of the time, with a ~2.3pp return advantage. A 2023 update extended through 2022 — same result, hit ratios 62-74%. The consensus in traditional finance is not controversial: LS wins.
The mechanism is mechanical. Markets rise more than they fall. Every day in cash is expected return forgone. DCA is not a strategy — it's a partial stay-in-cash strategy, and partial stay-in-cash is just a worse version of stay-fully-invested when the asset has positive drift.
Nobody had run this rigorously on BTC at scale, because its volatility makes people assume the answer must flip. It doesn't.
Same methodology, daily BTC prices 2013-2026, 5% APR on cash during DCA:

LS beats DCA in 58-72% of all historical entry dates, across every horizon and every DCA length. Longer DCA periods lose more often — because more time in cash means more expected return given up.
The Vanguard result transfers cleanly to BTC. DCA loses on average.
2. How badly does DCA lose?
The 60-70% win rate is the average case. The more interesting question is the magnitude.
Median LS returns minus median DCA returns, 5-year horizon:

A 12-month DCA on a 5-year hold costs the median investor +314pp of return vs LS. On a $10k deployment that's $31k left on the table at the median, not the best case. Even a "cautious" 3-month DCA costs +103pp — more than a full doubling.
DCA isn't free insurance. It's extremely expensive insurance.
3. But DCA actually works at the tail
What happens in the worst 5% of entry dates?

Two things:

1.) the worst 5% of LS entries on a 5Y hold still returned +120%. The bad case, across 13 years, was still more than a double. That's how asymmetric this asset has been.
2.) DCA 24m cushions the worst case meaningfully — +183% vs LS's +120%. On shorter horizons (1-2Y) the gap is wider and DCA genuinely protects. On 5Y+ it shrinks.
Honest framing: DCA buys downside protection on short horizons, paid for with expected return. On long horizons, both shrink.
Which means for most 5Y+ investors, the math says stop DCA'ing. But the aggregate heatmap hides something bigger.
4. The plot twist: drawdown-conditioned results break the rule
All of the above averages across every historical entry date. But "buying BTC at ATH" and "buying BTC at −70%" are obviously different decisions.
Bucket the entry dates by distance from ATH at that moment. The single most important chart in this article:

0-10% below ATH (near-ATH entries): LS wins 74-82%. Up-trending assets keep trending.10-20% below ATH: LS wins 76-87%. Still clearly LS.20-30% below ATH: LS wins 38-63%. Coin flip.30-50% below ATH: LS wins 46-68%. Coin flip.50-70% below ATH: LS wins 48-59%. Still mixed.70%+ below ATH: LS wins 60-100%. Full conviction.
The rule "just lump-sum, always" breaks in one specific zone: 20% to 70% below ATH. That's the band where forward return variance is so high that DCA over 12-24 months competes with immediate deployment.
Outside that band, LS wins clearly in both directions — at new highs and at capitulation lows. Inside it, outcomes are close to random.
There's a specific mechanism. BTC's worst drawdowns each cycle happened after a 30-50% correction. The first leg looks like a dip, then becomes a depression. Buying at −40% puts you directly in the path of the second leg about half the time. Meanwhile, buying at −70% means the second leg has mostly happened.
This is why buying BTC today is a worse idea than it feels.
BTC at $78k is −37% from the October 2025 ATH of $126k. Dead center of the worst zone for lump-sum buying in BTC's entire history. Every retail instinct says "40% off, back up the truck." The data says: about half the time, that truck gets flattened by the second leg.
https://x.com/nobrainflip/status/2044823156326596900
That’s where most people get trapped - and they’ll get trapped again this cycle:

They buy this zone with all their money because it “looks like the bottom” -> another leg down -> panic -> sell because they’re scared of ending up with nothing.
That’s why even if DCA isn’t mathematically optimal, at these levels it’s basically the only sane approach - hold/allocate only a portion of your intended total size.
5. Where BTC actually spends its time
To calibrate what's normal:

Most people assumes BTC spends most of its life near ATH. It doesn't.
Near ATH (0-10% DD): 25.8% of days.Shallow correction (10-30% DD): 17.6% of days.Coin-flip zone (30-70% DD): 46.3% of days. Almost half of BTC's history.Deep capitulation (70%+ DD): 10.3% of days.
BTC lives in the coin-flip zone more than it lives anywhere else. The drawdown band where lump-sum is actively worse than DCA is not a rare edge case — it's the modal state of the asset.
Two implications:
If you only deploy at ATH-ish levels, you'll compete for ~26% of days.If you only deploy at −70%+, you'll sit in cash most of your life and compete for ~10% of days.
Neither works as a standalone strategy. The playbook has to address all three zones, not just the comfortable ones.
6. Forward returns — the reward side
Win rate is one thing, payoff is another. Median 2Y and 5Y forward returns by entry drawdown:

