Everybody knows crypto is volatile, and almost nobody has changed their position sizing to account for the specific ways in which it is. The excess is not random and it is not sentiment. It comes from four features of how these markets are built, and each of them has a consequence you can act on.
Annualised volatility on major crypto assets has spent most of its history at several times that of a broad equity index, and comparisons usually stop there. The problem with stopping there is that the two numbers are computed from data with completely different properties, so putting them side by side implies a comparability that does not exist.
An equity index is a weighted basket of hundreds of businesses whose individual moves partly cancel, priced during defined hours, with a mechanism that halts trading when moves become extreme, and underpinned by cash flows that give participants something to anchor a valuation to. A single crypto asset is none of those things. It is one instrument, trading continuously, with no halt mechanism, and with no cash flow against which any participant can say the price is now clearly wrong. Comparing the volatility of those two objects tells you they are different objects, which was already known.
The comparison that carries information is narrower and more useful: crypto against a single volatile equity, or against a commodity with no yield. Done that way the excess is still there and it is much smaller than the headline suggests, and what remains is explained by the four structural features below rather than by anything about the asset class being irrational.
The volatility gap is real and most of the way it is quoted overstates it. What is left after an honest comparison is structural, and structure can be reasoned about.
These are properties of how the market is built rather than of how participants feel, which is what makes them worth understanding: sentiment changes and structure does not.
Equity markets stop overnight and halt when a move exceeds a threshold, and both mechanisms do the same thing: they insert time between a shock and the reaction to it. That pause lets participants read the news, recompute, and find capital, and it is the single largest reason equity moves resolve less violently than crypto ones. Crypto has neither. A shock at three in the morning meets whatever depth happens to be present at three in the morning, with no interruption and no obligation on anybody to provide a bid. The move that would have been distributed across a pause is instead compressed into the minutes it takes to consume the book.
Retail access to high leverage is a defining feature of crypto derivatives, and the consequence is that a much larger share of open interest sits at prices where a modest move triggers forced closure. Forced closure is a market order that has to execute, which pushes the price into the next set of liquidation prices. The largest single-day moves in this asset class are overwhelmingly cascades of this kind rather than reassessments of anything, and they can be identified after the fact by the shape: violent, brief, and followed by a partial recovery once the forced flow is exhausted.
A large fraction of most crypto assets does not trade. It sits in long-term holdings, in locked allocations, in staking contracts, or in wallets whose keys are lost. The price is therefore set by a much smaller pool of genuinely available units than the market capitalisation implies, and a given amount of buying or selling moves it correspondingly further. This is the same mechanism that makes a company with a small free float more volatile than its market capitalisation suggests, operating at a larger scale.
In an equity market, a sufficiently large fall eventually brings in buyers who can point to a valuation and say the price is now wrong relative to earnings. That argument can be right or wrong and its existence puts a floor under panic, because somebody with capital has a reason to act. Crypto assets produce no cash flow, so no such argument exists. There is no level at which anybody can say the price is objectively too low, which removes a stabilising force that other markets have and cannot be added by any amount of maturity.
Volatility is not constant and it is not independent from one day to the next. Large moves are followed by large moves, in both directions, and quiet periods are followed by quiet periods. This is one of the most robust empirical regularities in any market and it holds strongly in crypto.
The consequence is specific and expensive. A stop placed at a fixed distance, in percentage terms, is a different stop depending on when it is placed. During a quiet stretch it sits far outside normal movement and rarely triggers; during an active stretch it sits inside normal movement and triggers on noise. The trader has not changed anything and their stop has changed meaning entirely, which produces the experience of being repeatedly stopped out just before the move continued in the expected direction.
The correction is to express distances in units of current volatility rather than in percentages. Whether you use an average range over recent sessions or a standard deviation of returns matters less than the fact that the number moves with the market. A stop that adapts is placed in the same relative position every day, which is what a fixed percentage was trying and failing to achieve.
The common rule is to risk a fixed fraction of an account per position, with the distance to the stop determining the size. It is a sound rule and it was designed in markets where two assumptions held: that the stop would be filled near its price, and that the position could be exited without moving the market. Neither is reliable in crypto.
The first fails because stops trigger during exactly the conditions in which the book is thinnest, so the realised loss is regularly larger than the planned one. The gap is not an error, it is the mechanism working as designed, and it means that risking a fixed fraction produces a distribution of actual losses whose tail sits well beyond the intended figure. The second fails whenever position size is large relative to available depth, which happens at much smaller sizes than most traders assume on anything outside the largest pairs.
