Statistical Modeling Reveals the Keys to Success | By Sydney Rudko

CC photograph courtesy of Angus on Flickr (photograph taken on August 30 2009)

IN AUGUST OF 1891 MARK TWAIN PUBLISHED the short story “Luck” in Harpers Magazine. The story describes the rise of a successful young soldier in the Crimean War, and attributes his achievements purely to luck. The story is told from the perspective of a military instructor who happens upon a hopeless student and decides to tutor him. The teacher instructs the student on the history of Caesar, but teaches the student of nothing else. When the narrator recounts exam day the piece takes a comical turn as he describes, “By some strangely lucky accident–an accident not likely to happen twice in a century–he was asked no question outside of the narrow limits of his drill.” As the story continues the boy’s luck continues persists as he ascends the military ranks. In the story’s final and whimsical line, Twain asserts that luck is often more important than wits:

“Look at his breast; why, he is just clothed in domestic and foreign decorations.   Well, sir, every one of them is a record of some shouting stupidity or other; and, taken together, they are proof that the very best thing in all this world that can befall a man is to be born lucky.”

The piece was very poorly received at the time of publication despite it being a pleasant read. In fact, upon first reading it myself I didn’t like it, because it heavily correlated success with pure luck. I found this unrealistic, and yet I was reminded of this piece upon reading a recent paper published in Proclamations of the National Academy of Science, for in this research the authors provide some evidence for luck being more important than skill in achieving success under certain circumstances.


Success is the driving force in evolution. We all have the desire to compete and to win. We idolize those we perceive as successful, and aspire to be as skilled as the actors, authors, directors, musicians and businesspeople we hear about. The paper begins by demonstrating that people correlate great success with great skill. This is highly intuitive, and not a particularly remarkable discovery until you’re watching a Channing Tatum movie, and are asking yourself, “Why is this guy famous and successful? He’s a terrible actor!” Given this example, there must be some exception to that rule. The authors of the paper “Top performers are not the most impressive when extreme performance indicates unreliability” assert that under certain circumstances, outstanding performance is more strongly correlated with an average skill level and sheer luck. The researchers present two models to support their assertions. In both models they test their hypotheses using a game as a model and then apply various statistical methods to analyze the results.

In their first model, the researchers assert that ‘Matthew effects’, or more simply put, ‘the rich getting richer’ can have substantial effects on one’s performance. They found that in a scenario in which wins are dependent, meaning that the outcome of the first event will affect subsequent events, chance matters more than skill, because initial outcomes heavily influence subsequent ones. In other words, a lucky break early on in the game could carry you through to many successive wins. In fact, the researchers found that the players with fewer wins were on average more skillful than those with the highest number of wins. This means they found that the most skilled players were coming in second place instead of first, while an opponent with less skill was taking first place. For example the child of a rock star may have an easier time getting a record deal compared to a more talented youth without a parent in the music business. Furthermore, in their second model, they argue that extreme performance may be the result of extreme luck. To put this simply: no risk, no reward. The authors found that those players who demonstrated more risky behavior often were rewarded with more success.

Despite the clear circumstantial limitations to the paper, including quantifying success in real life, dependency in real-life scenarios, and the levels of skill required to actually attain success in certain fields (for instance while Channing Tatum may have the skill level to be a successful actor, would a comparable skill set in nanophysics yield him the same results?), the authors outline many practical implications of this observation. They focus mostly on implications in economics or business, but I think the most important implication is how this study relates to personal goal-setting. As humans we learn primarily through copying the behaviors and actions of those around us. If you require a case study for this, see my bookshelf. It is cluttered with the biographies of people who inspire me to be successful, but if those I perceive to be the most successful are not the most skillful, am I setting myself up for failure?


The models don’t make sense to me from an evolutionary standpoint. If our brains mistake success for skillfulness – despite the absence of the latter – and we learn through emulating the skills of those we perceive as the most successful, then in theory we would become increasingly less skillful and would ultimately become extinct. This may seem extreme, but consider the example of a female rhinoceros selecting a mate. She selects a mate based on those individuals who compete best in battle against other males. This is because those who are strongest are likely to be the best hunters, to defend their territory, and provide for themselves. She selects this mate because her offspring will potentially inherit these traits, and they will go on to survive, breed, and continue the species. This is the basis of evolution; it is survival of the fittest, or the most successful.

Assume now that this rhinoceros falls to the same pitfall the paper claims humans do, and she perceives the rhinoceros’s win to be a symbol of fitness, when in reality he landed an accidental blow with his horn while he was being distracted by a particularly attractive cloud in the sky. By reproducing with this male her offspring would not inherit the best traits, they might be absent minded, and would be less likely to survive and reproduce. This is not survival of the fittest. While this rhinoceros scenario may seem extreme given the context of our modern day society where success or failure might be defined as trivially as making the honor roll, it certainly demonstrates that the logic of these observations may be flawed. It is ultimately the ability to recognize success, emulate, and ultimately select a mate based on success that has allowed our species to evolve and prosper. It is for these reasons that I feel their model does not fully explore the parameters of what defines success.


The researchers suggest that “imitating exceptional performance could be detrimental. As their models show, the highest performers may both be less skilled and use methods with higher levels of risk. When exceptional performance is due to self-reinforcing processes and initial success, the exceptional performers may continue to perform well but imitators will likely be disappointed, because they can at best only replicate the practices, and thus the skill levels, of the high performers, but not their initial good fortune.” This is a very pessimistic view, and neglects the fact that the ability to distinguish success is important for the overall success of a species.

I believe it stems from a narrow definition of luck, or as they refer to it in their paper, ‘noise’. The researchers assume that random chance is in fact random, but in reality things such as meeting people, attending events, or even through the use of social media could be mistaken for luck. These decisions can be conscious or unconscious, but shouldn’t be chalked up to random occurrence. Consider again the example of a rhinoceros: even a weak rhinoceros that is regularly challenged by the strongest males to fight might ultimately become a better fighter, and may even win because over time he has learned how to fight. Do we define experience as luck, or is this an example of experience igniting with opportunity to create success?

This brings me to question how we define skill. For Channing Tatum, his skill and success as an actor doesn’t depend merely on his acting abilities. His personality, star appeal, connections, and, of course, those rock hard abs, have enabled him to achieve success. Do we define these characteristics as luck, or are they in fact skills? I don’t know about you, but I don’t think abs like that just happen by chance.

Even if we accept that luck and chance have something to do with success, obtaining a lucky break means nothing without having the courage to act. The data presented in this paper inspires me to take risks, as opposed to aiming for second best as the authors suggest. The paper, though the authors may not interpret their findings this way, also shows the importance of taking a chance and creating your own luck.

The next time I’m reading about someone who inspires me, I will stop and think not just upon the skill that got them there, but what they did with the opportunities they were given, and more importantly, how I will act upon those given to me.

Sydney Rudko is a fourth year Infection and Immunity student interested in sharing all things science with interested people!


Related posts: