Why losses and gains are not mirror images, and why this asymmetry is the most important force acting on your wealth
A 50% loss followed by a 50% gain leaves you at 75 cents on the dollar. Not back to even. Not close to even. Permanently poorer.
Most investors understand this fact in the abstract. They can follow the arithmetic: $100 falls 50% to $50, then rises 50% to $75. The numbers are simple. What almost no one internalises is that this is not a quirk of the example. It is the defining characteristic of multiplicative systems, and it has consequences that the investment industry systematically understates.
The asymmetry between losses and gains is not linear. It is not proportional. It is exponential. A 10% loss requires an 11% gain to recover. A 30% loss requires a 43% gain. A 50% loss requires a 100% gain. A 75% loss requires a 300% gain. The deeper the hole, the more violently the recovery requirement escalates. The relationship is a curve, not a line, and the curve steepens relentlessly with every additional percentage point of drawdown.
In Episode 1, we established that variance is a tax on compounding. This episode explores the mechanism by which that tax is collected. It is collected through drawdowns: the periods when a portfolio falls from its peak value. And the collection is not proportional. It is convex: each additional unit of loss costs exponentially more than the last. This convexity is the reason why, in a world where wealth compounds multiplicatively, the single most important property of any return stream is not its average, not its best year, and not its Sharpe ratio. It is the depth of its worst drawdown. Because the deeper that drawdown goes, the more catastrophically the compounding engine is damaged.
The Recovery Table
The following table should be committed to memory. It governs every investment decision you will ever make, whether you are aware of it or not.
Read the table from top to bottom and notice how the second column accelerates. The first 10% of loss requires 11.1% to recover: barely noticeable. But the relationship is not additive. It is governed by the formula: Required Recovery = 1/(1 – Loss) – 1. That fraction in the denominator is what produces the exponential character. As the loss approaches 100%, the denominator approaches zero, and the required recovery approaches infinity.
Now read the third column. It translates the required recovery into time, assuming the portfolio compounds at 8% per year after the drawdown, roughly the long-term CAGR of the S&P 500. A 20% drawdown costs about 3 years of compounding time. A 50% drawdown costs 9 years. A 75% drawdown costs 18 years. These are not years of negative returns. They are years spent climbing back to the peak from which the portfolio fell, during which the investor earns returns but creates no new wealth. The compounding engine is running, but it is running in place.
A drawdown does not just destroy capital. It destroys time. And time is the one ingredient of compounding that can never be replaced.
This is the asymmetry trap. The investment industry speaks of losses and gains as though they are symmetric: a 30% down year and a 30% up year, and you are roughly back to even. You are not. You are at 91 cents, and the 9% shortfall will take more than a year of compounding at 8% to reclaim. Scale that to a 50% loss, which the S&P 500 delivered during the Global Financial Crisis, and the investor is looking at nine years of compounding just to return to the starting point. The investor who avoided that drawdown entirely spent those nine years compounding forward from the peak. Not clawing back. Building.
Chart 4: The exponential recovery curve: required gain to recover from a given percentage loss. The S&P 500’s maximum drawdown (50.9%) is marked alongside the TTU TF Index’s (21.0%).
Chart 4 plots the recovery curve. The horizontal axis is the percentage loss. The vertical axis is the required gain. Below about 20%, the curve is gentle, nearly linear. Beyond 30%, it begins to steepen visibly. Beyond 50%, it becomes nearly vertical. This is the geometry of the asymmetry trap rendered as a picture: the left side of the curve is forgiving, the right side is catastrophic, and there is no sharp boundary between them, only a continuous and accelerating steepening.
Two benchmarks are marked on this curve. The S&P 500’s maximum drawdown of 50.9% required a gain of 104% to recover. The TTU Trend Following Index’s maximum drawdown of 21.0% required a gain of 27%. The S&P 500 sits deep in the steep region. The TF Index sits in the shallow region. The difference is not merely quantitative. It is qualitative: 104% and 27% are different categories of mathematical burden. One is a multi-year project requiring exceptional market conditions. The other is a speed bump.
The Gain That Does Not Match the Loss
The recovery table documents the cost of losses. There is a companion asymmetry that the industry almost never demonstrates: an equivalent gain does not add the same wealth that an equivalent loss destroys. The relationship is not symmetric in either direction.
Consider four investors, each starting with $100,000 and compounding at 8% per year for 26 years. Investor A compounds without interruption and arrives at $739,635. Investors B, C, and D each experience a single event in year 8: not a loss, but a gain of identical magnitude to the losses in the earlier table. Investor B gains 20%. Investor C gains 30%. Investor D gains 50%.
