One book has brought more reader questions than any other this year.

It's "The Little Book that Beats the Market" (Wiley, 2006, $19.95). In the book, money manager Joel Greenblatt provides a "magic formula" to beat the market. Better still, while other tomes clock in at 400 and 500 pages, the little book is just that -- 155 small pages. And it's an easy read.

Which leaves us with a single question: Does the magic formula work? Is it possible to use his formula (or the website that uses the formula for you, to create high performance stock portfolios?

The short answer is yes, with an important caveat.

Let's start with the yes part.

Mr. Greenblatt's magic formula is simple but powerful. First, he rank orders 3,500 publicly held companies by their return on capital. Not return on shareholder equity, but the return on the total capital employed by the company. This list tells us which companies have the greatest earning power. The greater the earning power, the more investors are willing to pay for the stock.

Why are investors willing to pay more?

Simple. If your $1,000 of capital earns 25 percent a year, it is earning $250. It will be worth a lot more than someone else's $1,000 of capital that earns only 10 percent a year or $100.

His second step is to rank order stocks by their earnings yield. This is the reverse of price to earnings ratio. If a stock is selling at a P/E of 10 it has an earnings yield of 10 percent. The higher the earnings yield, the less expensive the stock.

The third step is to add the two rank order numbers to get a rank ordered list of stocks with high earning power selling at a low price. Had you exercised that simple formula each year from 1988 through 2004, he says, your portfolio would have returned a compound annualized rate of 30.8 percent. That's well over the S&P 500 annualized return of 12.4 percent over the same period. You would have lost money in only one year -- instead of 3 in the S&P 500 -- and a $10,000 investment would have grown to a mind boggling $960,000 rather than the pedestrian $73,000 of the S&P 500.

In theory, all we have to do to enjoy a wildly superior return is follow the formula and buy a portfolio of 30 magic formula stocks each year.

Pretty nice, huh?

What does this tell us about deep analysis, management interviews, security analyst spreadsheets, and all the other treasured tools of investing?

It tells us that human judgment -- even expert human judgment -- should be a regular source of humility, not pride. The sorrowful truth is that in every field of human endeavor, simple decision rules---like the magic formula for stock picking -- have routinely produced better results, on average, than complex human judgment.

One of the most powerful examples of this is the rapidly accumulating research in behavioral economics and psychology. Repeatedly, researchers have found that a simple set of rules or algorithms will produce better results than idiosyncratic professional judgments. (For a quick introduction to this area, read the James Montier piece in the August 21 issue of John Mauldin's "Outside the Box" free online newsletter.)

This reality is the foundation of my preference for index investing and my interest in the new rules-based exchange traded funds.

OK, you ask--- so what's the caveat?

Execution, simple execution -- a subject barely mentioned in The Little Book.

While the Little Book seems to be a nice idea for the small investor, the reality is that it would be too expensive for most people to put in practice. Suppose, for instance, that you invested $333 in each of the 30 stocks recommended to have a magic formula portfolio. That would require $9,990. That's 3 or 4 times the minimum investment for most mutual funds.

If you dealt with a discount broker and could do your trades for $10 each, your commission costs would be $300 to buy the portfolio and $300 to sell it, which you would do once a year. That's a total cost of $600 -- 6 percent of the portfolio a year. The burden wouldn't kill the performance advantage but it would certainly dampen it.

To do this with minimal portfolio drag -- say  ½ of 1 percent a year in commission costs--- you'd need a starting portfolio of about $120,000. You could reach the same low cost level with a starting portfolio of about $60,000 by using a fixed annual cost outfit like foliofn (visit