BTC Holder Behavior

REALIZED CAPITALISATION values each unspent output at the price when it was created or last moved, so it is an aggregate on-chain cost basis. REALIZED PRICE is that figure divided by current supply - the derived per-coin number, which is worth getting the right way round.

MVRV is market capitalisation over realized capitalisation, equivalently market price over realized price. Above one, market value exceeds the aggregate last-movement basis, implying net unrealised gain under that assumption - it does not establish that any 'average holder' bought below today's price.

NUPL is (market cap - realized cap) / market cap. Do the algebra and NUPL = 1 - 1/MVRV: the two are ALGEBRAIC TRANSFORMS of one another, not independent pieces of evidence. That matters directly here: both are given to the narrator that writes the 0-100 score, so the same information can be counted twice.

SOPR compares spent outputs valued at the time of spending against those same outputs valued at creation. Above one, the spent cohort realised an aggregate gain relative to that inferred basis. Saying it shows 'what was paid' overstates what a chain can know - it knows when an output was created and at what price, not what any person paid.

LONG- AND SHORT-TERM HOLDER SUPPLY splits coins by time since last movement, using a convention around 155 days. That is Glassnode's empirical boundary, not a protocol definition and not a measurement of intent.

THE SHARED ASSUMPTION, and the first thing a student should question: every realized-cap-derived metric, SOPR, and every age cohort treats a coin MOVING on-chain as an economic transfer. Much movement is not - internal wallet management, exchange shuffling between hot and cold storage, change outputs returning to the sender - all of which reset a coin's basis or age without anyone transacting economically. These quantities are computed exactly and mean approximately. Entity-adjusted implementations attempt to filter that self-churn; a raw implementation does not.

Exchange reserves carry the address-versus-entity problem: an exchange address aggregates thousands of customers, and reserve changes can reflect operational transfers as well as customer intent. This analysis publishes no bitcoin reserve figure, because the exchange wallets tracked for bitcoin are cold storage and their balance is not the bitcoin held on exchanges.

Mempool backlog and average fee describe demand for block space - a statement about network congestion, not about price. Recent transaction activity is read alongside them, as the COUNT of transactions over 24 hours and not the VALUE transferred. Those are different quantities that move independently - a surge in small transfers raises one and not the other - so this figure says nothing about how much value moved.