(MONEY Magazine) – Everyone who has ever read a mutual fund ad recognizes this disclaimer: "Past performance is no guarantee of future results." But investors routinely ignore those eight little words. Consider:

--According to a survey conducted last October by Liberty Financial, an asset management company in Boston, 46% of investors said they choose their funds based solely on returns the portfolios have earned in the past.

--80% of the money that flowed into mutual funds in 1996 went into funds that had earned four or five stars for past performance from Morningstar, a mutual fund research firm in Chicago. However, Morningstar insists that star ratings predict nothing.

--The hottest funds cool just as quickly. For example, Alger Small Cap Fund shot the lights out in 1991, returning 54.7% that year. Investors swarmed in after that, swelling the fund from $81 million to $583 million today. But ever since that banner year, Alger has returned 11.7% annually, which is 1 1/2 points below the average for similar funds.

Clearly, fund investors have placed--or rather misplaced--considerable faith in the ability of past performance to shed light on the future. It's understandable, given the volume of fund rankings that sprout like mushrooms after each quarter's end. But while those numbers may tell you more than you could ever want to know about last year's results, they give you no information about future winners. In fact, statistical studies have long shown that stock returns from one period to the next are essentially random. The beginning of wisdom in fund investing is to recognize that no human being can predict next year's (or next month's or next decade's) top funds--and no analysis of past performance can change that.

But that doesn't mean you should ignore a fund's record. No professional adviser would think of evaluating a money manager or mutual fund without examining the record, says Charles Ellis, the founder of Greenwich Associates in Greenwich, Conn., a consulting firm for financial services companies. "If you are going to invest in actively managed funds, you need to know how to analyze their history correctly," he says. The key is to learn what knowledge the past does convey--and then to read no more into it than that.

To find out what you can learn from a fund's past performance, we interviewed more than two dozen finance professors, fund analysts and investment advisers. Their message: Returns may be random, but other aspects of a fund's past are not. Funds that were more volatile than average tend to remain so, for example, and funds that followed a particular investment approach last year are likely to do so this year too. Far from being a statistical dead end, a fund's past volatility and investment style can tell you a good deal about its likely behavior in upcoming years. There is even evidence that adjusting past returns for risk can give you a better-than-even chance of identifying the risk-adjusted winners of the future.

Here in more detail is what to look for when you examine a fund's history.

--How your fund did relative to its peers. Suppose your barber tells you that he is in Invesco Emerging Growth, which was up 11.6% last year. Unless you already know that the fund is a small-cap growth fund, you don't know whether your barber is bragging or looking for sympathy. In fact, last year the average small-cap growth fund was up 17.7%; the barber must be angling for a bigger tip.

Finding the right yardstick for your fund is not easy, however. Standard & Poor's 500-stock index is probably the most widely used benchmark, but it is appropriate only for a portfolio of stately domestic large-cap stocks. Funds that follow other investment styles should be measured against less well-known indexes. "You should compare a small-cap growth-stock fund to a benchmark like the Wilshire small cap growth index," says Roger Ibbotson, professor of finance at Yale and president of the investment research firm Ibbotson Associates in Chicago. To find the performance of the appropriate benchmarks for some 1,400 of the most widely owned funds, check Morningstar Mutual Funds, a publication carried by most major libraries. For an even wider choice of indexes, log on to Data Broadcasting Corp.'s Internet site (www.dbc.com).

Once you have settled on your fund's proper measuring stick, take a look at how closely the fund mirrored the benchmark's ups and downs--what statisticians call correlation, or r2. Correlation is frequently measured on a scale from 0 (for a fund whose movements bear no relation to its index) to 100 (for a fund that moved in perfect lockstep with its benchmark). You can find correlation figures in Morningstar.

Take two small-company funds, for example, both of which are usually benchmarked against the Russell 2000 index. Vanguard Explorer has a high correlation of 92, which tells you that its managers, Kenneth L. Abrams and John J. Granahan, seek out stocks in the mainstream of their investment style. They are likely to continue to shadow the benchmark in the future. By contrast, Heartland Value Plus' low correlation of 48 reflects the idiosyncratic low-volatility style of managers' Bill Nasgovitz and Ronald Saba. What does that tell you? Maybe a lot: If you're looking for a small-cap to diversify a portfolio that is heavy in blue chips, you would clearly prefer the Vanguard fund, with its conventional small-stock style. Heartland, on the other hand, would make sense if you want to put a toe in the small-stock pond but don't want the category's headlong risk.

--How volatile your fund was. Let's say that your fund's average return over the past three years was 10%. Now suppose that in the three-year period your returns varied widely, sometimes posting huge gains and other times suffering heart-stopping losses. That jumpiness surely wouldn't make you feel as good about your average.

