The editor in chief of the "Stock Trader's Almanac" is forecasting the Dow Jones Industrial Average at 38,820 by 2025. He's claiming that there will be an eight year boom starting in 2017.
Given that the Dow is at 10,800 today, this implies an annual return of (38820/10800)^(1/15) -1 = 8.9% annually. The Dow doesn't include dividends so adding in a 2% dividend yield would bring us to 10% per year. So here we have a news story about a guy who is predicting that markets will perform at their long term average. Genius.
Of course really what he is doing is trying to sell copies of the almanac.
In the story linked above there is a mention of Glassman and Hassett who wrote a book predicting that the Dow would hit 36,000 in 2005. The book called Dow 36,000 was a best seller, but is was based on completely dopey analysis. They argued that because stocks seemed to beat bonds in the long run, the risk premium on stocks should be very low. Therefore, assuming a low risk premium, you could basically get a very high valuation. It turns out stocks don't always beat bonds in the long run (consider the past 10 years or so). Furthermore, stocks are more risky than bonds - just look at the standard deviations. Finally stocks have the more insidious (and much more subtle) problem that we have uncertainty as to what their expected return really is.