Lucky Break? We Just Don’t Believe It

Imagine your company has a strong year and your CEO gets a significant bonus. The profits may have more to do with the weather or unexpected publicity than the CEO’s performance, so is the raise justified? Maybe not, but our brains may not be able to recognize what was due to luck as opposed to skill.

People are so predisposed to detect connections that sometimes we find patterns where they don’t exist. And research by Chicago Booth PhD student Russell Roberts, with Booth’s Reid Hastie and Alexander Todorov, finds that even when we know a given outcome is based purely on luck, we may still behave as if we have some influence over it.

The researchers set out to determine whether people would use something undeniably random—coin tosses—to make future predictions, conducting a series of experiments involving more than 12,000 participants who either flipped real coins in a lab or virtual coins online.

In one experiment, the researchers asked 1,000 participants to flip a coin five times and predict, before each flip, how it would land. After seeing how accurate their guesses had been, participants then had to state how many of the next 20 tosses they expected to correctly predict.

“The more (lucky) successes participants experienced, the more optimistic they were about future performance,” the researchers write, “whereas the more failures they experienced, the more pessimistic they became.” Even when people knew that their performance was purely due to chance, on some level they still trusted it to predict the future.

What’s more, the pessimism brought on by failure was more pronounced than the overconfidence resulting from success. “This suggests that bad luck may be even more potent than good luck when it comes to updating our beliefs about future performance,” says Roberts.

Another experiment demonstrated how these effects can actually change people’s behavior. Participants carried out similar tasks but with an added element: After an initial five flips of the coin, they had to bet on how many out of five future tosses they would correctly predict. If they met or exceeded their forecast, they could win increasing amounts of money, but if they failed to hit their forecast, they won nothing.

The more success they had in the first round, the riskier their bets, the researchers observed, while failure led to more caution.

What might explain this? When participants were asked whether their performance was due to luck or skill, those who had been luckier tended to say skill. And the more they attributed accurate predictions to skill, the better they thought their future performance would be. Moreover, “people who predicted random events accurately believed that they could convey information that would help others replicate their success.”

These findings suggest there may be some magical thinking at play, even among people who should know better, namely those with greater prior understanding of probability. People who knew more about odds were more pessimistic overall—yet were just as influenced by their performance as people who knew less.

We are wired to find connections, even where none exist, conclude Roberts, Hastie, and Todorov. This may also explain why we often overanalyze our own past behaviors to make sense of an outcome that was clearly due to chance. “People simply cannot help but use the past, good or bad, as a guide for the future,” the researchers write.