A single number is moving through crypto research channels this week: Bitcoin's long-term holders are realizing profit at roughly 72% of some undisclosed historical benchmark, measured against a prior peak of 350% logged in December 2024. The implied conclusion — selling pressure is moderate, the cohort is holding, the cycle still has room. Clean. Quantified. Actionable, apparently.
That cleanliness is the warning.
The metric ships with no named data vendor, no published formula, and no stated threshold for where "long-term" begins. Three separate constructions of that same headline figure produce three contradictory readings of the same market. That is not a data point. It is a decision wearing the costume of arithmetic.
For readers who arrived after 2021: LTH denotes a long-term holder, conventionally an address dormant for 155 days or more. The cohort functions as the market's ballast — historically 60% to 75% of circulating supply, low turnover, cycle-smoothing behavior. When it spends, the float loosens. When it doesn't, available supply compresses.
The indicator family is mature. Realized profit measures the dollar gain booked when an on-chain output moves, priced against its acquisition cost basis. SOPR renders a similar idea as a ratio, where readings above 1.0 mean coins are moving at a profit. MVRV compares market value to realized value and has historically topped out inside the 3x–4x band — 300% to 400%.
That band is the first thing that should catch your eye. The published 350% peak sits almost exactly inside the historical MVRV range. The current 72% is roughly one-fifth of it.
The messenger is Darkfost, an analyst publishing through a chain-analytics research platform. Not a protocol. Not a foundation. An individual with a proprietary model — which changes how much weight everything downstream deserves.
The platform matters too. Author programs inside analytics firms have proliferated since 2021, and quality varies enormously. Some contributors publish versioned, auditable methodologies. Others publish chart screenshots with a caption. The absence of a stated vendor here is not a minor omission. It is the difference between a dataset and a drawing.
Bitcoin has no team to evaluate and no governance to audit. The only reputational surface in this story is the analyst's own track record, and the source material supplies none of it — no historical hit rate, no methodology disclosure, no verifiable prior calls. When the subject of an analysis is an individual, the due-diligence obligation shifts entirely onto that individual. It has not been met.
I rebuilt my own data flows after the Compound flash-loan incidents in 2020, and the lesson that survived was methodological rather than technical: a single analyst's custom metric is a hypothesis, not an observation. Two independent sources minimum, or you are reading an opinion with decimal places attached.
Start with the arithmetic. The arithmetic is where the story breaks.
If 72% describes the average realized profit margin on each spent LTH coin, the cohort's aggregate cost basis sits near 58% of spot — one divided by 1.72. That would mean the average long-term holder carries a 72% cushion and could absorb a severe drawdown before capitulation pressure surfaces. A genuinely constructive reading.
If 72% instead describes realized profit as a share of realized market cap — a normalized flow metric — the number says nothing about cost basis and everything about distribution velocity. A 72% print on that construction would describe an extraordinarily active spending regime. That contradicts the bullish framing outright.
If 72% describes the proportion of LTH supply currently in profit, it is a solvency metric, not a behavior metric. Nearly every post-2023 cohort clears that bar. Least informative of the three.
Three definitions. Three conclusions. One headline. You don't build a position on a ratio whose denominator you cannot verify.
There is a fourth problem, deeper than the definitional ambiguity: the 155-day threshold itself. It is a convention, not a law, and it was calibrated for a market where a coin that moved was a coin that changed hands. Under ETF custody, enormous volumes of Bitcoin change beneficial ownership without moving on-chain at all. Shareholders buy and sell the trust; the underlying UTXO sits immobile. The clock keeps ticking on a coin whose economic owner may have turned over six times. The cohort definition is not wrong so much as increasingly irrelevant.
Now stress-test the anchor. The 350% reference is attributed to December 2024, when BTC printed near $100,000. Under the MVRV-adjacent reading, that December figure marked a genuine distribution peak — the moment long-term holders booked their largest gains of the cycle. The current 72% represents one-fifth of that intensity. The original framing uses this to argue selling pressure has cooled.
Fair. But note what the same comparison implies about the intervening months. If LTH realized profit fell from 350% to 72%, the cohort either stopped spending or the price stopped rewarding them for spending. Both describe a market that has already digested a significant move — not one that is preparing for the next.
The critical insight is this: realized profit is a behavioral derivative. It describes spending that has already occurred. It cannot lead price. It can only confirm what price has already done.
This is the trap embedded in every on-chain cohort metric. A 72% reading tells you long-term holders spent coins at a profit last week. It tells you nothing about next week. Presenting it as evidence of continued holding is a category error, and it flatters bullish priors precisely when certainty is scarcest.
The supply side of the argument is more defensible. Bitcoin issuance is rigid — 3.125 BTC per block post-halving, plus fees. No unlock schedule. No vesting cliff. No insider allocation. Every unit of selling pressure originates from an existing holder's decision, which means the entire question of available float reduces to how those holders behave. A moderate LTH spend rate therefore does signal constrained supply, provided two conditions hold: the reading is accurate, and the reading persists.
Persistence is the problem.
Long-term holder behavior is path-dependent. The cohort's willingness to hold is calibrated to a psychological price level nobody has disclosed. Cross it — a prior all-time high, a round number, an ETF-flow-driven breakout — and the 72% print can migrate toward 200% within weeks. Cycle history shows exactly this shape: a placid distribution regime that resolves into concentrated selling without an intervening announcement.