Their terminal values: $887,562 (Investor B, +20%), $961,526 (Investor C, +30%), $1,109,453 (Investor D, +50%). The windfall gain adds wealth, as expected. But now compare these gains against the losses of equivalent magnitude from the earlier table. A 20% loss destroyed $191,757 of terminal wealth relative to the uninterrupted baseline. A 20% gain adds only $147,927. The gain adds 77 cents for every dollar the loss destroys. A 50% loss destroyed $397,212. A 50% gain adds $369,818. The gain adds 93 cents for every dollar the loss destroys.
A 50% loss destroys more terminal wealth than a 50% gain creates. The damage and the windfall are not symmetric, even when their magnitude is identical. This is not a feature of bad markets. It is a feature of multiplication.
The reason is path dependency. The loss occurs when the capital base is at its peak, destroying the most compounding-capable dollars in the portfolio. The gain occurs after the loss, compounding a smaller base. Even at equal magnitude, the loss strikes first and strikes harder. In a multiplicative system, the sequence in which events arrive is not a detail. It is the determinant of terminal wealth.
The Compounding Interruption
The recovery table measures drawdown cost in terms of required return. But the true cost is larger, because while the damaged portfolio is recovering, an undamaged portfolio would be compounding forward. The drawdown does not merely pause the compounding engine. It forces the engine to run in reverse, spending energy to recover lost ground, while an undamaged engine is spending that same energy creating new wealth. The difference compounds forward through every remaining year.
Consider four investors, each earning 8% per year on $100,000. Investor A compounds without interruption for 26 years. Investors B, C, and D each suffer a single drawdown in year 8, of varying severity, then resume compounding at 8%.
A single 50% drawdown, occurring once in a 26-year investment lifetime, permanently destroys 54% of terminal wealth. Not 50%. Fifty-four percent. Because the drawdown destroys not only the capital lost in the event itself, but the future compounding on that capital, and the compounding on the compounding, cascading forward through every remaining year.
Investor B and Investor A experienced identical returns in 25 of their 26 years. They diverged in a single year. That single year of divergence cost Investor B nearly $400,000 of terminal wealth. This is the compounding interruption at its most stark: a momentary event with permanent consequences. The drawdown was over in months. The wealth it destroyed is gone forever.
Notice the convexity within the table itself. Moving from a 20% drawdown to a 50% drawdown adds 30 percentage points of loss, but the terminal wealth destruction jumps from 26% to 54%. The damage does not scale linearly with the severity of the drawdown. It accelerates. The same exponential relationship that governs the recovery table governs the destruction of terminal wealth. Compounding is cruel in both directions: it builds wealth exponentially when left alone, and it destroys wealth exponentially when interrupted.
The Underwater Evidence
Theory predicts that drawdown containment should be the primary driver of long-term geometric wealth. Let us now see what 26 years of real data show.
The chart below displays the underwater equity curve for three benchmarks: the S&P 500, Berkshire Hathaway, and the TTU Trend Following Index. An underwater curve plots the percentage distance below the all-time high at each point in time. When the line sits at zero, the portfolio is at or above its peak. When the line dips, the portfolio is in drawdown. The deeper the dip, the deeper the damage to the compounding base.
Chart 5: Underwater equity curves: S&P 500, Berkshire Hathaway, and TTU TF Index. January 2000 to January 2026. Crisis periods shaded.
The visual tells the story before the numbers do. The S&P 500’s curve is dominated by two enormous craters: the dot-com collapse, which dragged the index 44.7% below its peak and took over six years to recover; and the Global Financial Crisis, which took it to its deepest drawdown and required four and a half years to recover. Between these two events, there was only a brief window in 2006 and 2007 when the index touched its previous high before plunging again.
Berkshire’s curve shows a similar structure with similar depth, reaching 44.5% below peak during the GFC, though its higher CAGR enabled somewhat faster recoveries.
The TTU Trend Following Index inhabits a different visual world. Its maximum drawdown was 21.0%. It never breached 30%. It never approached 40%. The shape of its drawdown profile is qualitatively different: where the S&P 500 shows deep, prolonged craters that consume years of compounding, the TF Index shows shallow dips that resolve comparatively quickly.
The numbers quantify what the eye can see:
Read the S&P 500 column. It spent 38% of the entire 26-year period more than 10% below its previous high. More than a third of the time, the compounding engine was running at a deficit. It spent 5.6 years, 21% of the period, more than 20% below peak. It spent over two years below 30%. Nearly a full year below 40%. During all of those months, every dollar in the S&P 500 was recovering, not creating new wealth. The engine was paying down the debt imposed by the drawdown.
