This article is arithmetic: every number in it can be recomputed with a calculator from the stated assumptions. It recommends nothing. Nothing here is investment advice. Decisions are yours.
Volatility drag is the gap between average returns and compounded reality: a coin that gains 50% then loses 50% is not flat, it is down 25%. The wilder the swings, the wider the gap, and at crypto volatility the drag stops being a rounding error and becomes the main event.
The asymmetry underneath
Compounding multiplies; it does not add. Up 50% then down 50% is 1.5 times 0.5, which is 0.75. The average of the two returns is zero, the wealth outcome is -25%, and the difference is volatility drag. Its mirror is the recovery curve: a -50% loss needs +100% to cure, -80% needs +400%, -90% needs +900%, the non-linear staircase that makes crypto's drawdown base rates so punishing. For return streams the standing approximation is worth memorizing: compounded growth runs below average return by about half the variance, sigma squared over two. Small volatility, negligible correction. Crypto volatility, anything but.
The fair coin at crypto volatility
Run the cleanest possible demonstration. Flip a fair coin daily: heads the asset gains 8%, tails it loses 8%, magnitudes chosen to match the ordinary daily range of a liquid alt, which average true range measures directly. The expected value is exactly flat: average return is zero by construction. The experienced path is not. Each heads-tails pair multiplies wealth by 1.08 times 0.92, which is 0.9936, a loss of 0.64% per pair, about 0.32% per day. Compounded across a 365-day crypto year, the median outcome is roughly minus 69% while the expectation stays flat, the gap consisting of a tiny probability of enormous lucky streaks. That is volatility drag at this market's scale: an asset with no edge and no drift, held passively at alt volatility, has a typical-path bleed most participants would attribute to manipulation or bad luck. It is neither. It is sigma squared over two, and at sigma equals 8% daily it eats what a good year earns.
Leveraged tokens: drag as a product
Daily-rebalanced leveraged tokens industrialize the effect. A 3x token multiplies each DAY's return by three, which triples the drift term but multiplies the variance term, sigma squared over two, by nine. Worked example: the underlying gains 10% then gives it back the next day, a -9.09% print, finishing exactly flat. The 3x token gains 30% then loses 27.27%: 1.30 times 0.7273 is 0.9455, down 5.5% while the underlying went nowhere. String months of ordinary chop together and the token grinds toward zero with no help from direction at all, which is why these products behave as intended only during sustained one-way trends and decay through every sideways stretch, the same regime split that decides why breakouts fail in chop. The listed leverage is honest; the compounding is the fine print.
What drag implies for strategy
Three implications, each mechanical. First, volatility reduction is return, geometrically: cutting the variance of a return stream raises its compounded growth even when the average return is untouched, which is the arithmetic case for regime discipline. Stepping aside during the tape's most violent stretches does not need to time anything well to pay; it collects the drag it avoids, and that is one of the inputs behind how our dial is framed. Our own transplant experiment showed the destructive direction of the same coin: the ungated variant rode full alt volatility to a -74% drawdown, and the regime-gated variant's improvement came in large part from simply not being there for the worst variance.
Second, position sizing is drag management: half the exposure means a quarter of the variance term, which is why sizing rules bind hardest on the wildest assets. Third, leverage multiplies drag quadratically while multiplying drift linearly, so there exists a leverage level beyond which more exposure to a winning asset produces less compounded return, and at crypto volatility that level arrives early.
Reading performance claims with drag in mind
Drag literacy converts directly into skepticism. Average monthly returns quoted without compounding are the cheapest inflation in the industry: a feed advertising "average +5% a month" across a volatile track can sit meaningfully below water in wealth terms while the arithmetic mean smiles, a cousin of the 95% win-rate games. The audit-grade question is always the same: what did one unit of capital, compounded through the whole track, become? Geometric or it did not happen. That single habit, insisting on compounded wealth outcomes over averaged return outcomes, neutralizes most of the arithmetic theater this market produces, and it costs nothing but the asking. Decisions are yours.