PALO ALTO, Calif. (CNN/Money) - The preoccupation for all cautious investors these past few months has been figuring out how to avoid getting caught up in the madness. Again.

We've all admired the strength in our mutual fund accounts. But we've wanted to avoid diving head first back into the momentum-stock swimming pool.

And yet, over and over, certain stocks have soared to the kinds of silly valuations we last saw in 1999.

What to do?

I set out in early June to find four tech stocks (defining tech rather broadly) that might be good to avoid. Whether you short stocks -- betting they'll fall rather than rise -- or simply sell or decide not to own them is your business.

I asked a bunch of smart people what was most likely to go down in a period when everything else was running up.

Well, I'm pleased to report that in a period where the Nasdaq composite jumped about 6 percent (from June 9 to mid-day trading Friday), three of my four "avoids" are down and the fourth is up a merely a smidgen. All told, the group of four is down an average of 20 percent.

I'll get to a look at the individual stocks in a moment, but first consider what's not on the list. As I noted in an article in Fortune discussing these picks, perhaps the hardest stocks to bet against are the blue-chip highfliers like eBay or Yahoo! or Amazon.

Perhaps in part because of their media gloss, these kinds of companies get something of a fundamental valuation pass. You can be "right" all day long and still lose money by betting against them.

Instead, pros stick to stocks where there's a more obvious edge. An easy example was to look for biotechnology stocks that were too far from profitability but already benefiting from the glow of industry news like Genentech's announcement that its Avastin drug is indeed a cancer killer.

Celera Genomics (CRA: Research, Estimates), was worth about $900 million in early June at $13, but it is years from having a product on the market. And, as an analyst pointed out in my piece, it'll need more financing just to get its product out. The stock is off about 27 percent.

Recently by Adam Lashinsky • Sun still sets in the West • eMomentum.com! • Watch the tech leaders

By the way, Genentech (DNA: Research, Estimates), which had rocketed from the high $30s to the mid $50s on the Avastin news and was considered by some skeptics to be overvalued in the low $70s has gone up since early June. Stay away from the truly loved. They cause pain for naysayers.

Real Networks (RNWK: Research, Estimates) is a similar news-halo-to-valuation-concern story. At $8 when I wrote about it, it's fallen to $6, still nearly a billion-dollar valuation for an unprofitable company. Investors were jazzed then about Real collecting fees for online music in light of Apple's new iTunes project. They're still jazzed, just not as much.

Meantime, at Level 3 (LVLT: Research, Estimates), more losses and higher debt-reduction expenses have soured investors on an already shaky stock. It's down about 31 percent.

A tougher battle so far has been on Omnivision Technologies (OVTI: Research, Estimates), which makes chips that go into digital cameras and cell phones that embed digital cameras. I was alerted to this one by short sellers who think Omnivision will lose out in the long run to more entrenched players. It stock is up 4 percent, despite having priced a secondary offering that includes a sale of stock by several of the company's executives. Stay tuned on this one.

Bottom line: You can keep your head in an exuberant market.

Adam Lashinsky is a senior writer for Fortune magazine. Send e-mail to Adam at lashinskysbottomline@yahoo.com.

Sign up to receive The Bottom Line by e-mail.