Amos Ductan The Writing / Markets · Bitcoin
The Writing — Essay No. 06
Markets · Bitcoin

Bitcoin and the Four-Year Cycle

Two completed cycles bottomed at 363 and 376 days below the top. The third just confirmed its monthly low at day 268 — earlier and shallower than either. That's the interesting part, and it's not the part the calendar predicted.

Bitcoin's four-year cycle is the most over-narrated structure in markets and one of the least carefully measured. Everyone can recite the halving story. Far fewer can tell you how long the last two declines actually ran, or what specifically has to happen on a chart before a low stops being a guess. This essay does both — and then applies the same rule to the decline we're in, which as of August 21, 2026 has done something worth writing down.

One ground rule from the primer, and it's the same one that governs the gold and silver essays: "four-year" describes historical spacing, not a promise. Nothing owes you a low on an anniversary. Price action marks the low; the calendar only tells you when to start listening for one.

I.What the last two cycles actually did

Measured top to bottom on the daily, the two completed cycles are remarkably close in duration and remarkably brutal in depth.

CycleTop → lowDepth
Dec 2017 → Dec 2018363 days$19,892 → $3,129 · -84.3%
Nov 2021 → Nov 2022376 days$69,000 → $15,460 · -77.6%
Oct 2025 → Jul 2026 (so far)268 days$126,296 → $57,718 · -54.3%

Two declines within thirteen days of each other in length — 363 and 376 — is the kind of coincidence that makes people believe in clockwork. It's also the reason the popular version of this cycle is a calendar argument: count roughly a year down from the top and start buying.

The current decline breaks that symmetry in the direction nobody expected. It didn't run longer. It ran shorter — the low so far arrived at day 268, roughly three months earlier than either predecessor — and it was less than half as deep: -54% against -78% and -84%.

$10,000 $100,000 dec 2017 — $19,892 dec 2018 — $3,129 nov 2021 — $69,000 nov 2022 — $15,460 oct 2025 — $126,296 jul 2026 — $57,718 2018 2020 2022 2024 2026
fig. 1 — bitcoin monthly closes, log scale · three cycle tops and three lows · source: coinbase daily candles, August 21, 2026

II.The rule that decides it

Here's where I part company with the calendar crowd. A low is not a low because enough months have passed, or because it feels washed out, or because the drawdown matches a prior one. In the framework I trade, a swing low is a three-bar structure — a bar whose low undercuts the bars on either side of it — and it is confirmed only when the bar immediately after the low takes out that low bar's high. Not a higher low. Not a bounce. The high, taken out.

Why the high, specificallyAn inside bar with a higher low tells you sellers paused. Taking out the low bar's high tells you buyers went and got price that sellers previously controlled. One is an absence of supply; the other is the presence of demand. Only the second one is evidence, and the distinction is the whole discipline.

That rule is deliberately slow. It will never get you the exact low, and in a grinding bottom the monthly version lags by design — it can't confirm until a higher-high month prints. What it buys you is the thing that actually compounds: it refuses to fire on every premature "the bottom is in" call, of which the last two cycles produced roughly one a quarter.

The count tells you when to start listening. The takeout tells you it's time.

III.What just happened

Applying that rule, without adjusting it to suit the conclusion:

The weekly chart confirmed a swing low on the June 29, 2026 bar at $57,718 — the following week took out its high. The monthly chart confirmed the same low: July 2026's low of $57,718 sat below both its neighbors, and August has already traded through July's high of $66,924. That August bar is still forming as I write this, and I want to be precise about what that does and doesn't mean: a monthly close can still land anywhere, but a high that has been taken out cannot un-take-out. The confirmation is a completed fact, not a forecast.

Weekly and monthly agreeing on the same low is not a common configuration. In the two completed cycles, the grinding bottoms of December 2018 and November 2022 both confirmed on the weekly near the actual low, with the monthly following later. Getting both pointed at $57,718 is the strongest structural read this decline has produced.

The corroborating readings, all as of August 21, 2026: price $77,665, up 34.6% off the July low, back 21% above the 200-week moving average after months below it, with the Mayer multiple at 1.13 — over one for the first time since the decline began.

IV.What this is not

It's not a prediction, and it's specifically not the calendar argument dressed up in different clothes. If anything, the calendar argues the other way: at day 319 we are still short of the 363- and 376-day marks that ended the prior two declines, and this decline is far shallower than either. Someone who only trusted duration and depth would say this bottom hasn't earned the name yet. They'd have a real case.

What I'd say back is that duration and depth are descriptions of two events. Two. Any base rate built on a sample of two deserves to lose an argument with the tape in front of you — and the tape has now produced the specific structure I require, on two timeframes, at the same price.

So the honest statement is narrow, and I'd rather state it narrowly than sell it: the July 2026 low at $57,718 is confirmed by the rule. That changes my posture from waiting to positioned, with the low itself as the level that says I'm wrong. It does not tell me the next leg is up and away, it doesn't tell me the cycle top is in front of us, and it obligates nobody to agree.

The whole point of running a rule instead of a narrative is that the rule fires when it fires — including on a schedule you didn't expect, in a decline that was too short and too shallow to fit the story everybody already believed. This one fired early. That's not a reason to distrust it. That's the reason to have written it down in advance.

End — Essay No. 06
A.D.