There is a piece of arithmetic that almost every new investor gets wrong by instinct: the belief that a loss and a gain of the same percentage cancel out. They do not. A portfolio that falls 20% does not get back to even with a 20% gain. It needs 25%. A portfolio that falls 50% needs to double — a 100% gain — just to return to where it started. This asymmetry is not a quirk. It is a structural feature of how percentages compound, and it quietly shapes how much risk is actually worth taking.
The Asymmetry Nobody Mentions
Start with the mechanics. Percentage changes are multiplicative, and they apply to whatever balance currently exists, not to your original balance.
Say you start with $100,000. It falls 20%, so you lose $20,000 and you are left with $80,000. Now you gain 20%. But 20% of $80,000 is only $16,000 — not the $20,000 you lost. You end at $96,000, still down 4%. The gain was applied to a smaller base than the loss was, so the same percentage moves fewer dollars.
To recover, the gain has to be calculated against the reduced balance. To climb from $80,000 back to $100,000 is a $20,000 move on an $80,000 base, which is 25%. The formula is straightforward:
recovery gain required = loss / (1 − loss)
A 10% loss needs an 11.1% gain. A 20% loss needs 25%. A 33% loss needs 50%. A 50% loss needs 100%. The relationship is not linear — it accelerates, and it accelerates hard once losses get large.
A Small Table of the Asymmetry
| Drawdown | Gain required to recover |
|---|---|
| 10% | 11.1% |
| 20% | 25.0% |
| 33% | 49.3% |
| 50% | 100.0% |
| 67% | 203.0% |
| 80% | 400.0% |
| 90% | 900.0% |
The bottom of that table is the important part. A 90% loss is not "twice as bad" as a 45% loss. It requires a 900% gain to undo, against roughly 82% for a 45% loss. Deep losses are disproportionately destructive because the base they have to recover from is so small.
This is pure arithmetic, not a market prediction. The table says nothing about how long a recovery takes or whether it happens at all — only what gain would be arithmetically required if it did. Time is the other half of the cost: a deep drawdown that eventually recovers still consumed years of compounding you do not get back.
Why the Average Return Hides This
Here is where the asymmetry becomes a trap for anyone reasoning from average returns. Consider two years: one at +50%, one at −50%. The average annual return is 0%. It sounds like a wash.
It is not. Apply them in sequence to $100,000. Up 50% gives $150,000. Down 50% gives $75,000. You have lost a quarter of your money despite a 0% average return. Reverse the order and you get the same $75,000 — the order does not matter here, only that both happened.
The number that actually describes your outcome is the compound (geometric) return, not the average (arithmetic) return. The arithmetic mean of +50% and −50% is 0%. The geometric mean is the annual rate that actually produces the ending balance: about −13.4% per year. The arithmetic mean overstates what you earned, and it overstates it more the more volatile the returns are.
Any return figure quoted as an average is an arithmetic mean unless stated otherwise, and the arithmetic mean is always greater than or equal to the compound return for a volatile series. When you see "this strategy averaged 12% per year," the balance you would actually have ended with corresponds to a lower compound number. Ask which one is being quoted.
Volatility Drag: The Gap Between Average and Compound Return
The gap between the arithmetic average and the compound return has a name: volatility drag, sometimes called variance drain. The more a return series swings around, the larger the gap, and the gap is roughly proportional to the square of the volatility.
A useful approximation: the compound return is about the arithmetic return minus half the variance. Two portfolios can advertise the same average return, but the one with wider swings delivers a lower compound return — a smaller actual balance — because every loss has to be recovered from a diminished base. This is the precise, mathematical reason that "high average return" and "good investment" are not the same statement. Volatility is not just discomfort. It is a direct, quantifiable subtraction from your compounded result.
This also reframes diversification and risk control. Reducing the volatility of a portfolio, even if it slightly lowers the arithmetic average return, can raise the compound return by shrinking the drag. Smoother is not merely calmer; smoother compounds better.
What This Means in Practice
The recovery table is an argument for caring about the depth of potential losses, not just the expected return. A strategy with a higher average return but the capacity for a 70%+ drawdown is making a bet that is arithmetically much harder to recover from than its average suggests.
It is also an argument against the most expensive behavioral mistake: selling near the bottom of a deep drawdown. Doing so converts a recoverable paper loss into a permanent one and forfeits the disproportionately large rebound that the math says is required. The asymmetry cuts both ways — it makes deep losses costly, but it also means that the gains available after a deep loss are large by necessity.
When you evaluate any portfolio or strategy, look up its worst historical drawdown alongside its average return. Then apply the recovery formula. If the worst drawdown implies a recovery gain you would not have the patience or time horizon to wait for, the average return is not the relevant number — the drawdown is.
None of this tells you how much risk to take; that depends on your time horizon, your goals, and your tolerance for watching a balance fall. What the arithmetic does tell you is that losses and gains are not opposites of equal weight. A 50% loss is not half a problem. It is a hole that requires doubling your money to climb out of, and pricing risk without that fact built in leads to portfolios that look better on a brochure than they behave in a bad year.
Originally published at pickuma.com. Subscribe to the RSS or follow @pickuma.bsky.social for new reviews.
Top comments (0)