THERE'S NO POINT IN

regularly recommending stocks if no one keeps score. So at the end of every year, I take a look in the rearview mirror. I study a year's worth of Stockscreen columns and highlight what went right and what went wrong. I usually uncover some valuable lessons, often learning more from my failures than from my successes. In 2005 there weren't many failures to draw lessons from.

This is my 12th year of these annual performance reviews, and the results are among the best ever. On average, the companies that appeared in my tables are up an annualized 20 percent. That's almost three times the market's gain for comparable periods. I'm also pleased that none of my portfolios lost money. I managed that feat only once before back in 1997, when folks were quitting their real jobs to day-trade on the beach. Avoiding losers was considerably harder in 2005. Nearly half of all stocks now sell for less than they did 12 months ago.

Still, there are plenty of useful insights even if some come from things that went right. Before I dig into the numbers, here's a quick summary of where they come from. Every month I invest an imaginary $10,000 in the stocks that appear in my table. I make equal dollar-amount purchases at prices that appear in the magazine, and I exclude brokerage costs (which are different for every investor and steadily declining). Then I calculate the current value of the portfolios, adjusting for dividends, which I hold as cash.

This review covers a year of columns from September 2004 through August 2005. The average portfolio age is 12 months, but some have been around longer than others. So I annualize to provide comparability and because that's the way investors think about returns. This can distort short-term results, but I think it makes for better overall comparisons. These numbers appear in the fourth column.

No performance measurement makes sense without a benchmark, and I use the Dow Jones Wilshire 5000. It tracks the S&P 500 closely, but the Wilshire is a broader index that includes nearly all U.S. public companies, as well as REITs and partnerships. The sixth column compares the annualized performance of each portfolio with the Wilshire's performance over the same period. There's only one minus sign, which means I consistently beat the market by 13.1 percentage points a year. (For the record, my long-term average is closer to five percentage points.)

I'm pleased, of course. But I need to share the credit. Nearly all of my success comes because of someone else's hard work. I rely on academic research for most columns, and I have two goals. I want to introduce readers to ideas percolating among economists and money managers, and I want to illustrate them using my own stock screens an exercise that enterprising individuals can do at home. My portfolios are too small to be statistically significant, and I often muddy the waters with value-oriented tweaking. But not everything gets lost in the shuffle.

In this year's summary, two portfolios stand out. In October 2004, my colleague Ken Brown combined traditional value and momentum screens to generate spectacular gains. The stocks he selected, based on ideas used by two successful money managers, beat the market by more than 30 percentage points. My April search for companies with revenue and profit surprises, an attempt to piggy-back recent accounting research, was also a big winner. Both strategies can be duplicated, and the columns that explain them are worth rereading.

I used research about the dispersion effect to turn up winners in both November 2004 and August 2005. The idea behind dispersion is that investors should sell a stock short when analysts disagree over a company's prospects and buy when there's a tight consensus on the theory that management sends clear signals if the future looks bright and blows smoke if there's trouble ahead. Other research-based screens also did well: my focus in February on companies in which the founder is still chief executive and my search the following month for bargains based on cash flow from operations. Each approach should be a useful addition to your stock-picking tool kit.

Now for some less encouraging results. My September 2004 column about the guru-tracking Web site Validea.com featured a screen based on money manager Martin Zweig's strategy. He's still Validea's top-rated stock picker. But my portfolio beat the market by just 1.3 percentage points. Then, in June, I was impressed by research about wasteful capital spending, so I zeroed in on companies with spending cutbacks. Though that approach worked for the academics, my spending laggards are the only portfolio in this roster that trailed the market.

There's a common thread here. Scan across the table to the last column, and check the worst-performing stocks. Doral Financial, down 73 percent; American Italian Pasta, down 76 percent. These are phenomenal losses. Among companies with market values above $500 million (my cutoff for these columns), only about a dozen stocks have collapsed that resoundingly in the past year and I picked two of them. Without these duds, both screens handily trounced the market.

Could I could anyone have avoided such disasters? I'm not sure about Doral. It's a Puerto Rican bank that got in trouble with a mortgage-hedging strategy. Lots of big investors took losses including Fidelity Investments, which still owns nearly 10 percent of the company's stock. American Italian, on the other hand, is a roll-up of regional pasta brands that cooked the books, some shareholders claim. Short sellers were active, and there were critical press accounts when the company hit my screen. I ignored these danger signals, and I shouldn't have. Computers can sift stocks quickly and accurately. But they can't replace homework and healthy skepticism.

Not all of my columns are based on others' research. Occasionally, I stick my neck out and write about an industry that looks appealing. The utilities I wrote about in January are up more than 21 percent, as are the managed-care stocks I featured in July. I'm happy to hold these stocks today, but I'm not rushing to put new money in.

Things are different when it comes to pipeline partnerships, which I've discussed several times over the years. They've barely outpaced the market since my column in May. As I said then, the easy money here has already been made. But I still think these are underappreciated income investments. Rising interest rates are hurting share prices, but increasing payouts will help offset those declines. Sunoco Logistics, for example, just boosted its distribution twice in one quarter. And most closed-end pipeline mutual funds now trade at discounts to their holdings. Two worth watching are Energy Income & Growth and MLP Opportunity.

Now for a personal note. You may have noticed another byline on this column recently. While I have stepped back from writing Stockscreen every month, I will continue to write for SmartMoney, in this space and elsewhere. I'm still fascinated by the market and hope to find new ways to pass along that enthusiasm to readers. Meanwhile, thanks for more than a decade of kind words and constructive criticism. Until next time, the three most important things I've learned over the years: Diversify. Collect dividends. And invest, don't trade.