No one possesses a crystal ball, alas, not even investment analysts. But given the wealth of readily available market data, it’s tempting to look at past trends, whether in individual stocks or the market overall, and use them to predict the future.

In the 1980s, two quantitative analysts famously satirized such attempts when they ran regression analyses on various types of nonfinancial data and got this surprising result: Butter production in Bangladesh had the highest correlation of any data series to the S&P 500’s performance. The vagaries of demand for Bangladeshi butter could supposedly predict the market’s moves 75 percent of the time! While this highlights the pitfalls of data mining, it doesn’t mean there aren’t market indicators with some utility.

Seasonality, for example, has been an extraordinarily successful predictor of market performance. The best-known adage of this approach is “Sell in May and go away.” In 2012, investors who did just that avoided the market’s sharp selloff in May, but they also sat out the mid-summer rally that pushed the market to a four-year high.

While it’s hard to believe that there are regularly recurring patterns of bullishness in the market, numerous studies have shown that such trends do indeed exist. In fact, we’re on the cusp of a seasonally favorable period now, according to the aptly named “Halloween Indicator,” which signals strong stock market performance from November through April.

Ordinarily, investors exploit such widely observed phenomena until the effect dissipates and eventually no longer works. But the market’s seasonal strength during the winter months may actually be increasing.

The latest study on this issue is “The Halloween Effect: Everywhere and All the Time,” conducted by two academics in New Zealand, and certainly the most exhaustive such study to date. They surveyed all 108 global stock markets and all available historical data. Overall, their study includes more than 55,000 monthly observations, with historical data for some markets going back to 1693.

Their findings? The period from November through April produced returns that were 4.5 percentage points higher, on average, than the summer months. And over the past 50 years, the gulf in performance between these two periods has actually widened to 6.3 percentage points.

These full results, however, include less-developed markets where investors may not be taking advantage of this trend and, therefore, dampening its effect. When the data are narrowed to just the US market, for example, the discrepancy between the summer months and the winter months shrinks to 1.7 percentage points annually for the period 1791 through mid-2011. But in recent decades, the difference has been substantially greater than it was over the full period. Over the 10-year period ending mid-2011, for example, the winter months outperformed the summer months by an average of 5.7 percentage points annualized.

Although the Halloween Indicator works over the long term, there were exceptional periods when the winter months lagged the summer months. In other words, if you hope to use positive seasonality to boost your returns, you should take a long-term perspective. Consider this: In the UK market, over a 10 year period, there was a 92 percent chance that cashing out for the summer and going long in the winter would beat a buy-and-hold strategy. But over a one-year period, the odds dropped to 63 percent. Also, it should be noted that these results do not include the effect of trading costs and taxes.

The US market, however, had a robust rally this summer. What are the odds the next six months will continue to be bullish? Mark Hulbert, editor of The Hulbert Financial Digest, reviewed the November-to- April returns for the Dow Jones Industrial Average going back to 1896. He found that the US market tends to sustain its momentum following a strong summer, with returns averaging 1.7 percentage points higher per year than the average winter.

The bottom line: Seasonality indicators point to the next six months being bullish for stocks. But that doesn’t mean you should bet your portfolio on it.