Ten months after topping at $124,753, Bitcoin trades -48.0% below that high. Line up all four previous cycles at the same age and the typical reading is -69.2%. On the face of it, this is the mildest post-peak decline in the asset's recorded history — twenty-one percentage points shallower than the norm, and the slowest to arrive: it took 121 days to fall 40%, where the previous four cycles took 3, 12, 31 and 74.
We went looking for the bullish story in that number. We did not find one. What the same dataset shows, in the same four cycles, is that Bitcoin's rallies have been shrinking faster than its crashes — and a shallower drawdown is the second half of that sentence, not a rebuttal to it. Meanwhile a second clock, anchored to the April 2024 halving, is telling a much colder story: 841 days in, Bitcoin sits at 1.00× its halving-day price, where prior epochs were at a median of 9.83×.
Every Bitcoin cycle, redrawn as price divided by that cycle's peak price, and indexed by days since the peak. Stacking them this way is the only honest way to compare a crash in progress with crashes that finished: it asks what each one looked like at the same age, not what it looked like in the end.
At this point in every previous cycle, Bitcoin had already lost at least two thirds of its value. This time it has lost 48.0%.
The spread is not marginal. At day 307, the four prior cycles read −83.3%, −70.6%, −67.5% and −67.8%. The shallowest of them was still nineteen points below where Bitcoin is now. There is no overlap — the current cycle is not at the mild end of a distribution, it is outside it.
| Peak | Peak price | Rally into it | At day 307 | Trough | Trough day | Back to the old high |
|---|---|---|---|---|---|---|
| 2011-06-08 | $29.03 | — | -83.3% | -92.8% | 163 | day 622 |
| 2013-12-04 | $1,135 | 39.10× | -70.6% | -84.5% | 406 | day 1177 |
| 2017-12-16 | $19,641 | 17.31× | -67.5% | -83.8% | 364 | day 1080 |
| 2021-11-08 | $67,542 | 3.44× | -67.8% | -76.7% | 366 | day 847 |
| 2025-10-06 live | $124,753 | 1.85× | -48.0% | -53.1% running | 267 | — |
You are reading the free preview. Members get the rest — why the shallow drawdown and the shrinking rally are the same fact, the halving clock running at a tenth of its usual pace, where prior lows landed, and the cross-asset check that says this is not a Bitcoin story. Eludo is in private beta.
Depth is one measurement; velocity is another, and it moves in the same direction. We asked a simple question of each cycle: how many days after the peak did Bitcoin first close 40% down, and then 50% down?
Read the hollow markers top to bottom: 3 → 12 → 31 → 74 → 121 days. Bitcoin has taken longer to fall 40% in each successive cycle, and this time it took four months — forty times as long as it took in 2011, and 47 days longer than the cycle immediately before it. The 50% line tells the same story with one inversion: 26 → 14 → 47 → 182 → 242.
This is a real, monotonic trend across four observations, and it is the strongest thing in this report on the constructive side of the ledger. A market that takes eight months to give up half its value is behaving differently from one that does it in a fortnight. Deeper order books, a listed-fund bid that did not exist before 2024, and a holder base that no longer has to sell into every wick are all plausible reasons — none of which we can prove from a price series alone, and we are not going to pretend otherwise.
Here is where the bullish reading of section 01 falls apart. A drawdown is only meaningful relative to the rally that preceded it — and Bitcoin's rallies have been collapsing in size for a decade.
Both bars are shrinking, and the top one is shrinking faster. Bitcoin's rally into the 2013 top was worth 39.1× the previous peak. Into 2017, 17.3×. Into 2021, 3.44×. Into October 2025, 1.85× — the whole four-year advance was a doubling and a bit. Measured as total peak-to-trough amplitude in orders of magnitude, the four cycles run 2.73 → 2.05 → 1.33 → 0.90. The asset's swing has lost roughly two thirds of its range.
The shallow drawdown and the small rally are the same fact, measured from opposite ends.
This matters because of how the -48.0% figure is likely to be used. Read alone, it says Bitcoin is holding up unusually well. Read next to a 1.85× rally, it says something much more ordinary: an asset that went up less has, so far, come down less. Nothing in the pair implies the drawdown is finished, and nothing in it implies the next advance will be larger than the last. If anything, four consecutive observations point the other way.
We want to be precise about what this is and is not. It is not a claim that Bitcoin's returns are permanently smaller — that is a forecast, and four cycles cannot support one. It is a claim that this specific comparison, the one being made most often right now, is comparing a compressed downside to an uncompressed history and calling the difference strength.
The peak anchor asks how far below the top we are. It has an obvious weakness: it resets every time a new top is made, so a cycle that tops early looks young. The halving anchor has no such freedom — it starts on a date fixed by the protocol.
Bitcoin halved on 20 April 2024 at $64,908. As of 2026-08-09 it trades at $64,905. That is 1.00× — after 841 days, two and a half years of block rewards, and a round trip through an all-time high, the price is where it started, to within a rounding error. Prior epochs at the same age stood at 21.25×, 9.83× and 2.31×. Against a median of 9.83×, this cycle is running at 0.10× the typical result — a tenth.
And yet the timing is uncannily on schedule. Prior epochs put in their tops at day 371, 525 and 546. This one topped at day 534 — squarely between the last two. The clock kept time. Only the amplitude fell away.
The four-year rhythm did not break this cycle. It just stopped paying.
A question worth asking of the halving series: in the epochs that finished, when did the bottom arrive? Not the dip in the first weeks — the lowest close after the top, the one that ends the decline.
Measured forward from its own halving, today is day 841; the three prior lows came at days 777, 889 and 912. On that clock, we are inside the window. Measured backwards, today is 599 days from the next halving; the prior lows landed 542, 513 and 528 days out. On that clock, we are early — earlier than all three.
Both statements are arithmetically true and they point in opposite directions. The reason is mundane: epochs have lengthened, so a fixed day-count from the start and a fixed day-count to the end cannot both stay put. We are not going to resolve the disagreement by picking whichever clock reads better. With three completed observations, neither framing carries enough weight to time anything, and we flag the conflict rather than hide it behind the one that tells a cleaner story.
If shallow drawdowns were evidence of something specific to Bitcoin — a new class of buyer, an institutional floor — we would expect Bitcoin to stand out against its peers. We ran the same measurement across every asset with enough history to support it.
Bitcoin is shallower than its own history by 21 points. So is Ethereum, by 19, and Monero by 27 — both on more than one prior cycle. XRP is 13 points shallower on two. Solana's 21-point gap rests on a single earlier cycle and should be treated as an anecdote. The one asset sitting on its own history is BNB, at 3 points. Whatever is compressing drawdowns, it is not doing it to Bitcoin in particular — it is doing it to the majors as a group, which weakens the "institutional floor under BTC" reading considerably. A floor under one asset does not lift Monero.
Then there is the other half of the table, and it is the part most cycle commentary leaves out. Dogecoin, Litecoin and Cardano never made a new high after 2021. They are 1,920, 1,918 and 1,802 days past their peaks, down 89.7%, 88.1% and 93.3%. There is no median to compare them to because they have no second cycle to average. For a third of this sample, the cycle everyone is measuring the drawdown of never happened.
The compression thesis is measured on the coins that had a cycle to compress.
Findings that cannot be wrong are not findings. Here is what each claim in this report is betting on, and what would break it.
Read these before quoting any number above.