What the data actually shows
Economist Robert Frank, in 'Success and Luck' (2016), argues that in winner-take-all markets — where small advantages decide who captures an outsized share of the rewards — tiny early strokes of luck compound into large differences in final outcomes. When many highly capable people compete for a few top spots, skill becomes a near-prerequisite that most contenders share, and chance increasingly determines who actually wins. The more competitive and scalable the field, the larger luck's footprint on who ends up at the very top.
A widely discussed agent-based simulation by Pluchino, Biondo and Rapisarda (2018), often called the 'Talent versus Luck' model, illustrates the mechanism. The researchers modelled a population with talent distributed normally and then exposed agents to random lucky and unlucky events over a working life. The most successful individuals in their simulations were rarely the most talented; they were moderately talented people who happened to encounter more lucky events. Because talent is roughly bell-shaped while wealth and success are extremely skewed, the model suggests luck is doing much of the work that we tend to attribute to exceptional ability. It is a simulation, not a measurement of the real economy, so treat it as an illustration of a plausible mechanism rather than a precise estimate.
We also misjudge the mix in a predictable direction. The self-serving attribution bias, documented across decades of social psychology, leads people to credit their own skill and effort for successes while attributing failures to bad luck or outside circumstances. The same asymmetry quietly distorts the public stories we tell about success: winners narrate their path as earned and skilful, and the luck that helped them — and the equally talented people whom luck passed over — largely drops out of the telling.
Why this feels different from how it actually is
Success feels like it must be mostly skill because we experience our own effort directly and chance only faintly. You remember the late nights, the hard decisions, the things you got right; you do not feel the thousand small contingencies that had to break your way, because most of them are invisible and many never registered as decisions at all. From the inside, a fortunate path looks like a chosen one.
Survivorship bias compounds this. The success stories that reach us are, by definition, the ones that worked — we hear from the founder who made it, not the equally capable founders who hit a bad market and disappeared. Because the failures are filtered out, the surviving examples look like proof that skill reliably produces success, when really they are the lucky tail of a much wider distribution of similar effort.
Finally, the meritocratic story is emotionally appealing in both directions. It flatters the successful by framing their position as deserved, and it offers everyone else the promise that the right effort guarantees the right outcome. Acknowledging luck threatens both comforts, so the role of chance gets quietly edited out of the narrative even when people half-know it was there.
What the research says to do about it
Frank's own practical takeaway is gratitude rather than fatalism. Recognising the role of luck does not mean effort is pointless — effort is one of the few inputs you control, and it remains necessary. It means holding success and failure a little more loosely: being less self-congratulatory about wins, less self-flagellating about losses, and more generous toward people whose luck ran differently. There is some evidence that consciously acknowledging good fortune increases generosity and wellbeing.
Because chance plays a real role, the research-aligned strategy is to take more shots and stay in situations where luck can find you. If outcomes are partly random, increasing the number of opportunities — applications, introductions, projects, attempts — raises the odds that some of them break your way, without requiring you to predict which. Maintaining a wide and varied network matters here too, since many lucky breaks arrive through weak ties and chance encounters you could not have planned.
It also helps to judge yourself and others by process rather than purely by outcome. Since two people running the same sensible process can land in very different places for reasons outside their control, evaluating the quality of decisions — given what was known at the time — is fairer and more accurate than reading results backward as a verdict on ability.
What the research says does not help
Reading successful people's outcomes as a precise measure of their merit does not help, and it is one of the most common errors. Because survivorship bias hides the equally skilled people whom luck failed, copying the visible winners' specific choices is unreliable — you are studying the lottery winners' ticket numbers. The skill was usually real, but it was not the whole story, and the missing part is rarely reproducible.
Swinging to the cynical opposite — concluding that since it's all luck, skill and effort don't matter — is equally unsupported. The evidence is that talent and effort are necessary but not sufficient: they are the price of entry to a game that chance then helps decide. Treating effort as futile removes the one major input you actually control and lowers your number of shots, which makes good luck less likely to find you.
Harsh self-blame after failure, or unearned superiority after success, both rest on the same mistake: reading a single outcome as a clean signal of ability. The self-serving bias makes this feel natural, but it distorts the picture in both directions and tends to make people either complacent or unduly ashamed rather than clear-eyed.
Real numbers in context
It is hard to put a single clean percentage on luck versus skill, and anyone who offers a precise figure is overstating the certainty of the evidence. What the research supports is a direction, not a number: in competitive, winner-take-all fields, luck's share of who ends up at the very top is large and routinely underestimated. The Pluchino, Biondo and Rapisarda (2018) simulation is best read as an illustration of how moderately talented but lucky individuals can outpace the most talented — not as a measured statistic about the real world.
One useful anchor is the shape of the distributions. Talent and effort, when measured, tend to be roughly bell-shaped, clustered around a middle with no extreme outliers. Wealth and success are extremely skewed, with a tiny fraction capturing an outsized share. That mismatch is itself evidence: a roughly symmetric input cannot, on its own, produce a wildly skewed output, which is part of why researchers conclude that compounding luck and circumstance are filling the gap.