Bitcoin MVRV Z-Score today
Coin Metrics data, daily series since 2011.
Full historical series
Linear scale. The line at 0 marks the point where market
capitalization equals realized capitalization. Source: Coin Metrics,
derived from CapMVRVCur and CapMrktCurUSD.
What it actually measures
MVRV compares market capitalization and realized capitalization as a ratio, but the same ratio can mean different things depending on how much the market typically moves. The Z-Score corrects for that: instead of dividing, it subtracts the two capitalizations and divides that difference by the historical standard deviation of market capitalization.
Z = (market cap − realized cap) / standard deviation.
The standard deviation is calculated expansively — with the data
available up to each day, never with future information — so the
value for each date is the same one that would have been seen live
that day.
The idea behind it is that the same MVRV of 3 should weigh differently in 2011, when the market barely had any history and a small capitalization was enough to move the whole deviation, than in 2025, with fifteen years of data behind it. The Z-Score tries to make that comparison fair across cycles.
The all-time high wasn't in 2011
If the Z-Score delivered on what it promises, its single highest point should sit in 2011 — the cycle with the least history behind it, and therefore the smallest standard deviation. It doesn't. We calculated the real maximum for each year since 2011, with the price and the MVRV on the exact day it happened:
| Year | Maximum Z-Score | Date | Price that day |
|---|---|---|---|
| 2011 | 7.68 | Jun. 8, 2011 | US$ 29 |
| 2013 | 10.66 | Apr. 9, 2013 | US$ 231 |
| 2017 | 10.09 | Dec. 7, 2017 | US$ 17,032 |
| 2021 | 7.15 | Feb. 21, 2021 | US$ 57,501 |
| 2025 | 3.05 | Jan. 21, 2025 | US$ 105,988 |
The absolute maximum of the fifteen years is 10.66, and it didn't happen at the price top of any known cycle: it was on April 9, 2013, in an intermediate bubble that pushed the price to $231 and deflated almost as fast as it had risen, eight months before that year's final top at $1,150. The standard deviation was still small that week in April; by November, with additional months of accumulated history, it had already grown enough that the Z-Score at the definitive top (with the price nearly five times higher) never surpassed that April peak again.
Only from 2013 onward does the descending staircase that resembles raw MVRV's begin: 10.66 → 10.09 → 7.15 → 3.05. Four cycles in a row with a lower top than the previous one.
The peak almost never coincides with the price top
There's a second, even more pronounced distortion. In 4 of the 5 years in the table above, the Z-Score hit its maximum before that year's price top, not on the same day:
- 2021: the maximum Z-Score was on February 21, with the price at US$ 57,501 — two months before the year's first top (April 13, US$ 63,446) and nine before the final top (November 8, US$ 67,542).
- 2013: the April 9 maximum preceded the December top by eight months.
- 2017: the maximum was on December 7, nine days before the December 16 top — the only time it came close.
The reason is the same expansive standard deviation: as a rally rises, each new day of high prices widens the historical deviation, so capitalization has to grow ever faster just for the Z-Score to hold steady, let alone rise. At some point in the rally that race is lost, and the Z-Score starts falling while the price keeps rising for weeks or months more.
It's not just this indicator. Five independent metrics — MVRV, NUPL, Z-Score, Puell Multiple and Power Law — show the same cycle-by-cycle compression, and in 2025 none reached its classic threshold. See all five in one table →
That's why this site describes the current level instead of
announcing what comes next. The full methodology, with the code that
checks it, is published in the auditoria/ folder of the
repository.
Frequently asked questions
Is the Z-Score useful for anticipating the exact top?
No. In 4 of the 5 verified cycles, its maximum arrived before the real price top, sometimes months in advance. Confusing "the Z-Score went down" with "the top already happened" would have taken someone out of the market too early in both 2013 and 2021.
How often is it updated?
It's recalculated with Coin Metrics' daily series. This page fetches the latest available value every time it loads.
Can the data be verified?
Yes. It's derived from CapMVRVCur and
CapMrktCurUSD, from Coin Metrics' Community API, public
and free. Anyone can request the same series and reproduce the
expansive standard deviation calculation.
Does it work for other assets?
Coin Metrics publishes MVRV and market capitalization — the two inputs of the Z-Score — for Bitcoin and Ethereum. For the rest of the dashboard's assets there's no series available in the free API.
Follow MVRV without watching the screen
The dashboard lets you set up an email or Telegram alert for when MVRV — which this Z-Score is derived from — crosses a threshold you choose, with frequency control so it doesn't repeat the same alert on every run.
Related metrics
The Z-Score is a normalization of MVRV. The ones that read best alongside it:
- MVRV — the ratio without normalizing for volatility. See the dedicated page →
- NUPL — the same difference expressed as a fraction of capitalization. See the dedicated page →
- Realized price — the cost basis in dollars used by the denominator of all three metrics. See the dedicated page →
- Puell Multiple — looks at miner sell pressure instead of holders' cost basis. See the dedicated page →