Now read the TF Index column. It spent 23% of the period below 10%, which is not trivial and reflects the reality that trend following is far from a smooth experience. But it spent only two months out of 26 years below 20%. It never breached 30%. Never touched 40%. Never came close to 50%. The compounding engine was never severely damaged. The portfolio never entered the steep, punitive region of the recovery curve where every additional percentage point of loss becomes exponentially more expensive to reclaim.
The S&P 500 spent one-fifth of the last quarter-century more than 20% below its peak. The TF Index spent two months. This is not a marginal difference. It is a structural difference in how the compounding engine is treated.
The Time Cost of Deep Drawdowns
Let us express the damage in the currency that matters most for compounding: time.
The S&P 500 peaked in August 2000, fell 44.7% through the dot-com collapse, and did not recover its peak until October 2006. That is 74 months. Six years and two months during which every dollar invested in the index was climbing back to where it had been, generating no new wealth for the investor. Not losing money in absolute terms, perhaps. But losing the most valuable thing compounding requires: uninterrupted forward progress.
The index peaked again in October 2007, fell by more than half through the GFC, and did not recover until March 2012. Another 53 months. Four and a half years.
Combined, the S&P 500 investor spent more than ten years of a 26-year period recovering from just two drawdowns. Ten years out of twenty-six. During those ten years, the investor held the position, endured the volatility, paid the fees, and at the end of the recovery was precisely where they had been at the beginning. The compounding engine ran for a decade and produced nothing. Meanwhile, any portfolio that avoided those drawdowns was compounding forward from the peak through that entire period, building wealth upon wealth upon wealth.
This is the time cost of deep drawdowns. The drawdown itself may last months. The recovery may take years. But the true cost is neither the drawdown nor the recovery. It is the compounding that did not happen during both. Those years are gone, and the wealth they would have created is gone with them. This is what it means to say that a drawdown destroys time. The clock of compounding stops when the portfolio falls, and it does not restart until the peak is recovered. Every month spent below the peak is a month of compounding permanently forfeited.
The Eighth Wonder and Its Fragility
The quote attributed to Einstein, that compound interest is the eighth wonder of the world, may be apocryphal. No one has located it in his writings or recorded speeches. But the mathematics behind the sentiment are not apocryphal. They are demonstrable, and they reveal something the popular version of the quote obscures.
Compounding is not merely powerful. It is fragile. Its power is entirely conditional on continuity. Compound growth at 8% per year is extraordinary over 26 years, but only if it runs undisturbed for all 26 of them. Interrupt it once with a single 50% drawdown, and 54% of the terminal wealth vanishes. Interrupt it twice, as the S&P 500 was interrupted by the dot-com crash and the GFC, and the cumulative damage cascades through every subsequent year of the investor’s life.
The popular understanding of compounding is that it works for you: silently, automatically, inevitably. This is true only in the absence of severe drawdowns. In the presence of drawdowns, the same exponential mathematics that builds wealth during appreciation works against you during decline, and the destruction is always faster than the creation. A market can fall 50% in twelve months. It almost never rises 100% in twelve months. The eighth wonder of the world cuts both ways, and the cutting edge is sharper on the downside.
This reframes the entire question of investment strategy. The conventional question, the one the industry is built around, is: which assets should I select? The geometric question is different and more fundamental: which process best protects the compounding engine from interruption?
Because the selection of assets, however brilliant, cannot overcome the exponential cost of the drawdowns those assets produce along the way. A brilliant stock pick that falls 60% before it eventually recovers requires a 150% gain to return to its purchase price. The years spent recovering are years during which the compounding engine is running at a deficit. The pick may prove right in the end. But the geometry has already extracted its penalty, and the penalty is permanent. This is why the prevailing narrative of wealth creation, built around the identification of undervalued assets, around stock picks and sector bets and macro calls, is fundamentally incomplete. Selection tells you what to own. Geometry tells you how much of what you own you will actually keep. And keeping it, protecting the compounding base from the convex destruction of deep drawdowns, matters more for terminal wealth than any selection decision ever made.
The Geometric Case for Protection
All of this leads to a conclusion that is mathematically rigorous and deeply counterintuitive: in a multiplicative world, the investor who systematically protects against the worst outcomes will, given sufficient time, compound more wealth than the investor who systematically captures the best ones.
