HBS Prof Studies Gambler's Fallacy

A new paper by Harvard Business School professor Gregory M. Barron explores the role of experience in promoting the gambler’s fallacy, a notion that chance and randomness correct themselves in the short term.

The study, written by Barron and Harvard doctoral student Stephen Leider, argues that the fallacy has a heightened effect on the decision-making of people who experience a series of events in real time versus people who receive a complete description of the same events at a later point in time.

According to Barron, the fallacy applies in many situations, including simple scenarios such as casino games like roulette or even something as mundane as flipping a coin.

For instance, Barron said that if a fair coin were flipped five times and by chance landed on heads all five times, a person who had witnessed the series of flips would be more inclined to choose tails on the sixth flip.

This person thinks that since the outcome of tails is bound to occur eventually, there is a higher probability of tails occurring and thus bets on this outcome, even though there is no statistical change in tails actually coming up.

“This is a basic human idea that if things have equal probabilities, they have to even out in the short term as well as the long term, and this is the crux of the fallacy,” Barron said.

In the study, the researchers had their subjects predict the color outcomes of a roulette wheel, where one group saw the past outcomes all at once, and a second group saw them revealed in real time.

Barron and Leider found that the players who saw the outcomes in real time were much more likely to experience the fallacy because they overweighted the most recent outcomes in making their betting decision, while the players who were given all of the outcomes as a set of information were much more likely to take all of them into equal consideration when making a prediction.

This conclusion, along with Barron’s previous work, sheds considerable light on understanding the actions of investors in global financial markets, specifically in the realm of stocks.

Barron compared the hypothetical reactions of one investor who had experienced a significant downturn in the distant past with a new investor who hears about the downturn as part of a set of information.

“Both people would obviously have some concern, but this information is more salient for the new investor who receives a description of the event,” Barron said. “For the investor who experienced this in the past, it’s more important that since then, things have been better.”

As for why people place greater weight on more recent outcomes when they experience the events themselves, Barron provided some conjecture on potential evolutionary causes.

“Consider a mouse that finds a candy bar left in a corner,” Barron said. “The mouse would keep coming back to that location in the future, to see if there’s more food there. It’s natural to assume that what worked out well yesterday might work out well today, or that whatever I did that led to a bad outcome yesterday should not be repeated today.”

—Staff Writer Prateek Kumar can be reached at kumar@fas.harvard.edu.