Illusion of skill

The illusion of skill is the tendency to attribute success to one’s own ability when the success is actually due to luck. A fund manager beats the market for five years and is celebrated as a genius, even though statistical chance alone would produce some five-year winners among a large population of managers. The illusion of skill is especially pernicious in domains with significant random variation, like investing.

The mechanism

The illusion of skill arises from three sources:

Survivorship bias. Among a large population of managers, some will beat the market by chance. If you observe the survivors (managers with strong track records), you see only the winners, not the losers. This creates the illusion that outperformance is skill-driven, when actually it is chance at work in a large sample.

Small sample size. Five years of data is a small sample for measuring skill. A manager’s true skill-adjusted return might be 0.5% per year, but with market volatility of 15%, she could easily beat the market 4 out of 5 years by luck alone.

Fundamental attribution error. Humans default to attributing others’ behavior to their character or skill. If a manager beats the market, she is “skilled.” If she underperforms, she is “unlucky.” This asymmetry in attribution is the fundamental attribution error.

Illusion of skill in professional investing

The clearest example is fund managers. Studies by Burton Malkiel and others show that past performance of mutual funds does not predict future performance. A manager who beat the market for five years is no more likely to beat it in the next five years than a randomly selected manager.

Yet investors pour money into funds with strong track records. They attribute the past performance to skill (the illusion) rather than luck. The manager, having made money, might also suffer from the illusion, leading to overconfidence.

Illusion of skill and stock picking

Individual stock pickers often fall victim to the illusion. An investor makes several correct stock calls and feels like she has skill. But with thousands of stocks and daily movements, luck is a sufficient explanation. The illusion of skill drives overconfidence, which drives more concentrated bets, which eventually leads to large losses.

Illusion of skill and entrepreneurship

Entrepreneurs are especially susceptible. A startup succeeds, and the founder is celebrated as a genius. But most startups fail, and those that succeed often benefit from luck (good timing, fortunate hiring, market tailwinds). The illusion of skill leads entrepreneurs to believe they can repeat success, when the luck might not.

Distinguishing luck from skill

True skill has several hallmarks:

• Consistency. Skill shows up repeatedly, not just in a lucky streak. A manager with true skill beats the market consistently over decades.
• Replicability. Skill can be explained and replicated. A manager with true skill can articulate her process and it can be followed by others with similar results.
• Out-of-sample performance. Skill shows up in new data, not just the data used to build the strategy. Many strategies fail this test.

In investing, true skill is rare. Most apparent skill is luck.

Illusion of skill and overconfidence

The illusion of skill feeds overconfidence bias. A lucky performer believes she is skilled, which boosts her confidence in her decisions, which drives more aggressive bets, which eventually leads to a crash. The illusion of skill in the up phase creates the conditions for overconfidence-driven losses in the down phase.

Defenses against illusion of skill

• Use base rates. What fraction of managers beat the market over long periods? It is roughly the expected base rate from chance alone. Do not assume a track record is evidence of skill unless it is exceptionally rare.
• Check for replicability. Can the manager’s strategy be replicated by others? If not, it might be luck or insider information (both unsustainable).
• Prefer passive strategies. An index fund guarantees market returns (minus fees). An active manager might beat the market, but it is unlikely to be skill. The simpler strategy is more robust.
• Track backtests carefully. Many strategies look great on past data but fail on new data. If a strategy fits perfectly to the past, it has likely overfit and will not replicate.
• Remember: past performance does not predict future results. This disclaimer on every prospectus is true. Recent winners are not predictors of future winners.

See also