(MONEY Magazine) – Want to find a great mutual fund? No problem--just buy a mediocre one and hang on. That's the startling conclusion of a new study. While this may sound like an update of Aesop's fable about the tortoise and the hare, there's more to it than that.

It's Hard to Ignore the Hare

Ronald Wilcox, a University of Virginia marketing professor, found that when investors choose a fund, they place at least twice as much weight on past returns as they do on any other factor. And why not? The sun always rises in the east, so it makes no sense to look for first light in the western skies. Shouldn't the past predict the future for mutual funds in the same way it does for the dawn?

The short answer is no. The latest research shows that funds with middling returns in the short run come out on top in the long run. The study was done by the Brandes Institute, the research division of Brandes Investment Partners, a money manager that oversees roughly $100 billion. By looking at the results of hundreds of big funds from 1992 to 2003, Brandes concluded that a fund that consistently placed in the 43rd percentile (beating 57% of similar funds) would have moved into the top quarter after only five years.

That's hypothetical, of course; no actual fund ever beats 57% of its peers all the time. In the real world, moderately good results usually add up to great returns over time. Why? In investing as in sports, winning streaks rarely last long. Brandes Institute director Barry Gillman likens funds to a relay race: If your runners are second every leg, and rival runners sometimes finish first and sometimes last, you are likely to win. Plus, hot funds draw so much cash that they struggle to find enough good stocks--and end up like sprinters in cement sneakers.

It's also true in the real world that even "consistently good" funds falter, landing in the bottom quarter nearly 10% of the time, according to Gillman. Thus stalwarts like Selected American Shares and T. Rowe Price Equity Income looked bad in 1999, a time when racier offerings from Firsthand and Van Wagoner tripled in value. But Selected and T. Rowe bounced back, while the hotties flamed out, losing up to 80% in the next three years. The race goes to the patient.

You can't replicate the complex analysis in the Brandes study, but you can avoid the common mistake of merely looking at one year's performance or average returns over the past three or five. That "average" can be inflated by a short hot streak, so you want to look at how a fund did in each calendar year; at Morningstar.com, you can find how a fund fared relative to its peers each year back to 1998. Look for a fund whose percentile rank is rarely near the top (1) or bottom (100), and favor funds with yearly rankings usually in the top half. Over time these tortoises should end up passing the hares. Skip any U.S. stock fund with expenses above 1% (1.25% for small stocks). Finally, read the annual report. Can you understand what the managers say they are doing? Do they own up to mistakes--or blame "conditions"? Do they brag about high returns--or warn responsibly that the future won't be as rosy? If the tone is wrong, move on.