Source: Michael Mauboussin.
Mauboussin’s research on skill, luck, and statistics in sports produces some of the most directly applicable insights available for understanding investment performance. The core question his work addresses, how much of any observed result reflects genuine skill versus random variation, is the same question every investor and trader must answer when evaluating their own approach or any manager’s track record.
The fundamental insight is that outcomes in any competitive activity are produced by some combination of skill and luck. On the pure skill end of the spectrum, the best competitor wins almost every time. Chess grandmasters consistently beat weaker players. The luck component is near zero. On the pure luck end, outcomes are essentially random regardless of skill. Slot machines pay out based on probability, not player quality.
Most investment activities sit closer to the luck end than most participants acknowledge. Short-term market calls involve substantial luck because no one can consistently forecast where markets will go. This does not mean investing is entirely random. It means that the skill component is real but only reveals itself over large enough samples. A money manager who outperforms over three months has provided almost no statistical evidence of skill. The same manager outperforming consistently across ten years and multiple market environments has provided substantial evidence.
The sports connection sharpens this. Mauboussin found that hockey involves more luck than basketball because fewer scoring opportunities per game mean individual games are more random. Basketball’s higher scoring frequency produces more opportunities for skill to reveal itself within a single season. The lesson for investors: the fewer data points available, the less any individual result tells you about the underlying skill level. A manager with three months of returns has provided the investing equivalent of a low-scoring hockey game as evidence of their ability.
Reversion to the mean is the statistical consequence of this reality. When observed outcomes include a significant luck component, extreme results tend to move toward average over subsequent periods. The trader who made 150% last year is more likely to make average returns next year than to repeat 150%. Not because they lost their skill, but because last year’s result included a favorable luck component that will not systematically repeat. This is why chasing past performance is one of the most reliably losing investment strategies: it buys the luck component at the peak of its expression and holds through the reversion.
The practical implication for trend following investors is direct. Evaluating any systematic approach requires a sufficiently large sample, covering multiple market environments, to separate the skill in the system from the luck in the specific period. A system that works during a sustained trending environment is not necessarily a robust system. One that produces consistent results across trending periods, choppy periods, crises, and bull markets over many years has provided the sample size needed to make a reasonable assessment of its genuine edge.
Frequently Asked Questions
What is the core question Mauboussin’s sports statistics research addresses for investors?
How much of any observed investment performance reflects genuine skill versus random variation. The same statistical tools used to distinguish skill from luck in sports, measuring how much outcomes vary from what pure chance would produce, can be applied to investment track records to determine whether a manager’s performance reflects a systematic edge or a lucky period.
Why is reversion to the mean important for evaluating trading performance?
Because extreme results in any activity with a significant luck component tend to move toward average over subsequent periods. A manager who produced exceptional short-term returns likely benefited from favorable conditions that will not systematically repeat. Investing in that manager at the peak of their lucky period means paying for returns that are more likely to mean-revert than to continue.
Why do short-term investment results tell us little about skill?
Because over short periods, the luck component dominates the skill component. A small sample of results is consistent with both genuine skill and random variation. Only large samples across diverse market environments provide the statistical power to distinguish between them. This is why Mauboussin advises assessing skill over longer time horizons than the investment industry typically uses.
How should systematic trend following be evaluated using Mauboussin’s framework?
By examining audited performance records across multiple years and multiple market environments, including periods that favor the approach and periods that do not. A trend following system evaluated only during a sustained trending period is like a basketball player evaluated only against much weaker opponents. The genuine test requires varied conditions over a sufficient time horizon to distinguish the system’s edge from the favorable environment it was tested in.
Trend Following Systems
Want to learn more and start trading trend following systems? Start here.
