The Lie of the Average

What “On Average” Really Means in Finance

There is a moment when the phrase “on average” stops being informative and starts being slightly evasive. Not wrong, not misleading, just incomplete in a way that matters. You hear it in finance constantly, delivered with the confidence of someone who believes the hard thinking has already been done. On average, markets go up. On average, this asset class performs well. On average, patient investors are rewarded. The phrase does a particular kind of work: it compresses complexity into a single number that feels portable, reassuring, and final. It allows both the speaker and the listener to move on. And yet, if you slow down and sit with it for a moment, you realize that “on average” is never describing what anyone is about to experience. It’s describing what became true after everything was already over, after uncertainty collapsed into history, after the path no longer mattered because it had already been survived. The average is a summary, not a forecast, and certainly not a strategy.

What makes this dangerous in finance is that averages quietly erase sequence, timing, and survivability, which are precisely the dimensions that determine whether an investor can stay in the game long enough for the math to work. No one invests in an average. People invest at a specific moment, with finite capital, finite patience, and very real constraints. They experience markets one year at a time, sometimes one month at a time, carrying forward the psychological and financial residue of whatever came before. Losses early in a sequence are not emotionally equivalent to losses later. Long stretches of stagnation test resolve in a way a clean long-term number never captures. The phrase “on average” flattens all of this into something tidy, but tidiness is not the same thing as relevance.

What Investors Actually Live Through

Take the S&P 500, the most familiar example of long-term investing wisdom. You’ll often hear that it has delivered roughly 13 percent average annual returns over the past fifteen years. That statement is true. It’s also the source of an enormous amount of confusion. No investor ever received 13 percent per year, evenly distributed. What they received was a path: volatility, drawdowns, recoveries, and gains. But it is entirely plausible, and historically common, for stretches of mediocre or even negative returns to be followed by bursts of exceptional performance, and for that entire sequence to collapse neatly into an attractive average. The arithmetic works perfectly, but the experience does not.

An investor who entered during the wrong part of that sequence, or who needed liquidity during the flat years, or who simply wasn’t prepared psychologically to sit through extended drawdowns would never experience that average in any meaningful sense. They would exit early, often right before the part of the distribution that makes the long-term number look so compelling in hindsight. This is why so many people who “know the numbers” still fail to capture them. They weren’t wrong about the statistics; they were wrong about what those statistics demanded of them along the way. And this is why the average always survives, but the investor doesn’t.

Once you see this clearly, the right questions change. The question stops being “what’s the average return?” and becomes something more uncomfortable. How deep were the drawdowns before things recovered? How long did it take to get back to breakeven after major losses? How many years were spent underwater? What did an investor actually have to endure, financially and emotionally, to stay invested long enough for the average to materialize? Those questions don’t change the historical record; they change whether the record is usable for a real human making a real decision under constraint. The average is calculated from survivors. The decision has to be made before survival is guaranteed, and that gap is where most investment mistakes are born.

Imagine It’s the Last Day of 2020

You are preparing for New Year’s Eve on December 31st, 2020, you take a look at the data available for the S&P 500 and see that over the previous ten years the S&P 500 has delivered roughly a 12 percent annual return. It’s the kind of statistic people cite precisely because it feels reasonable. You decide to invest, not because you’re trying to time the market or outsmart anyone, but because the math suggests patience will be rewarded.

The first year feels validating. 2021 delivers a strong return, something in the high twenties that provides a cushion almost immediately. Then comes 2022 with a sharp decline. Not catastrophic, but enough to erase almost all of the previous year’s gains. Two years in, after having done everything right, you are essentially back where you started. The long-term math hasn’t changed, but the experience has.

If you stay invested, the story resolves itself. 2023, 2024, and 2025 are strong. Five years after that perfectly reasonable decision, you end up with an average annual return of roughly 14 percent. The number looks excellent. But notice what had to happen for that average to exist. You had to survive the second year psychologically. You had to tolerate a stretch where the math said you were right and your account balance said otherwise. Change the order of those same returns, remove the early cushion, or introduce a liquidity need at the wrong moment, and the average never gets the chance to become real.

This is what “on average” looked like in real time.

How Strategy Breaks When the Downside Dominates

The same logic applies even more harshly in entrepreneurship and business strategy, where liquidity is lower, recovery is slower, and mistakes don’t wash out as easily. Founders are often told that the upside justifies the risk, that while many ventures fail, the expected value can still be attractive. That framing already assumes the wrong lens. A business does not need to be a bad idea on average to be devastating in practice; it only needs a downside that overwhelms your ability to absorb it. Entrepreneurial outcomes are not normally distributed. They are brutally skewed. Most cluster near zero or outright loss, while a small fraction dominate the upside. That’s not a motivational slogan, it’s a distribution. And in a skewed distribution, the average tells you almost nothing about what you’re likely to experience.

What actually matters is not whether success exists somewhere in the tail, but what happens if you land anywhere else. Does failure leave you intact, or does it impair your capital, your reputation, your energy, your ability to try again? These are not abstract questions, but structural ones. If the downside outcome removes you from the game, then the average upside is irrelevant. You don’t get credit for being right in theory if you can’t stay solvent in practice.

The same mistake shows up at the company level when businesses evaluate expansions, acquisitions, or new product lines. Average-case projections look clean, revenues normalize, costs smooth out, and synergies arrive on schedule. On paper, the expected outcome is positive. But strategy doesn’t fail on average. It fails at the point of maximum stress, when timing turns out to be wrong, integration drags, demand softens, or fixed costs refuse to bend. One sufficiently bad outcome can dominate years of competent execution, not because of incompetence or bad luck, but because asymmetric systems work that way. This is why experienced operators spend so much time thinking about downside scenarios. Not because they are pessimistic, and not because they distrust growth, but because they understand something averages obscure. Survival comes first. Optimization only matters if you are still standing.

Averages are summaries of what happened once uncertainty resolved. Decisions live before that resolution, in a world where outcomes are uneven, timing matters, and some mistakes do not get diluted by time. Once you internalize this, “on average” stops sounding like reassurance and starts sounding like a prompt to slow down and ask harder questions about the full range of outcomes and whether you can survive the ones that don’t get mentioned when the average is quoted.