Microsoft Corp. is in big trouble. It's beset by competitors. It's too big to grow much anymore. There's the Y2K threat. There's a slackening of demand for personal computers. There's global uncertainty.

Sure, Bill Gates's software firm has had a long streak of success that has given it the nation's highest stock-market value, but face it, folks: The party's over.

That's what Microsoft executives regularly tell the analysts who follow its stock. "They warn everybody every quarter that the sky may be falling," says Massachusetts Institute of Technology professor Michael Cusumano, co-author of "Microsoft Secrets."

Instead, everything comes up roses. For every quarter since at least 1994, Microsoft has always done better than the analysts were forecasting, giving fresh impetus to the stock price. When the company reports its most recent earnings after the close of the market today, it is likely to add to a fabled winning streak.

Managing expectations is a game that many companies try to play. With the Internet allowing near-instantaneous availability of information, earnings shocks play out in minutes rather than weeks. The move to 24-hour trading is only accelerating the process. Hoping to minimize the carnage, companies now routinely "pre-announce" bad results.

It doesn't take much to cause trouble. Chipmaker Intel Corp. reported last week that earnings were up 21 percent. Good news, but a shade under what analysts had expected. The market immediately dragged the stock down 8 percent. For computer maker Unisys Corp., earnings were safely above forecasts but revenue was not. The stock plunged by a third.

Microsoft never gets trashed like this. "In the dance between analysts and companies to lower expectations, Microsoft is pretty close to the best there is," says Roger Kay, who follows the industry for International Data Corp.

While some Microsoft executives might be genuinely worried that the company will come up short at any moment, "they also relish the game," says Kay. "Look at the memos in the [Justice Department antitrust] trial--there are a lot of little interjections that are almost gleeful about how they maneuvered in a certain way. Gamesmanship is in their blood. They play as hard as they can."

The gamesmanship practiced by Microsoft and other companies worries Securities and Exchange Commission Chairman Arthur Levitt Jr. In a speech last night in New York, Levitt took issue with what he called "a web of dysfunctional relationships" between companies and analysts, including those "where companies' reported results are tailored more for the benefit of consensus estimates than to the reality of the ups and downs of business" and "where companies work to lower expectations when they fully expect they'll beat the estimates."

For Microsoft, the process began its most recent cycle at the company's annual meeting for Wall Street analysts, held here in late July. More than a hundred came from around the country to listen to Chief Financial Officer Greg Maffei and other executives.

From the start, the company seemed to playfully acknowledge its expectations game. In an electronic multiple-choice poll, more than 90 percent of the analysts said they began their day not with such things as coffee or ale but with "Prozac to counter Greg's annual 'woe is me' speech."

Director of Investor Relations Carla Lewis, who was conducting the poll, pretended dismay. "Am I the one to give Greg the news you're on to him?" she asked.

Then the analysts were asked how much they thought the company would grow over the next year. Forty-four percent chose the maximum: more than 25 percent.

For the rest of the day, the executives took issue with that number. "I was shocked at Carla's poll, shocked," said Microsoft President Steve Ballmer. "Twenty-five percent a year? Come on, it's outlandish, crazy . . . There's competition like we've never seen before. Y2K is looming."

Ballmer acknowledged that the company had just had a great year but said the good times were ending: "There's stronger competition than I've felt in my 19 years at Microsoft." He cited the Law of Large Numbers--that it's easier for a small company to grow at 25 percent a year than a big one.

Maffei noted that computer prices were sliding and said that would push down the price Microsoft could get for its flagship Windows operating system. The company's hardware and retail group? "I don't think either of these businesses will have huge top-line growth." Europe? "Not going to be the huge growth engine it's been."

Even the company's massive success in the past made the future more problematic. With software for small businesses, for example, Maffei noted that "there's no upside." With 49 million licenses, Microsoft controls 93 percent of the software market for office software.

At the end of the day, the analysts had to vote on their revenue projections again. The percentage voting for 25-plus-percent growth fell from 44 to 40. Clearly, the performance had been at least mildly convincing.

All this talk was an attempt by Microsoft to provide what is called "guidance." If a company is known for being very conservative in its counsel, that's called "guidance bias" and is supposedly taken into account by the analysts.

But adjusting for guidance bias can only do so much. "If a company is ultraconservative, it's hard to be ultra-aggressive in response," says Mike Kwatinetz, who covers Microsoft for CS First Boston. "They do know more than you about what they're looking at. You get a little nervous that this is the year things work differently."

Kwatinetz ascribes Microsoft's conservatism to paranoia: "Where other companies are thinking about what can go right, they're thinking about what can go wrong. They share their paranoia with Wall Street. It's one of the things that makes them a strong company. Paranoia is good."

The analysts' estimate for the quarter that will be announced today is 36 cents--"or better, as I like to think of it," Kwatinetz says. That's 2 cents over First Call's consensus. Kwatinetz says he can't recall an instance where a major analysts' estimate actually turned out to be higher than the final reported figure. "Microsoft," he adds, "could teach a PhD course in how to run a company and guide the Street."

Other observers are more critical. Cusumano, whose "Microsoft Secrets" is a generally favorable look at the company, sees it as being a bit more cagey than paranoid. "It's their strategy so the stock price doesn't get hit," he says.

The analysts, meanwhile, don't want the stock to get hit either. For one thing, most of them are recommending it. If the estimates prove to be too high, the stock will fall and they'll look bad.

"It behooves them all not to have a downside surprise," says Paul R. Brown, chairman of the accounting department at New York University's business school. He said the company and the analysts "both need each other to avoid bad surprises."

Complicating the issue is an ongoing SEC investigation into Microsoft's accounting practices. The investigation stems from a wrongful-dismissal lawsuit filed by one of the company's internal auditors, who charged that he was fired after uncovering evidence that Microsoft was manipulating its revenue reserves to keep its earnings in a steady upward progression.

The suit has since been settled privately, and Microsoft denies wrongdoing. But in a hearing last year, attorneys for the auditor, Charles Pancerzewski, introduced a 1995 electronic-mail message from the company's then-chief financial officer, Mike Brown, to Chairman Gates. "I believe we should do all we can to smooth our earnings and keep a steady state earnings model," Brown wrote.

While an attorney for the auditor argued that this was evidence of illegality, Brown, the NYU chairman, noted that "smoothing earnings in and of itself is not fraudulent--it just sounds like it is."

In fact, according to Brian Bruce, a Southern Methodist University professor who wrote "The Handbook of Corporate Earnings Analysis," the market's hair-trigger response to earnings disappointment is giving more impetus to smoothing.

Microsoft declined to comment for this article. But its history of earnings surprises, if not its many years of earnings growth, may be coming to an end. The Internet has facilitated the rise of the "whisper" number, off-the-record estimates that are usually more than the consensus estimates. (For the quarter being reported today, Microsoft's whisper numbers range from 34 to 39 cents a share.)

And all this could get even more complicated. Influential Merrill Lynch analyst Henry Blodget wrote in a report earlier this month on the Internet portal Yahoo that its earnings of 14 cents a share "exceeded consensus of $0.09 and whispers of $0.11-$0.12" and "even beat the double-secret, uber-whisper of $0.13."

Double-secret uber-whispers! If they catch on, even Microsoft may find itself someday failing to exceed estimates. Either that or the whole system will fall apart.

"Whisper upon whisper is getting to be exponentially crazy," says IDC analyst Kay. "But no one ever said it was a fair game."

Staff researcher Richard Drezen contributed to this report.

CAPTION: Chairman Bill Gates speaks at last week's Geneva telecom conference.

CAPTION: Never Disappointing (This graphic was not available)