World Series strikes out as a reliable predictor of elections
A Chicago Cubs win in this year’s World Series would do more than break the Curse of the Billy Goat. It would also help assure that Hillary Clinton becomes the next president of the United States.
You might wonder what the World Series has to do with politics. That must mean you haven’t heard of the so-called World Series Presidential Predictor, according to which a World Series win by the National League team presages a win by the Democratic candidate for president. An American League victory, in contrast, leads to a Republican electoral win.
This Predictor dates at least to the early 1970s. That’s when the Society for American Baseball Research (SABR) published an article reporting that the Predictor had a perfect record over the five presidential elections from 1952 through 1968.
As is so often the case, however, a funny thing happened on the way to the bank after this pattern’s existence was reported: Since then, the Predictor has failed on a number of occasions. Bill Staples Jr., a SABR member, recently reported that, based on all World Series since they were first held over a century ago, the Predictor’s track record is statistically no better than a coin flip.
This is a valuable lesson for investors since more alleged patterns have been “discovered” in the stock market than even in baseball.
In almost all cases, those stock market patterns turn out to be worthless.
Psychologists tell us that this in large part is because our brains are hardwired to “detect” patterns even when none exists. And with thousands of us mining the same historical data, it’s hardly surprising that a few patterns will emerge that — like the World Series Predictor — look uncannily accurate but in fact are bogus.
There are two steps investors should take to protect themselves from following such patterns. The first is to test their accuracy over different time periods than those used to initially “discover” their alleged existence. Upon doing that, most patterns disappear.
Had this step been followed in the early 1970s, the World Series Predictor would never have been proposed, since its “discovery” was based on the outcome of just the five previous presidential elections. If its author had considered all other elections as well, the pattern would never have emerged.
Most of the patterns on which Wall Street focuses fail this simple test from Statistics 101.
The second step investors should take to protect themselves from a bogus pattern is to insist that there be a plausible explanation for why it should exist in the first place. If none can be found, odds increase that the pattern is fictional — even if it is able to survive the first step. Needless to say, the World Series Predictor is unable to satisfy this second step, since there is no credible theory for why the outcome of a baseball playoff has anything to do with who wins a presidential election.
Perhaps the best illustration of this in the investment arena comes from David Leinweber of the Lawrence Berkeley National Laboratory.
Several years ago he searched through all the data on a United Nations CD data disk to find the indicator with the most statistically significant correlation with the S&P 500 index. His discovery: butter production in Bangladesh.
It’s easy for us to see that Bangladeshi butter production levels are of no real help to our U.S. stock market investments. But the same goes for most other alleged patterns on which we otherwise focus. The bottom line? There are no magic keys that can unlock the stock market’s secrets. The only people who reliably make money when we trade on bogus patterns are the brokers who earn a commission each time we buy or sell.