HODL Waves

HODL Waves stack the Bitcoin supply by coin age at every daily close. The sage band is long-term-holder supply — coins last spent at least 155 days ago. The rust band is short-term-holder supply — coins last spent within the trailing 155 days. The two sum to 100 % of circulating supply by construction; the hairline through the stack is the LTH dominance line itself, and Bitcoin’s USD price runs muted on the log left axis for cycle context.

Today the LTH band reads 82.5% of circulating supply, with a trailing 90-day change of +9.4 pts — the Accumulating regime. The series stretches from 18 Aug 2011 through to the most recent close on 14 Jun 2026. The 0.78 and 0.55 reference hairlines on the chart bracket the typical cycle-bottom and cycle-top LTH dominance readings on the post-2014 record.

For each daily close t:

LTH(t) = supply held in UTXOs last spent ≥ 155 days ago / circulating(t)
STH(t) = 1 − LTH(t)

The 155-day cohort boundary follows the convention Rafael Schultze-Kraft and Kilian Heeg established in their 2020 cohort-analysis paper, where they identified the day count at which an unspent transaction output’s conditional spend probability flattens to a near-constant baseline. Below 155 days a coin is statistically much more likely to be moved next week than next year; above it, the conditional spend probability barely changes with further age. The boundary is empirical, not arbitrary, and is the same threshold used for SOPR and cohort realized price.

The framework itself comes from Dhruv Bansal’s Bitcoin Data Science (Pt. 1): HODL Waves, published in April 2018. Bansal’s original chart slices supply into a dozen age bands; the two-bucket form on this page collapses those bands at the 155-day threshold and is sufficient for the cycle-level conclusions. The 90-day change in LTH dominance is the direction signal documented inline in the reading row. Reconstruction details live on the methodology page.

The legible cycle pattern is the LTH band swelling and contracting. During accumulation phases — the months and quarters following a cycle low — coins absorbed by patient holders age past 155 days and the LTH share rises. During distribution phases — the months and quarters around a cycle top — those same coins move and reset into the STH bucket. The 90-day change in LTH dominance is the direction signal; the absolute level shifts higher every cycle as the network matures, so a 70% reading in 2014 and a 70% reading in 2024 are not the same regime.

Reading every cycle anchor against the live series surfaces the regime-shift pattern with no further commentary needed. Cycle tops have printed progressively shallower distribution — the LTH band drained into the high 50s at the 2013 top and the low 50s at the 2017 top, but only into the mid 60s at the April 2021 peak, and held in the high 70s into the March 2024 pre-halving high. Cycle lows have, if anything, printed deeper accumulation each cycle — LTH sat in the low 70s through the 2015 and 2018 lows, and held near 80% through the late-2022 post-FTX trough. The topping extreme is compressing; the bottoming extreme is drifting higher.

The cleanest way to see the 2024 anomaly is the per-cycle peak-to-trough swing of LTH dominance. Each cycle has an accumulation peak (LTH share at its cycle maximum, typically 6–12 months before the price top) and a distribution trough (LTH share at its cycle minimum, typically within a quarter of the price top). The difference is the magnitude of cohort rotation in that cycle.

Three observations carry through the swing table. First, the broad direction is toward shallower distribution as the network matures. The 2017 distribution drained more than 25 points of dominance from accumulation peak to topping trough — by far the widest swing on the record; the 2021 cycle drained closer to 12; and the 2024 cycle to date has drained the narrowest band of any full cycle since cohort data has been available. The earliest 2013 window reads narrower still, but its accumulation peak sits near the 2011 start of the series and is truncated by it, so it is not directly comparable. The pattern fits the broader maturation story — MVRV, Puell, and NUPL all show the same shrinking-amplitude trajectory at the network scale.

Second, the institutional-custody era complicates the read. A spot-ETF custody wallet that holds coins for years on behalf of underlying shareholders ages those coins through the 155-day boundary identically to a self-custody cold-storage wallet. The on-chain record shows accumulation; the underlying ownership may have churned through hundreds of distinct shareholders without a single chain transaction. Cohort dominance is a supply-age statement, not an ownership statement, and the gap between the two has widened materially since January 2024.

Third, the 2024 reading has surfaced a new behaviour worth flagging. James Check has documented that long-term-holder supply remained anchored in the high 70s through a drawdown of roughly 21% from the all-time high — behaviour that prior cycles did not exhibit at comparable price drawdowns. Whether this reflects structurally stronger holder conviction, the new mechanics of ETF custody, or simply the not-yet-completed distribution phase of an in-progress cycle is the question this chart will answer over the next several quarters.

The spot-ETF wrapper severs dominance from ownership (January 2024 on). Since the US spot ETFs launched on 11 January 2024, hundreds of thousands of BTC have moved into a handful of custodian cold wallets — Coinbase Custody alone holds the bulk of IBIT and several peers. Those coins age past 155 days and read as long-term-holder supply, yet the underlying shares trade daily and rotate through thousands of distinct owners with zero on-chain footprint. The cohort line says “held”; the cap table says “churning.” Every reading after January 2024 is structurally less informative about distinct-holder conviction than the pre-2024 record was.

Lost coins permanently inflate the LTH floor. Every Patoshi-era coin untouched since 2010 sits forever in the long-term bucket. The ~1.1 million BTC isolated in independent Patoshi research, plus the broader 2.78–3.79 million BTC that Chainalysis estimates as lost, all count toward LTH dominance by construction. The structural baseline is a few percentage points above what an active-supply-only series would print — part of why every cycle’s LTH floor has crept higher even before accounting for genuine holder behaviour.

Old-coin reawakenings fire phantom distribution. When a dormant whale or a defunct exchange’s estate moves, the cohort line lurches without any live-market participant changing their mind. The Mt. Gox creditor distribution began moving roughly 142,000 BTC of 2014-vintage coins through trustee wallets in July 2024; the July 2024 transfer of ~50,000 BTC tied to the 2020 German authorities’ seizure did the same. Both events shoved multi-year-old coins across the boundary in days, printing as STH growth and reading like distribution that no holder actually chose. The 90-day change smooths most of it, but spot reads around named whale or estate movements should be treated as transitional.

The 155-day cut is a convenience, not a phase transition. A coin reads “long-term” on day 156 and “short-term” on day 154. Spending probability does flatten near 155 days, but the line is a step function laid over a continuum. In the months following a regime change — mid-2019 after the late-2018 capitulation, mid-2023 after the FTX trough — large cohorts of coins age through the boundary at once, jolting the dominance line for mechanical reasons rather than fresh conviction. Read transitions, not single prints.

If you accumulate on a schedule, sustained LTH-dominance readings above 78% have historically bracketed cycle-bottom regions. The signal is slow — the band moves on a quarterly timescale, not a daily one — and the modern compression toward shallower accumulation peaks means the high-78 region itself is hitting later in cycles. Treat a sustained 90-day-positive change in LTH dominance as a tactical accumulation accelerator; do not wait for an 80% print, because the modern peaks have not gone there.

If you are timing the cycle, the cleaner topping signal is the direction of the 90-day LTH change crossing from positive to sustained negative. Pair the read with cohort SOPR for the realised-profit cross-check and RHODL for the older-cohort tilt. HODL Waves alone is a slow signal; the modern compression means a peak-to-trough swing under 12 percentage points may not even register as “distribution” on the historical scale, even when a cycle top has passed.