BTC Auditor

On-chain analysis with verifiable data

The cycle-top indicators stopped working

Five different on-chain indicators —that measure different things, with different data— clearly marked the tops of 2011, 2013 and 2017. Since 2013, each one's peak has been lower than the previous cycle's. At the October 2025 top, none of them even reached its classic sell threshold.

This isn't anyone's opinion: they're five public series, and below is the table to verify them one by one.

Where each indicator stands today

MVRV classic top threshold: 3.5
NUPL classic top threshold: 0.75
MVRV Z-Score classic top threshold: 7
Puell Multiple classic top threshold: 4
Power Law extreme euphoria: +1.5σ
Price latest data point

Coin Metrics data, daily series.

The table

Each indicator's value at the top of each cycle. For MVRV, NUPL and Power Law it's the value on the day of the price maximum; for Z-Score and Puell Multiple it's the year's maximum, because their peaks tend to precede the price top by weeks or months.

Cycle MVRV NUPL Z-Score Puell Power Law
20117.390.8657.68+3.94σ
20134.720.78810.669.46+3.47σ
20174.430.77410.095.79+2.71σ
20212.850.6507.153.38+1.54σ
20252.290.5633.051.40+0.27σ
classic threshold3.50.7574+1.5σ

The last row is the one that matters: in 2025, all five ended up below their own threshold. Anyone who waited for the classic signal to sell wouldn't have sold.

Two clarifications, because detail is what makes a table verifiable. Puell 2011 doesn't exist: the indicator needs 365 days of prior price data and its series starts at the end of that year, so there's no cycle value. And Z-Score and Puell aren't monotonic since 2011: both had a higher peak in 2013 than in 2011. The sustained decline starts in 2013 for those two, and from 2011 for MVRV, NUPL and Power Law.

Why this happens

All five compare price against a base that grows along with the market: realized capitalization (MVRV, NUPL, Z-Score), the past year's miner revenue (Puell), or the price's own historical trend (Power Law).

When Bitcoin was worth US$ 10 billion, a few billion dollars of new capital could triple the multiple. With the market in the trillion-dollar range, that same multiple requires an order of magnitude more money. The relative enthusiasm can be identical and the number, much smaller.

That's why the indicators aren't broken: the thresholds are. A value calibrated on the 2013 cycle describes a market that no longer exists. What remains informative is the position relative to its own history —the percentile— which is what every page on this site shows.

The same pattern from the supply side

There's a sixth indicator that isn't in the table because it doesn't measure price, but it tells the same story: Stock-to-Flow. At every halving, issuance scarcity doubles almost exactly —3.99→7.99 in 2012, 11.98→23.97 in 2016, 27.97→55.94 in 2020, 59.93→119.86 in 2024—. The supply jump was always identical.

The price's response over the following ~18 months wasn't: +9,103% to the 2013 top, +2,913% to 2017, +686% to 2021 and +92% to 2025.

It's the same compression seen from the other side: scarcity rises the same amount every time; the demand needed to price it in doesn't.

What this does NOT say

It doesn't say Bitcoin no longer has cycles. 2021 and 2025 were real tops, followed by major drawdowns. What stopped anticipating them were the fixed thresholds, not the cycle.

Nor does it say where the price is going. Five indicators below their classic threshold describe where we stand relative to history; they aren't a buy signal, and extrapolating that "the next top will be even lower" is exactly the kind of inference this same table should make you suspicious of.

This site describes the level instead of announcing what comes next. The procedure, with the code that checks it, is in the auditoria/ folder of the repository — including a script that revalidates these tables against the source data and checks that the pages don't contradict each other.

See each indicator in detail

  • MVRV — the multiple over the network's cost basis.
  • NUPL — the same information as a fraction, with its 5 zones.
  • MVRV Z-Score — normalized for volatility. Its all-time high wasn't in 2011.
  • Puell Multiple — from miner revenue.
  • Power Law — price against its own 15-year trend.
  • Realized price — the cost basis in dollars.
  • Stock-to-Flow — deterministic scarcity.

Follow this without watching the screen

The dashboard alerts you by email or Telegram when a metric crosses a threshold you choose, with frequency control so it doesn't repeat the same alert.

Set up an alert Telegram channel