BTC Capitulation
The dollars destroyed each day by coins moving at a price BELOW what they last moved at, averaged over 30 days. Unrealised losses are paper pain; this only counts coins that actually changed hands, so it measures selling that hurt. The tall humps are the weeks holders gave up.
For every coin spent on a day, the indexer compares the USD price when it last moved with the price it moved at now. Where the second is lower, the difference is a **realized loss**, and the day's losses are summed. The chart plots the trailing 30-day mean of that total, because the raw daily series is far too spiky to read.
Sustained high readings are **capitulation phases** — not one bad day but weeks of holders selling into losses. Historically these clusters have overlapped with cycle lows, and the *easing* of one has mattered more than its peak.
Read it against price: losses realised while price is already far off its high mean late-stage selling, which is the exhaustion pattern. Losses realised near a high mean something broke.
**This is the AGGREGATE across every holder, not long-term holders only.** Our indexer emits realized loss without an age split, so short-term holders — who sell at a loss in every ordinary dip — are included and, being far more numerous, dominate the line. The rare long-term-holder capitulation that this chart's name evokes is *inside* this number, not isolated by it. Splitting by coin age at spend time is possible (the indexer already tracks age for CDD and the HODL bands) but needs a rollup change and a full history re-replay.
871 of 6,431 days read zero, nearly all of them 2009–2011 when barely any coins moved at a loss. Those are real observations and are carried into the average; dropping them would quietly inflate every early-cycle reading.
Computed in-house from our own Bitcoin node and UTXO indexer — cost basis fixed at the day each coin was minted · daily · history from 2009-01-03