As a diversified market index, the S&P 500 is less risky than most individual stocks, and, therefore, probably carries better "odds" of going up. The odds for any individual stock vary significantly and depend on the underlying company's maturity, business diversification, and industry, as well as general stock market trends (in a bull market, most stocks go up; in a bear market, most go down). The odds that the average profitless Internet or biotech stock will rise, in other words, are probably low, while the odds that the stock of the average mature, diversified company will rise are probably better than 50-50. The reason some Internet and biotech stocks are still worth buying—in well-diversified portfolios—is that the potential payoff can be much higher.

To test the assertion that the odds that an average stock will rise in any given year are close to 50-50, I looked at the daily advance/decline ratio on the Nasdaq for each year from 1999-2003 and then averaged the daily ratios to come up with a back-of-the-envelope average for the year. In each case, it was close to 50-50. Although this methodology obviously doesn't prove the theory (far from it), I think it provides a reasonable sanity check. To prove the theory, you would presumably have to look at the performance of every stock in the last, say, 30 years, including those that have subsequently been acquired, gone bankrupt, or disappeared. If such an analysis has been performed, I would be grateful if someone would refer it to me.