What Is Ergodicity? (Definition and Why It Matters)
Ergodicity is a property of stochastic (random) processes. A process is ergodic if the time average — what one participant experiences over a long enough period — and the ensemble average — what the average participant experiences across all paths at one point in time — converge.
In a perfectly random fair-coin process, ergodicity holds: over enough time, any individual's sequence of flips will average to approximately 50% heads, matching the 50% ensemble average. The lucky runs and unlucky runs even out over time.
Why This Matters for Performance Evaluation
In competitive environments with significant randomness — trading, venture investing, certain business domains — ergodicity means that lucky sequences will eventually revert. The outstanding run of a manager who has outperformed for five years in a game with meaningful randomness does not predict outstanding performance going forward. Their forward expected return is the same as any other participant's, because time tends to wash out the noise and reveal the underlying signal.
This is why Taleb's "Masters of the Universe" — traders celebrated at their peak — consistently end up in diminished circumstances later. Not because their skill disappears. Because the ergodic process absorbs the favorable sequence. The underlying distribution reasserts itself.
The Practical Application
For evaluating track records: outstanding performance over short periods in random-adjacent domains is weak evidence of genuine edge, because ergodicity guarantees that favorable sequences eventually revert. What distinguishes genuine edge from luck is:
- Long track records across multiple independent environments
- Performance that holds across regime changes (when the conditions that supported luck-based runs change)
- Consistency of process explanations (the same framework that worked in one regime continues working when conditions change)
Ergodicity is also a humility device for self-evaluation. If I'm in a good run in a domain with meaningful randomness, the intellectually honest question is: how much of this is ergodic luck that will eventually revert, versus genuine skill that will hold across environments?
For the full framework, read Fooled by Randomness: How Luck Masquerades as Skill.