mvrv

$BTC | MVRV Z-Score: A Walkthrough of the Most Misused On-Chain Indicator
The first time I watched the MVRV Z-Score on an institutional trading desk it was wrong, and the second time it was correct but in a way that wasn't useful. That tension - useful in retrospect, frustrating in real time - is the right starting point for understanding what this indicator actually does.
The standard formulation, often called the Awe & Wonder Z-Score after the analysts who popularised it on Glassnode, is straightforward:
Z = (Market Cap - Realised Cap) / standard deviation of Market Cap
Market Cap is trivially observable. Total BTC supply times current spot price. Realised Cap is the more interesting input. It values each UTXO at the price it last moved on-chain. So a coin sitting in a wallet since 2013 contributes its 2013 cost basis to Realised Cap, not today's spot value. The Z-Score normalises the gap between speculative valuation (Market Cap) and aggregate cost basis (Realised Cap) using an expanding-window standard deviation, so historical comparisons don't get skewed by recent volatility.
The thesis is straightforward. When the Z-Score is high - say above 7 historically - the market is so far above aggregate cost basis that holders are sitting on massive unrealised profit and are statistically likely to take some off the table. When it's negative, holders are underwater in aggregate and unlikely to capitulate further.
That thesis has worked at bottoms with reasonable consistency. It has worked less well at tops.
How it has performed across three completed cycles.
The 2013 December cycle top printed an MVRV Z-Score of approximately 9.3, well above the 7-zone that practitioners pointed at as the cycle-top trigger.
The 2017 December cycle top printed an MVRV Z-Score of approximately 11.5, again clearly inside the cycle-top zone.
The 2021 April top printed an MVRV Z-Score of approximately 6.3 - inside-the-zone if you used a 5-zone threshold, slightly below if you used the older 7-zone threshold.
The 2021 November echo top printed an MVRV Z-Score of approximately 2.9 - far below either threshold. The indicator essentially missed the second leg of the 2021 double-top entirely.
The bottoms have been more consistent. December 2018 printed a Z-Score of approximately -0.2, clearly in the negative zone that historically marks accumulation. November 2022 printed approximately -0.1, same zone. Both bottoms confirmed the negative-zone heuristic.
So the Z-Score is accurate at cycle bottoms with reasonable consistency, but the cycle-top threshold has drifted downward across cycles. 9.3 in 2013, 11.5 in 2017, 6.3 in 2021-04, missed in 2021-11. If you use the same fixed threshold across all four tops, you either accept false negatives (missed signals at later tops) or false positives (entries triggered before earlier tops were fully formed). Neither is acceptable for sizing capital.
Where the methodology gets fragile.
Three structural issues with MVRV Z-Score as practitioners commonly use it.
First, threshold drift across cycles. The supply has grown roughly 2x since 2013 due to halvings. The realised cap denominator scales differently from the market cap numerator in ways that compress the natural Z-Score range. So the same statistical "extreme" reads different across cycles. The 2021-11 miss is the obvious case. If we use a rolling threshold rather than a fixed one, we recover the signal but at the cost of look-ahead bias if not handled carefully.
Second, realised cap data quality. The pure-form MVRV Z-Score requires per-UTXO valuation at last-spent price, which only comes from on-chain data providers like Glassnode, CoinMetrics, or CryptoQuant. Free-tier access to that data is gated. The SMM model on satoshimacro.com currently uses a 4Y MA proxy for MVRV input - that proxy is documented honestly on the methodology page rather than dressed up as the real series. When the CoinMetrics community API was tested as a free source earlier in 2026, the endpoint blocked from Cloudflare build IPs and never returned. So the proxy stays in place. Indicator name on the panel ends with "(4Y MA proxy)" so readers can tell.
Third, position-classifier framing. The MVRV Z-Score does not forecast price. It positions you in a cycle zone. The output is "right now we are in accumulation / neutral / caution / distribution / cycle top". The next-12-month price path is not contained in that classification. People who treat it as a forecaster blow up. People who treat it as a position classifier compound.
That distinction is the single most important framing in cycle research. Forecasting price is a fool's errand at any time horizon shorter than the cycle itself. Positioning into the right cycle zone is achievable and creates durable alpha.
