Is It Luck or Skill, or Is That the Wrong Question?
From the Understanding Luck: Randomness, Risk, and the Real Sources of Success collection
Two stories get told about every outcome that matters. In one, the result was earned: the founder outworked rivals, the investor saw what others missed, the manager made the right call. In the other, the result was a gift of timing, circumstance, or a single conversation that happened to go well. Each story is partly true, and each becomes a distortion when it is told alone.
Writing luck out of the story produces overconfidence. People who credit every win to effort take bigger risks than their record justifies, and they judge those who failed as if the failure proved a lack of effort. Handing everything to luck produces the opposite error, a quiet resignation in which planning feels pointless.
Both errors share a hidden assumption: that the useful question is how much of an outcome was luck. That question matters for judging the past, but it gives almost no guidance for acting in the future. A more productive frame treats every outcome as the joint product of three things, skill, chosen risk, and randomness, and asks which of them a person can actually shape.
The answer is sharper than it first appears. Luck cannot be controlled, but exposure to it can be designed. Where any single draw lands is beyond reach; how many draws a person takes, how large each one is, how long they can afford to keep playing, and how ready they are when a good draw arrives are all choices. Effort does not shrink in this picture; it moves.
Luck lives in the gap between decision and result
A working definition helps. An outcome counts as luck when it matters to the person, lies outside their control, and could easily have gone otherwise. By that test, luck is not a mysterious force but a gap: the distance between the quality of a decision and the quality of its result.
A sound decision can lose, and a reckless one can win. A driver who runs a red light and arrives safely made a bad decision with a good outcome; a surgeon who follows every protocol and loses a patient to a rare complication made a good decision with a bad one. Judging decisions by their outcomes, a habit the decision strategist Annie Duke calls resulting, rewards the first and punishes the second.
The fix is procedural. Grade the decision on the information available when it was made, then grade the outcome separately. Only across many repetitions do the two grades converge, which is why a single result is weak evidence of anything.
Chance follows rules that intuition misreads
If a single outcome says little, the obvious move is to look at records. But records carry their own traps, because chance behaves in ways the mind consistently misjudges.
Small samples produce extremes in both directions. The statistician Howard Wainer showed that the US counties with the highest kidney-cancer rates were mostly small and rural, and so were the counties with the lowest rates. Nothing about rural life caused both; small populations simply swing further from the average. The same artifact crowds the top and bottom of school rankings, early sales figures, and A/B tests run on too little traffic.
Extremes also drift back. Any unusually good or bad result contains some luck, and luck does not repeat on command, so the next result tends to sit closer to average. Daniel Kahneman described Israeli Air Force flight instructors convinced that praise made cadets worse and criticism made them better. They were watching regression to the mean: praised cadets had just flown unusually well, criticized ones unusually badly, and both were drifting back toward their normal level.
These rules set a demanding standard for reading any track record. A record is evidence of skill only after a luck check: how many independent decisions it rests on, whether the performance persisted, and how many others tried the same thing and failed. Studying winners alone skips that check. Of the eleven companies profiled in Jim Collins' Good to Great in 2001, Circuit City had filed for bankruptcy and Fannie Mae had been placed under government conservatorship by the end of 2008.
The standard tightens as competition improves. Michael Mauboussin calls this the paradox of skill: when competitors get better and more alike, the spread of skill narrows, and random variation decides more of the ranking. Ted Williams hit .406 in 1941, and no major-league hitter has finished a full season at .400 since. Stephen Jay Gould argued that hitters had not gotten worse; the whole league had gotten better, leaving less room for anyone to stand far above it.
The draw and the ticket
If luck decides so much, the natural next step is to try harder to predict it. That is usually the wrong place to spend effort. The more useful move is to separate two things everyday language treats as synonyms: risk and luck.
They sit at different points in time. Risk is the ticket chosen before the draw: which bet, what size, what downside, how reversible. Luck is how that exposure resolves. Nobody can manage the draw, but anyone can manage the ticket, and the ticket decides how much any draw can help or hurt.
The arithmetic is unforgiving on the downside, since a 50 percent loss needs a 100 percent gain to recover. The physicist Ole Peters has made a deeper version of the point with a simple coin game in which each flip multiplies wealth by 1.5 on heads or by 0.6 on tails. The average payoff per flip is a 5 percent gain. Yet a single player who keeps flipping sees wealth shrink over time, because one head and one tail together leave 90 percent of the starting stake.
The lesson is that each person lives one path, not the average. A bet that looks attractive across a crowd can still ruin most of the individuals who keep taking it. Survival comes first because luck averages out only for those still playing, and sizing follows from the same logic: a commitment should reflect not only confidence in an edge but the likely error in estimating it.
A second distinction sharpens the first. Frank Knight drew it in 1921: risk has known odds that can be priced and insured, while uncertainty has odds nobody knows. Insurance answers the first. Flexibility, meaning reversible commitments and room to change direction, answers the second.
Tilting the odds
Once the ticket is in view, luck becomes partly a design problem. Good draws can be made more frequent, bad draws less damaging, and the decision to keep going less hostage to feeling.
More good draws come from more exposure. The more a person does, and the more people who hear about it, the more chances an unexpected opportunity has to find them; the programmer Jason Roberts called this luck surface area. Exposure pays best as many cheap, reversible experiments rather than a few expensive, irreversible ones, because small bets cap the downside while leaving the upside open.
Fewer ruinous draws come from safeguards, because much of what gets called bad luck is predictable failure without a buffer. When hospitals in eight cities adopted a simple surgical safety checklist, a 2009 study in the New England Journal of Medicine found that the death rate after surgery fell from 1.5 percent to 0.8 percent. Nothing about the operations became less random. The known failure points simply got a check.
Persistence needs the same design. Continuing is valuable only on a promising path, and sunk costs make that judgment hard in the moment, so setting stop conditions before starting moves the decision to a calmer time. Stewart Butterfield shut down the online game Glitch in 2012 and turned its internal chat tool into Slack, a stop that looks like luck from the outside and like discipline from the inside.
Earned and lucky at once
A full account of luck ends with a question the mathematics cannot answer: how to hold an outcome once it arrives. The accurate answer is that merit and fortune are partners. Most success is earned and lucky at the same time, and admitting the second does not diminish the first.
The accounting matters most when it is applied to other people. People tend to explain their own failures by circumstance and other people's failures by character. Asking what someone controlled and what they were dealt yields a fairer judgment, without excusing anyone from responsibility for their choices.
The practical stance fits in three short sentences. Respect the draw, because no amount of skill commands it. Choose the game, because the distribution matters more than any single outcome. Stay in it, because good luck can only arrive for those still standing when it comes.