Reading On-Chain Valuation: MVRV, NUPL and Realized Price Together

On-chain valuation metrics all rest on one idea: the blockchain records the price at which every coin last moved, so you can estimate what the market paid for its holdings and compare that with what the market is worth today. Everything else — MVRV, NUPL, realized price, SOPR — is that idea rearranged.

Understanding the shared foundation matters, because it means these metrics are not independent confirmations of each other. Citing three of them is often citing one fact three times.

Realized value, and why it is not cost basis

Realized value sums every coin at the price it last moved on-chain. It is usually described as the aggregate cost basis of the market, and it is a decent approximation — but the approximation leaks in three specific ways, and each one matters.

Moving coins between your own wallets resets their recorded price even though nothing was bought or sold. Trades inside an exchange never touch the chain at all, so a coin can change hands a hundred times without its on-chain price updating once. And coins that are permanently lost keep their ancient purchase price forever, quietly dragging the aggregate down.

None of that makes the metrics useless. It does mean they measure "the average price at which coins last moved on the chain", which is a different thing from what holders paid, and the gap between those two has almost certainly widened as custodial and ETF holdings have grown.

The family tree

MVRV is market value divided by realized value. Above 1, the average coin is held at a profit.

NUPL is (market value − realized value) divided by market value. It is the same two numbers in a different arrangement, and it carries no information MVRV does not. If MVRV is 2, NUPL is 0.5, always. Reading them as two opinions is a mistake.

Realized price is realized value divided by supply, split into short-term and long-term holder cohorts. This is where genuinely new information enters: the split tells you who is above or below water, which the whole-market averages cannot.

Long-Term SOPR is the flow rather than the stock — not what a cohort paid, but whether what they are spending today is above or below that. It is the only member of the family that measures a decision.

What each one has actually done

The whole-market ratios have decayed badly. MVRV's readings at the four cycle peaks run 4.72, 4.43, 2.86, 2.29. The widely quoted "sell above 3.7" threshold would have worked in 2013 and 2017 and then never fired again, through two subsequent tops and drawdowns of 77% and 53%. NUPL tells the identical story because it is the identical data: 0.79, 0.77, 0.65, 0.56.

The bottom side has held up much better. MVRV fell below 1 at every major low — 0.56, 0.69, 0.75 — and NUPL went outright negative at all three, meaning the market as a whole was sitting on a loss. Neither happened at the June 2026 low, where MVRV read 1.10 and NUPL 0.09.

Long-Term SOPR shows the sharpest version of the same asymmetry. Its peak readings collapsed from 16.1 and 20.6 in the early cycles to 2.9 and 2.4 in the recent ones — an artefact of long-term holders no longer sitting on hundredfold gains. Its trough readings barely moved: 0.50, 0.42, 0.44. Selling at a loss after holding for a year is a decision people only make under real pressure, and pressure looks the same in every cycle.

Realized price contributes the crispest single signal in the family. Short-term holders' cost basis has fallen below long-term holders' exactly four times, and each spell sits inside a bear-market bottom, including 109 days beginning on the exact day of the November 2022 low.

How to read them together

Three practical rules follow from all this.

Pick one whole-market ratio, not both. MVRV and NUPL are algebraically equivalent. Use whichever you find more legible and ignore the other.

Weight the downside more than the upside. Every one of these metrics has kept its bottom-side signal and lost most of its top-side one. That is not a coincidence — the upside readings depend on how large the holder base's gains are, which shrinks as the asset matures, while the downside depends on human behaviour under loss, which does not.

Confirm with something that shares no inputs. The strongest statement you can make from on-chain data is when a cost-basis metric agrees with something built from entirely different numbers — miner revenue, or price against its own 200-week average. In November 2022, MVRV read 0.75, Long-Term SOPR 0.44, the cohort cost bases crossed, and price was trading below its 200-week average for what became a 210-day stretch. Four measurements, three of them from different families, pointing the same way.

That is about as much as on-chain data can honestly offer: a description of where the market sits relative to what it paid, most reliable at the extremes, and worth very little in the middle.

Last updated 2026-08-10.