HomeLearn
News & Articles
Market
My Account
Tools
AboutNewsletter☕ Buy me a coffee
LearnManage Your Risk › Module 2

Module 2 · The Arithmetic of Drawdowns Foundation

This module is mostly arithmetic, and it is the arithmetic that everything else in the course rests on. Once you have felt how asymmetrical losses are, position limits stop sounding like timidity and start sounding like accounting.

~9 min read · Not started

By the end of this module you'll be able to

  • Compute the gain required to recover from any given loss, and do the common cases in your head.
  • Explain volatility drag and why an average annual return is not the return you actually earn.
  • Read maximum drawdown and time under water, and say what each one does and does not tell you.
  • Estimate the portfolio-level damage a single position can do, before you size it.
  • Explain why the same expected return with lower volatility compounds to more money.

The recovery formula

Lose a percentage of your money and you must gain a larger percentage to get back to where you started, because the gain is computed on a smaller base. The formula is one line:

Gain needed to recover = L ÷ (1 − L)   where L is the loss as a decimal

Lose 20% and $100 becomes $80. Getting $80 back to $100 requires $20 on a base of $80 — 25%, not 20%. The asymmetry is mild at first and then turns vicious:

LossGain needed to break evenAt 8%/yr, years to recover
−10%+11.1%~1.4
−20%+25.0%~2.9
−33%+50.0%~5.3
−50%+100.0%~9.0
−66%+200.0%~14.3
−80%+400.0%~20.9
−90%+900.0%~29.9

The third column is the one that should stay with you. A 50% loss is not “half your money” — it is nine years of good returns spent standing still. And those are years you do not get back, which matters enormously if you are 58 and not at all if you are 24. This is the arithmetic behind the claim that avoiding the deep hole matters more than catching the big winner: the hole compounds against you at exactly the rate the winner compounds for you.

Do these three in your head−20% needs +25%. −33% needs +50%. −50% needs +100%. Memorise those three and you can interpolate anything in between well enough to make a decision at a moment when you will not want to open a calculator.

Volatility drag

Here is a result that surprises almost everyone the first time. Two portfolios, both with an average annual return of 0%:

Portfolio B lost a quarter of its money while averaging zero. Nothing is wrong with the arithmetic — the problem is that the “average” being quoted is the arithmetic mean, and money compounds at the geometric mean. The gap between them is volatility drag, and it is real money.

Geometric return ≈ arithmetic return − (σ² ÷ 2)

The approximation is good for ordinary levels of volatility. Read it carefully: volatility subtracts from your compounded return even when it does not change your average return. A portfolio averaging 9% with 15% volatility compounds at roughly 9 − (0.15² ÷ 2) = 9 − 1.1 = 7.9%. Push volatility to 30% and the same 9% average compounds at roughly 4.5% — you lost nearly half your return to turbulence alone.

This is the mathematical reason risk management pays for itself rather than merely feeling safer. Two strategies with identical expected returns are not equivalent: the smoother one ends with more money. It is also why the marketing figure “average annual return” should always be checked against the compound annual growth rate (CAGR), which is the number you can actually spend.

Volatility drag: same average return, different ending wealth Two paths both average 8 percent per year. The smooth path compounds to a higher ending value than the volatile path, whose swings cost it several percentage points of compound growth. smooth volatile Identical arithmetic average. Different geometric outcome. The gap is volatility drag — roughly σ²/2 per year, every year.
Turbulence is not just uncomfortable; it is expensive. Two portfolios with the same average return end in different places, and the difference grows with the square of volatility.

Maximum drawdown & time under water

Maximum drawdown is the largest peak-to-trough decline over a period, measured from the highest value reached to the lowest value that followed before a new high. It answers the question standard deviation cannot: what is the worst this actually got?

It is the single most useful risk statistic for a private investor, for one reason: it is the number you have to live through. Nobody experiences a standard deviation. Everybody experiences watching their account fall by a third.

Time under water is its neglected twin — how long the portfolio stayed below its previous peak. It matters more than depth for anyone drawing an income, and it is psychologically brutal in a way that depth is not. A fast 40% crash that recovers in a year is easier to hold than a 25% decline that grinds sideways for six.

The limitation of maximum drawdown. It is a single realised observation from one particular history, not a bound. “The worst drawdown in the backtest was 22%” is not a promise that 22% is the worst possible — it is a statement that in this one sample, nothing worse happened yet. Treat any historical maximum as a floor for what to plan around, not a ceiling. A sensible planning assumption for a diversified equity portfolio is that a decline of roughly 50% will happen at some point in a multi-decade investing life, because it repeatedly has.

What real drawdowns have looked like

Approximate peak-to-trough declines in broad North American equity indices, on a price basis. Figures are rounded and illustrative — the precise number depends on the index, whether dividends are included, and the exact dates used.

EpisodeApprox. declineApprox. time to recover
1929–1932 (US)~86%Over two decades
1973–1974~48%Several years
2000–2002 (dot-com)~49%~5 years
2007–2009 (financial crisis)~57%~4 years
2020 (COVID)~34%~5 months
2022 (rates/inflation)~25%~2 years

Two things to take from the table. First, the frequency: a serious drawdown is not a freak event, it is a recurring feature. Anyone investing for thirty years should expect to sit through several. Second, the enormous variance in recovery time — five months in 2020, several years in 2000 and 2007. Recovery speed is not something you can plan on, which is exactly why the cash buffer in Module 8 exists.

What one holding can do to you

Scale the arithmetic down to a single position and you get the formula that Module 5 is built on:

Portfolio impact = position weight × position loss

A 5% position that falls 50% costs the portfolio 2.5%. A 25% position that falls 50% costs 12.5% — and now you need a 14.3% gain from everything else just to stand still. A 40% position that goes to zero costs 40%, which needs +67% to recover, which at 8% a year is a lost decade caused by one decision.

Worked example — two portfolios, one bad holdingBoth hold ten stocks and both have one blow up, falling 70%. Portfolio A is equal-weighted at 10% each: the damage is 0.10 × 0.70 = 7%. Portfolio B put 35% into its “best idea” and it was the one that blew up: 0.35 × 0.70 = 24.5%, requiring a 32% gain from what remains. Same stock-picking record. Wildly different outcome. The difference was sizing, not skill.

What the arithmetic implies

  1. Survival dominates. A strategy with a small chance of total loss has a poor long-run outcome no matter how good its average year is, because zero multiplied by anything is still zero. Module 8 makes this precise with leverage.
  2. Smoothness is worth paying for. Not because volatility is uncomfortable, but because volatility drag is a real cost measured in real dollars.
  3. The deep hole is the enemy. Losing less in bad years contributes more to a long-run result than winning more in good ones. This is not a philosophy; it is what L ÷ (1 − L) says.
  4. Size before you buy. Portfolio impact is weight × loss, and weight is the only one of those two you control.
💡 Want to see the compounding for yourself? The DCA calculator lets you run different return paths, and DCA vs lump sum covers the timing question this raises.

Educational purposes only; not financial advice. Historical figures are illustrative and past drawdowns are not a forecast of future ones. Always do your own research and consult a licensed advisor.