The adjustment that follows is not complicated. Size against the realistic fill rather than the stop price, using your own record of how far your stops have actually filled from their trigger. And cap size against depth as well as against account, so that exiting does not itself become the largest cost of the trade. Both changes make positions smaller, which is the correct direction and the reason they are unpopular.
A stop is a plan, and the fill is the outcome. Sizing on the plan rather than the outcome is the most common way a disciplined risk rule produces undisciplined losses.
Crypto trades continuously, which is usually described as an advantage and is a mixed one. Continuous trading means no overnight gaps, and it also means the thinnest books of the week occur while a large share of participants are not watching. A move that begins on a Saturday evening meets less depth than the same move on a Tuesday afternoon and travels further for the same amount of flow.
It also means there is no point at which a position is safely parked. In an equity market the close imposes a rest; in crypto the position is live at every hour, and a stop is the only thing standing between an unwatched account and a move. That is a meaningful operational difference and it is the reason automated protection matters more here than in markets that close.
The related mistake is treating the calendar as though it were irrelevant. It is not: liquidity has a weekly shape, that shape is stable enough to measure, and sizing that ignores it is sizing against an average that does not describe the moment you are trading in.
Implied volatility, derived from options prices, is what participants are collectively paying for protection over a future window. It is genuinely informative in one narrow way: it is the market's price for the uncertainty ahead, backed by money rather than by opinion, and it can be compared with what actually happened afterwards.
What it is not is a forecast of direction, and it is not a forecast of magnitude either. It is a price, and prices reflect demand for protection as much as they reflect expected movement. A rise in implied volatility frequently means participants are buying insurance rather than that a large move is coming, and the two are impossible to separate from the number alone.
Its most defensible use is comparative. Implied volatility persistently above what subsequently occurred tells you protection has been expensive; the reverse tells you it has been cheap. Those are statements about the past with some predictive content and no certainty, which is more than most indicators offer and considerably less than they are usually claimed to.
Very little of this is about predicting anything, and that is the point. The four structural sources are permanent features, and they imply a small number of adjustments that hold regardless of any view on where the price is going.
Express every distance in units of current volatility rather than in fixed percentages, so that a stop means the same thing in a quiet week and a violent one. Size on your realised fills rather than on your intended stops, using your own record. Cap size against available depth, not only against account balance. Treat the hours when the book is thinnest as a different market rather than the same one at a different time. And keep protection in place continuously, because the market does not stop and neither does the exposure.
None of these require an opinion, none of them are new, and all of them are routinely skipped by traders who arrived from a market where the assumptions underneath the standard rules happened to hold. The volatility is not the problem. Sizing built for a different structure is.
Yes, and the usual comparison overstates it. An equity index is a diversified basket with defined hours, halt mechanisms and cash flows behind it; a single crypto asset has none of those. Compared against a single volatile equity or a non-yielding commodity, the excess is still present and considerably smaller than the headline figure suggests.
Because that is when the book is thinnest. Crypto never closes, so a shock at three in the morning meets whatever depth is present at three in the morning, with no pause and no obligation on anybody to quote. The same flow travels further when there is less on the other side of it.
A chain reaction in leveraged positions. A move reaches liquidation prices, closing those positions requires market orders that must execute, and those orders push the price into the next set of liquidation prices. It explains most of the largest single-day moves in crypto, which is why they are violent, brief, and often partly recovered.
Usually because the stop distance is fixed in percentage terms while volatility is not constant. Large moves cluster, so a distance that sat outside normal movement last month sits inside it this month. Expressing the distance in units of current volatility places the stop in the same relative position every day.
The rule is sound and its assumptions are not reliable here. It assumes the stop fills near its price and the position can be exited without moving the market, and both fail in crypto at smaller sizes than most traders expect. Size on your own record of realised fills, and cap size against available depth.
No, and that framing prevents the useful analysis. The excess comes from four structural features: continuous trading with no halt mechanism, dense retail leverage, a float much smaller than the supply, and no cash flow to anchor a valuation argument. Each is a property of how the market is built rather than of how anybody feels.
What participants are collectively paying for protection over a future window, backed by money rather than opinion. It is not a forecast of direction and not reliably one of magnitude, because it reflects demand for insurance as much as expected movement. Its best use is comparative: whether protection has been expensive or cheap relative to what followed.
Some of the sources can ease and two of them cannot. Deeper markets and broader participation reduce impact and slow cascades, so those may improve. But there will still be no closing bell, and there will still be no cash flow to argue a valuation from, and those two are permanent features of the design rather than symptoms of immaturity.