A crypto asset that prints “+15%” on a green candle and “−15%” on the next red one has not broken even. It has lost money. The arithmetic seems trivial once stated, yet the financial commentary surrounding crypto—dominated by daily percentage changes, annualized extrapolations, and headline returns—routinely treats returns as if they were additive.
They are not. Returns compound multiplicatively, and once compounding, volatility, and the shape of the return distribution are taken seriously, the daily figures that dominate crypto discourse become not merely incomplete but actively misleading.
Bitcoin is math money 🧮 ⚡️ pic.twitter.com/EusQ9QpSqT
— Bitcoin Archive (@BitcoinArchive) November 25, 2025
The gap between what a daily return number says and what an account balance experiences is the subject of a precise and unforgiving mathematics.
The Arithmetic Mean Is Not the Return You Earn
The most fundamental distortion is the confusion between the arithmetic average of returns and the geometric, or compounded, return. The arithmetic mean answers the question: “If I added up the daily percentages and divided by the number of days, what would I get?” The geometric mean answers a different question: “What constant rate, applied to the starting capital, would have produced the actual ending balance?” Only the second corresponds to money in an account.
An asset rises 10% one day and falls 10% the next. The arithmetic average of +10% and −10% is zero. But $100 becomes $110 after the first day and $99 after the second—a 1% loss, not a flat outcome. The percentages are applied to different bases: the 10% gain is calculated on $100, while the 10% loss is calculated on the larger $110.
A volatile asset compounds more slowly than its average return suggests, and the drag is roughly half its variance: g≈μ−σ2/2. An asset with 80% annualized volatility—an entirely ordinary figure in crypto—carries a volatility drag of roughly 32% per year.
An asset can have a solidly positive arithmetic average return and still grind a buy-and-hold position downward over time, because the compounding process continuously subtracts the cost of its own fluctuations.
Volatility Drag: The Cost of Being Shaken
The formula g=μ−σ2/2 makes the mechanism explicit. The term σ2/2 is not a penalty imposed by fees or by bad luck; it is an algebraic consequence of compounding a multiplicative process. Under leverage, the relationship becomes even more punishing: g(L)=Lμ−L2σ2/2. Return scales linearly with leverage L, but drag scales with L2.
Doubling leverage doubles the expected arithmetic return but quadruples the volatility drag. This is why leverage does not raise the probability of long-term success; it raises the expected arithmetic payoff while deepening the geometric hole through which capital falls.
The practical implication is that volatility is not a neutral “risk” that can be tolerated in exchange for higher returns. It is a direct tax on compound growth. Two assets with identical arithmetic average returns but different volatilities will produce radically different terminal wealth, and the higher-volatility asset will be the poorer performer in geometric terms. In crypto, where daily swings of 5% to 10% are common, this tax can accumulate with remarkable speed.
Drawdowns Are Non-Linear, and Losses Demand More Than They Gave
Daily percentage changes also obscure the brutal asymmetry of drawdown recovery. A 10% loss requires an 11.11% gain to break even; a 20% loss requires 25%; a 30% loss requires 42.86%; a 50% loss requires 100%; and an 80% loss requires 500%. The recovery percentage is given by 1/(1−d)−1, where d is the drawdown fraction. This function is convex: as d grows, the required recovery accelerates without bound.
A portfolio that falls from $100 to $50 has halved its capital; a 50% gain on $50 returns only to $75, because the gain is calculated on the diminished base. The compounding process has “forgotten” the original capital.
Whose assets routinely experience drawdowns of 50%, 70%, or more, this mathematics means that the recovery phase must deliver returns that are not merely proportional but super-proportional to the preceding loss. A headline that reports a “50% crash” and a subsequent “50% rally” describes a round trip from $100 to $50 to $75—a 25% net loss, not a neutral event.
Fat Tails, Skewness, and Kurtosis: The Distribution Is Not Normal
Daily return figures implicitly invite the reader to imagine a bell curve of outcomes, with most days near the average and extremes fading symmetrically. Crypto returns do not oblige. Empirical studies of cryptocurrency returns consistently find high volatility, skewed distributions, and high kurtosis—fat tails in both directions, but often heavier on the left for major assets like Bitcoin and Ethereum.
Negative returns also exert stronger effects on higher-order moments than positive returns, a pronounced asymmetry that means “down days” are not simply mirror images of “up days.”
This distributional shape has direct consequences for the meaning of an average daily return. In a fat-tailed distribution, the sample mean is dominated by rare extreme observations, while the typical day is far less dramatic.
A strategy or asset that reports a strong average daily gain may be relying on a handful of outlier days; remove them and the picture reverses. Kurtosis measures the frequency of extreme deviations, and high kurtosis means that large moves occur far more often than a normal distribution would predict.
When returns are fat-tailed and negatively skewed, daily gain figures conceal a risk profile in which rare catastrophic losses can erase the accumulated effect of many ordinary gains.
