I'M ALMOST TEMPTED

to recommend

Starbucks

Also, screening metrics aside, Starbucks looks like a somewhat defensive business. It sells something that many people use every day. Four-dollar pumpkin lattes aside, that something is cheap enough so that even in an economic slowdown, many customers would be able to keep buying. Also, the company has met earnings estimates to the penny in each of its past four quarters (albeit, after previously reducing guidance at times). And there's only six cents separating Wall Street's lowest earnings per share estimate for next year ($1.02) from its highest one. Numbers like those tell me the company has excellent earnings visibility. Studies show that companies with tightly clustered earnings estimates tend to beat the market. Researchers figure that's because the most confident companies blab to anyone who'll listen about their prospects.

But there are a couple of problems with Starbucks. For one, it's no longer a blazing-fast grower, merely a peppy one. Analysts figure the company will increase its earnings by 19% this fiscal year (ended in September) and 21% next fiscal year. Since a prolonged acceleration in earnings growth for a company with $9 billion in yearly sales is rare, I have to figure Starbucks will at best average something like 18% earnings growth a year over the next several years. How much would you pay for that kind of growth? I'd pay a price that puts the stock's price/earnings ratio at no more than one and a half times its likely long-term growth rate. So Starbucks, to me, is worth 27 times this year's earnings forecast, tops. And since I'm looking for notably cheap stocks, and not fully priced ones, I'd be even keener on paying something like 20 times earnings. In fact, the company trades at 30 times this year's forecast. That's too much.

For me to be proven wrong in coming years, Starbucks will have to surpass analysts' earnings assumptions by a plump margin, something it hasn't done in recent quarters, recall. Of course, it could do that with runaway international growth, or by finding something new and particularly lucrative to sell at its stores. Five dollar pieces of chocolate have potential. Campbell Soup said in August that it will "explore strategic options" (read: sell) for its Godiva brand. Deutsche Bank analyst Marc Greenberg says the confectioner would be a good fit for Starbucks. Godiva has around 275 stores; Starbucks has more than 14,000. A few years ago it tinkered with selling Godiva cookies through its stores.

For now, though, I have to assume that the company's most reliable prospects for growth are baked into those aforementioned earnings estimates. So the stock still goes for a third more than I'd like to pay. Ten-year holders aren't complaining about their sixfold gain. But shares have lost nearly a third of their value over the past year. They're at $24 and change now. If they hit $17 and the business still looks healthy, I might change my tune.

Since you're likely more interested in stocks to buy than ones to avoid, have a look at two past Foxhole screen survivors that made the cut this time, too. Church & Dwight, which makes everything from Arm & Hammer baking soda to Trojan condoms, has tripled in price since I recommended it in March 2004. Chesapeake Energy, a natural gas producer, is up a quick 8%, or double the broad market's gain, since I endorsed the stock in August. (It has also tripled since I first recommended

Have a look at the full list of recent screen survivors if you please. And of course, you can run the search yourself anytime using the Foxhole screen recipe (or your modified version of it) and SmartMoney's stock screener.

Corrected on Oct. 31, 2007

Originally, the store count for Starbucks was understated. As of July 1, 2007, the company had more than 14,000 stores world-wide.

Jack Hough is associate editor at SmartMoney.com and the author of "Your Next Great Stock."