NEW YORK (MONEY Magazine) - In the Tennessee Williams play "Sweet Bird of Youth," an aging actress wakes from a drunken stupor to realize she's spent the night with a gigolo. Still clutching a sleep mask to her eyes, she asks him if he's attractive.

"I used to be the best-looking boy in this town," he says. The actress pauses, then asks, "How large is the town?"

Give her this much: She has a solid grasp of statistics. Whenever someone claims to be the best, the question to ask next is, "The best of what?"

This is especially critical for fund investors. We've all seen ads that boast of "top rated" funds or read about managers who beat 90 percent of their category. But what do these numbers really mean?

There are two ways that fund returns usually get sliced and diced.

Two firms, Lipper and Morningstar, produce most of the fund comparisons you see. Both sort stock funds into 30 to 40 categories, with diversified funds grouped by investment style -- that is, by the size and the valuation of their stocks.

So when Hotchkis & Wiley Large Cap Value is ranked No. 1 in total return over three years by Morningstar, it's being measured against other funds that buy cheap blue chips.

Related articles • The Ultimate Mutual Fund Guide 2004 • America's hottest fund company • More on mutual funds

Makes sense, but there are a couple of wrinkles. First, funds change styles, and that means they'll switch categories. (Russ Kinnel, Morningstar's director of fund research, says this happens to about 5 percent of U.S. stock funds every six months.) It may not make sense to compare their long-term records with their new peers.

Second, when was the last time you thought you needed a "midcap blend" fund? These fine distinctions mean a lot to investing pros. But if you just want to put most of your IRA into a decent stock fund, a better measure is how a fund fares against the no-brainer alternative -- a low-cost index fund like Vanguard Total Stock Market.

Getting a four- or five-star rating from Morningstar is marketing gold for fund companies. So what's behind this rating? Morningstar takes the same relative return data we've just discussed and adds a twist, factoring in sales charges and the volatility investors experience. That's why a high-risk fund that earned 70 percent last year may rate just one or two stars.

Funds are rated only against others in the same category, which helps you compare apples with apples. What stars can't tell you is whether you should invest in that category at all.

Most investors would be better off putting the bulk of their money in a three-star large-value fund than in, say, a risky tech-sector fund with five stars. What counts isn't who you beat but how much money you get to take home.

Seligman Communications & Information **** This fund earns high marks from Morningstar because it has shown both better returns and less risk than the average tech-sector fund. Well done.

But in this case, less risk still means a whole lot of volatility. In 2002 the fund lost more than 36 percent.

Franklin Small Cap Value ** According to Morningstar data, over the past three years this fund has been pretty volatile as small-value funds go -- but it lost only 9.6 percent in 2002.

An investor who put $10,000 in Seligman Communications & Information three years ago would have $9,697 today; the same amount in Franklin Small Cap Value would have grown to $13,514.

Vanguard Dividend Growth ranks in the bottom 7 percent That's how Morningstar says it compares with other large-value funds over the past five years.

It's not a fair comparison, though. Until December 2002 this was actually a utilities-sector fund. Utilities in general lost a lot of money in 2001 and 2002, but this fund did well enough on a relative basis that a Morningstar report called it "a fine choice."

Investors evaluating this fund today really have only its one-year record to go on. Its 33 percent return places it in the top half of its new pack.