In 2017, LTH distribution accelerated sharply in the final eight weeks, spending in a burst rather than a bleed. In 2021, the pattern repeated with two distinct peaks bracketing the November high. Neither episode was forecastable from a single quiet reading three months prior. The metric that actually carried predictive weight was not realized profit in isolation. It was realized profit combined with exchange netflow, funding rates, and stablecoin supply — four inputs, not one.
The current dataset supplies exactly one of those four.
My Terra/LUNA stress-test framework, assembled through May 2022, taught the same lesson in a harsher register. The failures that mattered were never linear. Peg mechanisms held, held, held, then stopped. Distribution regimes behave identically. Linearity is a comfort you purchase at the top.
And here is the detail buried in the original framing: the current reading is described as similar to levels seen during the previous bear market.
Read that twice.
If realized-profit intensity today resembles bear-market conditions, the honest description is that on-chain selling pressure is extraordinarily low. Quieter than mid-bull. Quieter than consolidation. That is a dormancy signal. It tells you the marginal long-term holder is asleep, not committed. Dormancy resolves in one of two directions, and the metric has no opinion about which.
The evidence chain is short in other ways too. No funding-rate input. No ETF flow cross-check. No short-term-holder supply breakdown. No open interest. In a market where leveraged positioning routinely anticipates spot moves by days, a behavior-only reading with zero derivatives context is a single instrument trying to carry an orchestra.
The exchange-side read is instructive as well. If long-term holders are genuinely holding, spot volumes compress and fee revenue across centralized venues decays — a pattern that surfaces in exchange equities before it surfaces in on-chain charts. My monitoring through 2024 and 2025 repeatedly found venue revenue lines front-running the on-chain distribution signal by roughly three weeks. That is not a coincidence. It is the same behavior, observed through a different instrument.
Now the angle almost nobody is pricing: the LTH metric itself is decaying.
Post-ETF approval, a structural share of Bitcoin's genuine long-term supply does not exist on-chain in the form the instrument measures. It sits in institutional custody — ETF trust holdings, corporate treasuries, custodial accounts — where it never generates a spent output, never registers as an LTH address, and never enters the realized-profit calculation at all.
The consequence is not a rounding error. It is a systematic truncation of the exact cohort the indicator claims to describe. The on-chain LTH now measures retail and self-custodied conviction while missing the institutional layer entirely. In 2017, those two populations were nearly identical. In 2026, they are not.
Which means the top signal is going stale. As ETF and treasury holdings grow as a share of total supply, the on-chain distribution peak will arrive earlier and read shallower than the true cycle top, because the largest holders are invisible to the instrument. Traders anchoring cycle calls on this metric are measuring the market minus its most consequential participant.
Strategic pivots aren't validated by instruments that cannot see the pivot.
There is a second-order consequence. If institutional custody absorbs latent distribution that on-chain LTHs would otherwise execute, the next top presents as a two-track event: on-chain cohort spending plus off-chain custodian redemption. The first is observable in real time. The second surfaces only indirectly, through ETF creation and redemption flows — data that publishes daily and that most on-chain analysts still file as a footnote.
Liquidity doesn't announce itself through instruments built before it existed.
For BTCFi, the distortion cuts the other way, and productively. If long-term holders will not sell, the rational next move is to make dormant coins productive — wrapped BTC, lending markets, restaking-adjacent structures. That migration would push Bitcoin from a sleeping asset toward a yield-bearing one, and it would generate on-chain activity the LTH instruments do detect. The irony is that the healthiest reading for the metric may come from the very mechanism that expands the float it is designed to measure.
Miner economics deserve a line as well. Rigid issuance plus a placid distribution regime equals a stable hashprice environment, which favors well-capitalized operators and squeezes marginal ones. A quiet spot tape is not neutral for the supply chain. It is selective.
Strip the framing and this is what remains. One analyst. One undisclosed formula. One comparison anchor. One narrative that happens to be internally self-consistent and happens to point up.
The defensible use of a reading like this is as a single input inside a multi-indicator framework, and only after the vendor and the formula are confirmed. Until then, discount it by a full confidence tier. Whether this tape is a mid-cycle consolidation or the opening of a drawdown is exactly the argument the metric cannot settle. Survival first, positioning second. In a defensive market, the cost of trusting a stale instrument is not a missed entry. It is an unhedged exit.
A note on positioning. If the dormancy reading is correct, realized volatility stays compressed and options carry cheapens — which makes volatility-long structures the asymmetric expression, not directional spot. That is where I would rather be paid to wait.
Watch three things from here. LTH realized profit migrating above 150–200%. Sustained net exchange inflows. And consecutive ETF inflow days that reverse. The first two warn of distribution. The third absorbs it. Which one moves first over the next six weeks determines whether this is accumulation or exit liquidity.
The structural shift is bigger than any single print. A market whose largest holders are custodial, and whose second-largest cohort is invisible to the dominant analytics stack, will misprice cycle risk for as long as the analytics stack lags custody. That gap is the actual trade.
You don't get to call a top with an instrument that cannot see its largest holder.