As a result, investors typically overpay for the most popular stocks, thinking that rapid earnings growth of 30 percent a year or more will always bail them out. Sometimes it does, but high-fliers with P/Es above 50 rarely keep outperforming the market for long. Even the best glamour stocks crash hard. In fact, the highest-priced growth stocks are often outpaced over a decade or longer by moderate-growth stocks with 12 percent to 18 percent annual earnings increases and P/Es below 25.

Investors make the same kind of mistake with value stocks, seeking out deep-discount situations in hopes of a big score. It's a risky strategy because a lot of cheap stocks are cheap for good reason -- the companies may be in big trouble. Instead of disasters, value investors should look for moderately depressed stocks, with below-average P/Es -- maybe 10 to 16 nowadays -- and 8 to 10 percent annual earnings growth. Even modestly better results will help those stocks bounce back.

Mistakes in valuation also occur on an industry-by-industry basis. In the early 1990s, for instance, the Clinton healthcare initiatives appeared likely to undermine the pharmaceutical industry. As a result, top drug stocks such as Merck were selling at P/Es as low as 11 or 12. But in the past few years, the drug ndustry has been one of the top-performing sectors.

To profit from major shifts in sentiment, you have to be mentally prepared to resist the temptation to join the herd. The simplest and safest strategy is to buy quality growth stocks when they are temporarily depressed. Develop a list of companies you like in a variety of sectors and track them. At any particular time, one of two of the stocks on your list will probably be underpriced. The key is to be contrarian about when you buy -- not what you buy. And you don't have to catch the exact bottom to get a terrific bargain.