Luck–Skill Continuum
Classify an activity before judging performance or predicting persistence
- Difficulty
- Moderate
- Time to result
- ~weeks to results
- Steps
- 6
- Confidence
- 99%
Place an activity on a continuum from pure luck to pure skill before interpreting its outcomes. In luck-heavy domains, short-term results carry little information and extreme performance should regress sharply toward the average. In skill-heavy domains, performance is more persistent. One diagnostic is whether a participant can intentionally lose under the same constraints: deliberate control implies some skill, while an inability to underperform on purpose reveals how strongly randomness and competition dominate. Evaluation should then shift from headline results toward sample size, repeatable behavior, decision process, organizational context, and persistence. This prevents hiring stars or selecting funds solely because recent outcomes were exceptional and improves expectations about what will transfer or continue.
Origin
Extracted from Masters in Business and Michael Mauboussin’s book The Success Equation, including a diagnostic borrowed from the poker community.
Core principles
- 01Most outcomes combine skill and luck in different proportions
- 02Short samples reveal less skill in luck-heavy activities
- 03Mean reversion strengthens as luck becomes more influential
- 04Persistent behavior is better evidence than a striking result
- 05The ability to lose deliberately reveals the presence of skill
How to run it
- 1
Classify the activity
Locate the activity qualitatively between pure luck and pure skill by examining how much deliberate action controls outcomes.
Pro tip Use clear anchors such as lotteries, chess, and sprinting.
Watch out Do not assume a difficult activity is necessarily skill-dominant.
- 2
Inspect the sample
Evaluate how many observations exist, how long the period is, and whether outcomes are independent or path-dependent.
Pro tip Demand longer records in luck-heavy fields.
Watch out A long calendar period may still contain few independent decisions.
- 3
Apply the lose-on-purpose test
Ask whether a participant could reliably produce a worse result while obeying the same rules and constraints.
Pro tip Design a symmetric test of intended outperformance and underperformance.
Watch out Artificial sabotage outside normal constraints does not demonstrate relevant skill.
- 4
Estimate mean reversion
Expect extreme outcomes to move toward the population average in proportion to the activity’s luck component.
Pro tip Use the population average rather than the latest exceptional result as the baseline.
Watch out Mean reversion does not guarantee immediate reversal.
- 5
Audit behaviors and context
Study how decisions were made and which systems, colleagues, or resources contributed to success.
Pro tip Ask behavioral interview questions that reach the nuts and bolts of the process.
Watch out Do not attribute an organization’s contribution entirely to one star.
- 6
Make a calibrated judgment
Combine process evidence, sample quality, context, and expected persistence before hiring, allocating, or rewarding.
Pro tip State confidence as a range rather than a binary verdict.
Watch out Exceptional outcomes alone are insufficient proof of transferable skill.
In the wild
Students select twenty-five stocks expected to beat the market and another twenty-five expected to underperform, freezing both lists for a year. If they cannot reliably do worse than the market under the same constraints, their apparent ability to outperform deserves skepticism.
→ The symmetric test exposes how difficult deliberate performance is in a highly competitive market.
A firm considers a celebrated analyst but examines the analyst’s research process, team support, data access, and decision behavior instead of extrapolating recent rankings. It identifies which capabilities are portable and which belonged to the prior organization.
→ The hiring decision reflects transferable skill rather than reputation and favorable context.
Common mistakes
Extrapolating a hot streak
Extreme short-term performance in a luck-heavy field is likely to contain substantial randomness.
Ignoring organizational contribution
Stars may attribute team systems, resources, and colleagues to their own portable ability.
Evaluating only outcomes
Results without process evidence cannot reliably distinguish repeatable skill from favorable luck.
Is it for you?
Best for
Investors, executives, recruiters, and coaches assessing performance in noisy fields.
Not ideal for
Activities whose outcomes and underlying behaviors cannot be observed with enough reliability to make even a qualitative classification.
From the transcript
“The first thing I would start with is sort of the centerpiece of the book. The success equation is what we call the luck skill…”
“when you're trying to determine if something is the results of skill or luck, ask yourself the simple question, can you lose on purpose?”
“The best estimate of the expected value some measure much closer to the average of the population.”
From the episode
Interview With Michael Mauboussin: Masters in Business (Audio)
Michael Mauboussin