The "ten best days" argument is a good argument against short-term timing (you could, at random, miss just those 10 days and nothing else if you're trying to guess good and bad days.) But it's not a very good argument against the particular type of long-term timing discussed in this sub-thread. The 10 best days tend to be clustered somewhat, but often interspersed with several bad days; if you're simply looking for a good price and then buying, you're not going to miss the 10 best days without also missing a large number of bad days. Missing both the best and worst 10 days of the decade gives you almost exactly average returns. Furthermore, the 10 best days tend to occur in the "short sprints" that follow a down market; the strategy described above says to buy in to a down market, which means you'd hit most of those 10 best days -- and, quite likely, miss at least a few of the 10 worst days, generating net above-average returns. (The parent post's graphs showed this exact result: buying into the market after a largeish downturn gets you great returns, long term.)

Why is it that no one ever does the same analysis for the 10 worst days?

Here you go: http://www.freemoneyfinance.com/2010/11/the-truth-about-mark...

  Invest in S&P 500 ETF (SPY) starting at inception
  Growth of $100,000 from 1/29/1993-8/30/2010
  Buy and Hold: $324,330.15
  10 Best Days Removed: $156,354.12
  10 Worst Days Removed: $692,693.90
  10 Best and 10 Worst Days Removed: not listed, but very close to Buy and Hold (the green line in the chart)

Rob Bennett's followup comments are quite good, as well.