On-Chain Analytics

Dormancy and HODL Waves

Dormancy and the HODL waves are built from the same raw material: the age of each unspent output. Dormancy weights spent outputs by how long they sat idle, and the HODL waves show how the supply is spread across age bands. Both describe the ledger's age distribution; both are read as evidence of holder conviction, which is an inference.

Last reviewed 2026-09-21Source: Glassnode dormancy and HODL waves methodology; Bitcoin Core UTXO setNo figure is quoted here. The measures depend on the ledger's output ages, which are a fact, and on a reading of conviction, which is not.

Written by: The Bitcoin Data Guide editorial team
Reviewed by: Data review desk
Dates: Published 2026-09-21 · Last reviewed 2026-09-21

Fact

The ledger records the creation time of every unspent output, so the age distribution of the supply is directly computable.

Forecast

Reading a growing share of old coins as long-term conviction, or a shrinking one as distribution, is an interpretation of holder intent that the age data alone does not establish.

The two measures

Dormancy is an average age weighted by value. For each output spent in a period, it multiplies the coins in the output by the number of days since the output was created, sums the results, and divides by the total coins spent. The result is the average age of the coins that moved, expressed in days. A high reading means the coins that moved were old; a low reading means they were young.

The HODL waves take the same raw material and present it as a distribution rather than an average. Every unspent output is placed in an age band — one day, one week, one month, one year, and so on — and the bands are stacked to show what share of the supply falls in each. The chart is a picture of how old the unspent supply is at a given moment.

The two measures are usually read together. A rising dormancy and a growing share of old coins are offered as evidence that holders are not moving their coins; a falling dormancy and a shrinking old band are offered as evidence that they are. The age data is a fact about the ledger; the reading of it as conviction is an inference.

What is observed and what is inferred

What the ledger directly records is the creation time of every unspent output and the moment each output is spent. From that, the age of any output and the age distribution of the supply are computable without any further assumption. That part is a fact about the chain.

The inference is the step from "these coins are old" to "their holders are committed". The ledger records the age of a coin and not the reason it has not moved. An old coin may be held deliberately, may be lost, or may sit in a wallet whose owner has forgotten the key. The age distribution cannot distinguish between those cases, and the reading of it as conviction assumes the first.

The measures are therefore statements about the age distribution of the unspent supply. They are facts about the ledger, and they are inferences about the behaviour of the people behind the outputs.

The assumptions they depend on

The measures depend on a definition of which outputs count and on how coins that are provably lost are treated. A coin sent to an unspendable output has an age but no holder, and whether it belongs in the distribution is a definitional choice. The age bands themselves are a choice: a different set of boundaries produces a different picture from the same data.

Where the measures are computed per entity, they depend on address clustering and entity labelling. Clustering decides which addresses belong to the same owner using heuristics about how wallets construct transactions, and labelling attaches a name to a cluster. Both are inferences, and a mislabel propagates into the distribution.

The most common distortion is internal shuffling. When an exchange moves coins between its own wallets, the outputs are spent and recreated, and their age resets to zero even though no economic transfer occurred. A large reshuffle can make the supply look younger than it is. The coin days destroyed page covers the related age-weighted measure, and the limits of blockchain analysis page explains why a cluster is an inference rather than a fact.

Sources and references

The description of the measures and their construction is taken from the published methodology of the providers that compute them, and from the ledger structure they read.

The age measures and their methodology. Glassnode, Dormancy and HODL Waves: documents the age-band definitions and the dormancy calculation.
The output model the measures read. Bitcoin, Bitcoin Developer Guide — Transactions: describes how outputs are created and spent.
The attribution assumptions behind ledger metrics. Meiklejohn et al., A Fistful of Bitcoins: characterises the clustering assumptions that entity-level metrics depend on.

Sources

Glassnode dormancy and HODL waves methodology: Documents the age-band definitions and the dormancy calculation built on spent-output ages.
Bitcoin Developer Guide — Transactions: Describes how outputs are created and spent, which is what the age measures read.
A Fistful of Bitcoins: Characterises the address clustering and attribution assumptions that ledger metrics depend on.