Can you predict the market?

The professional forecasters don't do a very good job, but you may be able to benefit from them anyway. Writing in The Wall Street Journal, Jason Zweig tells how.

Don't bet a year ahead.

Forecasts are more accurate in the short-term than in the long-term. The Federal Reserve Bank of Philadelphia keeps a database of decades' worth of median and individual forecasts from dozens of experts on a wide range of economic and financial variables. When it comes to predicting such variables as inflation, unemployment and economic growth, the survey is considerably more accurate than just extrapolating the latest trend—but only for one or two calendar quarters. Over the following few months, the accuracy of most forecasts falls very quickly.

Embrace error.

"When confronted with forecasts, people without Ph.D.s in economics tend to view them as very tight, as if you could take them to the bank," Mr. Stark says. "People don't understand the huge amount of uncertainty that's embedded in these forecasts." By seeing how wide the potential range of error is, you can adjust your own expectations to make sure you don't make an overconfident decision. Investors who like to hedge their bets, meanwhile, can get a better sense of how much protection they might want. Before relying on a forecast of any variable, check the "forecast error statistics" page on the Philadelphia Fed website to see how accurate the past predictions have been.

Listen to the markets.

"If there is a market with many traders, you shouldn't try to outguess the market," says J. Scott Armstrong, a professor at the Wharton School and author of "Principles of Forecasting." The S&P 500 tends to forecast recessions and recoveries fairly well in the short term, as it did when it started falling in late 2007 and again with its sharp rise in 2009. The bond market also can be a good indicator, relative to individual forecasts. Still, this kind of momentum surfing is best left to traders. If you are a long-term investor, you should accept that short-term market moves aren't predictable, and there is little point in adjusting your portfolio in response. In most cases, long-term averages of past returns—while far from perfect predictors—are probably a better guide than long-term forecasts of future returns.

Average many different forecasts.

Perhaps the most powerful tool for improving the quality of predictions is simply to combine several forecasts from a variety of independent sources. Prof. Armstrong has found that this technique reduces forecasting errors by up to 58%—a massive improvement over individual forecasts.