To describe how much a fund's return varies around its overall average return, statisticians use the words standard deviation. Sorry about that, but we have to use them too. To find out your fund's standard deviation, you can simply look it up in Morningstar. Then compare its standard deviation with that of funds with similar investment objectives and similar returns. If your fund's standard deviation is larger, then you've taken a bumpier ride than your peers to get to the same place. And if it was bumpy in the past, it'll probably be bumpy in the future. "There is a lot of consistency in a fund's volatility," says Martin J. Gruber, professor of finance at New York University.

--How well your fund performed on a risk-adjusted basis. Think of risk as the price you pay for fund performance. The more of it you assume, the higher the return you should expect in the long run. Thus most fund pros would say that the 12.3% annual returns delivered in the three years that ended Jan. 1 by Nicholas Equity Income is a more impressive effort than the 15% posted by the much riskier Marshall Mid-Cap Stock Fund. Why? According to Morningstar's risk rankings, Nicholas took on 41% less risk than the average fund, while Marshall loaded up with 46% more. "Looking at mutual fund returns without adjusting for risk is virtually useless," says Mark Wright, a Morningstar analyst.

There are several mathematical methods for adjusting a fund's past returns to reflect its risk level. Morningstar, for example, awards between one and five stars to more than 4,400 funds. Five-star funds have posted the biggest gains with the least risk over a variety of periods. (For more on risk-adjusted fund returns, see Fund Watch on page 49.)

Risk-adjusted returns are not, of course, a sneak preview of next year's winners. But it can distinguish between records that are built on bet-the-farm gambles and those created by consistent stock picking. At best, it may even give you a faint indication of a fund's future promise.

To reach that conclusion, Yale's Roger Ibbotson studied the performance of more than 1,000 funds between 1976 and 1995. He adjusted the trailing three-year averages for both risk and investment style, then ranked them from best to worst. He discovered that if a fund's three-year adjusted return is in the top half of its peers, then 62% of the time the fund will go on to finish in the top half for the following three years. Funds in the top quarter repeat their ranking 42% of the time. In other words, there is some consistency of performance. (If there were none, only about 25% of the top-quartile performers would repeat.) "Some fund managers--certainly not all--show they can consistently beat their peers," Ibbotson concludes.

One caution: Ibbotson's study says nothing about a top three-year performer's long-term promise. To replicate his study, you would have to turn over more than a third of your portfolio every three years to make sure you continually had funds from the top half. Ibbotson acknowledges that this is neither practical nor desirable for taxable investments. "Altering your portfolio every three years to capture that slight outperformance would create hefty capital-gains taxes," he says.

Even so, there's no reason not to try to choose the kind of three-year outperformers that Ibbotson identified for your long-term portfolio. "That way it's a little more likely that you'll beat the pack over the next three years," says Ibbotson's Yale colleague, finance professor William Goetzman. After that, you're on your own. For five equity funds that Ibbotson has tapped as outperformers, see the table below.

--Whether your fund performed badly on a risk-adjusted basis. New York University finance professors Martin J. Gruber and Edwin J. Elton have discovered a corollary to Ibbotson's findings: Poor performers on a risk-adjusted basis over the past three years will probably lag for the next three as well. Gruber and Elton examined risk- and style-adjusted returns for 270 funds averaged over three years from 1985 to 1995. They found that the bottom 10%, or decile, tended to underperform the average fund significantly over the next three years. So if one of your funds has been trailing its peers during the past three years--and for no good reason, like losing a big sector bet--then you should sell the laggard right away. "Three years is enough to let a fund manager work out the kinks," says Mark Bass, a financial planner in Lubbock, Texas. "Any longer than that and you're wasting your money."

--Whether any of this information is still relevant. Now you know the extent to which you can make inferences about a fund's future performance based on its past. But before you jump to even these modest conclusions, make sure your data really apply.

For starters, examine the fund's volatility and risk-adjusted return over a variety of time periods. That way, you won't be misled by a short-term anomaly. Take the now defunct Dreyfus Edison Electric Index Fund, for example. It ran an ad in the Wall Street Journal on July 20, 1995 calling itself "The No. 1 Utility Fund" and citing one-year returns ended June 30. For 12 months Dreyfus was indeed tops among the 75 utility funds with full-year records. But were you to move that one-year window back three months, you'd find the fund ranked 35th out of the then existing 72 with full-year records. Go back three more months to the year ended Dec. 31, 1994, and Dreyfus came in a dismal 62 out of 65.

Check too whether the fund has recently replaced its manager. Unless your portfolio is run by a team, a fund's track record is meaningless if the manager who compiled it is no longer there. So if you wanted to see how the Oppenheimer Main Street Income & Growth Fund has done, for example, look back no further than November 1995. That's when the current chief, Robert J. Milnamow took over.

Every seasoned fundholder has found some way to accommodate the uncertainty involved in fund investing. In that respect, investing is no different from many other financial endeavors. Buying a house, starting a new job, taking out a loan--all require you to make a commitment without knowing how things will turn out. The key to smart investing is to learn what you can from the information you have, without assuming it tells you more than you can really know. A little knowledge is a dangerous thing only if you think it's more knowledge than it is.