Coin Days Destroyed (CDD)
- 1M1M
- 3M3M
- 6M6M
- 1Y1Y
- ALLALL
- LinearLinear
- LogLog
About this metric
What it measures
Coin Days Destroyed measures how long the coins moved on a given day had sat unspent. Every coin accrues one coin day per day it stays put; when it moves, the amount moved times the days since its last move is counted as destroyed. The daily sum is shown in BTC-days.
How to read it
Moving a coin held for a year destroys 365 times as many coin days as moving one held for a day, so the metric mostly tracks long-term holders. Historically, readings have climbed in the later stages of bull markets as old coins are distributed, with spikes common near tops, while bear markets have brought long quiet stretches.
Daily values are noisy: exchanges and custodians reshuffling cold wallets, or holders moving coins between their own addresses, create spikes with no selling behind them, so check exchange flows before reading one as selling. Follow the 30-day average for the trend, and note that a growing supply lifts readings naturally over time.
- Frequency
- Daily (UTC 00:00)
- Asset
- BTC
- Category
- On-chain Indicators
- Type
- Single value
- Unit
- BTC-days
- Data since
- 2009-01-03
- Latest data
- 2026-10-27
- Data points
- 6,507
FAQ
How are coin days calculated?
A coin accrues one coin day for each day it sits in an address. If 2 BTC last moved 100 days ago and move today, 200 coin days are destroyed, and their age resets to zero.
Does a CDD spike always mean old coins are being sold?
No. A transfer is not a sale: exchanges and custodians switching wallets, or holders consolidating their own addresses, also destroy large amounts of coin days. Comparing the spike with Exchange Net Flow shows whether the coins actually went to exchanges.
Can early and recent CDD readings be compared directly?
Only loosely. Circulating supply and the stock of old coins keep growing, so the same degree of old-coin activity produces higher readings today. Across cycles, judge a reading against the current cycle's range or use a supply-adjusted version.
