Click here for the latest from James J. Cramer.

Is it better to buy a beat-up stock that has already preannounced a crummy quarter, or to stick with best of breed that hasn't screwed up -- betting that it won't?

We debate this dichotomy endlessly and we tend to settle it, as we do almost every difficult decision, with a little bit from column A and a little bit from column B because we simply don't know and we don't sweat the don't knows with doctrinaire thinking.

So we buy some Cisco (CSCO:Nasdaq - news - boards)and we buy some Compaq (CPQ:NYSE - news - boards). Cisco, where the numbers may yet turn out to be too high, is still expensive, but it has never been cheap. It seems ridiculous that this stock could sell at a market multiple given its growth. We don't think it goes there. But as Ron Insana asked me last night on CNBC, isn't it unreasonable still to be selling at 70 times earnings, as Cisco is? I shrug my shoulders and agree, but I can't wait until it sells at 35 times and then discover it never will.

Compaq's easier. We know Compaq is doing badly, but we are now in the "I know Compaq is worth something" mode, so we buy it with a core valuation that could sustain the franchise. Or, to put it in English, it is so cheap on old earnings that even if the new earnings are not so hot, the risk has been reduced.

When I talk about the dichotomy with good traders, we all realize that the one thing we now have on our side with stocks like Compaq is time. If we hold Compaq out seven, eighth months, we will begin to run into what we call "easy compares." That's where a company reports a number that, while not so hot, is still better than it reported last year. Right now we are facing hard comparisons, as business for Compaq had been booming. Until we lose the hard compares, Compaq might just be stuck in a trading range from here.

Isn't a trading range as risk better than one where Cisco could preannounce? Isn't there too much risk to Cisco in the $30s if it screws up? Undeniably there is a lot of risk. But, equally undeniably, the risk is much less than it was eight months ago. I think Cisco, while facing a slowdown in end markets, comes out as the winner and new champion in the sectors it works in during this downturn. That means, frankly, its earnings will turn out to be far higher than anybody thinks will happen in the out years. So, again, time moderates the risk.

And where do the 150-to-200 times earnings stocks come in? Those are the toughest. They are the ones that are owned by the wrong people -- mutual funds with redemptions -- but I can't advise selling them now because we are almost through the preannouncement season and if they haven't blown up yet they probably won't. I think that in January some of the selling pressure will be off these stocks and they will be just fine to lighten up on in a couple of weeks.

What's missing from this piece, I realize as I read it over, is the obvious Surgeon General's warning: none of this stuff is as good as owning a retailer or a financial going into a Fed ease. Remember, we at hedge funds are creatures of risk and reward. The risk to a Citicorp (C:NYSE - news - boards) here is much less than the reward. I regard the risk-reward in a Broadcom (BRCM:Nasdaq - news - boards), after it has moved 25 points up, as much more of a push, or a bet that won't pay off.

Random musings: I have been in touch with Bernadette.Sweeney@thestreet.com on the subject of gifting RealMoney and donating RealMoney to schools. I have heard that there are folks who are donating up to several thousand dollars worth of subs to kids who want to know the real deal. To which I say 'tis the season to give even more. Can I get you to give a few more subscriptions? Thanks. And have a great holiday if I don't see you again before Tuesday.