Sell your losers and let your winners run.

Great advice. But like other bits of Wall Street wisdom ("Buy low, sell high!"), it's easier to put on a bumper sticker than it is to execute.

Let's not even worry about the psychological barriers that make it hard for investors to get rid of a stock that's under water. Nobody likes to admit a mistake -- or to surrender to the fear that sometimes makes us sell too soon.

Instead, let's stick to an even more basic problem: How do you decide if a stock is a loser or a winner?

Seems simple enough, doesn't it? A loser is a stock that's gone down in price since you've owned it. A winner is one that's gone up since you tucked it into your portfolio.

But I'm afraid it's a lot more complicated than that. All winners go through periods when they look like losers. A stock can turn from a winner into a loser so gradually that an investor doesn't even notice. And simple mathematical formulas, like selling when a stock drops by 10%, don't work very well for investors -- whatever their usefulness to traders.

Take a look at

Puma Technology

(PUMA)

, a

Jubak's Pick on Dec. 14, 1999. For the next two months, the stock behaved like a loser. By Dec. 28, it was down almost 20 a share, a loss of slightly more than 20%. Time to sell that sucker, no? No! That day marked the low for the stock. Branding Puma a loser and selling then would have meant maximizing your loss and missing the explosive rally that began on Feb. 14 and has since taken the stock up to 133, which is 46% above the Dec. 14 recommendation price.

Or what about

Qualcomm

(QCOM) - Get Report

? Winner or loser? The stock was up 2,619% in 1999. "Winner! Let it ride." But the stock is down 17% since Jan. 3, 2000. "Loser! Sell and cut your losses." You can't have it both ways, of course, so Qualcomm can't be both a hold and a sell at the same time. Loser or winner? How do you decide?

I'm going to divide the problem in half. In this column, I'm going to take up the question of how to decide a stock that you own is a loser and that it's time to dump. Using examples such as Puma, Qualcomm,

America Online

(AOL)

,

Dell Computer

(DELL) - Get Report

,

Vitesse Semiconductor

(VTSS)

and

ARM Holdings

(ARMHY)

, I'll see if I can develop a general framework for defining a "loser." And in my next column, I'll follow the same process with a different set of examples to see if I can come up with a useful definition of a "winner."

1. No Stock Goes Straight Up Forever

Obvious, perhaps, but critical. Everything else builds on this fact.

Again, take a look at Puma. The stock began an astronomical rise on Aug. 30, 1999, taking it to 71 3/4 on Dec. 28. As the stock climbed, the price got further and further above the 50-day moving average -- an indication that the stock's gains were accelerating as it climbed. Then, on Dec. 29, the stock went vertical. It climbed to 131 for a gain of 60, or almost 85% in just three trading sessions.

This type of spike is almost never sustainable. Some investors always decide to take profits, figuring that much of the gain they projected is already in the stock. That initial profit-taking sends the stock price down and encourages other selling by investors who want to lock in their outsized gains. Typically, you'll see a strong surge in volume as the selling commences and a lot of investors head for the door at once.

That profit-taking can go on for a while, as other investors see the stock failing to go up as they'd hoped. And it's quite normal for a stock to retrace almost all of that last spike. By the end of January, for example, Puma was down to 76 3/4, within striking range of the 71 3/4 price on Dec. 28.

But as long as volume drops as this selling goes along, an investor usually has nothing to worry about. If fewer sellers appear as the price declines, it's a pretty good sign that the body of investors who hold this stock continues to believe in its future and still wants to keep the shares. At some point, indeed, the price starts moving sideways, marking time by moving up a few points, then down a few, but showing very little net change. In early February, Puma marked time in a range between 80 and 85. And a drop to just about 80 on Feb. 11 marked the turning point for the stock. The next day, on a significant increase in volume, Puma resumed the upward trend that the stock had established in 1999.

This isn't an unusual pattern, although it is unusually clear in the case of Puma. ARM Holdings, for example, shows a similar explosion from the 50-day moving average at 48 on Oct. 1 to a high of 200 a share on Dec. 28. The stock then pulls back to the 50-day moving average, moves up to a new high of 242, and then pulls back to the moving average again. None of this action, however, violates the long-term upward trend of the stock.

2. Time Counts

For a long-term investor, which I define as someone planning to hold a stock for five or more years, the December dips in Puma and ARM Holdings essentially are irrelevant. The best thing to do is hold right through them, as long as the stock doesn't violate its underlying upward trend. I think that's true as well for an investor with a holding period of a year to 18 months -- the holding period in Jubak's Picks.

But what if I stretched out exactly the same price decline for Puma over three months, instead of five weeks? Or nine months, instead of five weeks? Then, I think the same percentage price decline would be much more significant.