The logic flows directly from the convexity of the recovery curve. Avoiding a 50% drawdown is geometrically equivalent to earning a 100% return, because 100% is the return required to recover from the drawdown that was avoided. But the 100% gain that recovery demands is far harder to achieve than the 50% loss was to inflict. Drawdowns happen faster, more violently, and more frequently than equivalent recoveries. The asymmetry is not just mathematical. It is empirical.
A process that systematically prevents the portfolio from entering the steep region of the recovery curve, below 30%, below 40%, below 50%, confers a geometric advantage that no amount of upside capture can match. The advantage is structural. It does not depend on forecasting. It does not depend on selecting the right stocks. It does not depend on timing. It depends only on the mathematics of compounding in a world where drawdowns are asymmetric and the recovery curve is convex.
We have seen one investment process whose drawdown profile is categorically different from the S&P 500’s. The TTU Trend Following Index never breached 30%. It spent two months below 20%. Its maximum drawdown of 21% required only a 27% gain to recover, placing it permanently in the shallow, gentle portion of the recovery curve where the mathematics of compounding are forgiving rather than punitive.
We have not yet explored why the process produces this profile. That is the subject of the episodes that follow. But the geometric case is established: any process that keeps a portfolio in the shallow end of the recovery curve, systematically and across market cycles, possesses a compounding advantage that accrues quietly, relentlessly, over decades. It is an advantage that the arithmetic average will never show you, that the Sharpe ratio will never measure, and that the industry’s performance reporting infrastructure is not designed to reveal. But it is real, and it is where the money lives.
The Running Ledger
Our $100,000 invested on 1 January 2000. The terminal values are unchanged from Episode 1. This episode adds a new column: the recovery burden each portfolio carried from its worst drawdown.
The geometric burden is telling. The S&P 500’s worst drawdown required 104% to recover, a burden equivalent to 9 years of compounding at 8%. Berkshire’s required 80%, or roughly 8 years. The TF Index required 27%, achievable in about 3 years. Despite producing similar terminal wealth to the S&P 500 ($667K vs $748K), the TF Index achieved it while carrying roughly one-third the geometric recovery burden. It arrived at a comparable destination via a categorically easier path.
The Bridge
We have now established two structural truths about multiplicative wealth. Variance is a tax on compounding. The tax is collected convexly through drawdowns. Together, they invert the conventional hierarchy of what matters: protecting the compounding base from deep drawdowns matters more, geometrically, than maximising upside capture.
But the industry does not measure this. The industry’s dominant performance metric, the Sharpe ratio, treats upside volatility and downside volatility as equally dangerous. It penalises a portfolio for rising sharply in the same way it penalises one for falling sharply. For a return profile that is symmetrically distributed, this distinction is irrelevant. For a return profile that is positively skewed, which describes 37 of 41 long-term trend followers, it produces rankings that are not merely inaccurate but precisely backwards.
In Episode 3, we will dismantle the Sharpe ratio: the most widely used, most deeply trusted, and most fundamentally misleading risk metric in finance. We will show why it structurally penalises the return profiles that geometric compounding rewards, and introduce the tools that replace it when the goal is not to evaluate a Gaussian abstraction but to measure the creation of real, geometric, multiplicative wealth.
Data and Sources
All performance data from the NilssonHedge Trend Following Performance Database (January 2000 to January 2026). All returns are net of management and performance fees. The recovery formula is exact: Required Gain = 1/(1 – Loss) – 1. The “years at 8% CAGR” column computes ln(1 + Required Gain) / ln(1.08). The compounding interruption demonstration assumes a constant 8% annual return with a single drawdown event in year 8. Time-in-drawdown statistics are computed from monthly return data. S&P 500 drawdown events: dot-com peak August 2000, trough February 2003, recovery October 2006 (74 months); GFC peak October 2007, trough February 2009, recovery March 2012 (53 months). The Einstein attribution: no verified primary source exists; the earliest known appearance in print dates to the 1980s.
Want the theoretical foundation for why trend following works?
The Fractals of Finance: Determinism, Adaptation and the Geometry of Markets
The book explores the full architecture of feedback, fat tails, and fractal structure in financial markets, and what it means for how we trade, invest, and understand risk.
Available now on Amazon in paperback, hardcover, and Kindle.
Want a practical field manual for trading trends and capturing outliers?
The Aussie Turtles Trend Following Guide: A Field Manual for Hunting Outliers adapts the timeless principles of the original Turtle traders into a systematic, rules-based approach for modern markets. Co-authored with Adam Havryliv.
Available now on Amazon in paperback, hardcover, and Kindle.