Sequence Risk and Path Dependency: The Order Matters
Two investors can hold the same asset over the same calendar period, experience the same set of daily returns, and end with different balances if their entry points and cash-flow schedules differ. This is sequence risk, and it is especially acute in crypto because the asset class compresses deep drawdowns, violent recoveries, and long sideways ranges into short windows.
The mathematics of sequence risk is a direct consequence of compounding on a changing capital base. If an investor deploys capital at the start of a sustained decline, the subsequent recovery must overcome a larger percentage hurdle than if the same capital had been deployed gradually as prices fell.
Dollar-cost averaging (DCA) can therefore outperform lump-sum investing when prices spend time below the starting level, because later purchases are made at lower bases and participate more fully in the recovery.
When prices trend upward without revisiting the starting level, lump-sum investing typically wins. Neither outcome represents a failure of discipline; both are expressions of path dependency.
Across hundreds of thousands of simulations, lump-sum investing outperformed DCA in a majority of scenarios, but DCA gained a distinct advantage when entries occurred in the 20%–70% drawdown band—a range in which Bitcoin has historically spent a large fraction of its days.
The “best” strategy depends not on a universal rule but on the shape of the price path, which is unknowable in advance. Daily return reports, which aggregate over these paths into a single number, cannot capture this dependency.
Leverage and Volatility Decay: When “3x” Does Not Mean 3x
The misleading nature of daily figures becomes most extreme in leveraged crypto products. A 3x leveraged Bitcoin ETF or token is designed to deliver three times the daily return of Bitcoin, not three times its cumulative return over weeks or months.
Because the leverage is reset every day, the product rebalances its exposure on a shifting net asset value, and the compounding of daily multiples produces a result that can diverge dramatically from the multiple of the cumulative return.
The canonical example: Bitcoin rises 10% on Day 1 and falls 9.09% on Day 2, returning to its starting price. A 3x product gains 30% on Day 1, turning $100 into $130, then loses 27.27% of $130 on Day 2, leaving $94.55—a loss of more than 5% even though the underlying asset finished flat. The product has “bled” capital through the mechanics of daily resets.
This is volatility decay, also called beta slippage, and its severity increases with the volatility of the underlying asset, the leverage multiple, and the holding period. In choppy, sideways markets, a leveraged product can lose value even when the underlying asset ends exactly where it started.
Leveraged crypto products are therefore short-term tactical instruments, not long-term substitutes for direct exposure, and a daily return figure for such a product is almost meaningless as a guide to multi-day or multi-week outcomes.
Simple Returns, Log Returns, and the Asymmetry of Percentages
A further layer of distortion comes from the choice of return definition. The simple or linear return is R=(Pt−Pt−1)/Pt−1, while the log return is r=ln(Pt/Pt−1). Log returns are additive across time: the log return over two periods is the sum of the log returns of each period. Simple returns are not additive across time, but they are additive across assets in a portfolio.
The convenience of log returns in modeling has led to their widespread use in quantitative finance, but they can create a false sense of symmetry. A log return of −50% and a log return of +50% do not cancel: e−0.5×e0.5=1, but the price path is 1→e−0.5≈0.6065→1.
The arithmetic of log returns is symmetric, but the arithmetic of wealth is not.
A day with +10% and a day with −10% produce a net loss in simple-return space, and the effect compounds over sequences. The larger the daily swings, the more the compounded outcome deviates from the arithmetic average, and the more the headline daily number ceases to represent the experience of a holder.
What the Mathematics Implies
The hidden mathematics of crypto returns does not yield a single trading rule, but it does impose constraints on how return information should be interpreted. Arithmetic averages of daily returns overstate compound growth, and the overstatement grows with volatility.
Drawdowns require super-proportional recoveries, so the downside is mathematically heavier than the upside is helpful. Fat tails and negative skewness mean that averages are dominated by rare events, while typical outcomes are less impressive.
Sequence risk means that the same asset can produce different outcomes for different investors depending on entry timing and cash-flow schedules. Leverage introduces a drag that scales with the square of both volatility and leverage, so daily leveraged returns are especially poor guides to multi-period performance.
The corrective is not to abandon daily data but to translate it into the quantities that actually govern wealth: geometric returns, realized volatility, drawdown recovery requirements, and distributional shape. A crypto investor who reads “+12% today” and imagines a 12% richer account is performing an arithmetic operation on a multiplicative process.
The mathematics of compounding performs a different operation, and over time it is the mathematics of compounding that determines the balance.
Liam Foster follows crypto markets from France, with a close eye on Bitcoin, Ethereum, Layer 1 assets, derivatives, sentiment and smart money flows. His coverage looks past the daily price move to understand what is happening underneath: liquidity, leverage, positioning and the behavior of larger players.
Liam’s style is calm, sharp and deliberately anti-crystal-ball. He does not frame every market move as a breakout or collapse. Instead, he focuses on what the data can actually support, where traders may be crowded and when a signal deserves attention without becoming a prediction.