The arrival goes like this. The price makes a new high, and the new high makes the news. It’s on the front pages, it’s in the group chats, it’s the thing someone mentions at work with that half-smile that says I got in early. The number is enormous and still climbing. And somewhere in those weeks — not the year before, not at the bottom, those exact weeks — a person who had never seriously considered it downloads an exchange app, verifies their identity, and buys.
This is not a story about one impatient person. It is the measured shape of how retail arrives. The Bank for International Settlements — the central bank of central banks, an institution with no coins to sell you — studied crypto exchange-app adoption across 95 countries. Two findings draw the arrival’s outline. New users didn’t trickle in evenly: 73% of them downloaded their app with Bitcoin above $20,000 — a price it has spent most of its existence below. And the inflow wasn’t even fast: user numbers rose about two months behind the price rises that attracted them. The crowd does not find Bitcoin at the bottom, when it is cheap and nobody cares. The crowd arrives two months of headlines after the move that made the headlines — which is to say, late by construction.
And when the arrival goes wrong, the standard story blames the asset: retail loses money because Bitcoin is too risky. Hold that story in mind as you read. By the end, the data will have taken it apart.
If that was you, this article is not going to call you foolish. The arrival feels rational from the inside — that is precisely what makes it a problem worth taking apart. So let’s take apart a week like the one you arrived in. As it happens, I am writing this inside one.
What actually made the headline
I am writing in the fourth week of August 2026. For months, Bitcoin traded in a range on thin volume — the kind of market that generates no headlines and no downloads. Then the daily close rose 16.1% in four days (August 16 to August 20, on my own computed price index). That is the kind of move that makes the news, and the news is what makes arrivals.
Here is what the data of those four days actually shows.
In the twenty-four hours read on August 20, $1.45 billion of leveraged positions were forcibly closed — 97% of them were bets against the price. That was roughly forty times the normal daily pace, measured against the quiet days before the move began. A liquidation is not a decision; it is the absence of one. A trader borrowed to bet on a direction, the price moved against them far enough, and the exchange closed the position automatically — which, for a liquidated short — a bet against the price — means the exchange bought on their behalf, at market, right now. Thousands of those forced buys, stacked in hours, push price into the next cluster of forced buys. The move feeds itself.
Now the tell. While price rose through those days, the momentum of open interest — a reading of whether the market’s total open positions are growing or shrinking — went down through those same days. Sit with that for a second, because it is the whole lesson. A rally driven by conviction shows positions being opened — people choosing to buy. A rally driven by force shows positions being closed — people being bought out of bets against it. This one, on the days I watched it, showed the second signature.
No villain is required for any of this. There is no smoky room. The clusters of leveraged positions are publicly visible to every serious participant; price gravitates toward liquidity because that is where large orders can fill; and the cascade, once started, runs on its own mechanics. Nobody needs to plan it. It simply happens, every cycle, to whoever is standing in it with borrowed money.
And sentiment did what sentiment does: the market’s most-watched fear gauge went from 31 — fear — to 62 — greed — in three days. The facts did not change. The crowd’s story about the facts changed.
I am deliberately not telling you what this move does next. That is not modesty; it is the discipline this entire article is about. A move fed by forced buying has a different structure than a move fed by demand — you have now seen how to tell them apart. What any single move does next is exactly the thing nobody can tell you, including the people who will claim, next week, that they did.
Who was on the other side
Every buy has a seller. So when the crowd arrives in a loud week, it is worth asking the question almost nobody asks on the way in: who is selling to me?
The BIS looked at that too, around the shock episodes of 2022. Their data show that the largest wallets — holders of more than 1,000 bitcoin, the “whales” — reduced their bitcoin in the days after the shocks hit, while small holders (under one bitcoin, the “krill”) bought the dip. The researchers are careful with their words, and I will be careful with mine: they write that larger investors probably cashed out at the expense of smaller holders — the wallet data shows who grew and who shrank, and the price pattern is consistent with the transfer, but nobody can observe the exact counterparty of a trade. Their own chart title is less careful. It reads: “the whales eat the krill.”
There is no villain in this beat either, and that is what makes it worth printing. Nobody dumped on you personally. The market did what markets do: it moved the asset from those who arrived early and cheap toward those who arrived late and expensive, and it did most of the moving in the loudest weeks — the same weeks the download data says the crowd walks in.
The news trap
Which brings us to the belief that causes the arrival timing in the first place: the idea that you can act on news.