Key numbers:
Buying near ATH (0-10% DD): median +700% over 5Y. The feared "bought the top" scenario across 13 years delivered a 7x on a 5Y hold.Buying at −50 to −70%: median +1,963% over 5Y. ~20x.Buying at −70%+: median +3,403% over 5Y. ~34x.
Watch the 2Y column. It's not monotonic. At −20-30% DD, 2Y forward return is lower than at 0-10%, because you bought into the middle of a bear leg and needed time to recover. The coin-flip zone shows up in returns, not just win rates.
Combined read: LS is almost always fine on 5Y. At −20-70% DD, 2Y return is compromised. If your real horizon is shorter than 5Y, the coin-flip zone is more dangerous than the heatmap alone suggests.
7. When to buy — the framework
Everything above is descriptive. The rules:
Rule 1. BTC within 20% of ATH → lump sum.74-87% historical win rate. Strong median outperformance. No real downside on 5Y. The only reason not to is behavioral — if a 30% drawdown after buying will make you panic-sell, you need a smaller position, not DCA.
Rule 2. BTC 20-50% below ATH (where we are now) → DCA 12-24 months.This is the only zone where math actively favors spreading. DCA 18-24m cuts tail risk by ~60pp at the 5th percentile while costing <1-2% in median vs LS. Outside emotion, it's the only drawdown band where DCA is rational.
Rule 3. BTC below −50% → tiered aggressive LS.
At −50% deploy 40% of reserved capital.At −65% deploy another 30%.At −70%+ deploy the rest. P(LS > DCA) at −70%+ is 95-100% on 12-24m DCA. Median 5Y forward return ~34x. This is the only zone where the math unambiguously says back up the truck.
Compressing drawdowns caveat. Cycle-over-cycle, BTC bear lows have gone −93% → −86% → −84% → −77%. Next capitulation, if it happens, is likely −70 to −76%. But it might not happen. Which is why Rule 2 matters: you can't sit in cash waiting for −70% and miss a rally if the floor forms at −55%. DCA'ing through the coin-flip zone guarantees exposure either way.
8. Why HODL is slowly dying as a strategy
Historical HODL returns, measured ATH to next ATH (the full cycle a buyer-at-the-top actually lives through):

2013 → 2017: $1,163 to $19,650. 16.9× over 4 years. 101% CAGR.2017 → 2021: $19,650 to $69,000. 3.51× over 4 years. 38% CAGR.2021 → 2025: $69,000 to $126,296. 1.83× over 4 years. 17% CAGR.
That's an 89% collapse in HODL returns across two full cycles.

Project that pattern forward. If the next cycle (2025 → 2029) delivers even 100% of the last cycle's return, HODL gives you 1.83× over 4 years — 17% CAGR.
The Nasdaq-100 has returned ~14% CAGR over the last 20 years. S&P 500, ~10%. MAG7 basket, ~25%. You are now paying an 80% drawdown for returns that barely edge out index ETFs.
This is where the compound-interest math becomes terminal.

Three paths, $100 starting, 12 years:
HODL through declining cycles (10×, 3.5×, 1.8×, each with 80% drawdown): ends at ~$280.Stable 15% compounder (think a disciplined Nasdaq/MAG7 allocation, no drawdowns >35%): ends at ~$535. Beats HODL by 1.9× with no −80% drawdowns.Sell-and-reenter BTC at −50% DD: ends at ~$2,800. 10× HODL, 5× the stable path.
volatile assets need higher CAGR than stable assets just to break even because recovery from a drawdown is geometrically expensive. −80% requires +400% to recover. −50% requires +100%. Every cycle, HODL burns most of its 3-year gains in the bear market, then has to rebuild from a lower base. Stable 15% just keeps compounding.
This is not a bearish thesis on Bitcoin. It's a bearish thesis on holding through drawdowns as a strategy. The insight is that BTC's volatility has always been the feature, not the bug — but only if you actually respond to it.
What this means practically:
If Bitcoin delivers a −70% drawdown this cycle (from $126k to ~$38k) - deploy aggressively, ride it back up, exit at the next cycle top (+50-100% to prior ath). Historical 3/3.If Bitcoin doesn't deliver a −50% drawdown this cycle? BTC in general becomes a slightly-better-than-index asset with extra volatility. Still holdable, but no longer the life-changing bet it was.Either way, pure HODL from current levels ($74k-$79k, −41% from ATH) has negative expected edge vs waiting. The math from Section 4 still applies: you're in the coin-flip zone. The math from this section compounds on top: even if you catch the upside, the upside is now small. The combined EV of lump-summing here against alternatives is bad move
9. Ethereum and alts - a different game
Alts look like BTC but the math works differently:
Bull phases they beat BTC 3-10x.Bear phases they lag BTC 2-5x.Across full cycles, most alts underperform BTC. The ones that didn't (2017 ETH, 2020-21 SOL) are survivor-bias picks that can't be reliably identified in advance.
Translation: lump-sum-and-hold on alts is structurally worse than on BTC. What works is narrow rotation windows during confirmed altseason, then back to BTC or stables.

Here's my article on that matter (must-read)

https://x.com/nobrainflip/status/2013617784647778701
10. The answer: what to do today (April 2026)
Don't lump-sum here. One of the only times in the cycle where DCA is mathematically superior to LS. The data says the second leg of a drawdown starts from exactly this depth about half the time. DCA over 12-18 months.Reserve 30-40% of deployable capital for lower levels. −55% would be $56k. −70% would be $38k.Don't buy alts for long-term yet. ETH/BTC < 0.035 weekly close = negative EV. Wait for the trigger.HODL alone is no longer enough. With last cycle's 1.83× return and the ongoing degradation trend, pure HODL from $74k into the next cycle's top offers ~15% CAGR at best - Nasdaq-100 territory with 3× the drawdowns. The capital allocation decision has changed: BTC exposure only makes sense if you're willing to exit into strength and re-enter into weakness, or if you're sizing it as a small satellite allocation next to stable compounders.The 4-year cycle probably isn't dead. ETF flows compressed volatility, maybe dampened the drawdown magnitude. Every analyst calling "super-cycle" or "cycle broken" was wrong in every prior cycle BUT it still works, with smaller amplitude.