How BlockShield builds Bitcoin Supply in Loss: a measured identity from 2022, a reconstruction re-anchored to realized price back to 2010, and the lambda band.
A coin's cost basis is the market price on the day it last moved on-chain. If that cost basis sits above spot, the coin is held at a loss. Supply in Loss is the sum of every such coin, expressed in BTC.
It describes the position of holders, not a flow. It says how much of the circulating supply is currently underwater — not how much of it moved, and not what anyone did about it.
Two series, never blended. The MEASURED series (2022 onward) is an identity rather than an estimate: circulating supply minus supply in profit, where supply in profit is measured across the whole supply by a keyless UTXO provider. That is all the free measured history that exists.
The RECONSTRUCTED series (2010 onward) is modelled, because the true figure before the measured window needs a full UTXO-set scan. The shape is rebuilt from the network itself: each day the coins that moved are re-stamped at that day's price, newly mined coins enter at theirs, and the probability that a given vintage moves decays with its age.
The raw rebuild drifts. Run against the measured realized price over sixteen years, no single decay constant tracks it, so the reconstruction is re-anchored daily to the measured realized price: shape from the network, level from a real measurement.
The decay constant stays unpinned. Three values are run and the spread between them is published as a band rather than hidden inside one line. The width of that band is the metric's stated uncertainty.
The chart draws the two series separately and never bridges the measured line across its own gap. Where the two overlap, the gap between them is recomputed on every run and published with the payload — never quoted from a past measurement.
Read the timing and the shape of the reconstructed years. Do not read their absolute level as if it were measured — that is the one reading the band exists to prevent.
The band belongs to the reconstruction. A measured identity has no decay constant, and the payload says so rather than hanging a model's uncertainty on a measured point.
The 5,000,000 BTC line is drawn because people ask for it, and it is not a signal. The series sits above that line on a large share of all days, so it marks ordinary corrections as readily as bottoms — and nothing on this platform alerts on it. Its caveat is computed from our own run rather than repeated from the claim.
The measured window is short by the standards of the question. Anything said about how this series behaved in 2015 is said by the model.
Open the live Supply in Loss tab →