Yesterday, on CNBC, Sam Stovall, chief investment strategist at CFRA Research, noted that the S&P 500 has advanced in both January and February 27 times since 1945. All 27 years ended with positive returns when including dividends, the data show. So, all investors need to do is to wait until February ends, and if the S&P 500 had a positive month, pile money into the stock market. It is a sure thing. It has never failed.
Not so fast. In another arena we just witnessed an unprecedented event. Prior to ten days ago, no team had ever come from more than ten points down to win a Super Bowl. FiveThirtyEight's Nate Silver points out that of all 364 games where a team was down between 24 and 27 points in the third quarter, the losing team only came back 0.8 percent of the time. Heck, President Donald Trump, a Patriots' fan, even left his own Super Bowl watch party when the score was 28-3.
Academicians pejoratively describe data mining as the process of digging through numbers to discover hidden connections and predict future trends. In the era of big data, it is easy to find two sets of data that are highly correlated. The problem is that many of the relationships are spurious -- that is, have no cause and effect relationship. One of my favorites is the fact that the number of people who drowned falling into a pool in any given year during the prior decade highly correlates with the number of films Nicolas Cage appeared in that year.
Trying to draw inferences from past stock market experiences is a waste of time and can be hazardous to one's wealth. What patterns have existed in the past in the stock market aren't a guarantee of future performance. Statistical oddities are fun. Just don't bet your wealth on them.