(FORTUNE Magazine) – Stargazing is one of the great pastimes in the mutual fund industry. Investors have come to expect nothing short of prescience from the ratings awarded by investment advisory Morningstar. Marketers and managers, meanwhile, crave the prestige and fund flows that follow sparkling ratings. So there has been an understandable amount of fear and trembling in the buildup to July 1, when some 55% of the 10,000 funds ranked by Morningstar will suddenly gain or lose coveted "stars" without so much as a nip or tuck in performance.

The blockbuster change has supporters and detractors. But for all the controversy--and there's plenty--it raises a more fundamental question: Do the ratings actually tell investors anything meaningful about the future performance of mutual funds?

Answering that question requires a quick look at the star system, old and new. Until this month Morningstar graded funds against a curve in four broad categories: U.S. stocks, international stocks, taxable bonds, and municipal bonds. The 10% that performed best for various time periods (adjusting for sales fee and risk) were awarded five stars; 22.5% received four stars, 35% three stars, 22.5% two stars, and 10% one star. But lumping together so many disparate funds meant that apples were not being compared with apples. Small-cap funds competed with large caps even though history suggests the two styles succeed in different periods. The system was designed to rate the small 1980s-era universe of all-purpose portfolios. But since then the industry had grown exponentially and segmented its funds into endless permutations.

To accommodate this segmentation, Morningstar is slicing and dicing the fund universe into a dozen times as many groupings as before. There will now be 48 distinct sectors. Each of the 48 will have the same proportion of stars--regardless of the sector's overall performance. That means hot sectors will lose most of their five-stars, while ice-cold segments will suddenly be festooned with more brass than a meeting of the Joint Chiefs of Staff. As of the end of May, for example, Morningstar had lumped 190 small-cap value funds within that one ranking of U.S. stocks. Of those, 111 received the highest ratings; under the new system, which carves out a separate category for small caps, only 21 would do so. Similarly, a meager five out of 684 large-cap growth funds earned five stars under the old system; 54 would under the new.

Some fund denizens hail the fact that being in a hot sector no longer guarantees a good rating. Fund managers, says Bob Batchelor of Strong Funds, will "now have to earn those stars." But others say Morningstar has simply shifted from lauding one sort of mediocrity to rewarding another. Since even the weakest sector is now guaranteed its share of five-star funds, some funds whose investors have taken massive losses could win gaudy ratings.

Still others argue that the new sector focus ties the hands of managers who risk losing stars if they don't tailor their investment strategies to Morningstar's categories. (One of the paradoxes of the ratings is that they're so influential they end up affecting fund managers' choices rather than simply assessing them.)

No matter how effective the changes are, however, they won't alter the most basic and obvious--and yet most forgotten--fact about the star system. The ratings, as Morningstar has always acknowledged, tell you way more about the past than they do about the future. They have little if any predictive value.

Consider this simple but extraordinary example, which comes from Morningstar itself. One-star funds have outperformed five-star funds by 45% since 1995. If you invested $1,000 across the universe of one-star funds in January 1995 and then moved the proceeds each year to the then-current crop of one-star funds, your investment would have been worth $2,948.54 at the end of 2001. If you'd pursued the equivalent five-star strategy, you would have had $2,030.48. Ooof.

"By the time a fund reached five stars," says fund consultant Geoff Bobroff of the old ratings, "its style had usually crested, and it was the worst time to be getting in." Before the tech plunge in 2000, for example, most five-star-rated funds were either growth- or tech-oriented. Those funds dropped 14% that year, while one-stars, then dominated by value and small caps, gained nearly 6%.

In theory, the new star allocation makes it less likely that the one-star funds will outperform the five-stars. Since both groups will include every style of fund, they'll be much more diversified, and thus the performance gap will probably be smaller. Morningstar managing director Don Phillips argues that the revamped rules will help investors find good funds in out-of-favor sectors. To the extent that that encourages diversification, it's commendable and could help protect investors' portfolios during difficult markets. But investors take note: The predictive power of the stars will remain what it always was.

REPORTER ASSOCIATE Ellen Florian