Why gamblers and investors make mistakes

How many times have you assumed that what’s happened in the past will continue in the future? For instance, if your trash collector has been showing up early every morning, it seems a pretty good bet that you better get the garbage out the night before, rather than waiting until after breakfast on pickup day. He won’t necessarily show up again at 6 a.m. at the next scheduled pickup—a change in routes or a traffic jam might delay the pickup until late morning. There will be no harm done if you got the garbage out early and the trend didn’t repeat.

But this common cognitive error—known as “representativeness”—can have devastating effects on an investor. It means that we make a judgment based on a stereotype or strong image held in our minds. The recent past often is the strongest image, so we tend to extrapolate the future from the past.

For an investor this is a guarantee of infinite future surprise. Markets and investments change directions all the time and no trend lasts forever. Representativeness makes us more optimistic about recent winners than losers, even though the markets are ruled by “regression to the mean,” which says all investment returns will eventually be pulled down—or up—to their long-term averages. Tech stocks may have offered triple digit returns in 1999, but this wasn’t normal, and it took years of horrific declines after 1999 to push returns back to their historic averages.

Unfortunately, those who are not as prone to representativeness often fall victim to its opposite cognitive error, the “gambler’s fallacy.” This fallacy correctly predicts that recent trends will someday reverse, but it expects the reversal to happen almost immediately. Gambler’s fallacy is a misunderstanding of the law of large numbers, which says it can take huge amounts of data to move toward a long-term average. For instance, we all know probability says that fair coin flips should average equally among heads and tails. This makes us surprised when someone flips heads 10 times in a row. However, the law of large numbers states that you may need thousands or tens of thousands of coin flips to reach the 50-50 average. This fallacy misunderstands that regression to the mean does not necessarily mean a sharp reversal: it may mean a long period of returns that are closer to the mean instead. Gambler’s fallacy leaves investors surprised by long-term investment trends, such as the dominance of growth and tech stocks throughout the second half of the 1990s.