By the time a move has a headline, the move is the reason the headline exists. Newsrooms do not schedule markets; markets schedule news coverage. The story runs because the price already moved. Genuine surprises exist — but the headline-driven arrival, the one the download data describes with its two-month lag, is not a response to surprise. It is a response to attention, and attention peaks after price peaks, because the price is the story.
Which means the person who buys on the news is, structurally, buying at the meeting point of maximum attention and maximum price — the place where an earlier holder’s exit is easiest to fill. Retail’s primary information source is not early. It is the lagging indicator wearing the costume of a starting gun. This is why the arrival problem is systematic rather than unlucky: the mechanism that informs the crowd is the same mechanism that marks the moment the crowd is most useful to everyone already in.
Two years underwater
Here is what the arrival costs. It is not measured in percent. It is measured in time.
The last full cycle, from my own price series: the 2021 top closed at $67,551 on November 8. The trough closed at $15,770 on November 21, 2022 — a 77% drawdown, 378 days after the top. The first daily close back above the old top came on March 4, 2024. From top to recovery: 847 days. Two years and four months in which every buyer of that loud 2021 autumn was underwater, every day, with no way to know from inside it whether the recovery was coming or not.
The BIS put numbers on what that did to the app-era arrivals, and the detail is the uncomfortable part. Measured against late-2022 prices, a user who made a single purchase on the day they downloaded their app would have lost money 73% of the time (yes, another 73% — a coincidence of numbers; these are the same users from the download study, now measured on their losses). A user who bought $100 in their download month and every month after — the disciplined one, the one doing what much good advice says to do — would have lost money 81% of the time. Read that pair carefully. Measured at the bottom of the cycle, the steady monthly buyer was underwater more often than the one-shot buyer — because every additional month of buying through the decline added coins above the bottom. The problem was never the cadence. The problem was that the whole program was anchored to an arrival at the top, and then measured at the worst moment a market can choose — which is exactly the moment most people give up.
The Fed’s own household survey shows the aftermath in the usage data: US adults using cryptocurrency fell from 12% in 2021 to 10% in 2022 to 7% in 2023. That fall is consistent with the crowd that arrived in the loud year leaving through the quiet ones — and anyone who sold on the way out converted a paper loss into a permanent one. The arrival problem, running in reverse.
The asset, through all of this, did what it has always done: fell hard, found a floor, recovered, and eventually closed above its old high. Bitcoin was never the part that failed.
The part you control
So the standard story — retail loses money because Bitcoin is risky — does not survive contact with the data. The volatility is real, but it is symmetric; it punished and rewarded with the same indifference. What the data actually shows is two specific, human failures: when people arrived, and what they did in the two years after. The first is a trap built out of headlines and attention and a two-month lag. The second is the absence of a plan, discovered at the worst possible moment to be improvising one.
You cannot un-arrive. Whenever you got here — last cycle’s top, last week’s rally, or years ago — that part is written. But the data’s two failures — the arrival and the two years after — differ in one way that matters: the second is behavior, and behavior is the one thing in this entire market that is fully yours. You can decide, in advance, on a quiet day, by rule: what you buy, when you buy it, what you do in a drawdown, and what would ever make you sell. Written down before the next loud week — because the next loud week is exactly when you will not be able to think clearly, and the whole point of a rule is that it does the thinking when you can’t.
The market will keep manufacturing arrivals. The headlines will keep marking the tops. The forced buyers will keep feeding the moves, and the early holders will keep exiting into the noise. None of that is yours to fix.
When you act is yours. It was never the market’s decision — it only felt that way.
Sources. BIS Working Paper No 1049, Crypto trading and Bitcoin prices (Auer, Cornelli, Doerr, Frost, Gambacorta; Nov 2022, rev. Jul 2023) — the 95-country app study, the 73%-above-$20,000 download finding, the 73%/81% loss estimates and their methods. BIS Bulletin No 69, Crypto shocks and retail losses (Cornelli, Doerr, Frost, Gambacorta; Feb 2023) — the two-month lag, the whale/krill wallet flows around the 2022 shocks. Federal Reserve, Survey of Household Economics and Decisionmaking (SHED) — crypto use among US adults, 2021–2023. August 2026 market readings and the 2021–2024 price sequence: PRIORAXIS’s own computed price index and dated data records. These studies describe crypto app users and US adults in their respective periods; they are cited for what they measured, in their own terms.