Notes · Sep-13 2026
Five-year-old coins hold 81 % of all the waiting
Count bitcoin's coins and about a third of them are old. Weigh the same coins by how long they have sat still and the picture flips: on Sep-12 2026, coins older than five years were 33.6 % of the supply our ledger sees, but 81.16 % of all the waiting on the chain.
"Waiting" is coin-days: every coin multiplied by the whole days since it last moved. A coin untouched for ten years weighs about 3,650 coins that moved yesterday. Sort all 20 million coins into thirteen age bands and count them one for one, and the bands look ordinary. Sort the same coins into the same bands by age, and the two oldest bands swallow the chart: ten-year-and-up money alone is 17.8 % of the coins but 57.0 % of the coin-days, a 3.2× gap.
The number grinds upward almost every day, because it counts the age of everything that did not move. That is the whole reason to read it. It is not coin-days destroyed, which counts the age of what moved and resets it. This one does not fall until old money actually moves, and when it drops, the page next door on old coins moving says which band did it.
Where it comes from: our own copy of the UTXO ledger, outputs of 0.01 BTC and up, the same file every age page on this site is built from. The thirteen bands add back to the whole on every build (largest difference 7.6e-6 BTC-days over 6,462 days, floating-point noise on numbers in the tens of billions), and the same stock is rebuilt by a second, independent calculator on the dormancy page (largest relative difference 1.4e-6). This page claims arithmetic against itself and nothing else.
What it does and does not do next, measured 2026-09-13 on our own data since 2013: the share of coins older than a year has run with the next 90 days (rank +0.25, three cycles, too few to call it more than a cycle marker). We call it a marker and nothing more, and we will re-measure it on 2027-09-13 against the file we pre-registered today.
What we measure next: whether the ten-year band's share of the waiting keeps rising through the coming year, or whether the first drop in the headline comes from that band.
We grade this on 2027-09-13: re-measure the ≥1y share against the next 90 days on the pre-registered file
Built Oct-23 2026 21:59Z · all notes · RSS · all metrics