Ergodicity in Risk: Why Ruin Changes Everything

Key Takeaways

• Ergodicity economics risk analysis separates the average outcome over time from the average across many parallel bets.
• When ruin is possible, the time-average outcome is worse than the expected value, so positive-expectation bets can still ruin you.
• The common mistake is sizing positions by expected value, which ignores the path and the absorbing barrier of ruin.
• This view favors smaller bets and survival, directly informing position sizing and the use of the Kelly criterion.

What It Is

Ergodicity is a property of a random process. A process is ergodic if the average of one path followed over a long time equals the average across many paths at one moment. For ergodic processes, the expected value is a reliable guide to what any individual will experience.

Most investment returns are not ergodic, because they compound multiplicatively and because ruin is an absorbing state you cannot return from. In that setting, ergodicity economics risk analysis argues that the relevant quantity is the time-average growth rate of a single trajectory, not the expected value averaged across hypothetical parallel investors.

The framework was developed by physicist Ole Peters. Its central message for risk is simple and sharp: you only get one trajectory through time, so you should optimize the outcome along that single path, not the average across worlds you never inhabit.

The Intuition

Imagine a bet that wins 50 percent on heads and loses 40 percent on tails, fifty-fifty. The expected value per round is positive: average of plus 50 and minus 40 is plus 5 percent. By that logic, you should bet repeatedly.

But play it out over time on one bankroll. A win then a loss multiplies your money by 1.5 then 0.6, which is 0.9, a 10 percent loss. Repeat this and the single path grinds toward zero even though the expected value is positive each round. The ensemble average rises while the typical trajectory falls.

That is the heart of ergodicity economics risk thinking. The expected value describes the average across many gamblers; the time average describes the one gambler living through the sequence. When the two disagree, the time average is what actually happens to you.

How It Works

For multiplicative dynamics, the right object is the expected log growth rate, not the expected return. The time-average growth rate is:

g_time = E[ ln(1 + r) ]

The ensemble-average growth rate, which ordinary expected value tracks, is:

g_ensemble = ln(1 + E[r])

Where:

r           = single-period return, a random variable
E[ ]        = expected value
ln          = natural logarithm

By the mathematics of concave functions, g_time is less than or equal to g_ensemble, and they are equal only when there is no volatility. The gap between them grows with volatility, which is why aggressive, high-variance strategies can have a positive expected value yet a negative time-average growth rate. Maximizing g_time leads directly to logarithmic, ruin-aware position sizing.

Worked Example

Take the coin bet above: plus 50 percent or minus 40 percent, equal odds, on your full bankroll each round.

The ensemble average growth per round is positive 5 percent, so an expected-value optimizer bets the maximum. The time-average growth is the average of ln(1.5) and ln(0.6), which is the average of plus 0.405 and minus 0.511, equal to minus 0.053. That is a negative 5.3 percent compound rate per round. Over many rounds, the single bankroll heads to zero.

Now bet only a fraction of the bankroll each round instead of all of it. Sizing the bet small enough turns the time-average growth positive, capturing the edge without the multiplicative decay toward ruin. The expected value never told you to do this. The ergodicity view did, and it is the same logic behind the Kelly criterion.

Common Mistakes

1. Optimizing expected value with money you cannot replace. Expected value is an ensemble average. With one bankroll and a ruin barrier, it can recommend bets that destroy you over time.

2. Ignoring the absorbing barrier. Ruin is permanent. Once your capital hits zero you are out, so any non-zero ruin probability eventually dominates a single long path.

3. Confusing high expected return with high time-average growth. Volatility drags down compound growth. A higher-variance strategy can have a worse time average despite a better expected value.

4. Betting the same fraction regardless of edge and odds. Time-average optimization scales bet size to the edge. Fixed large bets ignore the multiplicative penalty of losses.

5. Assuming diversification fully solves it. Spreading bets helps, but if positions crash together the combined path can still face multiplicative ruin. Path and co-movement still matter.

Frequently Asked Questions

What is ergodicity economics risk in simple terms? Ergodicity economics risk analysis asks whether the average result across many gamblers matches what one gambler gets over time. When losses compound and ruin is possible, the single person's long-run outcome can be far worse than the average suggests.

How does ergodicity economics risk affect investment decisions? It pushes you to size bets for survival of your one trajectory, not for the highest expected value. Even a positive-expectation strategy gets sized small enough that compounding losses cannot drag your capital toward zero.

What is a real-world example of ergodicity in risk? A bet that pays plus 50 percent or minus 40 percent on equal odds has positive expected value, yet betting your whole bankroll each round drives you toward zero over time. Sizing the bet small turns the long-run path positive.

How can investors apply ergodicity thinking effectively? Optimize the expected logarithm of wealth rather than the raw expected value, keep any single loss survivable, and use fractional position sizing such as the Kelly criterion to capture an edge without courting ruin.

How is ergodicity in risk different from ergodicity economics? Ergodicity in risk is the applied lesson for sizing bets and avoiding ruin. Ergodicity economics is the broader theoretical framework, developed by Ole Peters, that rebuilds decision theory around time averages instead of expected utility.

Sources

1. Peters, O. (2019). "The ergodicity problem in economics." Nature Physics, 15, 1216-1221. https://www.nature.com/articles/s41567-019-0732-0
2. Ergodicity Economics. "Ergodicity, jail, and time scales." https://ergodicityeconomics.com/2019/05/16/ergodicity-jail-and-time-scales/
3. ScienceDaily. "This 'fix' for economic theory changes everything from gambles to Ponzi schemes." https://www.sciencedaily.com/releases/2019/12/191202113024.htm
4. Peters, O., and Adamou, A. "The Two Growth Rates of the Economy." arXiv:2009.10451. https://arxiv.org/pdf/2009.10451

Disclaimer

This article is educational content only and is not financial advice. Nothing here is a recommendation to buy, sell, or hold any security. Consult a licensed advisor before making investment decisions.