Each bull run gains less than the last — and almost on a straight line.
Start with the plainest measure of a Bitcoin bull market: how many times bigger the top was than the bottom it climbed from. Across the four cycles in the record, that number falls off a cliff — about 604× bigger in 2013, then 107× in 2017, 21× in 2021, and just 7.9× by the 2025 top.
What makes it more than a curiosity is the regularity. On a multiply-by-ten scale, the four gains drop by a near-constant fraction each cycle — a steady fade, not a random walk down. That fade is the whole report.
cycle_decay
It is not just the booms. The busts are getting milder too.
A shrinking boom on its own could just mean Bitcoin is topping out. But the other half of each cycle — the crash that follows the top — is easing on the same schedule. The 2014–15 bust took roughly 85% off the top; the 2022 bust, the mildest of the three, took about 77%.
So the right picture is not “the rocket is running out of fuel.” It is a whole swing — up and down — that compresses each cycle. That is what lets the forward projection in Part 3 narrow the downside as well as the upside, instead of bracing for another 85% wipeout.
drawdown_decay
Where the fading cycle points from here — leaning mildly up, held as a wide range.
Today Bitcoin is around $62,000, about eight and a half months past the October-2025 top, in the early bear-and-recovery leg. To project forward we replay each prior cycle from the same point in its life, rebased to today’s price — then shrink the up-swing to match the fading cycles (and the milder busts from Part 2).
Counter to a naïve “crash repeat,” two of the three analogues were already recovering by this phase. So the middle of the range tilts mildly up — but the three paths plainly disagree, and that disagreement is the honest signal.
forward_cone
The honest win, and the honest split — stated side by side.
A pattern that fits the past always looks convincing. The real test is held-back: stand before a cycle’s top, predict it using only the cycles that came before, and check the error. We run that test twice — once for the cycle’s size, once for the price a year or two out — and the two come back with opposite verdicts.
That split is the most important thing in this report, so we show both and let neither hide the other.
amplitude_backtest
price_skill
A real, fading pattern — solid on size, silent on next year's price.
This is not advice to buy or sell, and it is not a price target. It is a measurement: Bitcoin’s cycles are shrinking on a remarkably steady schedule, that shrink is a genuine edge for guessing the next boom’s size, and it buys nothing for guessing next year’s price. Holding both of those at once is the whole point.
Every Bitcoin bull run has gained less than the last — about 604× bigger in 2013, then 107×, 21×, and just 7.9× by 2025 — and almost on a straight line on a log scale. The log-gain fades by a near-constant fraction (~0.69×) each cycle. It is the single most regular feature of Bitcoin's cycles in this data.
It is not just the booms. The crashes are easing too — about −85%, −84%, −77% — so the whole four-year swing is compressing, up and down. That is why the forward range can narrow both edges rather than brace for another 85% wipeout.
Held back, the fading law predicts the next cycle's SIZE about 2.6× more accurately than 'just repeat the last cycle' — the very assumption that overshot every cycle. That is the report's real contribution: it quantifies the shrink that the full-amplitude cycle read keeps missing.
Knowing the size shrinks does NOT give a near-term PRICE edge. The cycle-analogue path loses to a flat coin-flip at one year and only ties it at two. So this is a direction-and-size story — never a price target.
Even shrunk, the projected up-leg leans mildly up: a 1-year middle near $69K ($45K–$107K) and a 2-year middle near $130K ($90K–$187K), held as wide ranges. The bullish P90 tail is real; so is the downside edge. A calibrated band, not a call.
- —The whole study rests on four completed cycles and, where it is back-tested, just two real size transitions. Four points can show a direction; they cannot pin down a law. Today's cycle would be the fifth.
- —The fade must eventually flatten — a steady shrink keeps heading toward a near-1× 'cycle,' which is really the four-year boom fading out. The one-step projection of the next gain sits at the edge of where the pattern still makes sense, so read it as 'much smaller,' not as a literal 1.5×.
- —The near-term price path is the weakest piece here: out-of-sample it loses to assuming today's price simply holds, over a 12-month horizon. We surface it only as a wide range, and say plainly it carries no edge.
- —With only two or three out-of-sample windows per horizon, there are too few points to calibrate an honest coverage number for the price band. We report that as not-estimable, not as a passed test.
- —Cycle structure can break. Spot ETFs, a maturing market, and a larger base may be flattening the cycle for reasons no four-point fit can foresee. This is descriptive research about a fading pattern, not financial advice.
Descriptive research, not financial advice. Bitcoin is volatile and has lost three-quarters or more of its value four times in this dataset. A pattern measured on four cycles is an observation about the past, not a promise about the next one.
Appendix — all 5 charts & the method →