How SMM treats MVRV Z-Score in its multi-factor architecture.
The SatoshiMacro Model (SMM) puts MVRV Z-Score inside Tier 1 (Valuation), which carries a 25 percent weight in the composite. Tier 1 contains six signals: MVRV Z-Score, Power Law deviation, NVT ratio, Mayer Multiple, Pi Cycle ratio, and a long-window moving-average premium. The MVRV input is one of six, weighted proportionally inside the tier.
That construction matters. If MVRV misses a cycle top in isolation (as it did in 2021-11), the other five signals in the same tier can still fire in the correct direction. The tier-level output dampens the single-signal failure mode. And Tier 1 is one of six tiers in the full model (Cycle Timing 30, Valuation 25, Sentiment 20, Rotation 10, Miner 10, Macro 5). So even if all of Tier 1 falters, the other 25 percentage points of the composite can still register the cycle position correctly through cross-tier confirmation.
The 7-of-7 in-zone calibration on historical BTC inflections - 2013-12, 2017-12, 2021-04, 2021-11 for tops, 2015-01, 2018-12, 2022-11 for bottoms - is what comes out the other side of that diversified construction. No single indicator including MVRV would survive that test. The composite does.
This is the point most cycle-research consumers miss. They take one indicator, calibrate it against three tops, and treat it as decisive. When the fourth top arrives and the indicator misses (which it will, statistically, because three data points cannot characterise a tail distribution), they revise their conviction in the indicator rather than recognising that no single indicator should ever have carried that weight to begin with.
Current MVRV Z-Score reading.
As of 25 May 2026, the SMM composite reads 65.8 calibrated (Caution zone, edging toward Distribution). Tier 1 Valuation reads 45.7. The MVRV input is one of the components that pulls Tier 1 down off the cycle-top zone - we are still meaningfully off the 2025-10 ATH and the unrealised-profit math has compressed since then. That reading is consistent with a mid-drawdown post-cycle-top position. Not a pre-cycle-top setup.
The honest limitations to call out.
The SMM-side MVRV input is a 4Y MA proxy until a free-tier realised-cap source is wired in. The historical calibration accuracy was preserved despite this because the proxy series tracks the real Z-Score series with reasonable directional consistency across cycles - peaks and troughs align even when the absolute magnitude differs.
For Australian-resident readers specifically: if you intend to act on a Z-Score-flavoured cycle call, the AUD CGT framework matters. Selling more than 12 months after acquisition currently qualifies for the 50 percent CGT discount under the existing CGT framework (under review for non-super-fund holders from 1 July 2027 onward per the 2026 Federal Budget). Selling inside 12 months means no discount and full marginal-rate taxation under s6-5 ITAA 1997. The practical CGT classification depends on the investor-vs-trader test under TR 97/11. Any cycle-call execution needs to account for the holding-period asymmetry, not just the price-zone read.
What I actually do with this on real capital.
My read is that MVRV Z-Score is a corroborating signal, not a primary one. I size positions against a multi-tier confluence reading, not a single indicator. The reading I take from MVRV is whether the cycle has compressed unrealised profit enough that further downside is unlikely to be driven by holder distribution. We are not there yet for the current cycle, but we are closer than the SMM Distribution zone reading on its own would suggest.
Practically, the way I use this on real capital is to overlay the MVRV reading against the Pi Cycle position, the Mayer Multiple, the funding-rate state, and the rotation-tier read on ETH/BTC ratio. When 3 to 4 of those align directionally, I size in. When they disagree, I stay neutral. The composite read is the position; the individual indicators are diagnostic, not actionable on their own.
The single biggest mistake I see in cycle-research consumption is people who treat MVRV (or any indicator) as binary. The market does not produce binary states. It produces probability distributions over states, and your portfolio construction should respect that.
Full MVRV Z-Score chart with every historical fire date and the methodology callout: https://satoshimacro.com/tools/crypto/cycle-indicators/bitcoin-market-value-z-score/
Reading as of 25 May 2026.
#BitcoinCycleAnalysis #OnChain #MVRV #SatoshiMacro