From Feb. 2, 1997 to Feb. 2, 1999, Dell's stock gained 1,207%. But since the stock hit 54 on Feb. 2, 1999, after a sharp spike that resembles those I've pointed out in the cases of Puma and ARM Holdings, Dell has been stuck in a trading range between 40 and 50. Total loss from that high to the Feb. 23 price of just a tad less than 40 is about 25%. That's not a huge drop in comparison to the 41% Puma fell in the month after Dec. 31. But the duration of the decline makes it much more significant.

3. Forget Lifetime Performance

An investor who bought Dell in February 1997 would have been up 1,207% in February 1999. An investor who held from Feb. 2, 1997, through Feb. 23, 2000, would be showing a gain of 860%. I certainly would never sneer at an 860% return. But the life of the investment gain disguises the fact that this long-term winner has been a loser for more than a year now.

This is exactly what I was afraid was happening to Vitesse Semiconductor in December and January. From Nov. 10, 1998, through Nov. 10, 1999, the stock gained 172%. But then it stalled. What had initially seemed to me to be a healthy rest started to look like a decisive break in the stock's upward trend. That's what led me to put Vitesse on my list of stocks to sell in my Feb. 4 column,

Finding the Upside in an Up-and-Down Market.

But on Feb. 9, the stock broke decisively above the 50 level. And it has since moved up to near 70 before settling into another range. Is Vitesse a winner or a loser? I don't think you can tell right now by looking at past price movements. And that brings me to my next observation.

4. A Loser With a Great Future Isn't a Loser

Is Vitesse a loser that should be sold now? How about Dell? Well, it depends not on the stocks' past, but on their future. A stock with Dell's past is a sell if you think the future is full of the negative surprises on earnings and revenue that have hurt the company over the last year. It's a hold or a buy if you think the company's new products and initiatives are going to reignite growth.

I admit that analysis is harder than just looking for patterns in a chart, but I don't think there's any other way to avoid selling a stock that you should have held just as business is about to pick up. If you don't own a stock, you've got the luxury of waiting to buy until the charts show signs of an upturn. But if you already own it, you're torn between holding and selling precisely because you don't see any pattern that might indicate an upturn.

For example, the argument against selling Vitesse here -- and the one that has been made by the analysts who have slapped 12-month target prices of 100 on the stock -- is that the problems caused by inventory reductions at

Lucent Technologies

(LU)

are over. The company has, in fact, scored key design wins in Lucent's new WaveStar optical networking system. In addition, these analysts point out, Vitesse has added

Nortel Networks

(NT)

as a significant customer. Three recent design wins at Nortel should turn into significant revenue for Vitesse in the current quarter. And by the end of 2000, sales to Nortel could be responsible for 10% of Vitesse's revenue. Frankly, I'm skeptical about Lucent. But the design wins at Nortel, the top dog in optical networking at the moment, are significant. These fundamental factors about the future incline me to hold Vitesse rather than sell it -- at least until I see how these trends play out in the quarter.

How about other potential sells, such as Dell, Qualcomm and America Online? Dell seems to face tough going for at least the next two quarters. Parts remain expensive because supplies are extremely tight, and that's likely to cut into margins at all the PC makers. I like Dell's prospects in the last half of the year, but not over the next six months. The stock is a sell rather than a hold here. (I'm selling it out of Jubak's Picks with publication of this column.)

Qualcomm, by contrast, strikes me as a hold on its near-term future prospects. The company has been telling analysts recently that it expects to pick up market share in the CDMA wireless-phone chip business. Just as the growth rate in Korea, the traditional growth driver for CDMA phones, is slowing due to market saturation, growth in Japan and Latin America is picking up. The decline from near 180 at the end of December to a trading range of 130 to 140 in February should make it possible for positive news in these areas to move the stock up. (If you like the general stock market more than I do right now and you don't already own Qualcomm, it's a buy at current prices.)

And America Online? The analyst community finally has started to get comfortable with the business of a combined America Online and

Time Warner

(TWX)

. Much of the stock's fall to a low below 50 a share was the result of silence from every Wall Street investment house. In this context,

Merrill Lynch's

recent report on the deal, which recommended both Time Warner and America Online, is a significant breakthrough. The other major brokerages won't lag far behind now. Merrill's 90 target price on America Online immediately moved the stock. I'd call this one a hold as well if you already own it.

So much for the "Sell your losers" half of the adage. I've deliberately tackled the easier half first. Even though selling a stock that's under water can be a blow to the ego, I don't think it causes anywhere near the anxiety that holding onto a winner does. I'm not sure what that says about investors -- perhaps we all think disaster must follow hard on the heels of any good fortune. Next column: How to hold onto your winners.

Jim Jubak is senior markets editor for MSN MoneyCentral. At time of publication, he was long American International Group, America Online, Commerce One, Nortel Networks, Puma Technology and Vitesse Semiconductor, although holdings can change at any time. Under no circumstances does the information in this column represent a recommendation to buy or sell stocks. He welcomes your feedback at

mctsc@microsoft.com.

More from

MSN MoneyCentral

Jubak's Picks

Markman's SuperModels

Rowland's Watch Portfolio