It's been a tough three years for most stocks. Since April 4, 2000, this bear market has taken a 25.9% bite out of the

Dow Jones Industrial Average

, gouged 41.2% from the

S&P 500

and ripped the

Nasdaq Composite

by 66.7%.

But it hasn't been a terrible three years for all stocks. The top-performing 25% of stocks have returned between 13% and 202% for the period, according to Morningstar. Returns drop off dramatically from there, with the second-best 25% returning a range between a gain of 13% and a loss of 11%. The worst 25% returned losses ranging from 43% to 97%.

So what made this top 25% so special? And is there a way to identify stocks with a better-than-even chance to make it into this group in the future?

After poring through the data on these market outperformers and writing more than a few screens to pinpoint the cause of these returns, I have an idea of how to look for the kind of stocks that might do well in this kind of market in the next year or more.

Start With Free Cash Flow

My method isn't very complicated. It begins by looking at a measure called free cash flow, to identify the kind of cash-generating businesses that are especially attractive to own when the economy and the stock market aren't going gangbusters. And it ends with adding a value measure to provide some insurance against overpaying.

I didn't go through all this trouble simply as an exercise in nostalgia. I'm interested in what the performance of this group of stocks over the last three years might be able to tell us to look for in our stock picks in the future.

Maybe the economy is about to kick into high gear, fueled by the sudden release of spending pent up by uncertainty about the war in Iraq. If that happens, the postwar stock market won't look very much like the bear market we've come to know so well. In that case, what's worked in the stock market over the last three years probably won't work very well in the stock market going forward.

But if that hoped-for economic recovery doesn't arrive on schedule or is weaker than expected, then the lessons of the last three years might be useful. Stocks that worked well in the low-growth, no-growth economy and the skittish investing economy might work well for months to come.

An additional point: The longer it takes for solid and reassuring growth to arrive, the better chance there is that the type of stocks that worked in the last three years will continue to make profits for investors.

Unlike the list of short-term top performers for 2003's first quarter, the list of the top-performing 25% over the last three years includes some names you'll recognize even near the very top end of it.

And by the time you've moved down the list to stocks returning 30% or less for the period, the territory should seem positively welcoming to even big-company growth stock investors.

This list feels less random than the list of first-quarter winners with its largely unfamiliar stocks. And that makes sense: The twists and turns of news can propel obscure and thinly traded companies to huge gains over the short term. In fact, the more obscure and thinly traded a stock is, the easier it may be to generate a huge short-term pop in it.

But over a longer period of, say, three years, top-performing stocks need to be more than just one-trick ponies or overnight success stories. The companies behind these stocks need to be fundamentally solid to compete successfully for the long haul.

Identify Fundamentals

Or at least we'd like to believe this. So let's take a look at which fundamentals seem most important to putting a stock in the top 25% of all performers over the last three years.

Forget growth in earnings per share. Of the stocks that grew earnings per share faster than the S&P 500 for the three years, just 16% made it into the top 25% on performance.

Revenue growth works better. About 35% of the companies that grew revenue faster than the S&P 500 managed to make it into the best-performing 25% of all stocks.

But my efforts to improve those odds by raising the bar on this fundamental didn't help the results. If you look at companies that grew revenue twice as fast, or more, than the S&P 500 over the last three years, only 31% of these stocks made it into the top 25% of price performers.

I have a theory on why earnings per share and revenue growth show this behavior. At the time the bubble popped in 2000, investors had bid up fast earnings growers and put an even bigger premium on fast revenue growers without earnings. The bear market for the last three years has crushed the multiples of these stocks. So even though growth may have remained above the market average, these stocks still fell because they'd been trading at such extreme multiples of earnings and revenue.

A lot of investors, fed up with the hyping of stocks with revenue but no earnings, and especially fond of the games that companies and analysts play with earnings numbers, have turned to cash flow, and especially free cash flow -- operating cash flow less capital expenditures.

By looking at how much actual cash a business generates, these investors figure they can avoid the worst accounting abuses and be reasonably certain they're buying a solid business and not just a stock certificate.

By itself, cash flow does a pretty good job of picking winners -- but not significantly better than simple revenue growth. You'll find very few stocks that have been able to increase free cash flow in each of the last three years. That's not surprising. It's a pretty tough test for a company and its management. They'd been operating at a time when the lure of cheap capital hooked many companies on what would turn out to be overexpansion. And all that was followed up by a huge contraction in orders and revenue.

Refine the Strategy

But there are companies that have pulled off this trick despite the economy and the capital that markets have thrown at them. For example,

Electronic Arts

( ERTS) has managed to increase free cash flow in each of the last three years, and generated $531 million in free cash flow during the last 12 months. Or how about

TST Recommends

GTech Holdings

( GTK), with $211 million in free cash flow on just $983 million in revenue in the trailing 12 months? Or

Stryker

(SYK) - Get Report

, which generated $311 million in cash over the same period?

About 33% of the stocks of companies that increased cash flow in each of the last three years worked their way into the top 25% of all stocks over the period. That just about matches the results from picking on revenue growth alone during this period.

But I did find one significant advantage to using free cash flow as the fundamental first cut at picking stocks. Sticking to companies with solid and increasing cash flows makes losing money more difficult. About 54% of all the companies I found that increased free cash flow in each of the last three years turned in a positive return for the three-year period.

Granted, making 3% on

Illinois Tool Works

(ITW) - Get Report

or 0.6% on

Staples

(SPLS)

over three years isn't very exciting or profitable, but even those meager returns are better than the losses being generated all over this market.

Is there a way to improve the stock-picking accuracy of free cash flow? Not without adding another variable to the system, in my opinion. I tried extending the time period to four years, looking for companies that showed annual increases in free cash flow during each year, and that actually lowered my results, with only 31% of the stocks of those companies making it into the best-performing 25% of all stocks. Looking for reversals -- surprise stocks that managed to start a string of positive increases in free cash flow after showing a dip in that financial measure -- didn't do the trick either.

Again, I have a theory why extending the time period didn't improve the results. At some point, as a company continues to generate huge and predictable free cash flows, investors catch on. And then they bid the stock up in price. The increase in multiple becomes a limiting factor on the stock's performance over time.

Reality Checker

So, to improve the performance of the stock-picking system based on free cash flow, I think an investor will have to add some value measures. That will help the investor buy these very attractive businesses at attractive prices, before the financial markets have priced the stocks for the premium goods that they are.

I kept my value measure in the cash flow family by looking for stocks that were trading back in 2000 at a price-to-free-cash-flow multiple that was below the average for the next three years. With that added to the system, almost 40% of the stocks of companies with three years of increasing free cash flow wound up among the 25% of best-performing stocks over the last three years. And 57% of the group showed positive returns for the three-year period.

Let me share some of the names that this free cash flow plus price-to-free-cash-flow system churned out. No screen should be the endpoint of your investment research, but these are all stocks that I think will pay a positive return on your research.

So take a look at these five:

Laboratory Corp. of America

(LH) - Get Report

, Stryker,

TransCanada PipeLines

(TRP) - Get Report

, Sysco and

Procter & Gamble

(PG) - Get Report

.

You can find this information on each of these companies' stock quotes pages by clicking on "cash flow." You'll be able to get complete statements for the last five years. They read nicely indeed.

Jim Jubak appears Wednesdays on CNBC's "Business Center" at 6 p.m. EST. At the time of publication, Jim Jubak owned or controlled shares in none of the equities mentioned in this column. He does not own short positions in any stock mentioned in this column.